Citation
Employee or independent contractor

Material Information

Title:
Employee or independent contractor : an examination of the relevant variables employed by the Federal courts in deciding the question
Creator:
Stewart, Dave Nelson, 1952- ( Dissertant )
Kramer, John L. ( Thesis advisor )
Collins, William A. ( Reviewer )
Calfee, Dennis A. ( Reviewer )
Henretta, John C. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1980
Language:
English
Physical Description:
ix, 187 leaves.

Subjects

Subjects / Keywords:
Common law ( jstor )
Discriminant analysis ( jstor )
Discriminants ( jstor )
Employees ( jstor )
Employment ( jstor )
Independent contractors ( jstor )
Labor ( jstor )
Mathematical variables ( jstor )
Modeling ( jstor )
Taxes ( jstor )
Accounting thesis, Ph.D.
Contractors -- Taxation -- United States
Dissertations, Academic -- Accounting -- UF
Independent contractors -- Taxation -- United States
Payroll tax -- Law and legislation -- United States
Withholding tax -- Law and legislation -- United States
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
The distinction between an employee and an independent contractor is crucial in the application of the federal employment tax laws. If a worker is classified as an employee, the employer must (1) withhold income taxes, (2) withhold social security taxes, (3) match the social security taxes paid by the employee, and (4) pay unemployment taxes. If the worker is classified as an independent contractor, the employer need not pay or withhold any of the above mentioned taxes. Instead, the independent contractor is responsible for paying his own income and self-employment taxes. The Internal Revenue Code statutorily defines several occupational categories that are treated as employees for employment tax purposes. The most prominent of these categories is that of a "common law" employee. However, the scope of the Code and Regulations is sufficiently limited to leave the common law employee definition shrouded in uncertainty. It is this ambiguity, coupled with the IRS's increased enforcement of employment taxes, that has resulted in pressure on Congress to enact legislation that will add a degree of certainty to the definition of a common law employee. The legislation being considered by Congress would result in several "safe harbor" criteria that if satisfied would automatically indicate employee status. The purpose of this study is to develop a mathematical model that will empirically identify those variables used by the federal courts in employee versus independent contractor cases in hopes that these variables may provide direction to tax policymakers. A sample of 148 District Court and Court of Claims cases was identified and used as the data with which the mathematical models were estimated. Discriminant analysis, OLS regression analysis, and a non-linear logit analysis were all used to identify those variables deemed most important by federal judges. A forward stepwise discriminant procedure resulted in a 6-variable model that was able to correctly classify 96.6 percent of the cases. A forward stepwise OLS regression procedure resulted in a 7-variable model that correctly classified 95.3 percent of the cases and explained 80 percent of the variance of the model. The non-linear stepwise logit procedure resulted in a 5-variable model that was able to correctly classify 97.3 percent of the cases. Both the linear and non-linear techniques produced substantially identical results. When compared with the proposals presently being considered by Congress, three of the variables recommended by the Congressional subcommittee are also included in this study. Using these three variables as the model, approximately 90 percent of the cases were correctly classified. A secondary question addressed in this study is whether the choice of legal forum is important in an employment status tax case. Through a comparison of the discriminant classification accuracy rates and the Chow test applied to the OLS results, it was concluded that for employment tax cases the two Federal Courts use very similar decision-making models. The findings of this study suggest several conclusions. First, the use of the non-linear analysis did not produce substantially different results than did the linear regression or discriminant analysis procedures. Second, there are factors that have been used successfully by federal judges in employment tax cases that could provide direction to tax policymakers attempting to statutorily define a common law employee. Third, the decision-making models employed by the District Courts and the Court of Claims are not significantly different for federal employment tax cases.
Thesis:
Thesis (Ph.D.)--University of Florida, 1980.
Bibliography:
Includes bibliographical references (leaves 177-186).
General Note:
Vita.
General Note:
Typescript.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
06800179 ( OCLC )
AAL4039 ( LTUF )
0023382484 ( ALEPH )

Downloads

This item has the following downloads:


Full Text












EMPLOYEE OR INDEPENDENT CONTRACTOR:
AN EXAMINATION OF THE RELEVANT VARIABLES EMPLOYED
BY THE FEDERAL COURTS IN DECIDING THE QUESTION














BY

DAVE NELSON STEWART


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1980





































Copyright 1980

by

Dave Nelson Stewart















ACKNOWLEDGEMENTS


I wish to thank John L. Kramer, the chairman of this dissertation

committee, for his guidance and support throughout the course of

this study. Professor Kramer has given unselfishly of his time,

especially in the reading and commenting on different drafts of

this dissertation.

My entire dissertation committee has contributed greatly to

this project. Professor William A. Collins, Dennis A. Calfee, and

John C. Henretta have each shared freely of their knowledge and

expertise. They have never been too busy to sit down and discuss

the progress of this study. Valuable criticisms and suggestions

were also received from the Ph.D. students and faculty at the

University of Florida Accounting Workshop.

Finally, I would like to thank my dear wife, Jane. She not

only provided constant encouragement and support, but her typing and

editing skills were invaluable to the production of the finished paper.
















TABLE OF CONTENTS


PAGE

ACKNOWLEDGEMENTS . . . . ... . .iii

LIST OF TABLES . . . ... ...... vi

ABSTRACT . . . ... . . .vii

CHAPTER

I. INTRODUCTION . . . .. .. ... 1

Significance of the Problem . . . 2
Research Methodology . . . . 7
Research Questions . . . . 7
Data Collection . . . . 8
Research Question 1 . . . . 9
Research Question 2 . . . .. 10
Research Question 3 . . . . 11
Research Question 4 . . . ... 12
Limitations . . . . ... . 13
Summary of Chapter Contents . . ... .14
Notes . . . . . ... 15

II. HISTORICAL BACKGROUND . . . ... 19

Development of Legislative History . . .. 19
Social Security Act of 1935 . . ... 19
Common Law Definition of Master-
Servant Relationship ............. 21
Social Security Amendments of 1939 . ... .24
Income Tax Withholding Provisions . ... 26
Judicial Interpretation . . .. 27
Social Security Amendments of 1950 . ... .34
Current Legislative Status . . . ... 38
Revenue Act of 1978 . . . ... 43
Current Proposed Solutions . . . ... 44
General Accounting Office . . . 44
H.R. 3245; Safe Harbor Remedy . . ... 46
Department of the Treasury . . ... .49
Notes . . . . . ... 50









PAGE


CHAPTER


III. RESEARCH METHODOLOGY . . . .


Research Question 1 . .
Variable Identification
Operational Definitions of
Coding of the Variables
Research Question 2 . .
Mathematical Techniques in
Discriminant Analysis .
Logit Analysis . .
Research Question 3 . .
Research Question 4 . .
Notes . . . .

IV. EMPIRICAL RESULTS . .


Variables



Tax Research




. .


Research Question 2 . . .
Discriminant Analysis Results .
OLS Regression Results .. ...
Logit Analysis Results . .
Comparisons of Discriminant, OLS
Regression, and Logit Analyses .
Research Question 3 . . .
Research Question 4 . . .
Notes . . . . .

V. SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS .
Summary and Conclusions . . .
Historical Background of Common Law
Variable Identification . .
Empirical Results . . .
Temporal Stability of the Model .
Current Legislative Proposals .
Choice of Legal Forum . .
Recommendations . . . .
Suggestions for Future Research . .
Notes . . . . .


57
57
58
89
93
94
98
104
111
113
114


. . 121

. . 121
. . 121
. . 136
. .. 142


Employee


.. .

. .
. .
. .


146
151
153
159

160
161
162
163
164
166
167
167
168
169
171


APPENDIX


A. LIST OF DISTRICT COURT CASES INVOLVING THE EMPLOYEE
VERSUS INDEPENDENT CONTRACTOR QUESTION . .


. 172


B. LIST OF COURT OF CLAIMS CASES INVOLVING THE EMPLOYEE
VERSUS INDEPENDENT CONTRACTOR QUESTION . ... 176

BIBLIOGRAPHY . . . . ... ... 177

BIOGRAPHICAL SKETCH . . . . ... . 187


. 57
















LIST OF TABLES


PAGE


TABLE


1.1 A COMPARISON OF THE FICA AND SECA TAXES
FOR 1978 THRU 1987 . . .

4.1 BMDP 6-VARIABLE MODEL . . .

4.2 SPSS 9-VARIABLE MODEL . . .

4.3 COMPARISON OF CLASSIFICATION ACCURACY .

4.4 DISCRIMINANT & CLASSIFICATION COEFFICIENTS
FOR THE 6-VARIABLE MODEL . .

4.5 CLASSIFICATION ACCURACY MATRICES .

4.6 MODEL BASED ON 22 DUMMY VARIABLES .

4.7 CLASSIFICATION ACCURACY MATRIX ..

4.8 TEMPORAL STABILITY TEST OF
CLASSIFICATION ACCURACY . . .

4.9 OLS REGRESSION RESULTS . . .

4.10 OLS STEPWISE REGRESSION RESULTS . .

4.11 COEFFICIENTS AND CLASSIFICATION
MATRIX FOR FULL 11-VARIABLE MODEL .

4.12 COEFFICIENTS AND CLASSIFICATION MATRIX
FOR THE BMDP 5-VARIABLE MODEL . .

4.13 COMPARISON OF STEPWISE RESULTS . .

4.14 CLASSIFICATION ACCURACY FOR
CHOICE OF FORUM TEST . . .

4.15 FACTORS CONTAINED IN ALL THREE PROPOSALS

4.16 CLASSIFICATION ACCURACY MATRICES
FOR THE 3-VARIABLE MODELS . . .


4

. . 122

. . 124

. . 126


. . 127

. . 129

. . 131

. . 132



. . 135

. . 137

. . 139



. . 143


. . 145

. . 147


. . 152

. 155


. 157















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


EMPLOYEE OR INDEPENDENT CONTRACTOR:
AN EXAMINATION OF THE RELEVANT VARIABLES EMPLOYED
BY THE FEDERAL COURTS IN DECIDING THE QUESTION

By

Dave Nelson Stewart

June 1980

Chairman: John L. Kramer
Major Department: Accounting

The distinction between an employee and an independent contractor

is crucial in the application of the federal employment tax laws. If

a worker is classified as an employee, the employer must (1) withhold

income taxes, (2) withhold social security taxes, (3) match the social

security taxes paid by the employee, and (4) pay unemployment taxes.

If the worker is classified as an independent contractor, the employer

need not pay or withhold any of the above mentioned taxes. Instead,

the independent contractor is responsible for paying his own income

and self-employment taxes.

The Internal Revenue Code statutorily defines several occupational

categories that are treated as employees for employment tax purposes.

The most prominent of these categories is that of a "common law"

employee. However, the scope of the Code and Regulations is sufficiently

limited to leave the common law employee definition shrouded in uncer-

tainty. It is this ambiguity, coupled with the IRS's increased

vii








enforcement of employment taxes, that has resulted in pressure on

Congress to enact legislation that will add a degree of certainty

to the definition of a common law employee. The legislation being

considered by Congress would result in several "safe harbor" criteria

that if satisfied would automatically indicate employee status.

The purpose of this study is to develop a mathematical model

that will empirically identify those variables used by the federal

courts in employee versus independent contractor cases in hopes

that these variables may provide direction to tax policymakers.

A sample of 148 District Court and Court of Claims cases was iden-

tified and used as the data with which the mathematical models were

estimated. Discriminant analysis, OLS regression analysis, and a

non-linear logit analysis were all used to identify those variables

deemed most important by federal judges.

A forward stepwise discriminant procedure resulted in a 6-variable

model that was able to correctly classify 96.6 percent of the cases.

A forward stepwise OLS regression procedure resulted in a 7-variable

model that correctly classified 95.3 percent of the cases and explained

80 percent of the variance of the model. The non-linear stepwise

logit procedure resulted in a 5-variable model that was able to correctly

classify 97.3 percent of the cases. Both the linear and non-linear

techniques produced substantially identical results.

When compared with the proposals presently being considered by

Congress, three of the variables recommended by the Congressional

subcommittee are also included in this study. Using these three vari-

ables as the model, approximately 90 percent of the cases were correctly

classified.
viii









A secondary question addressed in this study is whether the choice

of legal forum is important in an employment status tax case. Through

a comparison of the discriminant classification accuracy rates and the

Chow test applied to the OLS results, it was concluded that for employ-

ment tax cases the two Federal Courts use very similar decision-making

models.

The findings of this study suggest several conclusions. First,

the use of the non-linear analysis did not produce substantially

different results than did the linear regression or discriminant

analysis procedures. Second, there are factors that have been used

successfully by federal judges in employment tax cases that could

provide direction to tax policymakers attempting to statutorily define

a common law employee. Third, the decision-making models employed by

the District Courts and the Court of Claims are not significantly

different for federal employment tax cases.














CHAPTER I

INTRODUCTION


Until the 1960's, the collection of federal employment taxes--

withholding, social security, and unemployment--followed a relatively

smooth course for employers and other taxpayers. However, since the

late 1960's, the Internal Revenue Service (IRS) has increased its en-
2
forcement of the employment tax laws. The question of employment

status generally centers around the classification of whether a worker is

an independent contractor or an employee. Because so few categories

of employees are statutorily defined, the question of whether a worker

is an independent contractor or an employee generally centers around the

issue of what constitutes a "common law" employee. The common law

definition of employee has evolved from a substantial body of case law.3

Congress initially felt that all uncertainty in the classification of an

individual as an employee or an independent contractor would "vanish if

only common law principles distilled from many decisions by many courts

were applied to factual situations as they arose." The assumption that

there is some simple, uniform and easily applicable
test which the courts have used unfortunately .
is not true. Few problems in the law have greater variety
of application and conflict in results than the cases arising
in the borderland between what is clearly an employer-employee
relationship and5what is clearly one of independent-entrepre-
neurial dealing.
6
The substantial amount of litigation in the federal employment tax

area is evidence that the common law definition of an employee is easier

stated than applied. The determination of employment status is an






2

issue on which both the taxpayer and the government desire some degree

of closure. The taxpayers complain that the retroactive reclassification

of workers as employees and the resulting taxes, interest, and penalties

are causing interruption of business activities and financial hardship

through severe cash flow demands.7 The IRS is claiming that with the

uncertainty surrounding the common law rules that taxpayers are able to
8
easily avoid employee status. Since this results in no withholding of

taxes by the employer, independent contractors are easier able to

escape detection should they fail to file their tax returns.

Consequently, proposals have been forthcoming from a Congressional
9 10 II
subcommittee, the General Accounting Office, and the IRS that

recommend the solidification of the many factors that determine a common

law employee into a smaller number of factors that provide a more clear-

cut test. However, these factors lack theoretical support and in some

cases even lack the desirable tax characteristics of simplicity and

certainty. The purpose of this study is to determine those factors that

the courts have found significant in defining employment status with

the hope that these variables may prove helpful to tax administrators

as they search for an appropriate definition of the employee-employer

relationship. Such a study could provide valuable insight into what

factors can be consistently applied when individuals are faced with the

task of determining if a worker is an employee or independent contractor.


Significance of the Problem


Once the court determination is made that a certain individual is

an "employee," the employer may have the following responsibilities:

(1) withhold social security taxes under the Federal Income Contributions







12 13
Act (FICA),12 (2) withhold federal income taxes, 13 (3) match the FICA
14
taxes paid by the employee, and (4) pay unemployment taxes under the

Federal Unemployment Tax Act (FUTA).15 When an individual is clas-

sified as an independent contractor the purchaser of his services

need not pay or withhold any of the above mentioned taxes. The pur-

chaser of the services may only need file an information return (Form

1099).16 It is then the responsibility of the independent contractor

to pay self-employment taxes under the Self-Employment Contributions

Act (SECA)17 and pay his own income taxes through the use of the

estimated tax procedures.8 With inflation pushing wages into higher

tax brackets and prospective increases in both the FICA and SECA wage
19
bases and rates,9 the amount of employment taxes can vary substantially

according to one's employment status.

Table 1.1 represents the present and proposed FICA and SECA tax

rates and wage bases from 1978 to 1987. Columns (5) and (6) illustrate

that the social security taxes are substantial and will drastically

increase in the next ten years. Column (7) shows the differences in the

actual tax that goes into the Social Security Fund depending on the

employment status. The difference in what an employer must pay is even

more pronounced when one includes the FUTA taxes (3.4% of the first

$6,000 of wages per employee). For example, using the maximum wages

applicable to employment taxes for 1980, Column (7) shows that an ad-

ditional $1,077.44 of taxes are paid into the federal treasuries if a

worker is an employee versus an independent contractor. This difference

of $1,077.44 should be increased by the FUTA taxes of $204 (3.4% x

6,000) that are also imposed thereby resulting in a total difference

of $1,281.44. Consequently, Column (7) probably represents for most









TABLE 1.1

A COMPARISON OF THE FICA AND SECA TAXES
FOR 1978 THRU 1987


(1) (2) (3) (4) (5) (6) (7)

Max. Max. Excess FICA
Maximum over
FICA SECA Wage FICA SECA Maximum over
SECA Tax
Year Ratea Rate Basec Tax Tax Maximumd


7 % $ $ $ $


1978 12.10 8.10 17,700 2,141.70 1,433.70 708.00

1979 12.26 8.10 22,900 2,807.54 1,854.90 952.10

1980 12.26 8.10 25,900 3,175.34 2,097.90 1,077.44

1981 13.30 9.30 29,700 3,950.10 2,762.10 1,188.00

1982 13.40 9.35 31,800 4,261.20 2,973.30 1,287.90

1983 13.40 9.35 33,900 4,542.60 3,169.65 1,372.95

1984 13.40 9.35 36,000 4,824.00 3,366.00 1,458.00

1985 14.10 9.90 38,100 5,372.10 3,771.90 1,600.20

1986 14.30 10.00 40,200 5,748.60 4,020.00 1,728.60

1987 14.30 10.00 42,600 6,091.80 4,260.00 1,831.80



aThis percentage includes both the employer's and the employee's
portion of the FICA tax. This percentage includes the social security
rate and hospital insurance rates as contained in Section 3101.

This percentage includes the self-employment rate and the hospital
insurance rate as contained in Section 1401.

CAny wages above these specified amounts are not subject to FICA
or SECA tax.

This amount is obtained by subtracting column (6) from column
(5).






5

taxpayers a conservative estimate of the total additional taxes paid

when a worker is deemed to be an employee.

The preceding illustration is based simply on the differences in

employment taxes between an independent contractor (self-employed) and

an employee. The determination of employment status can be much more

critical if the IRS should choose to reclassify an independent contrac-

tor as an employee. The employer would then be assessed for the income
20
taxes he failed to withhold, both the employer and employee FICA
21
taxes that should have been paid by the employer, and the FUTA taxes

that should have been paid,22 plus the appropriate interest and penal-
23
ties. Such assessment can be made retroactively for all tax years

that remain open under the statute of limitations (generally three

years under Section 6501). Only if the employer can furnish certifi-

cates signed by his employees stating that they have paid the proper

amount of income tax, can the employer possibly have the assessment
24
reduced.24 The employer can have the employee's share of the FICA tax

abated by the SECA tax already paid only if the employee is prevented
25
by law from filing for a refund of the SECA tax that he paid in error.2

This situation obviously results in a double payment of the employee's

share of the FICA tax.

A reclassification by the IRS causes even more problems for

employers because of the Tax Court's lack of jurisdiction over employ-
26
ment taxes. Therefore, a suit cannot be instigated by a taxpayer

without payment of the assessed tax because the District Courts and

Court of Claims require the payment of the assessment prior to their

judicial review. To require a business to pay three years worth of

income, FICA, and FUTA taxes while instituting a legal action against








the government could impose severe hardship and cost upon a firm.

However, since employment taxes are generally paid on a quarterly basis,

the taxpayerordinarily files a return covering one quarter's employment

taxes and then sues for a refund of that amount and abatement of the

yet unpaid amount.27 This partial payment procedure can still involve

a considerable amount of funds when dealing with industries that have

a large number of workers that would be affected by such a determination.

Jerome Kurtz, Commissioner of the Internal Revenue Service, in a speech

before the American Bar Association pointed out:

The employee independent contractor issue has generated a
staggering problem for employers assessed with huge liabili-
ties for past noncompliance. At present, in two districts
of one region alone there are three potential assessments
against taxpayers ranging from $6 million to $60 million .
and reclassification of independent contractors as employees
result in deficiency assessments estimated at $107,000,000
annually.28

Even if the taxpayer only has to pay one quarter's worth of employment

taxes to initiate a suit for refund and abatement of the IRS assessment,

the IRS is not precluded from seizing a taxpayer's property in this

type of partial payment litigation.29 This action is not taken in most

situations unless the IRS feels there are circumstances which might

jeopardize the availability of the taxpayer's funds subsequent to the
30
litigation.30 This threat exists in employment tax cases as evidenced

by American Consulting Corporation v. U.S.31 where the IRS seized a bank

account owned by the plaintiff whose employment status was being ques-

tioned.

The determination whether a worker is an employee or an independent

contractor is far more reaching than simply the employment taxes in-

volved. The reclassification of an independent contractor can have

adverse effects on a firm if it has a qualified pension, profit-sharing,








or stock bonus plan. Such a reclassification may cause the plan to

violate the minimum participation requirements. This could cause the

disallowance of deductions taken for contributions to the plan and

possibly even taxation of the trust's income.32 This would most

likely result in termination of the plan by the employer because of the

increased expense or at best the effect would be to reduce the benefits

available to the employees. Workers who consider themselves self-

employed and establish Keogh plans for their own retirement may find

that prior contributions are disallowed or the plan frozen with

respect to any future tax exempt contributions.

Reclassification of workers as employees instead of independent

contractors affects various other fringe benefits. Employees exclude

from income any contribution paid by their employer for accident and
33 34
health plans,33 employee education assistance plans,34 certain employer-
35
provided life insurance and death benefit programs, and the value of

any meals or lodging furnished for the convenience of the employer.36

Independent contractors are not afforded these privileges. In addition,

other sections of the Code that apply specifically to employees would

be affected by any reclassifications by the IRS. For example, the

WIN tax credit for wages paid to certain employees hired under the
37
Federal Work Incentive program37 and the targeted jobs credit for wages

paid to certain new employees are all examples of areas of the tax

law affected by reclassifying independent contractors as employees.


Research Methodology


Research Questions


The four major research questions that are studied in this









dissertation are:

Research Question 1.


Research Question 2.




Research Question 3.





Research Question 4.


What factors or variables have been used in
determining employment status?

In constructing a mathematical model of the
courts' decision-making behavior, which of the
variables espoused in Research Question
I prove to be significant?

In determining the employment status of the
cases examined, do the different courts of
original jurisdiction (District Courts versus
Court of Claims) use similar factors in the
rendering of decisions?


a) Do the variables determined in Research
Question 2 have any similarity with the
criteria in the current proposals before
Congress?

b) What possible recommendations might these
results suggest?


Data Collection


The Regulations specifically state that a common law employee is to
39
be determined according to the facts of each case. The facts of a

case are determined in the original court of jurisdiction. In this type

of study, the assumption must be made that because of the possibility

of appeal to a higher court, judges attempt to include in their opinions

a discussion of the relevant facts in their decision-making process.

With this assumption, the Lexis computer data bank was used to search
40
for all employee versus independent contractor cases.0 From Lexis and

other manual research methods a total of 179 cases were identified for

the 1940 through 1979 period. The year 1940 represents the first year

in which a significant number of cases were tried that dealt with the

employment status question in connection with federal employment taxes.








The study also includes all the cases that could be found through

the end of the 1979 calender year.

Of the 179 cases, 24 were tried with a jury. In this situa-

tion, the written opinions contain only the instructions to the jury

and the actual verdict. Therefore, these 24 cases were eliminated

from the study. Sixteen cases also were eliminated because there was

not sufficient information in the opinion to determine what factors

the judge considered in his decision. This leaves a total of 139

cases of which 119 were tried in the District Courts and 20 were tried

in the Court of Claims (for a list of cases see Appendices A and B).

A small number of cases were discovered at the appellate level that

had not been reported at the District Court or Court of Claims level.

These cases were also excluded because the facts of a case are determined

in the court of original jurisdiction.


Research Question 1


The sources examined to determine the relevant factors in deciding

employment tax cases for federal tax purposes include the following:

(1) the Internal Revenue Code, (2) the corresponding Treasury Regula-
41
tions, (3) certain landmark judicial decisions,1 and (4) a BNA Tax
42
Management portfolio entitled Employee Defined.42 Another source that

proved helpful was the Employment Tax Procedure audit guidelines
43
published by the IRS for their agents.43 The list of factors that was

compiled was reduced as the court cases were read and it became evident

that certain factors seldom were discussed in the court opinions or

that two variables were so highly correlated that they were combined

and represented as only one variable. A list of the complete set of






10

variables used in this study is contained in Chapter III with additional

explanation of the selection process, operational definitions of the

respective variables, and how they were measured for this study.


Research Question 2


In constructing a mathematical model of a court's decision-making

behavior, three different approaches were taken. First, a discrimi-

nant analysis model was built using the variables espoused in the

results of Research Question I and the model was used to classify the

court cases as either an employee or an independent contractor outcome.

The P7M program of the Biomedical Computer Programs P-Series (BMDP)44

computer software package was used to derive the discriminant function.

From this discriminant function, classification functions for each of

the two groups were derived. Each case was then assigned to the group

in which the highest probability of membership existed.

A frequently recommended technique to test the accuracy of the

classification function is that of using a holdout sample. However,

with the limited number of cases available in this study, an alternate

test was applied.45 This technique calculated n-1 discriminant func-

tions where n is equal to the number of cases in the sample. The

procedure involved calculating a discriminant function with all but

one case, and then that function was used to classify the left out

case. The procedure was repeated n-1 times, and the number of incor-

rectly classified cases was used as an estimate of the probability of

misclassifying an employee versus independent contractor case.

Second, an ordinary least squares (OLS) regression model was built

using the variables espoused in the results of Research Question 1.








The OLS model was used to generate a probability of an employee

finding versus an independent contractor finding for each case.

These probabilities were used to classify each case. These classi-

fications were compared with those obtained from the discriminant

model to determine if these two linear statistical techniques gener-

ate different results.

Certain statistical assumptions are violated by using either of

the preceding linear modeling techniques. Therefore, the third

approach taken was to use logit analysis which is a non-linear es-

timation technique. Logit analysis assumes that the underlying func-

tional form of the relationship being studied is curvilinear and

estimation of the model allows for interaction of the independent

variables. The logit model involves a transformation of the depen-

dent variable from a binary to a continuous random variable. The
46
logit model was estimated using the Nerlove and Press and the
47
BMDP stepwise logistic regression programs which generate maximum

likelihood (ML) estimates. The ML estimates were used to generate

probabilities for each case by which they could then be assigned

to either the employee or independent contractor groups. These

results were compared with the discriminant analysis and OLS re-

gression results to determine if linear and non-linear estimation

techniques produced different results.


Research Question 3


To determine if the District Courts and Court of Claims apply

different factors in deciding an employee versus independent contrac-

tor case, the first approach used in this study was to build a







discriminant function using only the District Court cases. The

resulting classification functions were used to classify both the

District Court and Court of Claims cases. If the misclassification

rates are significantly higher for the Court of Claims cases, this

will be taken as evidence that the two courts use different decision-

making models.

The second approach used to determine if the District Courts

and Court of Claims use different decision-making models was to

build two additional OLS regression models. The first model was

based exclusively on the District Court cases and the second model

was based exclusively on the Court of Claims cases. The Chow test4

was used to determine if the two regressions came from the same

population. The results of this test along with the results of the

discriminant analysis test were used to draw conclusions pertaining

to the similarity of the two courts with respect to deciding common

law employee versus independent contractor cases.


Research Question 4


The preceding research questions were designed to discover which

variables are significant to the courts in deciding the employee

versus independent contractor question. The results are based on

the facts of the cases as discussed by the respective judges. The

logical extension of this research is to determine if the variables

found to be significant in Research Question 2 are related to the

criteria contained in the proposals before Congress. If the criteria

are quite different from those determined in this study, tax policy

makers may want to reconsider those variables that have been suggested





13

in order to provide a more certain definition of an independent con-

tractor. If the results of this study produce variables that have

been consistently used by judges in rendering decisions in employ-

ment status cases, it would seem that these variables would be a

good starting point for legislation. Therefore, the factors ob-

tained from Research Question 2 were compared with the current

proposals before Congress. Similarities were noted and recommen-

dations made of additional factors that Congress may want to con-

sider in the actual drafting of legislation in the employment tax

area.


Limitations


This study attempts to identify those factors that a diverse

set of judges have consistently used in deciding the employee versus

independent contractor question. To do so, a research methodology

was used that provides for a more systematic approach than the

legal analysis that is generally used to study problems of this

type. While the advantage may be that the approach is more objec-

tive than legal analysis which depends on the researcher's own

heuristics, it also suffers from certain limitations. The greatest

limitation of this type of study is the reliance on the fact that

the judge has actually included in his written opinion all of the

significant variables he used in his decision process. If a sig-

nificant variable is not included in the written opinion, it can

not be measured and considered in this type of study. This problem

is somewhat mitigated by the fact that the judge knows his opinion is

subject to reversal for failure to consider all relevant facts.







The ability to reduce the data in the written opinions into a

numerical form also presents some problems. In an attempt to re-

duce the subjectivity involved in actually obtaining the data, a

variable generally was not assumed present unless it was explicitly

noted in the text of the case. Each case was read several times

by the author in an attempt to be consistent in the actual coding

of the variables.

It should also be noted that because of the dynamic structure

of the tax law and the legal process, that the results of this study

are not predictive in a future sense unless one assumes stability

for all the factors affecting the employee versus independent con-

tractor question. The fact that the data are based only on cases

that were actually litigated also produces a potential sample bias.

The majority of cases that are actually litigated are cases in which

both parties feel they have a reasonable chance of winning. It

is not clear if this bias is very crucial because it is those very

situations where the decision is in doubt that this study is in-

terested in investigating.


Summary of Chapter Contents


This study is divided into five chapters. Chapter I is the

introduction to the study which includes a brief discussion of the

nature and scope of the problem, the research questions to be

addressed, the basic research methodology to be used, and possible

limitations of the study.

Chapter II provides a historical perspective of the common

law definition of employee by tracing its use through the Code and








corresponding regulations. This historical background is supple-

mented by a discussion of the current legislative status of common

law employees and also proposed legislation that is currently being

considered by Congress.

Chapter III presents a more detailed outline of the research

methodology that was used. Chapter III also examines Research

Question I which outlines the selection of the variables used in

the study along with their operational definitions.

Chapter IV describes the discriminant, OLS regression, and

logit analyses that were used to model the courts' decision-making

behavior. The empirical results of each of the three techniques

are analyzed. The differences and similarities of the three

techniques are then evaluated.

Chapter V summarizes the preceding chapters. In addition,

implications of this study are discussed and recommendations for

future research are made.


Notes


ISmith, Independent Contractor or Employee?--That is the
Question, 33 N.Y.U. INSTITUTE ON FEDERAL TAXATION 578 (1975).

JOINT COMMITTEE ON TAXATION, ISSUES IN THE CLASSIFICATION
OF INDIVIDUALS AS EMPLOYEES OR INDEPENDENT CONTRACTORS 7 (1979).

GENERAL ACCOUNTING OFFICE, TAX TREATMENT OF EMPLOYEES
AND SELF-EMPLOYED PERSONS BY THE INTERNAL REVENUE SERVICE: PROBLEMS
AND SOLUTIONS 5 (1977).

S. REP. No. 1255, 80th Cong., 2d Sess., p. 7 as cited
in M.F.A. Mutual Insurance Co. v. U.S., 314 F. Supp. 595 (W.D. Mo.
1970).

5National Labor Relations Board v. Hearst Publications,
322 U.S. 120,121 (1944).








See the list of cases in Marmoll, 391 TAX MNGM'T (BNA),
'EMPLOYEE' DEFINED, pp. c-6 thru c-10 (1979).

GENERAL ACCOUNTING OFFICE, supra note 3, at 13.

Id. at 23.

H.R. 3245, 96th Cong., 1st Sess., 125 CONG. REC. E 1342
(daily ed. March 27, 1979) (remarks of Rep. Gephardt).

GENERAL ACCOUNTING OFFICE, supra note 3.

1 Hearings on H.R. 3245 Before the Subcommittee on Select
Revenue Measures, 96th Cong., Ist Sess. 5 (1979) (statement of
Donald C. Lubick).
12 Section 3102(a).
13I.R.C. Section 3102(a).

14I.R.C. Section 3402(a).


15I.R.C. Section 3301.
16I.R.C. Section 6041(a).

I7.R.C. Section 1401.

18I.R.C. Section 6015.


I91.R.C. Sections 1401, 3101, and 3111.
20
I.R.C. Section 6205(a)(I) and Treas. Reg. Section 31.620
l(a)(l)(iii).
21
I.R.C. Section 6205(a)(1) and Treas. Reg. Section 31.620
l(a)(1)(i).
22
22I.R.C. Section 6157(a).
23
2.R.C. Sections 6601,6651, and 6653.

2I.R.C. Section 3402(d).

25
I.R.C. Section 6521(a).

26
I.R.C. Section 7442 and Shaw v. U.S., 331 F. 2d. 493
(CA 9th 1964).
27
2See Marvel v. U.S., 548 F. 2d. 295 (10th Cir. 1977) as
cited in JOINT COMMITTEE ON TAXATION, supra note 2, at 8.


'5-


5-








28
2North, The Employment Tax Morass, 11 CREIGHTON LAW REVIEW
797 (1978).
29
29I.R.C. Sections 6321, 6331, and 6862.

30JOINT COMMITTEE ON TAXATION, supra note 2, at 9.

311 F. Supp. 715 (W.D. Pa. 1970).

I3.R.C. Section 404(a)(5).

33
I.R.C. Section 106.
34
34I.R.C. Section 127.
35.R.C. Section 79.

36
I.R.C. Section 119.
37
37I.R.C. Sections 40, 50A and 50B.
38
38I.R.C. Section 51.

39Treas. Reg. Section 31.3121(d)-l(c)(3).

40Lexis computer data bank is simply a huge data base con-
sisting of several libraries, one of which is the Tax Library. In
this library, among other things, all reported cases since 1960
for the District Courts and all cases since 1942 for the Court of
Claims are stored in its memory. Through a determination of words
and phrases common to a selected subject, the researcher can obtain
a listing of all cases relevant to that issue. For the years not
contained in the Lexis system a manual search was performed to
identify the relevant cases.

4U.S. v. Silk, 331 U.S. 704 (1947) and Bartels v. Birm-
ingham, 332 U.S. 126 (1947).
42
4Marmoll, supra note 6.

INTERNAL REVENUE SERVICE, INTERNAL REVENUE MANUAL CCH,
ch. 4600, pp. 8463-1 thru 8463-5.

44BMDP BIOMEDICAL COMPUTER PROGRAMS P-SERIES 711 (W. Dixon
& M. Brown ed. 1979).
45
4Lachenbruch, On Expected Probabilities of Misclassifi-
cation in Discriminant Analysis, Necessary Sample Size and a Relation
with the Multiple Correlation Coefficients, 24 BIOMETRICS 823-834
(1968).
46
M. Nerlove & M. Press, Univariate and Multivariate Log-linear
and Logistic Models, RAND CORPORATION TECHNICAL REPORT R-1306-EDA/NIH
(1973).






18

47
4BMDP, supra note 44, at 517.
48
4Chow, Tests of Equality between Sets of Coefficients
in Two Linear Regressions, 28 ECONOMETRICA 591 (1960).















CHAPTER II

HISTORICAL BACKGROUND


Development of Legislative History


Social Security Act of 1935


With the devastating effects of the great depression on

unemployment, attention was keenly focused on establishing the

means by which such a calamity might be avoided in the future.1

On January 28, 1934, an Executive Order of the President of the

United States created the Committee on Economic Security.2 On

January 15, 1935, the Committee submitted its report which Congress
3
promptly used as the substance for the drafting of legislation.

The Social Security Act of 19354 came after long deliberation by

both the President and the Congress on "the evil of the burdens

that rest upon large numbers of our people because of the insecur-

ities of modern life, particularly old age and unemployment."5

The original Social Security Act included eleven "Titles." Of

principal concern for this study is Title VIII which created provi-

sions dealing with old age benefits, Title IX which created provi-

sions dealing with unemployment compensation, and Title XI which

provided the definitions to be used in the other ten Titles. The

principal solution adopted by Titles VIII and IX was to provide for

periodic payments similar to annuities to the elderly and to provide

19






20

for compensation to workers during sustained periods of unemployment.

The enactment of employment taxes was the method adopted by Congress

to fund payments to the elderly and the unemployed.

The primary purpose of this study is to determine the parties

to whom the employment taxes apply. Title VIII of the original

Social Security Act, provided for a tax on both the employee and

employer of percent of the employee's "wages." Wages were defined

in Section 811 of the Social Security Act as

all remuneration for employment, including the cash
value of all remuneration paid in any medium other than
cash; except that such term shall not include that part of
the remuneration which, after remuneration equal to $3,000
has been paid to an individual by an employer with respect
to employment during any calendar year, is paid to such
individual by such employer with respect to employment
during such calendar year. (Emphasis added)

One of the key terms used in this definition of wages was "employ-

ment" and Section 811 defined employment as "any service, of what-

ever nature, performed within the United States by an employee or

his employer."

The unemployment provisions of Title IX basically provided for

an excise tax to be levied on employees of 1 percent of the employees'

wages if the employer had eight or more employees. The determination

of whether an individual qualifies as an employee is critical in the

application of the Title IX provisions. Title XI simply defines an

employee as "an officer of a corporation." The absence of an elab-

orate definition of employee in either title seems to imply a "will-

ingness of Congress to allow the courts to determine to whom the

statute should apply."6

Regulations were issued in 1936 by the Treasury that expanded

the employee definition for both the old-age and the unemployment







tax provisions. The Regulations provide in part that

within the meaning of the Act, any person is an
employee if he is in the service of one or more employers
for compensation. An individual is in the service
of an employer if he is subject to the continuing authority
of the employer to supervise and direct the manner in
which he renders services for compensation. It is not
necessary that the employer actually direct or control
the manner in which the services are performed; it is
sufficient if the employer has the right to do so.
The right of an employer to discharge an individual is
also an important factor indicating that the individual
is an employee. Other factors indicating that an indivi-
dual is an employee are the furnishing of tools and the
furnishing of a place to work by the employer to the
individual who performs the services. In general, if an
individual is subject to the control or direction of an
employer merely as to the result to be accomplished by
the work and not as to the means and methods for accom-
plishing the result, he is an independent contractor.
An individual performing services as an independent
contractor is not, as to such services, an employee
within the meaning of the Act.

A comparison of this Regulation to the "definition" found in leading

cases and legal authorities will disclose that the Regulation is

substantially a brief statement of the common law definition of

the master and servant relationship.


Common Law Definition of Master-Servant Relationship


Maitland provides the following as a reasonable definition of

the common law:

The Common Law (called also the unwritten law, or from
its mode of development, the customary law) embraces
those rules of civil conduct which originated in the
common wisdom and experience of society, became in time
established customs, and finally received judicial
sanction and affirmance in the decisions of the courts
of last resort.0

This system of common law was originally established through the

decisions reached by the English common law courts of Exchequer,

Common Pleas, and the King's Bench as early as the year 1300.11





22

What helped promote the advancement of the common law was the

increase in the recording of court decisions and the resulting

increased availability of admissible legal precedent.12 At first,

court proceedings were circulated primarily only to judges and
13
select lawyers.3 But as the availability of court decisions

became more widespread, citation of precedent also became more
14
prevalent by lawyers and more accepted by the courts. From

this beginning, "the common law system was refined, worker over,

modified, improved, and increased as the centuries passed until

there developed an entire body of law."15

Even though the acceptance of the English system of law by

the early American colonists was not immediate, the English system

eventually did triumph as the principal source of American law.1

With England's control over the original colonies, legal matters
17
generally dealt with England and the English system. This also

resulted in the training received by the early American lawyers to
18
be heavily concentrated in English law. In addition to England's

influence on the colonies, another significant reason for the accep-

tance of the English system was the fact that English was the primary

language of Americans and the majority of early legal writing was
19
done in English.9 For whatever reasons, the English system of law

(including its common law) was the basis of the early American system

of law.

The common law is a "comprehensive body of law" providing a

"comprehensive system of remedies."20 Contained within this broad

system of law is a body of case law and legal thought dealing with

the master and servant relationship. The determination of whether






23

or not a master-servant relationship existed was of primary impor-

tance under English common law in assessing the "liability of a

master for a tort committed by a servant who was acting within the

scope of his employment.21 The question of a master-servant rela-

tionship also has become important in cases involving workmen's

compensation, contracts, and of primary importance to this study,

the area of employment taxes.

The 1936 Treasury Regulations dealing with employment taxes

were promulgated pursuant to the common law definition of the

master-servant relationship as defined by the American judicial

system. An early example of the application of the common law

definition of the master-servant relationship was in an 1889

Supreme Court case dealing with whether a corporation which had

hired a commission salesman was liable for a negligent act of that
22
same salesman.22 The salesman was furnished with a wagon which

was to be used exclusively in canvassing for sales.23 The salesman

was convicted of driving the corporation's horse and wagon care-

lessly which resulted in personal injury to Katie Rohn as she
24
crossed a local street. The Supreme Court held that the salesman

was a servant of the corporation and that such corporation was
25
responsible for Katie Rohn's injuries.25 In arriving at this

decision, the Supreme Court cited Railroad Co. v. Hanning as

authority for the definition of a master-servant relationship

that follows:

S. the relation of master and servant exists when-
ever the employer retains the right to direct the manner
in which the business shall be done, as well as the
result to be accomplished, or, in other words, "not
only what shall be done, but how it shall be done."26









In a later case, the Supreme Court used this same definition

in a situation dealing specifically with employees and independent

contractors.27 In concluding that the workers involved in this

case were independent contractors, the Supreme Court held that the

parties had the "liberty of action which excludes the idea of that

control or right to control by the employer which characterizes

the relation of employer and employee and differentiates the

employee or servant from the independent contractor."28

From this evolution of the common law comes the concept of

the master-servant relationship that is contained in the two

Supreme Court decisions discussed previously. Even though the

term common law employee is not specifically mentioned in the

1936 employment tax Treasury Regulations, the definition attri-

buted to an employee by the Treasury Department is similar to that

espoused by the two decisions discussed above. These two decisions

are merely representative of many early decisions that use this

same common law definition.29 Consequently, at the very incep-

tion of the social security provisions, the notion of a common law

employee assumed a position of importance as far as determining

to whom the provisions applied.


Social Security Amendments of 1939


The social security provisions were re-examined four years

after the enactment of the original Social Security Act and prior

to the codification of the existing tax law into the 1939 Code.

The Social Security Board recommended that the coverage of the

social security provisions be expanded.30 Arthur J. Altmeyer.






25

Chairman of the Social Security Board, in hearings before the House

Ways and Means Committee recommended that the social security pro-

visions

S. be expanded to the extent feasible to cover more
of the persons who furnish primarily personal service.
The intention of such an amendment would be to cover
persons who are for all practical purposes employees,
but whose legal status may not be that of an employee.

Using essentially the same language, the House adopted the

Board's recommendation for the old-age insurance portion of the

32
social security provisions.32 However, for some unexplained reason

the House failed to include this expanded definition for the unem-

ployment tax provisions.33 When the House proposal went to the

Senate Finance Committee, the expanded definition of employee was

deleted from the bill because it was "believed inexpedient to
1,34
change the existing law which limits coverage to employees."

Therefore, the employee definition that was codified in the 1939

Code was substantially unchanged from the original Social Security

Act of 1935. Title VIII was codified into the 1939 Code as Sec-

35
tions 1400 through 1432, referred to as the Federal Income Con-

tributions Act (FICA). The unemployment tax provisions were codi-

fied into the 1939 Code as Sections 1600 through 1611, referred

36
to as the Federal Unemployment Tax Act (FUTA). The 1939 FICA

Treasury Regulations were also substantially unchanged from the

language contained in the original 1936 employment tax Treasury

.37
Regulations.







Income Tax Withholding Provisions


In 1942, in an attempt to raise additional funds to support the

war effort, Congress enacted what was known as the Victory Tax.38

This tax was a temporary income tax that was levied in addition to

the normal tax and surtax. The Victory tax amounted to 5 percent

of "Victory Tax net income," but was limited in as much as it could

not result in a total tax greater than "90 percent of the indivi-

dual net income."39

Because of the substantial tax rates that existed during World

War II, the Treasury suggested for the first time that income taxes

be withheld at the source.40 The Treasury recommended this pro-

cedure as a method of aiding taxpayers in meeting their tax payments

with a minimum strain.

Greater equity and fairness is gained for the great body
of income-tax payers by more uniform application of the
law, and the administrative problem of collection, espe-
cially in the case of taxpayers who move from one locality
to another, will be made easier.4

The House adopted the recommendation of the Treasury and intro-

duced legislation that would require withholding of income taxes at
42
the source on dividends, bond interest, and wages. As with the

FICA and FUTA provisions, the question of whether the withholding

requirements applied to a given individual had to be answered. The

House expressed the view that the common law rules which developed

largely in the field of tort law would be "unnecessarily restric-
143
tive if they stood alone as the tests for withholding tax purposes.

The definition of employee is accordingly drawn to cover
all common law servants, and in addition such other indi-
viduals performing services as are not in reality inde-
pendent businessmen or independent practitioners of a






27

profession, whether or not they are "independent contrac-
tors" at common law. Thus under this definition an indi-
vidual may be an employee even though he is not subject
to control."4

The Senate for reasons of "simplicity" and "ease of administra-

tion" required that collection at the source of payment be limited

to salaries, wages and other forms of compensation for personal

services."45 The more surprising result, and of greater impor-

tance to the tracing of the common law definition of an employee,

was the fact that in the Conference Committee the scope of the

withholding provisions was limited to those individuals coming

under the common law definition of an employee.46

A year later, the definition of employee that became incor-

porated in the Victory Tax provisions was extended to the regular

income tax area by the Current Tax Payment Act of 1943.47 By so

doing, Congress expanded the application of the common law employee

definition from the social security tax area to the income tax

area while at the same time resisting pressure to expand the def-

inition of employee beyond that encompassed by the common law rules.

Therefore, it was not until 1947 and the introduction of the eco-

nomic reality test that the common law definition of employee was

again challenged.


Judicial Interpretation


Economic Reality Test. In 1947, the Supreme Court considered

three federal employment tax cases which gave rise to what has been

termed the "economic reality test.48 Of these three cases, the

Supreme Court decided both the Silk and Greyvan Lines cases together

on June 16, 1947.








The Albert Silk Coal Company was in the business of selling

coal at retail in the city of Topeka, Kansas. The question being

considered by the Supreme Court in this case was whether workmen

engaged in unloading railroad cars and truck drivers who made

retail deliveries of the coal were employees for purposes of the

employment tax provisions. The unloaders were paid an agreed

price per ton to unload coal from the railroad cars. The men

provided their own tools, worked when they desired, and were free

to work for other employers.49 When an unloader came to the yard

he was assigned a car to unload and told where it was to be unloaded.

Some of the unloaders were quite regular workers while others were

"floaters"and came to the yards only intermittently.50

Since Silk owned no trucks of his own, he contracted with

individuals who owned their own trucks to deliver the coal at an

agreed upon price per ton. The truckers were able to come and go

as they pleased, were able to haul for others if they desired,

paid all their own operating expenses, furnished their own helpers

when needed in delivering the coal, and were able to refuse to make
51
certain deliveries without suffering any penalties. The truckers

were not instructed in how to perform their services, but simply

were given a delivery ticket and told whether the purchase was cash
52
or charge.52 No record was kept of the truckdrivers' time and they

settled up with Silk Coal Company by the trip, by the day, or by

the week according to the truckdriver's wishes.

Greyvan Lines, Inc.,was a household-furniture trucker operating

in thirty-eight states and part of Canada.54 Similar to Silk, the








question that was decided by the Supreme Court was whether the

truckmen who hauled the household furniture were employees under

the employment tax provisions. These truckers were required to

furnish their own trucks, haul exclusively for Greyvan Lines, pro-

vide any assistants needed to deliver the furniture, pay all oper-

ating expenses, carry the specified insurance coverage, drive their

own trucks, and follow any regulations and instructions issued by

Greyvan Lines.55 These instructions covered directions as to where

and when to load and deliver the furniture. Compensation was deter-

mined on a commission basis with a bonus for successful completion

of a job.

The Supreme Court held that the unloaders in the Silk case

were employees.56 The Court explained its decision as follows:

(1) the tools the unloaders provided were only picks and shovels,

(2) they had no opportunity for gain or loss except through their

own manual labor, (3) the fact that many did not work regularly

was insignificant, and (4) since the unloaders did work in the

course of the employer's regular trade or business, they should

come under the employment tax provisions.5

On the other hand, the Supreme Court held that the truckers

in both cases were independent contractors. Even though on the

surface, the description of the relationship of the truckers to

their respective employers seemed quite different, the Supreme

Court's opinion did not distinguish between the two cases with

respect to the trucker's verdict. The Court indicated that even

though the Greyvan Line truckers hauled just for one firm. this





30

distinction was not controlling in the rendering of its decision.59

The Court held that the truckers were small businessmen who owned

their own trucks and that "it is the total situation, including

the risk undertaken, the control exercised, the opportunity for

profit, from sound management, that marks these driver-owners as

independent contractors.

The Silk and Greyvan decisions themselves are not the most

significant factors of these two cases. Despite the Congressional

mandate that the common law rules were to determine if the employ-

ment relationship existed in other than the statutorily defined

cases,61 the Supreme Court in Silk opted for an expanded defini-

tion of employee.

As the federal social security legislation is an attack
on recognized evils in our national economy, a con-
stricted interpretation of the phrasing by the courts
would not comport with its purpose. Here the
District Courts and the Circuit Courts of Appeals deter-
mined the cages largely if not indeed exclusively by
applying the so-called "common law control" test as
the criterion. This was clearly wrong in view of the
Court's present ruling.62 (Emphasis added)

The Supreme Court in Silk also listed the following factors as

important in determining whether employment status actually is

present: (1) degree of control, (2) opportunities for profit or

loss, (3) investment in facilities, (4) permanency of relation,

and (5) skill required.63 With respect to these factors, the

Court states that "no one factoriess controlling nor is the list

complete."64 The Supreme Court's opinion definitely speaks of

the degree of control as simply one of several factors to be

applied and not as the overall test.









The Supreme Court upheld both the Silk and Greyvan Lines

opinions a week later in another employment status case, Bartels

v. Birmingham.65 This case deals with orchestras hired to play

for limited engagements; usually one-night stands. The question

being decided by the Supreme Court in this case is whether the

band leaders and other members of the band were the employees of

the dance establishments.

These bands are "built around a leader whose name, and

distinctive style in the presentation and rendition of dance

music, is intended to give each band a marked individuality."66

The leader hires and fires the musicians, fixes their salaries,

tells them what and how to play, and pays all of their transpor-
67
station and operating expenses. It is also the band leader who

contracts with the dance establishments to play at a specified
68
price.

The Supreme Court held that the band leaders were the employers

of the individual musicians and in turn the band leaders were

independent contractors with respect to the dance establishments.69

In explaining its opinion, the Supreme Court stated that a band

leader organizes and trains the band, selects the members, and

bears the risk of losses that might occur.70 The relationship

between the band leader and his orchestra is relatively permanent,

while the relationship between the band and the dance establish-

ments is only temporary.7

In issuing the Bartels opinion, the Supreme Court cited the

Silks case and restated the notion that the common law employee

definition for employment tax purposes was "not to be determined









solely by the idea of control which an alleged employer may or

could exercise over the details of the service rendered to his
,,72
business by the worker or workers.72 The Court goes on to suggest

that determination of an employer-employee relationship is a matter

of "economic reality."73 Again, the Supreme Court was emphasizing

that it is the total situation that controls, not simply the

the common law control test.

These three 1947 cases were not the first instancesin which

the economic reality test was espoused by the Supreme Court.

Before the test was introduced into the employment tax area, the

Supreme Court had already applied the economic reality doctrine
74
in National Labor Relations Board v. Hearst Publications. In

Hearst Publications, the Supreme Court held that newsboys were

employees under the National Labor Relations Act.7 In interpreting

the Act, the court stated that "Congress had in mind a wider field

than the narrow technical legal relation of master and servant"

and because of the broad language in the Act's definitions, there

is "no doubt that its applicability is to be determined broadly,

in doubtful situations, by underlying economic facts rather than

technically and exclusively by previously established legal classi-

fications."76 After the economic reality test was introduced in

the labor relations area, the Supreme Court held in the Silk case

that "application of the social security legislation should follow

the same rules that we applied to the National Labor Relations
",77
Act in the Hearst case. This then completes a brief history

of the advent of the economic reality test and its integration

into the federal employment tax area.









Status Quo Amendment. With the success of the Supreme Court

in altering the common law employee definition first in the labor

relations area and then in the federal employment tax area, the

Treasury Department quickly attempted to alter its employee defi-

nition by issuing Proposed Regulations in 1947.78 These Regula-

tions attempted to substitute the economic reality test for the

common law control test in determining employment status for FICA

and FUTA purposes.79 These Regulations would have reduced the

"right to control the performance of services by an individual"

to simply one of several factors that should be examined in deter-

mining employment status instead of the overall test as envisioned

by Congress in the original enactment of the social security pro-

visions.80

Congress viewed the Proposed Regulations as an attempt by the

Treasury Department to usurp the authority of Congress in making
81
law.

The issue involved in the proposed regulations is whether
the scope of social-security coverage should be deter-
mined by Congress or by other branches of the Government
accordingly, under the proposed regulations, the
question of coverage will be determined, not by Congress,
but by the Social Security Agency, the Treasury, and the
courts.82

By joint resolution, Congress responded to the Proposed Regulations

by passing legislation that would maintain the "status quo" of the

current statutes and regulations and reaffirm the legislative intent

of the original Social Security Act of 1935. The intent was that

the usual common law rules, realistically applied, would continue

to be used to determine employment status for social security pur-

poses.83 The accompanying Senate Report stated emphatically that









the "end point determination" is whether there is an absence or

existence of control and that whatever factors are pertinent to

this determination should be used.84 But the final determination

is whether or not control actually exists.

To further emphasize their desire to have the common law

control test be the primary test of employment status, Congress

added the following language to the FICA and FUTA provisions:

such term does not include (1) any individual who,
under the usual common law rules applicable in determining
the employer-employee relationship, has the status of an
independent contractor or (2) any individual (except an
officer of a corporation) who is not an employee under
such common law rules.

This is the first time that the common law definition of employee

was officially included in the statute. Previously, support for

its application had been obtained either from legislative history

of the social security provisions or the employment tax Treasury

Regulations.


Social Security Amendments of 1950


Congressional support for the Status Quo Amendment was strong

enough that it was passed over the veto of President Truman.86

Representative Gearhart explained the actions of Congress as

follows:

If this Congress had not interfered, tens of thousands
of people in America who never dreamed they were employed
by anybody and never for one moment thought they were
covered by social security or subject to payroll taxes
would have found that they had been swept into the social
security system by bureaucratic ukase. In other words,
they would suddenly have found that they had more employers
than a dog has fleas. So, to end this confusion, Congress
acted promptly and, after thorough-going debate, by a








vote of nearly 7 to 1, proceeded by legislation to put
the matter in order once again by restoring the ancient
doctrine of the common law defining the relation of
master and servant, employer and employee.87

Despite this strong Congressional reaction to the Treasury's

Proposed Regulations, in the very next Congressional session the

Social Security Administration recommended the repeal of the Status
88
Quo Amendment and enactment of the economic reality test.8 The

House adopted the proposal and passed H.R. 6000 which would have

incorporated the economic reality test as a general definition of
89
an employee for social security purposes.8 Also the bill stat-

utorily determined that the following seven occupational groups

would be classified as employees for the withholding and payment

of social security taxes: (1) an outside salesman in the manufac-

turing or wholesale trade, (2) a full-time life insurance salesman,

(3) a driver-lessee or a taxicab, (4) a home worker, (5) a contract

logger, (6) a mining lessee or licensee, and (7) a house-to-house

salesman.90

The Senate did not accept H.R. 6000 as it was passed by the

House. Instead, the Senate deleted the economic reality test,

reaffirmed the common law definition of employee, and reduced the

number of occupational categories to certain agent or commission

drivers and full-time insurance salesmen.9 Therefore, the economic

reality test was again specifically rejected by Congress for use in

resolving the employment status question. For FICA purposes, the
92
same definition of employee that was enacted in 19502 has not

been substantially changed to date. In the next section of this

chapter, this definition is reproduced in its full text as Section

3121(d) of the Internal Revenue Code of 1954.






36

Before analyzing the current legislative status of the common

law definition of employee, it is appropriate to consider another

relevant piece of legislation that was introduced by the Social

Security Amendments of 1950. To this point, this entire histor-

ical background discussion has been concerned with the FICA and

FUTA provisions as they relate to the overall social security package

adopted by Congress. No mention has been made of the status of

the self-employed worker who did not come under the social security

provisions discussed thus far. Another significant aspect of the

1950 amendments was the extension of the social security provisions

to self-employed individuals.93

The self-employment tax provisions were codified in the 1939

Code as Sections 480-482. Section 482(b) states that subchapter

E (which contains Section 480-482)may be cited as the Self Employ-

ment Contributions Act (SECA). The original SECA provisions pro-

vided for a tax on "self-employment income."95 Self-employment

income was defined as "net earnings from self-employment derived

by an individual" except (1) those net earnings from self-employment

that are less than $400 and (2) those net earnings from self-

employment in excess of $3,600 (which first had to be reduced by

any wages received that had been subjected to the FICA taxes under

Section 1426(b) of the 1939 Code).96

The term "nct earnings from self-employment" were defined as

the "gross income derived by an individual from any trade or busi-

ness carried on by such individual," less certain deductions.97

The term trade or business was defined to have the same meaning

as it did in Section 23 of the 1939 Code which dealt with the






37

ordinary and necessary expenses incurred in connection with a trade

or business that could be deducted in arriving at net income.98

Section 481(c) contained an except clause that applied to the

Section 23 definition of a trade or business which resulted in

the exempting of certain activities from the self-employment tax.

The most important 481(c) exception for this study was the "per-

"99
formance of services by an individual as an employee.99 Section

481(d) referred to the FICA provisions to define the term employee.

The FICA definition of employee, as has been discussed in considerable

detail, means a common law employee. With this link to the FICA

provisions, Congress had effectively extended the coverage of the

social security provisions beyond merely employees and provided the

means by which self-employed persons would now be subject to employ-

ment taxes.

Sections 1401-1403 are the 1954 Code counterparts of Sections

481-483 of the 1939 Code. In substance, the self-employment tax

provisions have been altered slightly in the past 30 years, but

essentially their import is the same.00 The most salient aspect

of the SECA provisions for this study is that because of their

construction, they provide a clear distinction between how an

employee and and independent contractor will be treated for employ-
101
ment tax purposes. It is this distinction that has contributed

to the increased importance of being able to determine when a

common law employer-employee relationship actually exists.








Current Legislative Status


For FICA purposes, the Internal Revenue Code of 1954 defines

an employee substantially as it was defined after the 1950 amend-

ments. Section 3121(d) defines an employee for FICA purposes as:

(1) any officer of a corporation; or
(2) any individual who, under the usual common law
rules applicable in determining the employer-employee rela-
tionship, has the status of an employee; or
(3) any individual (other than an individual who is an
employee under paragraph (I) or (2)) who performs services for
renumeration for any person--
(A) as an agent-driver or commission-driver engaged
in distributing meat products, vegetable products, bakery
products, beverages (other than milk), or laundry or dry-
cleaning services, for his principal;
(B) as a full-time life insurance salesman;
(C) as a home worker performing work, according to
specifications furnished by the person for whom the services
are performed, on materials or goods furnished by such a person
which are required to be returned to such person or a person
designated by him; or
(D) as a traveling or city salesman, other than as
an agent-driver, engaged upon a full-time basis in the
solcitation on behalf of, and the transmission to, his
principal (except for side-line sales activities on
behalf of some other person) of orders from wholesalers,
retailers, contractors, or operators of hotels, restaurants,
or other similar establishments for merchandise for
resale or supplies for use in their business operations;
if the contract of service contemplated that substantially all
of such services are to be performed personally by such in-
dividual; except that an individual shall.not be included in
the term "employee" under the provisions of this paragraph if
such individual has a substantial investment in facilities
used in connection with the performance of such services
(other than in facilities for transportation), or if the
services are in the nature of a single transaction not part of
a continuing relationship with the person for whom the services
are performed. (Emphasis added)

It is Section 3121(d)(2) dealing with common law employees

that is of principal interest in this study. A common law employee

is one of the classes of employees that is subject to the FICA

withholding and payment requirements. However, all six categories

of employees defined in paragraph (d) are subject to the if clause








that follows the Section 3121(d)(3)(D) definition of a traveling or

city salesman. Ignoring the language not pertinent to this study,

paragraph (d) could be paraphrased as follows: an individual is

treated as an employee- for the FICA provisions if he is considered

an employee under the usual common law rules applicable in determining

the employer-employee relationship if the contract of service

contemplates that all the services would be performed personally

by the common law employee in question; except if such common law

employee has a substantial investment in facilities used in con-

nection with the performance of such services or if the services

are in the nature of a single transaction and not part of a con-

tinuing relationship with the person for whom the services are

performed.

This interpretation of the statute would mean that a common

law employee is subject to the FICA withholding and payment re-

quirements only if he satisfies the "if clause" discussed above and

also does not come within either of the situations described in the

"except clause." Even though personal performance of the services,

substantial investment, and a continuing relationship are all

factors that are considered in applying the usual common law

definition of employee, this statutory construction places these

three factors as additional requirements that must be satisfied

before an adjudged common law employee can be subjected to the FICA

provisions.

In the 139 cases included in this study, the argument never

was presented to the courts that a common law employee should be

excepted from the FICA withholding and payment requirements due to








the statutory exceptions contained in paragraph (d) of Section

3121. Marmoll concludes that "there has been a tendency in the

authorities to de-emphasize--primarily through silence--the statu-

tory exceptions set forth in Section 3121(d)."I02 Additional dis-

cussion is devoted to these statutory exceptions in Chapter III

when the factors are defined that are used in building the mathe-

matical models developed in this study. The additional discussion

in Chapter III is included in the definitions of factor 3 (right

to delegate) and factor 4 (continuing relationship).

Treasury Regulation Section 31.3121(d)-l(c) further elaborates

on the term common law employee for FICA purposes.

(c) Common law employees. (I) Every individual is an
employee if under the usual common law rules the relationship
between him and the person for whom he performs services
is the legal relationship of employer and employee.
(2) Generally such relationship exists when the person
for whom services are performed has the right to control and
direct the individual who performs the services, not only as
to the result to be accomplished by the work but also as to the
details and means by which that result is accomplished. That
is, an employee is subject to the will and control of the
employer not only as to what shall be done but how it shall be
done. In this connection, it is not necessary that the employer
actually direct or control the manner in which the services
are performed; it is sufficient if he has the right to do so.
The right to discharge is also an important factor indicating
that the person possessing that right is an employer. Other
factors characteristic of an employer, but not necessarily
present in every case, are the furnishing of tools and the
furnishing of a place to work, to the individual who performs
the services. In general, if an individual is subject to the
control or direction of another merely as to the result to be
accomplished by the work and not as to the means and methods
for accomplishing the result, he is an independent contractor.
An individual performing services as an independent contractor
is not as to such services an employee under the usual common
law rules. Individuals such as physicians, lawyers, dentists,
veterinarians, construction contractors, public stenographers,
and auctioneers, engaged in the pursuit of an independent
trade, business, or profession, in which they offer their
services to the public, are independent contractors and not
employees.








(3) Whether the relationship of employer and employee
exists under the usual common law rules will in doubtful cases
be determined upon an examination of the particular facts of
each case. (Emphasis added)

The FICA Regulation pertaining to the common law definition of

an employee states that an individual is an employee for FICA

purposes if a legal relationship of employer and employee exists

under the usual common law rules. The legal relationship of employer

and employee generally exists when the following factors are pre-

sent: (1) the right to control and direct the individual who

performs the services, (2) the right to discharge, and (3) the

furnishing of tools and a place to work by the employer.

Treasury Regulation Section 31.3121(d)-l(c) further states

that an individual performing services as an independent contractor

is not considered a common law employee. Instead of defining

factors that are characteristic of an independent contractor, the

Regulation simply gives examples of several professions that are

in the pursuit of an independent trade and who offer their services

to the public; such as, lawyers, physicians, dentists, etc. Finally,

the Regulation states that determining an employer-employee rela-

tionship under the usual common law rules is a question of fact

and must be determined for each case individually.

While Section 3121(d) and its corresponding Treasury Regulation

provide some elaboration on the definition of a common law employee

for FICA purposes, the incompleteness of the statutory authority is

evident by the substantial amount of litigation in this area.103

It has been left largely to the courts to determine when a common

law employer-employee relationship exists within the broad guidelines

contained in Section 3121(d) and Treasury Regulation Section 31.3121

(d)-l(c).








For FUTA purposes, Section 3306(i) states that "the term

employee has the meaning assigned to it by Section 3121(d), except

that subparagraphs (B) and (C) of paragraph (3) shall not apply."

This means that with the exception of certain types of home workers

and full-time insurance sales persons, the FICA and FUTA definition

of employees are identical.

For the withholding of income taxes, Section 3401(c) simply

provides:

For purposes of this chapter, the term "employee" includes
an officer, employee, or elected official of the United
States, a State, or any political subdivision thereof,
or the District of Columbia, or any agency or instru-
mentality of any one or more of the foregoing. The term
"employee" also includes an officer of a corporation.

The Treasury Regulations for Section 3306(i) and 3401(c) are

substantially the same as the FICA Regulations already reproduced

above and consequently need not be included. So even though

Section 3401(c) does not mention common law employee specifically,

the Section 3401 regulation that defines employee is synonomous

with the other regulations that do define a common law employee

for FICA purposes.

While "employee" has not been defined for income tax
withholding purposes as any particular type of "employee,"
the authorities have merely assumed, sub silentio in
most cases, that the word "employee" as used in the income
tax withholding statutes means "common law employee".
Because of this historical treatment of the income tax
withholding provisions and, further, because the statu-
tory scheme would seem to support such an inference, it
may be inferred that an "employee" for income tax with-
holding purposes means a "common law employ e", as that
phrase is defined in the FICA regulations.'0

Consequently, under the current employment tax provisions,

when a taxpayer, an administrator, or a court is trying to determine

if an employer-employee relationship actually exists, unless








specifically exempted by the statutes, such a determination must

be decided under the common law rules. In litigation, the court

generally consolidates the FICA, FUTA, and income tax withholding

issues into one question and that is whether or not the worker is

an employee of the plaintiff under the common law rules (Treas.

Reg. Section 31.3121(d)-(1)(c)(3)).


Revenue Act of 1978


With the increased enforcement of the employment tax laws

by the IRS, many taxpayers have complained that proposed reclas-

sifications of their workers represent a change in the IRS's
105
position.05 Some taxpayers have relied on private letter rulings,

technical advice memoranda, and results of prior audits for deter-

mining employment status but nevertheless have found themselves the

subjects of reclassifications by the Service.06 Congress found

it appropriate to provide interim relief to taxpayers involved in

employment tax status conflicts with the IRS until it had adequate
107
time to resolve the issue. As a result, Section 530 of the

Revenue Act of 1978 was enacted to provide relief for the 1979
108
tax year. Congress was unable to resolve the employment tax

status issue during 1979, so H.R. 5505 extended the relief pro-
109
visions of Section 530 until the end of 1980.1

In general, Section 530 accomplishes two primary objectives.

First, Section 530(a)(1) provides that if the taxpayer did not

treat an individual as an employee for any period ending before

January 1, 1980, and all necessary Federal tax returns were cor-

rectly filed in 1979, then the individual will not be treated as








an employee unless there is no reasonable basis for such treatment.

Section 530(a)(2) stipulates that a reasonable basis exists for not

treating a worker as an employee regardless of the circumstances

if the taxpayer has relied on any of the following:

(1) judicial precedent, published rulings, technical
advice with respect to the taxpayer, or a letter
ruling to the taxpayer,

(2) a past Internal Revenue Service audit of the tax-
payer in which there was no assessment attributable
to the treatment (for employment tax purposes)
of the individuals holding positions substantially
similar to the position held by this individual, or

(3) long-standing the recognized practice of a signifi-
cant segment of the industry in which such individual
was engaged.

Second, Section 530(b) prohibits the Treasury Department from

publishing any Regulation or Revenue Ruling after the enactment of

the Revenue Act of 1978 (November 6, 1978) dealing with the employ-

ment status of any individual for purposes of employment taxes.

With the interim relief provided by Section 530 (and extended

by H.R. 5505), Congress has had time to consider many suggestions

as to the remedies for the complex issues that beset the employment

tax area. The next section of this study will briefly introduce

the principal recommendations that have been presented to Congress.


Current Proposed Solutions


General Accounting Office


The General Accounting Office (GAO), at the request of the Joint

Committee on Taxation, conducted a study of the problems with ad-

ministration of the social security tax and the definition of a

common law employee.0 In reviewing the problem of classification







of workers as employees or independent contractors, the GAO con-

cluded that the principal problem involved uncertainty in the

application of the common law control rules. Numerous examples

were discovered where the IRS is inconsistently applying the common

law rules in different geographical areas.ll The GAO report

states that the "major cause of the employee/self-employed contro-

versy involves those cases in which the reclassified workers operate

a business separate from the one that the IRS considers to be the

employer.112

The primary recommendation of the GAO is that a clear cut test

needs to be devised to introduce a greater degree of certainty

into the employment tax area for both the taxpayer and the IRS.

The approach taken by the GAO is to use four basic tests in deter-

mining whether a true separate business entity actually exists.

If a worker can satisfy the following four tests, the IRS will

be prohibited from reclassifying the worker as an employee. The

individual must:

(1) Have a principal place of business other than that
furnished by the person or persons for whom he or
she performs or furnishes services;

(2) Maintain a separate set of books and records that
reflect all items of income and expense of his or
her trade or business;

(3) Bear the risk of suffering a loss and the oppor-
tunity of making a profit; and

(4) Hold himself or herself out in his or her own name
as self-employed and/or makes his or her services
generally available to the public.

The report argues that the four tests are "more precise, easier to

understand, and can be applied more accurately and consistently

than the common law rules."3 However, failure to meet all four








criteria will not result in automatic employee status. Instead,

if the worker satisfies three of the four criteria this will still

be a possible indication of self-employment and such a situation

will revert to the regular common law rules. If only two of the

four criteria are satisfied, then the worker will be presumed to

be an employee. The GAO concludes that the number of cases that

will require litigation to determine common law employee status will

be greatly reduced by this type of an approach.

In addition to the tests, the GAO proposal prohibits retro-

active determinations (in absence of fraud) if a business (1) obtains

annually from the individual whom the business classifies as self-

employed a signed certificate stating that the individual meets

all the separate business entity tests, and (2) annually provides

the IRS with the name and employer identification number or social

security number of each such certificate signer. To increase the

effectiveness of these criteria the report also proposes that the

taxpayer should sign the certificate under penalty of perjury.

Another problem perceived in the GAO study is the problem of

double collection of taxes on the same income as mentioned in the

introduction of this study. The GAO report proposes that the IRS

amend Section 6521 to allow an offset against the employees' por-

tion of the FICA tax by the appropriate portion of the SECA tax

already paid by the worker.114


H.R. 3245; Safe Harbor Remedy


A similar approach to the GAO proposal was introduced by
115
Representative Gephardt on March 27, 1979, as H.R. 3245. The

overriding purpose of this bill is to preserve the independent





47

contractor status of workers in industries that have traditionally

enjoyed such status and are currently being reclassified by the
116
IRS. This bill, like the GAO proposal, does not replace the

common law rules. Those workers who can not satisfy the five

"safe harbor" criteria for independent contractor status will

then come under the common law rules.

The five safe harbor rules are similar to several of those

proposed by the GAO.

(1) An independent contractor must control the hours he
works. This rule implies a two pronged test. The
worker must control not only the aggregate number of
hours actually worked, but he also must control
substantially all the scheduling of the hours worked.

(2) Independent contractors provide their own place of
business. Renting from the person for whom the
worker provides services would not be prohibited
as long as the rent is reasonable.

(3) An independent contractor experiences investment
or income fluctuations. His compensation is neither
fixed nor guaranteed. Substantial investments
in assets introduce the risk of business failure
and again this is a characteristic of self-employment.

(4) The worker, as an independent contractor, must
perform services pursuant to a written contract
that spell out the employment status and the tax
consequences that result. This test is fairly
mechanical, but it alerts both parties to their
relationship for both business and tax purposes.

(5) The person for whom the services are provided must
file the required tax returns. The worker must re-
ceive the necessary information to file his tax
returns and the IRS must receive the necessary
information to permit collection of the taxes due.

The satisfaction of the five safe-harbor criteria by a worker

would result in automatic independent contractor status. If the

criteria could not be satisfied then the worker would fall under the

usual common law rules. This notion is corroborated by Rep. Gcphardt







in the following summary of the purpose and application of H.R.

3245.

This bill creates a statutory scheme that promises to
end the confusion and uncertainty created by IRS enforce-
ment practices that have characterized employment tax
status determinations for nearly a decade. It is re-
cognized, however, that the tests cannot govern all
cases. Therefore, if any one of the tests is not met,
the common law test will be applied to determine the
worker's status as an independent contractor or an
employee. 117


Department of the Treasury


The Treasury's principal arguments against the GAO and

Gephardt proposals are that

(1) the taxpayer will be able to manipulate the tests
simply by changing the form of the business relation-
ship without changing the substance,

(2) the tests are unclear and difficult to administer,

(3) the proposals would continue the need for the use
of the common law rules in many situations, and

(4) the Service does not feel that the proposals will
deter the large number of independent contractors
who willfully fail to comply with the employment tax
provisions. 118

However, the solutions offered by the Treasury take an entirely

different approach to the situation because they view the primary
119
problem as one of noncompliance.9 This assessment of the problem

is largely due to a recent study conducted by the IRS where they

concluded that 47 percent of the workers treated as independent

contractors failed to declare any of their compensation for income
120
tax purposes and 62 percent paid none of the social security tax due.

Since the Treasury considers noncompliance of independent

contractors as the most serious problem, their solutions have nothing










to do with the definitional aspects of the common law rules.

Instead, the Treasury supports a flat 10 percent withholding on

compensation paid to independent contractors. This aids in the

collection process for the IRS and also provides more information

with which they can monitor independent contractors and the taxes

they should be paying. In addition to the 10 percent withholding,

the Treasury Department has proposed that instead of a one dollar

penalty per information return not filed (with a maximum penalty

of $1,00), that the penalty should be increased to 5 percent of

payments not reported (with a $50 minimum penalty).121

An interesting note to the Treasury proposal is that the

taxpayer who is paying a worker must still determine if he is an

employee and subject to the graduated withholding requirements

or if he is an independent contractor and subject to the flat

10 percent withholding requirements. The issue of noncompliance

is obviously an important concern, but the fundamental problem that

still remains is the need to differentiate between an independent

contractor and an employee.


H.R. 5460


The fact that both noncompliance and the lack of certainty

with the employee definition are important problems is manifest

in the compromise bill, H.R. 5460, that was submitted to the
122
House Ways and Means Committee shortly before the end of 1979.1

Briefly, the compromise bill incorporated the five safe harbor

criteria that are contained in H.R. 3245 and the flat 10 percent








withholding from compensation paid to independent contractors

as proposed by the IRS. The only exceptions to the 10 percent

withholding requirements are that no withholding need occur from

a person who (1) performs similar services for five or more

unrelated service-recipients; or (2) certifies that withholding

would be excessive; or (3) makes direct sales for a supplier who

provides no renumeration other than a volume sales bonus.123

Therefore, unless a person qualifies under one of the above excep-

tions, regardless of a person's employment status some sort of

withholding will be applied to any compensation received. Until

this or similar legislation is enacted, the relief provisions of

Section 530 of the Revenue Act of 1978 will continue to be in force

awaiting Congressional enactment of permanent legislation in the

employment tax area.


Notes



IJ. HUGHES, THE FEDERAL SOCIAL SECURITY TAX 1 (1941).

Id.

3R. COMPTON, THE SOCIAL SECURITY PAYROLL TAXES 4 (1940).

4Social Security Act of 1935, ch. 531, 49 Stat. 620 (1935).

5U.S. v. Silk, 67 S.Ct. 1463 (1947).

R. COMPTON, supra note 3, at 129.

7Who Are Employees, Title 26 C.F.R. Sec 401.3 (1939) and
Employed Individuals, Title C.F.R. Sec. 400.205 (1939).

Id.

9R. COMPTON, supra note 3, at 130.

1F. MAITLAND & F. MONTAGUE, A SKETCH OF ENGLISH LEGAL HISTORY
213 (1915).








11
C. KINNANE, ANGLO-AMERICAN LAW 284 (2nd edition, 1952).

I2d. at 281.

131d.
Id.


15d.


1Id. at 471.
1 Id.

1581d

19Id.
16







20Id. at 281.

2F. POLLOCK & F. MAITLAND, THE HISTORY OF ENGLISH LAW, vol.2,
p. 528 (1911).

22Singer Manufacturing Co. v. Rohn, 132 U.S. 518 (1899).
2317






Id. at 521.

24Id at 518.

2d. at 524.


2615 Wall. 649, 656 as cited in Singer Manufacturing Co.,
supra note 22, at 523.

27Metcalf & Eddy v. Mitchell, 269 U.S. 514 (1925).
19





















2Id. at 521.
21









See fn. 18 in J. HUGHES, supra note 1, at 48.

30Hearings Relative to the Social Security Act Amendments of
1939 Before the Committee on Ways and Means, 76th Cong., 1st Sess.
8, 11 (1939) (statement of Arthur J. Altmeyer).



2SH.R. NO. 728, 76th Cong., 1st Sess. 61-62 (1939).

33Id. at 74.

34S.REP. NO. 734, 76th Cong., st Sess. 75 (1939).
S.REP. NO. 734, 76th Cong., Ist Sess. 75 (1939).









3Sections 1400-1432 (Subchapter A), Chapter 9, Internal Revenue
Code of 1939.
36
3Sections 1600-1611 (Subchapter C), Chapter 9, Internal Revenue
Code of 1939.

37Smith, Independent Contractor or Employee?--That is the Ques-
tion, 33 NEW YORK UNIVERSITY INSTITUTE ON FEDERAL TAXATION 581 (1975).

38S.REP. NO. 1631, 77th Cong., 2d Sess. (1942).

I3d. at 8.

40Hearings on Revenue Revision of 1942 Before the Committee on
Ways and Means, 77th Cong., 2d Sess. 81 (1942) (statement of Randolph
Paul).

41H.R. NO. 2333, 77th Cong., 2d Sess. 14-15 (1942).
42
Id. at 125.
43
43Id. at 127.

44Id. at 127-128.

45S.REP. NO. 1631, supra note II, at 165.

46H.R. NO. 2586, 77th Cong., 2d Sess. 56 (1942).

47H.R. NO 510, 78th Cong., Ist Sess. 30 (1943); Current Tax
Payment Act of 1943, ch. 120, par. 2, 57 Stat. 126 (1943).

4U.S. v. Silk, 67 S.Ct. 1463 (1947); Harrison v. Greyvan
Lines, Inc., 67 S.Ct. 1463 (1947); and Bartels v. Birmingham,
67 S.Ct. 1547 (1947).

49U.S. v. Silk, supra note 5, at 1465.

50d.

Id.

52Id.

53Id.

54Id.

55
Id. at 1466.

56Id. at 1470


57d.
Id.








58
5Id. at 1471.

59
59Id.

60
6Id.

61
See previous discussion dealing with the Social Security
Act Amendments of 1939 and the introduction to the income tax
withholding provisions.

62U.S. v. Silk, supra note 5, at 1467,1472.

63d. at 1469.

64
Id.

667 S.Ct. 1547 (1947).

66Id. at 1548.

671d.
Id.
681d
Id.
69
Id. at 1551.

70
Id.

71Id.

72d. at 1550.

73
73Id.

74322 U.S. Ill (1944).

75Id.

76Id. at 124,129.

7U.S. v. Silk, supra note 5, at 1468.
78
712 Fed. Reg. 7966-69 (1947) as cited in Smith, supra note
37, at 584.

79Id.

801d.
Id.

8H.R. NO. 1319, 80th Cong., 2d Sess. 3 (1948).

82Id. at 304.








83Act of June 1, 1948, Pub. L. No. 80-642, ch. 469, 62 Stat.
468.

84S.REP. NO. 1255, 80th Cong., 2d. Sess. 11 (1948).

85Act of June 1, 1948, supra note 83, at 438.

86North, The Employment Tax Morass, 11 CREIGHTON LAW REVIEW
784 (1978).

87See fn. 2 in Bonney Motor Express v. U.S., 10 AFTR 2d 5233
(E.D. Va. 1962).

88Hearings on H.R. 2893 Before the Committee on Ways and
Means, 81st Cong., Ist Sess. 1087-1088 (1949).

8H.R. NO. 1300, 81st Cong., Ist Sess. 14-15 (1949).

90
Id. at 81.

91S.REP. NO. 1669. 81st Cong., 2d Sess. 95-97 (1950).

92Social Security Act Amendments of 1950, Pub. L. No. 81-734,
ch. 809, Sections 104(a) and 205(a), 64 Stat. 500,536 (1950).

93Id., Sections 211-213, at 502.

9426 U.S.C. Sections 480-482, pp. 3556-3558 (1952).

95Id., Section 480, at 3557.

961d., Section 481(b), at 3557-3558.

97Id., Section 481(c), at 3558.

981d., Section 23(a)(1)(A), at 3295.

99Id., Section 481(c)(3), at 3558.

100Compare the language of Sections 1401-1403 of the 1954
Code with Sections 480-482 of the 1939 Code.
1010utside of the employment tax area, the notion of employee
can become equally confusing. For example, the benefits of deferred
taxation of income were extended to self-employed individuals
through the Keogh or H.R. 10 plans. For purposes of qualified
pension, profit-sharing, or stock bonus plans, a shareholder-
employee of a Subchapter-S corporation is interestingly not treated
as an employee, but rather is subject to the limitations imposed
on the self-employed retirement plans Section 401(a)(17) On
the other hand, partners who are treated as self-employed persons
for purposes of retirement benefits may be treated as employees
for purposes of Section 119 (exclusion from gross income for meals







and lodging provided to an employee for the convenience of the
employer) according to Armstrong v. Phinney, 394 F. 2d 661
(5th Cir. 1968).

02Marmoll, 'Employee Defined', TAX MNGM'T (BNA) No. 391,
p. A-3 (1979).
103
0For an example of the numerous cases litigated in this
area see Marmoll, supra note 102, at C-6 through C-10.
104
0Marmoll, supra note 102, at A-5.

5S.REP. NO. 1263, 95th Cong., 2d Sess. 210 (1978).

JOINT COMMITTEE ON TAXATION, ISSUES IN THE CLASSIFICATION
OF INDIVIDUALS AS EMPLOYEES OR INDEPENDENT CONTRACTORS 8 (1979).

S.REP. NO. 1263, supra note 105.

0Revenue Act of 1978, Pub. L. No. 95-600, Sec. 530, 92
Stat. 2885 (1978).
109
0Act of December 29, 1979, Pub. L. No. 96-167, 93 Stat.
1278.

10GENERAL ACCOUNTING OFFICE, TAX TREATMENT OF EMPLOYEES AND
SELF-EMPLOYED PERSONS BY THE INTERNAL REVENUE SERVICE: PROBLEMS
AND SOLUTIONS 2 (1977).

IIId. at 9.
112
1Id. at 19.

SId. at 21.

4Id. at 45.

15H.R. NO. 3245, 96th Cong., 1st Sess., 125 CONG. REC. E1342
(daily ed. March 27, 1979) (remarks of Rep. Gephardt).
116
1125 CONG. REC. H2140 (daily ed. April 9, 1979) (remarks
of Rep. Gephardt).

117Id. at H2142.

Hearings on H.R. NO. 3245 Before the Subcommittee on Select
Revenue Measures, 96th Cong., Ist Sess. 5 (1979) (statement of
Donald Lubick).

19Id. at 7.


10d. at 14.





56

121
Id. at 12.

H.R. NO. 5460, 96th Cong., Ist Sess., 125 CONG. REC. H8769
(daily ed. September 28, 1979) (remarks of Rep. Rostenkowski).
23Id. at H8774.















CHAPTER III

RESEARCH METHODOLOGY


Research Question 1


Variable Identification


The purpose of Research Question 1 is to identify those factors

or variables that are relevant in determining whether or not a

worker is a common law employee. Once these factors are identified,

the court cases can be examined to determine which factors were

mentioned by judges in their written opinions. These data will then

be used to build the models needed to answer the remaining three

research questions.

The first source of relevant factors was contributed by the

appropriate statutory authority. To supplement the limited

statutory guidelines a literature review was conducted. The most

significant work in the common law employee area as far as identify-

ing relevant definitional factors was a BNA Tax Management port-

folio.2 The Tax Management portfolio heavily cited both the IRS

audit manual3 and the Restatement of Agency (hereinafter referred

to as the Restatement). Consequently, these two sources were

studied in detail and are relied on heavily in the subsequent ex-

planation and definition of the variables.

57








Operational Definition of Variables.


The following is a list of the variables that were obtained

from the above mentioned sources. Each of these eleven variables

is discussed in detail in the pages that follow.

1. Supervision.

2. Integration.

3. Right to delegate.

4. Continuing relationship.

5. Set hours of work.

6. Having control over the place of work.

7. Independent trade.

8. Method of payment.

9. Payment of business and/or travel expenses.

10. Furnishing of tools and materials.

11. Realization of profit or loss.

1. Supervision. Supervision as defined in this study,

centers on the ability of a would be employer to control "how"

a worker accomplishes his assignment. Even though it does not use

the term supervision, Treasury Regulation Section 31.3121(d)-l(c)(2)

states that a common law employee relationship generally "exists

when the person for whom services are performed has the right to

control and direct the individual who performs the services,

not only as to the result to be accomplished by the work but also

as to the details and means by which that result is accomplished"

(Emphasis added).

The IRS audit manual discusses supervision in terms of three

separate criteria. The first criterion is referred to as







instructions and is defined as--

A person who is required to comply with instructions
about when, where, and how he is to work is ordinarily
an employee. Some employees may work without receiving
instructions because they are highly proficient and
conscientious workers. The instructions which show how
to reach the desired result may be oral or written.5

The IRS definition of instructions closely parallels the earlier

definition of supervision taken from the Treasury Regulations. Both

sources refer to the notion that an employer is concerned with

more than simply the end results of a job. Instead,an employer

generally desires to control how the job is to be performed. The

IRS definition also relates the need for supervision to the amount

of skill the worker possesses. This same point is made in Ben v. U.S.6

where the District Court stated that "the element of freedom from

control in the manner of the performance of the work, emphasized

in some decisions, loses much of the significance when the skill

of the worker is relied upon in the accomplishment of the par-

ticular task."

The second criterion is referred to by the IRS as training and

is defined as--

Training a person by an experienced employee working with
him, by correspondence, by required attendance at meet-
ings, and by other methods indicates that the employer
wants the services performed in a particular method or
manner. This is especially true if the training is given
periodically or at frequent intervals. An independent
contractor ordinarily uses his own methods and receives
no training from the purchaser of his services. In fact,
it is usually his methods which bring him to the attention
of the purchaser.8

If the employer requires an individual to attend meetings or work

with an experienced employee, this is further evidence of a desire





60
9
to control how the work is to be accomplished. In McCombs v. U.S.,

salaried supervisors were employed to observe workers who had

contracted to apply aluminum siding to residential housing. The

purpose of the supervisor was to "instruct the workers in certain

techniques and procedures" and "train people who had no previous

experience. The Court of Claims held that implicit in this find-

ing is the desire of the plaintiff to control not only the result

but also the manner of achieving it."l0

The ability to "direct the order or sequence in which the work

must be done" is the third criterion used by the IRS that relates

directly to the supervision factor. This criterion is defined

in the IRS audit manual as--

If a person must perform services in the order or sequence
set for him by the employer, it shows that the worker
is not free to follow his own pattern of work but must
follow the established routines and schedules of the em-
ployer. Often, because of the nature of an occupation,
the employer either does not set the order of the services
or sets them infrequently. It is sufficient to show
control, however, if he retains the right to do so.11

Again, the reasoning is that if a purchaser of someone's services

is able to control the order or sequence that must be followed by

the worker, then such an individual is controlling the details of

how the job is to be accomplished and is assuming an employer role.

The home improvement industry is an example of how the super-

vision factor has been applied in deciding employee versus indepen-

dent contractor cases. A firm hires salesmen to solicit home

repair jobs. As the jobs are acquired, an "applicator" is given

the job of affixing the roofing or siding materials. The question

is the applicator's employment status. The home improvement





61

company generally has supervisors to keep track of the applicators

progress and make sure the customers are satisfied with the com-

pleted job. The court has the task of determining the degree of

control these supervisors have exercised over the applicators.

In certain situations the supervisor only appears at the job site

12
to verify progress and offer advice if it is solicited.1 In

these cases, the courts have held that the lack of supervision

tends to support an independent contractor finding.13 On the other

hand, some courts have held that the supervisors had significant

control over the applicators and through considerable training and

instruction by the supervisors these workers become competent
14
applicators. In such cases, the courts held that the presence

of supervision supported an employee finding.15 Security Roofing and

Construction Co. v. U.S.16 discounted the notion that supervision

is an important factor when the work is largely routine.

Although it is true that the applicator is normally
unsupervised, he is checked on from time to time .
If he needs instruction, as, for instance, in how to
conserve material, he receives it. After this, the
work being largely routine, supervision is unnecessary.
I do not regard its absence of controlling significance.
One can be an employee even though he supervises himself
as to the details.17 (Emphasis added)

The presence or absence of supervision is a determination of fact

that must be made by the trier of fact and is often a difficult

distinction to make.

The Restatement defines the overall test that should be

applied in determining a master-servant relationship is whether

"a person employed to perform services in the affairs of another and

who with respect to the physical conduct in the performance of the

services is subject to the other's control or right to control."l8








However, the Restatement then lists "the extent of control which,

by agreement, the master may exercise over the details of the work"

as one of the criteria to be used in applying the above described

test.19 It appears that the Restatement considers control of the

physical conduct of the worker as the overall test to be used in

determining employment status while it also considers control over

the details of the work to simply be one of the factors to be

looked at in applying the overall test. This apparent dual role

is somewhat confusing and is reasonably similar to the question
20
that surrounded the economic reality test.20 The economic reality

test attempted to relegate control from the overall test that should

be applied in determining a common law employee to a lesser role

as simply a single factor out of many that would be used in de-

fining employment status.

2. Integration. Whether the services performed by a worker

are an integral part of the business is closely related to the

independent-trade factor, variable number seven, to be discussed

later. However, the IRS audit manual treats integration as a

unique factor--

Integration of the person's services into the business
operations generally shows that he is subject to direction
and control. In applying the integration test, first
determine the scope and function of the business and
then whether the services of the individual are merged
into it. When the success or continuation of a business
depends to an appreciable degree upon the performance
of certain services, the people who perform those services
must necessarily be subject to a certain amount of control
by the owner of the business.21

A worker is considered to be a integral part of the taxpayer's

business if the services performed by the worker have "been molded

into he taxpayer's s overall business as one integrated operation
into the taxpayer's overall business as one integrated operation.






63

A contra example to the notion of being integrated into the business
23
is found in Aparacor v. U.S.23 which is a case dealing with the

employment status of commission salespersons or distributors.

Aparacor was a large corporation in the business of designing

and selling women's apparel which contracted with various indivi-

duals to distribute and sell its products at retail prices. The

distributors furnished their own office facilities, hired their

own assistants, and were free to engage in other work or the sale

of other products. They were neither reimbursed for expenses nor

trained by Aparacor. The Court further held that the retail dis-

tribution performed by the salespersons was not part of Aparacor's

"regular business of designing and supplying merchandise."24

Aparacor was distinguished from the cases relied on by the IRS

because the court held that those cases involved "typical (albeit

temporary) master servant relationships, that is, the performance

of relatively simple labor, in the course of and as an integral

part of the employer's business."25

Entertainers are an example of individuals that have been held

to be an integral part of the business for which they perform their

services.26 One Court relied on the fact that before certain belly

dancers were featured as part of the lounge show entertainment, the
27
business had been losing money.27 After the belly dancers and

Greek music were added the establishment became very profitable.

In a similar case, another Court held that go-go dancers were an

integral part of the business.28

The Court finds that the evidence does not support the tax-
payer's contention that the entertainment provided by
these individuals was merely ancillary and subordinate to
the main business. On the contrary, we must conclude that
the go-go dancers were an integral part of the business.29







In some situations, integration is considered an important

factor in employment status cases. However, in the 139 cases

(since several of the cases had multiple verdicts, the total

number of decisions rendered was 148) used in this study, inte-

gration was not explicitly mentioned in 120 of the decisions.

In twenty-six decisions integration was specifically mentioned

in favor of an employee finding and in only two decisions was the

lack of integration mentioned in favor of an independent contractor

finding. It appears that generally the integration factor either

does not apply to the situation or it is a difficult factor to

apply. A businessman is generally not going to hire a worker

to perform services that are unnecessary to the successful run-

ning of his business. To determine if workers are an integral

part of the business is a fact judgment that must be made by the

court. Only where a court explicitly mentions integration in a

written opinion is it considered as evidence supporting either

an employee or independent contractor finding for purposes of this

study.

3. Right to Delegate. If a worker is able to hire, pay, and

fire his own helpers this evidences a certain degree of freedom

characteristic of an independent contractor. On the other hand,

an employer-employee relationship is generally indicated if a

worker is unable to delegate work to another person, or if the

purchaser of his services controls the hiring and firing of as-

sistants.

The IRS audit manual defines the "right to delegate" in terms

of the freedom to "employ assistants." The right to delegate is








defined in the manual as--

If the services must be rendered personally, presumably
the employer is interested in the methods as well as
the results. He is interested in not only the result
but also the worker.30

The ability of a worker to employ his own assistants is

defined in the IRS audit manual as--

Hiring, supervising, and paying assistants by the employer
generally shows control over the men on the job. Some-
times one worker may hire, supervise, and pay the other
workmen. He may do so as the result of a contract
under which he is responsible for only the attainment
of a result. In this case he is an independent contrac-
tor. On the other hand, if he hires, supervises, and pays
workmen at the direction of the employer, he may be
an employee acting in the capacity of a foreman for or
representative of the employer.31

The Internal Revenue Code is not specific about what consti-

tutes a common law employee. For FICA and FUTA purposes, Section

3121(d) states that an individual is an employee under the usual

common law rules if the contract of service contemplated that sub-

stantially all of such services are to be performed personally

by such individual. This "if" clause seems to imply that it is not

enough for a person to be adjudged a common law employee, but

in addition the contract must contemplate that the services be

performed personally by the worker in question. Thus, according

to this statutory construction, the FICA withholding and payment

requirements cannot be applied to a common law employee unless

the right-to-delegate factor is first satisfied. It is interesting

to note that even though this statutory construction exists no

argument was ever made in the 139 cases included in this study to

avoid the FICA requirements due to the fact that the contract

involved contemplated personal performance of the services. Marmoll








concludes that "there has been a tendency in the authorities to

de-emphasize--primarily through silence--the statutory exceptions

set forth in Section 3121(d)."32

The presence or absence of the right-to-delegate factor was

a factor that was reasonably easy to elicit from the cases read.

Of the 148 judgements made by the District Courts and Court of

Claims, ninety-three decisions incorporated the ability to delegate

by the worker as a factor in favor of an independent contractor

finding. An additional fifteen decisions determined that the worker

did not have the freedom to delegate work or hire his own assis-

tants which tended to indicate an employer-employee relationship.

In the remaining forty decisions the right-to-delegate factor

was not relevant, or its effect was not determinable by the courts.

4. Continuing Relationship. In general, this factor supports

an employer-employee finding if the relationship between the

prospective employer and the worker is found to be of a permanent

nature. Exactly what factors must be present to constitute a

permanent relationship is not easy to pinpoint from the statutory

law or cases.

The Restatement's explanation of this factor is rather cursory.

"If the time of employment is short, the worker is less apt to

subject himself to control as to details and the job is more

likely to be considered his job than the job of the one employing

him."33 The problem with this definition is that it does not give

adequate guidance as to what should be considered permanent em-

ployment. If a worker performs services on a sporadic job-by-job

basis with no one job lasting over a few days, is this then








considered a non-permanent relationship? If this sporadic type

of relationship continues for several years does this then indi-

cate a continuing relationship? If the relationship is contractual

by nature, how long must the contract extend for a permanent rela-

tionship to exist? These are all questions that can arise when

attempting to use the continuing-relationship factor to aid in

determining an employee versus independent contractor case.

The IRS audit manual defines this factor as--

A continuing relationship between an individual and the
person for whom he performs services is a factor which
indicates that an employer-employee relationship exists.
Continuing services may include work performed at fre-
quently recurring though somewhat irregular intervals
either on call of the employer or whenever the work
is available. If the arrangement contemplates continuing
or recurring work, the relationship is considered per-
manent even if the services are part-time, seasonal,or
of short duration.34

The IRS is fairly clear in its interpretation of how to handle

sporadic employment that lasts for a reasonable length of time.

Their view is upheld in Tapager v. Birmingham35 where the District

Court held that even though certain salespersons were part-time

workers because their relationship was regularly recurrent this

was an indication of an employer-employee relationship. The Court

defined the continuing-relationship factor as--

The relationship of an independent contractor generally
contemplates the obtaining of an agreed end and usually
contemplates the obtaining of that end within a stipu-
lated period of time. The relationship of employment
generally contemplates a continuous and indefinite
rendering of services which relationship is terminable
either at the option of the employer or employee without
contractual liability.36

This definition implies that if a definiteness as to the time of

the termination of the relationship exists, then a non-continuing







relationship exists and we have an independent contractor finding.

On the other hand, a continuing relationship or employee finding

is indicated by an indefiniteness as to the time of the termination

of the relationship.

A somewhat contrary finding to Tapager was indicated in
37
Silver v. U.S. This case involved the home improvement industry

and the classification of an applicator of roofing and siding

materials as an employee or independent contractor. The applicator

in question had performed several jobs during the three year period

under question. The Court held that "permanency of relationship

can hardly be said to exist or be a weighty element when each

obligation was of comparatively short duration and the worker was

free to accept or reject the offer of a new or similar obligation."38

Therefore, when services are performed sporadically, but over a

reasonable length of time, it is not completely clear how a court

will interpret this situation in terms of the continuing-relationship

factor.

It is interesting to note that while the continuing-relationship

factor is mentioned in various sources (Restatement, IRS audit

manual, and court cases), that none of these sources refer to any

statutory support for this factor. The continuing-relationship

factor is listed in Section 3121(d) as an exception to the ap-

plication of the FICA provisions to a common law employee. Para-

graph (d) states that if a person is adjudged a common law employee

and satisfies the "if clause" discussed in conjunction with factor

three (right to delegate), then such a common law employee will be

considered an employee for the FICA provisions "except if the







services are in the nature of a single transaction not part of

a continuing relationship with the person for whom the services

are performed" (Emphasis added). The statute quite clearly

provides that a worker who is classified as a common law employee,

but whose services are not part of a continuing relationship is

excepted from the FICA withholding and payment requirements which

for all intents and purposes treats the worker as an independent

contractor. As was mentioned previously, the exceptions of Section

3121 have not been argued in the cases included in this study

and appear to be substantially ignored in the application of the

common law employee rules for FICA purposes.

For purposes of this study, the continuing-relationship

factor was fairly prevalent in the written opinions of the cases

read. In forty-seven decisions the absence of a continuing rela-

tionship was indicated which favored an independent contractor

finding. In an additional sixty-eight decisions the presence of

a continuing relationship was found to exist which favored an

employee finding. In the remaining thirty-three decisions the

courts did not mention the continuing-relationship factor in their

written opinions.

5. Set Hours of Work. The ability of a worker to set the

aggregate number of hours he will work and also determine when he

will work those hours is usually an indication of the independence

possessed by a self-employed person. Whether a person can control

the hours he works is generally a readily determinable fact. This

is somewhat evidenced by the fact that of the 148 decisions con-

sidered in this study, only eight times did the court fail to








consider setting the hours of work as a relevant factor in their

written opinion. In the other 140 decisions ninety-nine times

the ability to set one's hours of work was mentioned in favor

of an independent contractor finding while in the other forty-one

decisions the ability of the prospective employer to control the

hours worked was considered to favor a finding of employee status.

In regards to the ability to control the hours an individual

works, the IRS audit manual provides that--

The establishment of set hours of work by the employer is
a factor indicating control. This condition bars the
worker from being master of his own time, which is the
right of the independent contractor. If the nature of
the occupation makes fixed hours impractical, a require-
ment that the worker work at certain times is an element
of control.39
40
In H.R. 3245 as proposed by Representative Gephardt, the first

of the "safe-harbor" provisions is controlling the hours worked.

In hearings before the subcommittee on Select Revenue Measures,

Representative Gephardt proposed that this factor contain a two-

prong attack.41 First, in order to satisfy the safe-harbor pro-

visions, a worker must be able to control the aggregate hours he

works. Second, the worker must also be able to control the

schedule upon which he works these hours. If a worker meets these

two requirements then a reasonable degree of freedom appears to

exist favoring an independent contractor finding.

The real crux of this factor that has been mentioned by both

the IRS audit manual and H.R. 3245 is that: if a worker can control

how much, when, and how hard he works, then this is a strong indi-

cation of self-employment. If on the other hand, a person has to

work a 9 to 5 job due to his employer's desire, then the worker

is not master of his own time.








6. Having Control Over the Place of Work. The general

premise of this factor is that independent contractors generally

provide their own place of business while an employee is provided

a place to work by his employer. This general distinction, however,

is somewhat of an oversimplification. There are certain kinds of

work that must be done on the employer's premises, but such work

need not be done by an employee. For example, an electrician

must come to the employer's place of business, but by no means

necessitates that the electrician is an employee.

The IRS audit manual defines controlling-the-place-of-work

factor as--

Doing the work on the employer's premises in itself is not
control. However, it does imply that the employer has
control, especially when the work is the kind that could
be done elsewhere. A person working in the employer's
place of business is physically within the employer's
direction and supervision. The use of desk space and
telephone and stenographic services provided by an em-
ployer places the worker within the employer's direction
and supervision.42

According to this definition, the Service's position seems to be

that a fairly strong indication of an employer-employee relation-

ship exists if a worker performs his services on an individual's

business premises. However, Marmoll suggests that the Service's

current position is that "furnishing a place of work and of

supplies--while factors to be considered in determining an employer-

employee relationship--are not alone sufficient to establish the

relationship."43

Controlling the place of work is a factor included in both

the GAO proposal and H.R. 3245. In the GAO proposal, the second

of the four basic tests which must be met to be guaranteed








self-employment status requires the worker to have a "principal

place of business other than that furnished by the persons for

whom he or she performs or furnishes services."44 As stated, this

test seems to eliminate anyone being automatically considered an

independent contractor if they happen to work on the employer's

premises. However, in further explanation of this test, the GAO

proposal states that the principal place of business can be rented

from the would be employer if the rent if fair and reasonable4 5

While the ability to rent from the individual to whom the services

are rendered seems to provide more flexibility, the GAO proposal

states clearly that using a person's own residence for the princi-
46
pal place of work will not satisfy this test.4

Providing one's own place of business is also the second

requirement of the safe-harbor criteria suggested in H.R. 3245.

However, the bill does recognize that certain individuals, by the

very nature of the services they perform, do not have places of

business and should not be held to this requirement.47 For example,

insurance agents, door-to-door salesmen, real estate agents, truckers,

and loggers do not perform their services at a single location.48

The notion of salesmen and truckers introduces another element

into controlling the place of work. It is possible that a salesman

may be required to work in a certain area and that truckers may be

required to drive certain routes. According to the IRS audit manual,

this type of situation is also evidence of controlling the place of
49
work. The IRS audit manual states that--

Control over the place of work is indicated when the
employer has the right to compel a person to travel a
designated route, to canvass a territory within a certain








time or to work at specific places as required. In
some occupations services must be performed away from
the premises of the employer, for example, employees
of construction contractors or taxicab drivers.5

In most situations, the place-of-work factor is fairly easy

to determine. This partially explains why it was chosen by both

the GAO and Representative Gephardt in drafting proposals that

would hopefully add certainty to the common law employee definition.

In the cases included in this study, the employer controlling the

place of work was mentioned seventy-seven times in favor of an

employee finding. In another forty-nine decisions, the worker

was determined to control the place of work which favored an inde-

pendent contractor finding. In only twenty-six decisions did the

courts consider the place-of-work factor irrelevant.

7. Independent Trade. Whether or not a worker is engaged

in an independent trade or business from the person for whom he

is providing services is a very fundamental element of the common

law control test used in defining employment status. The Restate-

ment lists as one of the facts to be determined in establishing

whether a worker is an independent contractor or employee "whether

or not the one employed is engaged in a distinct occupation or

business."51

In discussing common law employees, the Treasury Regulations

state that individuals performing services as independent contrac-
52
tors are not common law employees.52 As an example of these types

of individuals, the Regulation lists "physicians, lawyers, dentists,

veterinarians, construction contractors, public stenographers,

and auctioneers, engaged in the pursuit of an independent trade,








business, or profession ."53 (Emphasis added). The entire idea

of an independent contractor is closely tied to the existence of

a separate or independent trade or business.

The GAO, in their report to the Joint Committee on Taxation

discussed in Chapter II, directed the employment status test at

determining whether a true separate business entity actually
54
exists.54 From this study, the GAO concluded that the major prob-

lem centered around those situations where the worker had a sep-

arate trade or business from the prospective employer, but the IRS
55
still reclassified the worker as an employee. Because of the

GAO's belief that many of these separate businesses should indeed

by treated as independent contractors, their proposal was structured

such that a trade or business not only had to exist, but that trade

or business had to satisfy four basic tests to be considered as

a self-employed entity.56

The IRS audit manual provides substantial detail in how to

determine if a worker is engaged in a separate trade or business.

The manual divides the independent-trade factor into the following

three separate categories:

(a) Can the individual providing the services work for
a number of firms at the same time?

(b) Does the individual make his services available to
the general public?

(c) Is the individual required to devote his full time
to the person for whom he performs the services?

The audit manual defines the separate categories as--

Working for more than one firm at a time. A person
who works for a number of persons or firms at the same
time is generally an independent contractor because he
is usually free from control by any of the firms. It









is possible, however, for a person to work for a number
of people or firms and be an employee of one or all of
them.
Making service available to general public. The
fact that a person makes his services available to the
general public usually indicates an independent contrac-
tor relationship. An individual may hold his services
out to the public in a number of ways; he may have his
own office and assistants; he may hang out a "shingle"
in front of his home or office; he may hold business
licenses; he may be listed in business directories or
maintain business listings in telephone directories;
or he may advertise in newspapers, trade journals, maga-
zines, etc.
Full time required. If the worker must devote his
full time to the business of the employer, the employer
has control over the amount of time the worker spends
working and impliedly restricts him from doing other
gainful work. An independent contractor, on the other
hand, is free to work when and for whom he chooses.
Full time does not necessarily mean an 8-hour day or a
5- or 6-day week. Its meaning may vary with the intent
of the parties, the nature of the occupation, and customs
in the locality. These conditions should be considered
in defining "full time."
Full-time services may be required even though not
specified in writing or orally. For example, to produce
a required minimum volume of business may compel a person
to devote all of his working time to that business; or
he may not be permitted to work for anyone else, and to
earn a living he necessarily must work full time.58

It is very difficult to identify the distinction between working

for more than one firm at a time and having to devote one's full time

to one employer. Simply stated, if a worker works for other emp-

loyers and is not required to answer to just one boss, a certain

degree of independence is shown which is consistent with being

self-employed. For example, this type of independence is charac-

teristic of a worker known as a "gypsy chaser." A gypsy chaser

is an individual who contracts with a truck driver to unload his

truck which may contain furniture, fruit, etc. Gypsy chasers

generally frequent truck stops and attempt to solicit jobs as








truckers enter the stop. Generally, a gypsy chaser is engaged by

many different truck lines in a single day and is not considered

to work full time for any one firm.59 The courts have generally

interpreted this as significant evidence in favor of an indepen-

dent contractor finding.6

The occupation of a telephone solicitor is an example of a

different situation where a worker performs services for one per-

son and is required to work a set number of hours per day. In

Lieb v. U.S., telephone solicitiors who worked for an exterminator

company were classified as employees by the IRS. Among other fac-

tors discussed by the Court was the fact that the solicitiors were

required to work their entire shift in the employer's office for

five or six days a week.62 Such a finding was evidence that the

basic relationship between the telephone solicitors and Lieb was

that of an employer-employee.

The distinction between working for more that one firm at

a time, or offering services to the general public, is also very

slight. The nature of the services performed by a worker apparently

determines whether an individual can provide his services to the

public at large, or whether he can provide his services to a number

of different firms, and thus whether an employer-employee relation-

ship exists. For example, a steel consultant's services are of

fairly limited application to the general public and would there-

fore probably be limited to certain engineering or steel-related

industrial firms. In contrast, an insurance salesman's services

are of the nature that are more applicable to the general public.








It is this type of service that the IRS audit manual is referring

to when it states that a worker may advertise in newspapers,

magazines, or be listed in the telephone directory. In reading

the court cases that comprise the data for this study, it was found

that working for more than one firm and working for the general

public measured the same basic factor, but simply applied in dif-

ferent factual situations.

The independent-trade factor was considered present in the

cases read for this study if it was manifest in any of the three

separate categories as defined by the IRS. In the majority of the

decisions (ninety-four), an independent trade or business was deter-

mined to exist. In twenty-five decisions the courts held that the

worker was not involved in a separate independent trade or bus-

iness. In the remaining twenty-nine decisions, whether or not

an independent trade existed was not a relevant factor according

to the court's written opinion or if it was relevant it was inde-

terminable whether or not it favored an independent contractor

versus an employee finding.

8. Method of Payment. The method of payment is the seventh

fact that the Restatement mentions should be considered in determining

whether "one acting for another is a servant or an independent con-

tractor."63 The Restatement fails to elaborate on what methods of

payment distinguish an independent contractor finding from an em-

ployee finding. The IRS audit manual's definition, however, is

quite helpful in adding some substance to this factor. The IRS

audit manual defines the method of payment as--

Payment by the hour, week or month generally points to
an employer-employee relationship, provided that this








method of payment if not just a convenient way of paying
a lump sum agreed upon as the cost of doing a job. The
payment by a firm of regular amounts at stated intervals
to a worker strongly indicated an employer-employee
relationship. The firm assumes the hazard that the
services of the worker will be proportionate to the
regular payments. This action warrants the assumption
that, to protect its investment, the firm has the right
to direct and control the performance of the worker.
It is also assumed in absence of evidence to the con-
trary that the worker, by accepting payment upon such
basis, has agreed that the firm shall have such right
of control. Obviously, the firm expects the worker to
give a day's work for a day's pay. Generally, a person
is an employee if he is guaranteed a minimum salary or
is given a drawing account of a specified amount at
stated intervals and is not required to repay any excess
drawn over commissions earned.
Payment made by the job or on a straight commission
generally indicated that the person is an independent
contractor. Payment by the job includes a lump sum
computed by the number of hours required to do the job
at a fixed rate per hour. Such a payment should not
be confused with payment by the hour.64

The explanation of the method-of-payment factor provided by the

IRS indicates that the basic distinction is that employees are

paid regular amounts at stated intervals regardless of the employee's

output, whereas an independent contractor is paid on a lump-sum

or per-job basis and his compensation does depend on his completion

of certain requirements.

This basic distinction makes sense when one assumes that if a

worker is to be paid by the hour his employer is interested in

the worker completing the task in the shortest reasonable amount

of time. Since the employer must pay such a worker by the hour,

he likely will provide more detailed instructions and supervision

to ensure the quickest possible completion of the job.

On the other hand, if the worker is paid by the job in a lump

sum payment, this tends to indicate that the purchaser of the

services is only interested in the final results to be accomplished.








The amount of time to complete the job is irrelevant in this type

of situation as long as the results are those that both parties

agreed upon. In fact, it would be incumbent upon the worker to

determine the quickest method to complete a job because his com-

pensation is directly related to the number of jobs that he can

successfully complete. Also, the fact that a worker may receive

payment in several installments before the job is completed is

not an indication of employer-employee relationship. According

to the IRS audit manual, if the receipt of intermittent payments

is merely a convenient method to disperse a lump sum, then such

an arrangement still favors an independent contractor finding.65

The rule that is being applied throughout the entire analysis of

this factor is whether the worker's compensation is contingent on

his performance of certain requirements (meaning no guarantee of

profit) or simply the number of hours he works.

In general, the determination of how a worker is paid is

rather straightforward. In the cases included in this study,

128 times the judge determined that the worker was being paid in

a method that indicated an independent contractor finding. In an

additional seventeen cases, the judges felt the method of payment

seemed to indicate an employer-employee relationship. In only

three cases was the method-of-payment factor not considered in the

judge's written opinion. From this it is evident that the method-

of-payment factor has been applied in most of the cases included

in this study and the information needed to apply this factor in

actual court case situations is usually readily available.

9. Payment of Business and/or Traveling Expenses. The

reimbursement of expenses is concerned in general with the







method-of-payment factor as discussed above, but is unique enough

to be considered a separate factor. Reimbursement of expenses has

a similar advantage to the method-of-payment factor; namely, it is

a reasonably simple factor to apply. Either a worker is being re-

imbursed for his business expenses or he is not. This is not

particularly consistent with the fact that in fifty-seven of the

decisions analyzed in this study, the court did not appear to

consider reimbursement of expenses in their written opinions.

This is most likely attributable to the fact that a reimbursement

of expenses was not mentioned specifically in the Restatement or

in one of the landmark cases as was the method-of-payment factor.66

This may result in the situation where reimbursement of expenses

is simply considered jointly with the method-of-payment factor as

does Marmoll in the Tax Management portfolio.67

This study takes the same approach as the IRS audit manual

by treating reimbursement of expenses as a separate factor. The

audit manual defines reimbursement of expenses as--

If the employer pays the person's business and/or traveling
expenses, the person is ordinarily an employee. The
employer, to be able to control expenses, must retain
the right to regulate and direct the person's business
activities.
Conversely, a person who is paid on a job basis and
who has to take care of all incidental expenses is
generally an independent contractor. Since he is account-
able only to himself for his expenses, he is free to work
according to his own methods and means.68

The audit manual implies that a worker who has to depend on an

employer for reimbursement of expenses is more subject to that

employer's control than is a worker who pays his own expenses.

In other words, reimbursement of a worker's expenses is simply








another method of controlling that worker's business activities.

An employer-employee relationship tends to be indicated when a

purchaser of services pays for the worker's business licences,

meals, transportation, or allows the worker to use the company's

charge account.

The Service's position as discussed above is somewhat inflex-

ible. Marmoll states that--

The Service's position, as set forth in the manual, on
the payment of business or travel expenses ignores the
possibility of a cost plus fixed fee contract as well
as the modern day custom of charging itemized expenses
to the customer or client in addition to the fee for the
work done. The clients of lawyers and accountants, for
example, are customarily charged for the attorney's or
the accountant's travel expenses and, yet, the client
does not control the business activities of the attorney
or accountant, the relationship clearly being one of
client and independent contractor.69

The lawyer or accountant example above appears to suggest that the

reimbursement-of-expenses factor needs to be considered in terms

of the overall compensation agreement. If the reimbursement of

expenses is in addition to an agreed upon lump sum fee, then Mar-

moll seems to suggest that the lump sum payment which supports

an independent contractor finding somehow overrides the fact that

expenses are reimbursed which of course favors an employee finding.

While it is true that the method-of-payment factor may carry more

weight than the reimbursement-of-expenses factor when they have

conflicting results, the fact still remains that reimbursement

of expenses by the employer still favors an employee finding.

In the cases included in this study, judges were able to apply

the reimbursement-of-expenses factor on its own merits in ninety-

one decisions. In seventy-four of these decisions the factor








favored an independent contractor finding while in the other

seventeen decisions the factor favored an employee finding.

10. Furnishing of Tools and Materials. The more substantial

the investment in tools and materials by a worker, the more likely

the worker is to be considered an independent contractor. If the

employer furnishes the tools and materials necessary for the

accomplishment of the worker's assignments, then this tends to

indicate an employer-employee relationship. There are certain

kinds of jobs for which the tools needed to perform the assigned

task are inconsequential. For example, a gypsy chaser70 who unloads

trucks needs practically no tools to perform his services. Since

the work is primarily manual, the furnishing of tools would not

be a particularly significant factor in a case involving the

employment status of a gypsy chaser. On the other hand, an appli-

cator of siding and roofing materials must have a rather complete

set of hand tools, ladders, scaffolding, etc. and the furnishing

of tools would be a much more significant factor in this type of

case.

Both Treasury Regulation Section 31.3121(d)-l(c)(2) and the

Restatement include the furnishing of tools as a factor that should

be included in the definition of a common law employee. The Restate-

ment further states that if the tools are provided by the employer

and are of substantial value, it is understood that the worker will

follow the directions of the owner in their use and this indicates

the owner is master.71 The IRS audit manual provides additional

detail as to how the furnishing-of-tools-and-materials factor







72
should be interpreted.72 The audit manual describes this factor

as--

The fact that an employer furnishes tools, materials,
etc., tends to show the existence of an employer-emp-
loyee relationship. Such an employer can determine
which tools the person is to use and to some extent,
in what order and how they shall be used.
An independent contractor ordinarily furnishes his
own tools. However, in some occupational fields, e.g.,
skilled workmen, workers customarily furnish their own
tools. They are usually small hand tools. Such a prac-
tice does not necessarily indicate a lack of control over
the services of the worker.73

Application of the furnishing-of-tools-and-materials factor

is relatively straightforward which is evidenced by the fact that

in 131 out of 148 decisions included in this study the judges men-

tioned this factor in their written opinions. In eighty of those

decisions, the courts held that the worker provided his own tools

and/or materials which favored an independent contractor finding.

In another fifty-one decisions the courts held that the employer

provided the tools and/or materials which favored an employee finding.

In only seventeen decisions did the judges fail to discuss the fur-

nishing-of-tools-and-materials factor as part of their written

opinion.

11. Realization of Profit or Loss. The worker who can realize

a profit or sustain a loss is generally an independent contractor

and the worker who is prevented from earning a profit in excess

of his stated compensation or is insulated against sustaining

losses, is generally considered an employee. For example, in
74
Nevin Inc. v. Rothensies the court held that a lessee of a drug-

store was an independent contractor. In this case, the parties

involved established a separate (though not a fully independent)

business where the lessee had a considerable personal financial








stake in the business.

He could make or lose money in the business. The pos-
sibility of profits could be enhanced by efficiency,
economy and skill. He took the risk of considerable per-
sonal loss in case the business was not successful a
risk which involved a good deal more than the mere
possibility of loss of position. He obtained complete
freedom in connection with his own time, working hours,
etc., and an opportunity to increase his income not only
by sales records (as in the case of a commission salesman)
but by business judgment in most of the matters which
are in the province of any independent retailer.75

Consequently, the lessee could affect the amount of profits he was

able to earn through his own managerial efficiency and skill.

The IRS audit manual also discusses the importance of managerial

skill in the realization of profits and avoidance of losses.76

The manual states that "whether a profit is realized or loss

suffered generally depends upon management decisions; that is,

the one responsible for a profit or loss can use his own ingenuity,

initiative, and judgment in conducting his business or enterprise."77

Another important component of the realization-of-profit-or-

loss factor is determining whether the worker has a significant

capital investment in the business. The IRS audit manual provides

that for a significant investment to exist, the investment must

be real, essential, and adequate.78 Such an investment is made

generally in equipment, furniture, or real property necessary to run

the business.

The question of whether an investment in equipment is real

seems to be concerned with whether the worker really has an equity

investment, or is the worker simply buying the equipment on time

from the person for whom he is providing services. The audit manual

illustrates this point in the following illustration.







Little weight can be accorded to a worker's investment .
if the worker purchases equipment from his employer
on a time basis but the employer retains title to the
equipment, has the option of retaining legal ownership
by paying the worker the amount of his equity in the
equipment at any time before the equipment is fully paid
for, requires its exclusive use in the operation of
his business, and directs the worker in its use. Such
investments are not "real".79

To be essential the investment must be in equipment or facil-

ities necessary for the performance of the worker's services.

The example given in the IRS audit manual concerns a model who has

a large investment in her wardrobe, but poses for a photographer

that ordinarily provides the wardrobe and strictly as a matter

of indulgence the photographer allows the model to use her own

clothes. Such an investment in clothes is not essential to the

services that she performs. "The photographer hires her only for

her photogenic qualities and her ability to pose; it is not required

that she furnish her own wardrobe."80

The critical test suggested by the IRS audit manual in deter-

mining if an investment is adequate is to compare the value of the

investment of the worker to the total value of all the facilities

needed to perform the work.

An investment in facilities is not adequate if the
worker must rely appreciably on the facilities of others
to perform the services. For instance, an individual
who is engaged to perform a machine operation on his own
premises and who furnishes his own equipment of sub-
stantial value may be a self-employed contractor instead
of an employee of the manufacturer.81

The realization-or-profit-or-loss factor is not only a major

part of the IRS audit guidelines, but it is also a factor included

in both the GAO proposal and H.R. 3245. In the GAO proposal,

"the risk of suffering a loss and opportunity of making a profit"







is one of the four basic tests used to determine if an independent

contractor engaged in a trade or business should be considered

self-employed.82 The only explanation of the factor provided by

the proposal is that there should be a "real possibility that

expenses directly related to the business will exceed business

income."83

On the other hand, H.R. 3245 provides considerable detail in

describing its third safe-harbor criteria which it calls "invest-

ment or income fluctuation."84 The bill proposes that a worker who

has a substantial investment or who risks income fluctuations should

satisfy this safe-harbor criteria.85 It treats investment and

income fluctuations as two separate components of an overall test

which can be satisfied by the presence of either a substantial

investment or risk of income fluctuation.

The income fluctuation test is satisfied if the income level

is neither fixed nor guaranteed, but rather directly depends on

the out-put of the worker. This can be exemplified by a salesman

who can make several sales presentations and not make a single sale.

This is not the same as an employee who is basically assured that

if he works a certain number of hours at a reasonably acceptable

level, he will receive a specified amount of compensation.8

Application of the substantial-investment-in-assets test is

similar to the income fluctuation test. The percent of the in-

vestment is not considered an important criterion despite the IRS

audit manual, but rather the worker must "invest a significant

amount of money or his time as in the income fluctuation test

with no assurance of success."87
with no assurance of success.


1








With the importance placed on the realization-of-profit-or-

loss factor by the IRS and the inclusion of the factor in both

the GAO proposal and H.R. 3245, it is not surprising that the

factor was discussed by judges in 119 of the 148 decisions included

in this study. In only twenty-nine decisions did the courts

fail to consider the realization-of-profit-or-loss factor in their

written opinions.

Before discussing the methods employed in the coding of the

variables, those factors that were deleted from the study due to

their infrequent occurrence in the written opinions will be pre-

sented. The following is a list of those seven factors discovered

in the literature review, but not included in the analysis of this

study.

(1) Did the worker possess a special skill?

(2) Was the taxpayer's claim consistent with industry
custom?

(3) If required, who supplied special uniforms?

(4) What was the intent of the employer and worker?

(5) Did the worker receive any fringe benefits?

(6) Was the worker required to submit written or oral
reports?

(7) Can the worker terminate his relationship with the
employer at any time he wishes without incurring
liability?

Since none of these factors were included in more than 10 percent

of the court cases, this was interpreted as evidence that these

factors were not particularly relevant in determining what consti-

tutes a common law employee.








The only other factor that was discovered in the search for

relevant variables, but not used in this study was the "right to

discharge." Treasury Regulation Section 31.3121(d)-I(c)(2) mentions

the right to discharge as "an important factor indicating that the

person possessing that right is an employer." The IRS audit manual

discusses the right to discharge in the following manner--

The right to discharge is an important factor in indicating
that the person possessing the right is an employer. He
exercises control through the ever-present threat of dis-
missal, which causes the worker to obey his instructions.
An independent contractor, on the other hand, cannot be
fired so long as he produces a result which meets his
contract specifications.88

The IRS audit manual's definition, simply stated, is that an employee

can be fired anytime, but an independent contractor cannot be dis-

charged except under the terms of his contract. This rather sim-

plistic approach ignores the presence of union contracts that can

often limit the possibility of discharge of an employee. This

argument is supported by Marmoll as follows--

Since the law has developed in such a manner that most
employees have some protection against an arbitrary right
of discharge, and since many contractually agree with
their employers to give a certain amount of notice before
terminating the employment relationship, the restate-
ment's explanation of who is a "servant" or "employee"
does not mention the right to discharge. Moreover,
there are courts which, while considering the factor
of right to discharge or right to terminate, have placed
the factor in perspective, not allowing it to govern
the case despite the exaggerated importance given to the
factor in the regulations.89

One District Court has held that simply because the taxpayer

had the right to discharge the worker if his services were not

satisfactory in "no way" meant that an employer-employee relation-

ship existed.90


___j








This same right exists in every contractor who had
employed a subcontractor when the latter's performance
falls below standard. An applicator who did not like a
particular assignment was free to refuse it and in such
a case the plaintiff would attempt to give him another
job.91

This same position is taken again by another District Court in a

similar case.

While it is true that he could terminate an applicator's
work before the completion of a particular job if the
desired result was not being accomplished, this power
was no different than that of a general contractor to
declare a breach of his contract had been committed by
a sub-contractor.92

What makes this factor even more difficult to understand in

this type of study is that the written opinions of the judges

ordinarily fail to provide enough information to adequately

determine if this so-called right-to-discharge factor is really

being applied to an independent contractor relationship or an

employer-employee relationship. Because of the inability to

ascertain when the right-to-discharge factor was present in the

written opinions, it was decided to delete this factor from the

study.

With the selection of the eleven factors completed, the

mathematical models necessary to answer Research Question 2 can

be constructed. Before developing the methodology to be used in

studying Research Question 2, it is first necessary to discuss the

coding procedures used in the analysis of the data.


Coding of the Variables


Each of the eleven variables previously defined was treated

as a trichotomous random variable when collecting the data from








the court cases. There are alternate methods of coding tricho-

tomous independent variables in building mathematical models.

To obtain the maximum amount of information contained in this

type of data, each variable should be represented by two dummy

variables. For example, factor 8, method of payment, would be

coded as follows: (1) if the worker was paid by the job, variable

XI would be coded as a 1 and 0 otherwise; (2) if the worker was

paid by the hour, variable X2 would be coded as a I and 0 other-

wise; and (3) if the method-of-payment factor was not mentioned

by the court, then both XI and X2 would be equal to zero.

This simple model could be expressed as

Y = B + BIXI + B22 + U, (3.1)

where

Y is the probability of a worker being classified
as an employee,

U is the error term,

B's are the independent variable coefficients.

Taking the expected values of this model results in

E(Y1IXI = 0, X2 = 0) = B0 + BI

E(Y IXI = 0, X2 = 1) = B + B2

E(YIIXI = 0, X2 = 0) = B .

The interpretation of these coefficients would be that B0 equals

the "probability of an employee finding" (hereafter referred to

as Y) if the method-of-payment factor was not mentioned in the

case. If BI is added to BO, this represents Y when a worker is







paid by the job while B0 + B2 equals Y if the work is paid by the

hour. The difference in Y according to whether a worker is paid

by the hour or by the job is


B + BI (B + B2 = BI B2.

A different approach using a trichotomous independent variable

in building a mathematical model would be to simply represent

the method-of-payment factor with three values. For example,

0 could represent the worker being paid by the job, I could re-

present the method-of-payment factor not being mentioned, and 2

could represent a worker being paid by the hour. This simple model

could be represented as

Y = B0 + BIXI + U. (3.2)

Taking the expected value would give

E(YIXI = 0) = BO,

E(YIXI = 1) = B0 + BI,

E(YIX1 = 2) = B0 + 2B2.

At first glance the only difference in the expected values for

model 3.2 and those calculated for model 3.1 would be the expected

values of Y given the worker is being paid by the hour. However,

a more important difference is that the expected values for model

3.2 imply that the difference in Y when the method-of-payment factor

is not mentioned and when the worker is paid by the job is

(B0 + BI) B0 = BI

while the difference in Y when the method-of-payment factor is not




Full Text

PAGE 1

EMPLOYEE OR INDEPEND EN T CONTRACTOR: Ai~ EX A MINATION OF THE RELEV At~ T VARL'J3LI:S EMP LO YE D BY THE FEDERAL COURTS IN DEC I DING THE QUESTIOtl BY DAVE NELSON STEWART A DISSERL\TI O N PRESENTED TO TH E GRADUA TE COUNCIL OF THE UN IVERSITY OF FLORIDA IN PARTIAL FUL FILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UN IV ER SITY OF FLORIDA 1980

PAGE 2

Copyri g ht 1980 b y Dave Nelson Stewart

PAGE 3

ACKNOWLEDGEMENTS I wish to thank John L. Kramer, the chairman of this dissertation committee, for his guidance and support throughout the course of this study. Professor Kramer has given unselfishly of his time, especially in the reading and commenting on different drafts of this dissertation. My entire dissertation committee has contributed greatly to this project. Professor William A, Collins, Dennis A. Calfee, and John C. Henretta have each shared freely of their knowledge and expertise. They have never been too busy to sit down and discuss the progress of this study. Valuable criticisms and suggestions were also received from the Ph.D. students and faculty at the University of Florida Accounting Workshop. Finally, I would like to thank my dear wife, Jane. She not only provided constant encouragement and support, but her typing and editing skills were invaluable to the production of the finished paper.

PAGE 4

TABLE OF CONTE NT S ACKNOWLEDGEMENTS LIST OF TABLES ABSTRACT CHAPTER I. INTRODUCTION Significance of the Problem Research Methodology . Research Questions Data Collection. Research Question Research Ques tion Research Question Research Question 1 2 3 4 Limitations Summary of Chapter Cont e nts Notes II. HISTORICAL BACKGROUND Dev e lo pment of L egisla tiv e History Social Securi t y Act of 1935 Common Law Definition of Master Servant Relationship . . S oc ial Security Amendments of 1 939 Income Tax Withholding Provisions Judicial Interpretation ... Social Security Am e ndments of 1 950 Current Le g islative St atus R c vL'IlllL' \ c t of 1978 ... Current Pr opose d Solution s . Notes C><.:' n e r ~l A CC'Lll ll1ting Offic e H.R. 3245; S3fe Harb o r R e medy Departm e nt of th e Treasury iv PAGE iii vi vii l 2 7 7 8 9 10 11 12 13 14 15 19 19 19 21 24 26 27 34 38 4J 44 44 46 49 50

PAGE 5

CHAPTER III. IV. V AP PEN DIX RESEARCH METHODOLOGY Research Q u es ti o n 1 Vari a ble Id e ntifi ca ti o n Op e r a ti o nal Definiti o ns of Variables Codin g o f the Variables Res earc h Q u es tion 2 . . .. ~~th e matical Techni q ues in Tax Research Dis c riminant Anal ys is Lo g it Ana l y sis Research Q ue s t i o n 3 Res ea rch Q u es tion 4 Notes ..... EHPIRICAL RESULTS Resear c h Q u es tion 2 Discrimin a nt An alys i s R es ults OLS Re g r ess ion R es ults . Lo g it A n a l ysis Res ult s .... Co m p a ri so ns of Dis c rimin a nt, O L S Regr ess ion, a nd L og it Analyses Res ea r c h Q u es tion 3 Resea r c h Q u es tion 4 Notes ..... SUHH ARY CONCLUSIONS AN D RECOMN EN DATimlS Summ ary a nd Co nclusi o n s ..... Hist o rical B ac k g r o und o f Co mmon L aw Emp lo ye e Va ri able Id e ntification ... Empir i ca l Results ... Tempor a l St a bilit y of th e Mode l C urr ent L eg isl a ti v e P r oposals Q1oi ce of Legal F o rum Recomme nd a ti o n s ....... S u gges ti o n s for Future Resea r c h No t es . . . . A LIST OF DISTRIC T COURT C..\S[S INVOLVING T l!E EM PLOY EE PAGE 57 57 57 5 8 89 93 94 98 10. 4 111 113 11 4 121 1 2 1 1 21 13 6 1 42 1 46 1 5 1 15 3 159 1 60 1 6 1 1 62 163 164 1 66 1 6 7 1 6 7 1 68 1 6 9 171 VERSG S INDEP EN DENT CO NTR, \ CTOR QUES TIO N . . 172 B. LIST OF COURT OF CLAHIS CASES I NVO LVI NG THE EMP LOYEE VERSUS INDEPENDENT CONTRi \CTOR QUES TIO N 176 BIBLIO GRAP HY BIO GRAP HICAL SKETCH V 177 187

PAGE 6

LIST OF TABLES TABLE 1.1 A COMPARISON OF THE FICA MD SECA TAXES FOR 1978 THRU 1987 4.1 BMDP 6-VARIABLE MODEL 4.2 SPSS 9-VARIABLE MODEL 4. 3 COMPARISON OF CLASSIFICATION ACCURACY PAGE 4 122 124 126 4.4 DISCRIMI NAN T & CLASSIFICATION COEFFICIENTS FOR THE 6 -V AR IABLE MODEL 4. 5 CLASSIFICATION ACCURACY MATRICES 4.6 HODEL BASED ON 22 DUMHY VARIABLES 4. 7 CLASSIFICATION ACCURACY t~TRIX 4.8 TE~WORAL STABILITY TEST OF CLASSIFICATION ACCURACY 4.9 OLS REGRESSION RESULTS 4.10 OLS STEPWISE REGRESSION RESULTS 4 .11 COEFFICIENTS A..'lD CLASSIFICATION ~ ~TRIX FOR FULL 11-VARIABLE ~~DEL 4 .1 2 COEFFICIENTS A.'-lD CLASSIFICATION r-~TRIX FOR THE BMDP 5 -VARIABLE MODEL 4.13 COMPARISON OF STEPWISE RESULTS 4 .14 C LA SS IFI CA TIO :-J ACCUR ACY FOR CHOICE OF FORU~! TE S T 4.15 F,\CTORS CONT, \INED rn ALL TIIRCE PRO!'OS \LS 4 .16 CLASS IFICATI ON ACC U RACY MATRICES 12 7 12 9 131 132 1 35 137 139 . . .... 14 3 145 14 7 155 FOR TIIE 3-VARIABLE MODELS ................ 157 vi

PAGE 7

Abstract of Dissertation Presented t o the Graduate Council of th e University of Florida in Pa rti a l Fulfillm e nt of the Requirements for the Degree o f Doctor of Philosophy EHPLOYEE OR IND EPENDENT CONTRAC T OR: AN EXAMINATION OF THE RELEVANT VAR I AB LES H!PLOYE D BY THE FED ERAL COURTS I N D EC IDI NG THE QUESTION B y Dave Nelson Stewart Chairman: J o hn L. K r a m e r Major Department: Accounting Jun e 1 980 The distinction between an employee a nd an ind epe nd e nt contractor is crucial in th e application of th e federal e mp loyme nt t ax laws. If a worker is classified as an employee, th e employe r mu s t (1) wi thh o ld income ta xe s, (2) withh old s ocial sec urit y taxes, (3) match th e social securit y ta xes pai d b y the e mploy ee and (4) pay un emp l oyme nt t axe s. If th e worker is cl ass ified a s a n ind e pend e nt co ntr ac t o r, th e emp l oye r ne e d n o t pay or withhold a n y of th e ab ove m e nti o n ed t axes. Instead, the independent contractor is re spo nsible for paying his own income and self-employm e nt t axes The Internal Revenue Code s t a tut o ril y defi n es seve r a l occupational categories that a r e treat e d as emplo yee s for e mplo ymen t tax purposes The most promin e nt of th ese ca t egories is th a t of a "c o mmon l. : 1w" employee. However, th e scope of the Code a nd Reg ul a t io ns i s sufficiently limited to leav e th e com m on law emp lo ye e d ef inition shrouded in uncer taint y It is this ambigu it y, coupl ed with th e IRS's increas ed vii

PAGE 8

enforcement of employment taxes, that has resulted in pressure on Congress to enact legislation that will add a degree of certainty to the definition of a common law employee. The legislation being considered by Congress would result in several "safe harbor" criteria that if satisfied would automatically indicate employee status. The purpose of this study is to develop a mathematical model that will empirically identify those variables used by the federal courts in employee versus independent contractor cases in hopes that these variables may provide direction to tax policymakers A sample of 148 District Court and Court of Claims cases was iden tified and used as the data with which the mathematical models were estimated. Discriminant analysis, OLS regression analysis, and a non-linear logit analysis were all used to identify those variables deemed most important by federal judges. A forward stepwise discriminant procedure result e d in a 6-variable model that was able to correctly classify 96.6 percent of the cases. A forward stepwis e 01S regression procedure result e d in a 7-variable model that correctly classified 95.3 percent of the cas e s and explained 80 percent of the variance of the model. The non-lin e ar stepwise logit procedure resulted in a 5-variable model that was able to correctly classify 97.3 percent of the cases. Both th e lin e ar and non-linear techniques produced substantially identical results. When compared with the proposals pr e s e ntly b e in g consider e d by Congress, three o f the variables recomm e nd e d by th e Co n g ressional s ubcommitt ee are also included in this study Usin g these three vari ables as the mod e l, approximately 90 p e rcent of the cases were correctly classified. Vlll

PAGE 9

A secondary question addressed in this study is whether the choice of legal forum is important in an employment status tax case. Through a comparison of the discriminant classification accuracy rates and the Chow test applied to the OLS results, it was concluded that for employ ment tax cases the two Federal Courts use very similar decision-making models. The findings of this study suggest several conclusions. First, the use of the non-linear analysis did not produce substantially different results than did the linear regression or discriminant analysis procedures. Second, there are factors that have been used successfully by federal judges in employment tax cases that could provirl c dir e ction to tax policymakers attempting to statutorily define a common law employee. Third, th e decision-making models employed by the District Courts and the Court of Claims are not significantly different for federal employment tax cases.

PAGE 10

CHAPTER I INTRO DU CTI ON Until the 1960's, the collection of federal em ployment taxeswithholding, social security, and unemployment--followed a relatively I smooth course for employers and other taxpayers. However, since the late 1960's, the Internal Revenue Service (IRS) has increased its en2 forcement of the employment tax laws. The question of employment status generally centers around the classification of whether a worker is an independent contractor or an employee. Because so few cate g ories of employees are statutorily defin e d, the question of wh e ther a worker is an independent contractor or an employee g enerally centers around the issue of what constitutes a "common law" employee. The common law definition of employee has evolved from a substantial body of case law. 3 Congress initially felt that all unc er tainty in the classification of an individual as an employee or an independ e nt contractor would "vanish if only corrnnon law principles distilled from many decisions by many courts 4 were applied to factual situations as they arose." The assumption that th e re is some simpl e uni f orm and easily applicable t e st which the courts have used unf o rtunately 1.s not tru e Few probl e ms in the law have greater variety of application a nd conflict in r e sults th an the cas es arising in th e bord e rland betw ee n wh3t is clearly an e mploy er em ploy ee r e lationship and 5 what is cl ea rly on e of ind e pendent-entrepre neurial dealing. The substantial amount of liti g ation 6 1.n the federal employment tax area is evidence that the common law definition of an employee is easier stated than appli ed The determination of employment status is an

PAGE 11

2 issue on which both the taxpayer and the government desire some degree of closure. The taxpayers complain that the retroactive reclassification of workers as employees and the resulting taxes, interest, and penalties are causing interruption of business activities and financial hardship 7 through severe cash flow demands. The IRS is claiming that with the uncertainty surrounding the common law rules that taxpayers are able to easily avoid employee status. 8 Since this results in no withholding of taxes by the employer, independent contractors are easier able to escape detection should they fail to file their tax returns. Consequently, proposals have been forthcoming from a Congressional subcommittee, 9 the General Accounting Office, IO and the IRS 11 that recommend the solidification of the many factors that determine a common low employee into a smaller number of factors that provide a more clear cut test. However, these factors lack theoretical support and in some cases even lack the desirable tax characteristics of simplicity and certainty. The purpose of this study is to determine those factors that the courts have found significant in defining employment status with the hope that these variables may prove helpful to tax administrators as they search for an appropriate definition of the employee-employer relationship. Such a study could provide valuable insight into what factors can be consistently applied when individuals are faced with the task of determining if a worker is an employee or independent contractor. Significance of the Problem Once the court determination is made that a certain individual is an "employee," the employer may have the fol lowing responsibilities: (J) withhold social security taxes under the Federal Income Contributions

PAGE 12

3 Act (FICA), 12 (2) withhold federal income taxes, 13 (3) match the FICA 14 taxes paid by the employee, and (4) pay unemployment taxes under the Federal Unemployment Tax Act (FUTA). 15 When an individual is classified as an independent contractor the purchaser of his services need not pay or withhold any of the above mentioned taxes. The purchaser of the services may only need file an information return (Form 1099). 16 It is then the responsibility of the independent contractor to pay self-employment taxes under the Self-Employment Contributions Act (SECA) 17 and pay his own income taxes through the use of the 18 estimated tax procedures. With inflation pushing wages into higher tax brackets and prospective increases in both the FICA and SECA wage 19 bases and rates, the amount of employment taxes can vary substantially according to one's employment status. Table I. I represents the present and proposed FICA and SECA tax rates and wage bases from 1978 to 1987. Columns (5) and (6) illustrate that the social security taxes are substantial and will drastically increase 1n the next ten years. Column (7) shows the differences in the actual tax that goes into the Social Security Fund depending on the employment status. The difference in what an employer must pay is even more pronounced when one includes the FUTA taxes (3.4% of the first $6,000 of wages per employee). For example, using the maximum wages applicable to employment taxes for 1980, Column (7) shows that an ad ditional $1,077.44 of taxes are paid into the federal treasuries if a worker is an employee versus an independent contractor. This difference of $1,077.44 should be increased by the FUTA taxes of $204 (3.4% x 6,000) that are also imposed thereby resulting in a total difference of $1,281.44. Consequently, Column (7) probably represents for most

PAGE 13

4 TABLE I I A COMPARISON OF THE FICA AND SECA TAXES FOR 1978 THRU 1987 ( I) (2) (3) (4) (5) (6) ( 7) Max. Max. Excess FICA FICA SECA Wage FICA SECA Maximum over b SECA Tax Year Ratea Rate Base C Tax Tax Maximum d % % $ $ $ $ 1978 12. 10 8. I 0 17,700 2,141.70 1,433.70 708.00 1979 12.26 8. I 0 22,900 2,807.54 1,854.90 952. 10 1980 12.26 8. 10 25,900 3, 17 5. 34 2,097.90 1,077.44 1981 13.30 9.30 29,700 3,950.10 2,762. JO I, 188. 00 1982 13 .40 9.35 31,800 4,261.20 2,973.30 1,287.90 1983 13,40 9,35 33,900 4,542.60 3,169.65 1,372.95 1984 13.40 9.35 36,000 4,824.00 3,366.00 I ,458 .00 1985 14. 10 9.90 38, JOO 5,372.10 3,771.90 1,600.20 1986 14 .30 10.00 40,200 5,748.60 4,020.00 1,728.60 1987 14 .30 10.00 42,600 6,091.80 4,260.00 I 83 I 80 aThis percentage includes both the employer s and the employee 's portion of the FICA tax. This perc e ntage includ e s the social security rate and hospital insurance rates as contained in Section 3101. bThis perc e nta ge includ e s the self e mploym e nt rate and the hospital insurance rate as contained in Section 1401. C Any wages above these specified amounts are not subject to FICA or SECA tax. dThis amount is obtained by subtracting column (6) from column ( 5).

PAGE 14

5 taxpayers a conservative estimate of the total additional taxes paid when a worker is deemed to be an employee. The preceding illustration is based simply on the differences in employment taxes between an independent contractor (self-employed) and an employee. The determination of employment status can be much more critical if the IRS should choose to reclassify an independent contrac tor as an employee. The employer would then be assessed for the income taxes he failed to withhold, 20 both the employer and employee FICA 21 taxes that should have been paid by the employer, and the FUTA taxes that should have been paid, 22 plus the appropriate interest and penal. 23 h ties. Sue assessment can be made retroactively for all tax years that remain open under the statute of limitations (generally three years under Section 6501). Only if the employer can furnish certifi cates signed by his employees stating that they have paid the proper amount of income tax, can the employer possibly have the assessment 24 reduced. The employer can have the employee's share of the FICA tax abated by the SECA tax already paid only if the employee is prevented by law from filing for a refund of the SECA tax that he paid in error. 25 This situation obviously results in a double payment of the employee's share of the FICA tax. A reclassification by the IRS causes even more problems for employers because of the Tax Court's lack of jurisdiction over employ26 ment taxes. Therefore, a suit cannot be instigated by a taxpayer without payment of the assessed tax because the District Courts and Court of Claims require the payment of the assessment prior to their judicial review. To require a business to pay three years worth of income, FICA, and FUTA taxes while instituting a legal action against

PAGE 15

6 the government could impose severe hardship and cost upon a firm. However, since employment taxes are generally paid on a quarterly basis, the taxpayer ordinarily files a return covering one quarter's employment taxes and then sues for a refund of that amount and abatement of the d 27 yet unpai amount. This partial payment procedure can still involve a considerable amount of funds when dealing with industries that have a large number of workers that would be affected by such a determination. Jerome Kurtz, Commissioner of the Internal Revenue Service, in a speech before the American Bar Association pointed out : Even if taxes to the IRS type of The employee independent contractor issue has generated a staggering problem for employers assessed with huge liabili ties for past noncompliance. At present, in two districts of one region alone there are three potential assessments against taxpayers ranging from $6 million to $60 million and reclassification of independent contractors as employees result in deficiency assessments estimated at $107,000,000 annually ,2 8 the taxpayer only has to pay one quarter's worth of employment initiate a suit for refund and abatement of the lRS assessment, is not precluded from seizing a taxpayer's property in this partial payment litigation. 29 This action not taken in most is situations unless the IRS feels there are circumstances which might jeopardize the availability of the taxpayer's funds subsequent to the 1 . 30 itigation. This threat exists in employment tax cases as evidenced b 1 . 3 I h h IRS d b k y American Consu ting Corporation v. U.S. were t e seize a an account owned by the plaintiff whose employment status was being ques tioned. The determination whether a worker is an employee or an independent contractor is far more reaching than simply the employment taxes in volved. The reclassification of an independent contractor can have adverse effects on a firm if it has a qualified pension, profit-sharing,

PAGE 16

7 or stock bonus plan. Such a reclassification may cause the plan to violate the minimum participation requirements. This could cause the disallowance of deductions taken for contributions to the plan and "bl 32 possi y even taxation of the trusts income. This would most likely result in termination of the plan by the employer because of the increased expense or at best the effect would be to reduce the benefits available to the employees. Workers who consider themselves self employed and establish Keogh plans for their own retirement may find that prior contributions are disallowed or the plan frozen with respect to any future tax exempt contributions. Reclassification of workers as employees instead of independent contractors affects various other fringe benefits. Employees exclude from income any contribution paid by their employer for accident and 33 34 health plans, employee education assistance plans, certain employerprovided life insurance and death benefit programs, 35 and the value of 36 any meals or lodging furnished for the convenience of the employer. Independent contractors are not afforded these privileges. In addition, other sections of the Code that apply specifically to employees would be affected by any reclassifications by the IRS. For example, the WIN tax credit for wages paid to certain employees hired under the 37 Federal Work Incentive program and the targeted jobs credit for wages 38 paid to certain new employees are all examples of areas of the tax law affected by reclassifying independent contractors as employees. Research Methodology Research Questions The four major research questions that are studied in this

PAGE 17

8 dissertation are: Research Question I. What factors or variabl es hav e been used 1n determining employment status? Research Question 2. In constructing a mathematical model of the courts' decision-making behavior, which of the variables espoused in Research Question I prove to be significant? Research Question 3. In determinin g the employment status of the cases examined, do the different courts of original jurisdiction (District Courts versus Court of Claims) use similar factors in the rendering of decisions? Research Qu e stion 4, Data Collection a) Do th e variables determined in Research Question 2 hav e any similarity with the criteria in the current proposals before Congress? b) What possible r e commendations might these results suggest? The Regulations specifically state that a common law employee is to 39 be determined according to the facts of each case. The facts of a case are determin e d in the original court of jurisdiction. In this type of study, the assumption must be made that because of the possibility of appeal to a high e r court, judges attempt to includ e in their opinions a discussion of th e r e levant facts in their decision-making process. With this assumption, the Lexis computer data bank was used to search for all employee versus independent contractor cases. 4 From Lexis and other manual res e arch methods a total of 179 c ases were id e ntified for the 1940 through 1979 period. The year 1940 repres e nts the first year in which a significant number of cases w e re tried that dealt with the employment status question in connection with f e deral employment taxes.

PAGE 18

The study also includes all the cases that could be found through the end of the 1979 calender year. 9 Of the 179 cases, 24 were tried with a Jury. In this situation, the written opinions contain only the instructions to the jury and the actual verdict. Therefore, these 24 cases were eliminated from the study. Sixteen cases also were eliminated because there was not sufficient information in the opinion to determine what factors the judge considered in his decision. This leaves a total of 139 cases of which I 19 were tried in the District Courts and 20 were tried in the Court of Claims (for a list of cases see Appendices A and B). A small number of cases were discovered at the appellate level that had not been reported at the District Court or Court of Claims level. These cases were also excluded because the facts of a case are determined in the court of original jurisdiction. Research Question I The sources examined to determine the relevant factors in deciding employment tax cases for federal tax purposes include the following: ( I) the Internal Revenue Code, (2) the corresponding Treasury Regula. ( 3) 1 k d. 1 d . 4 I d (4) BNA T tions, certain andmar JU icia ecis1ons, an a ax Management portfolio entitled Employee Defined. 42 Another source that proved helpful was the Employment Tax Procedure audit guidelines 43 published by the IRS for their agents. The list of factors that was compiled was reduced as the court cases were read and it became evident that certain factors seldom were discussed in the court opinions or that two variables were so highly correlated that they were combined and represented as only one variable. A list of the complete set of

PAGE 19

10 variables used in this study 1s contained in Chapter III with additional explanation of the selection process, operational definitions of the respectivevariable~ and how they were measured for this study. Research Question 2 In constructing a mathematical model of a court's decision-making behavior, three different approaches were taken. First, a discrimi nant analysis model was built using the variables espoused in the results of Res e arch Question I and the model was us e d to classify the court cases as either an employee or an ind e pendent contractor outcome. The P7M program of the Biom e dical Comput e r Programs P-S e ries (BMDP) 44 computer softwar e packa g e was used to derive the discriminant function. From this discriminant function, classification functions for each of the two groups w e re derived. Each case was then assi g n e d to the g roup in which the hi g hest probability o f membership existed. A frequently recommended technique to t e st th e accuracy of the classification function 1s that of using a holdout sampl e However, with the limited number of cases available in this stud y an alternate 1 d 45 test was app 1e This techniqu e calculat e d n-1 discriminant functions where n is e qual to the numb e r of cases in the sample. The procedure involved calculating a discriminant function with all but one case, and then that function was used to classify the left out case. Th e proc e dure was r e p e ated n-1 tim e s, a nd th e numb e r of incor rectly classified cas e s was used as an estimate of th e probability of misclassifying an employe e versus ind e pend e nt contractor case. Second, an ordinary least squares (OLS) regression model was built using the variabl e s espoused in the results of Research Question I.

PAGE 20

11 The OLS model was used to generate a probability of an employee finding versus an independent contractor finding for each case. These probabilities were used to classify each case. These classi fications were compared with those obtained from the discriminant model to determine if these two linear statistical techniques generate different results. Certain statistical assumptions are violated by using either of the preceding linear modeling techniques. Therefore, the third approach taken was to use logit analysis which is a non-linear es timation technique. Logit analysis assumes that the underlying func tional form of the relationship being studied is curvilinear and estimation of the model allows for interaction of the independent variables. The logit model involves a transformation of the depen dent variable from a binary to a continuous random variable. The 46 logit model was estimated using the Nerlove and Press and the BMDP stepwise logistic regression programs 47 which generate maximum liklihood (ML) estimates. The ML estimates were used to generate probabilities for each case by which they could then be assigned to either the employee or independent contractor groups. These results were compared with the discriminant analysis and OLS re gression results to determine if linear and non-linear estimation techniques produced different results. Research Question 3 To determine if the District Courts and Court of Claims apply different factors 1n deciding an employee versus independent contrac tor case, the first approach used 1n this study was to build a

PAGE 21

12 discriminant function using only the District Court cases. The resulting classification functions were used to classify both the District Court and Court of Claims cases. If the misclassification rates are significantly higher for the Court of Claims cases, this will be taken as evidence that the two courts use different decision making models. The second approach used to determine if the District Courts and Court of Claims use different decision-making models was to build two additional OLS regression models. The first model was based exclusively on the District Court cases and the second model 48 was based exclusively on the Court of Claims cases. The Chow test was used to determine if the two regressions came from the same population. The results of this test along with the results of the discriminant analysis test were used to draw conclusions pertaining to the similarity of the two courts with r e spect to deciding connnon law employee versus independent contractor cases. Research Question 4 The preceding research questions were designed to discover which variables are significant to the courts 1n deciding the employee versus independent contractor question. The results are based on the facts of the cases as discussed by the respective judges. The logical extension of this research is to determine if the variables found to be si g nificant in Research Question 2 are relat e d to the criteria contained in the proposals before Congress. If the criteria are quite different from those determined in this study, tax policy makers may want to reconsider those variables that have been suggested

PAGE 22

13 in order to provide a more certain definition of an independent con tractor. If the results of this study produce variables that have been consistently used by judges in rendering decisions in employ ment status cases, it would seem that these variables would be a good starting point for legislation. Therefore, the factors ob tained from Research Question 2 were compared with the current proposals before Congress. Similarities were noted and reconunen dations made of additional factors that Congress may want to con sider in the actual drafting of legislation in the employment tax area. Limitations This study attempts to identify those factors that a diverse set of judges have consistently used in deciding the employee versus independent contractor question. To do so, a research methodology was used that provides for a more systematic approach than the legal analysis that is generally used to study problems of this type. While the advantage may be that the approach is more objec tive than legal analysis which depends on the researcher's own heuristics, it also suffers from certain limitations. The greatest limitation of this type of study is the reliance on the fact that the judge has actually included in his written opinion all of the significant variables he used in his decision process. If a sig nificant variable is not included 1n the written opinion, it can not be measured and considered in this type of study. This problem is somewhat mitigated by the fact that the judge knows his opinion is subject to reversal for failure to consider all relevant facts.

PAGE 23

14 The ability to reduce the data in the written opinions into a numerical form also presents some problems. In an attempt to re duce the subjectivity involved in actually obtaining the data, a variable generally was not assumed present unless it was explicitly noted in the text of the case. Each case was read several times by the author in an attempt to be consistent in the actual coding of the variables. It should also be noted that because of the dynamic structure of the tax law and the legal process, that the results of this study are not predictive in a future sense unless one assumes stability for all the factors affecting the employee versus independent con tractor question. The fact that the data are based only on cases that were actually litigated also produces a potential sample bias. The majority of cases that are actually litigated are cases in which both parties feel they have a reasonable chance of winning. It is not clear if this bias is very crucial because it is those very situations where the decision is in doubt that this study is in terested in investigating. Summary of Chapter Contents This study 1.s divided into five chapters. Chapter I is the introduction to the study which includes a brief discussion of the nature and scope of the problem, the research questions to be addressed, the basic research methodology to be used, and possible limitations of the study. Chapter II provides a historical perspective of the common law definition of employee by tracing its use through the Code and

PAGE 24

15 corresponding regulations. This historical background is supple mented by a discussion of the current legislative status of common law employees and also proposed legislation that is currently being considered by Congress. Chapter III presents a more detailed outline of the research methodology that was used. Chapter III also examines Research Question I which outlines the selection of the variables used in the study along with their operational definitions. Chapter IV describes the discriminant, OLS regression, and logit analyses that were used to model the courts' decision-making behavior. The empirical results of each of the three techniques are analyzed. The differences and similarities of the three techniques are then evaluated. Chapter V summarizes the preceding chapters. In addition, implications of this study are discussed and reconnn e ndations for future research are made. Notes 1 smith, Ind e p e ndent Contractor or Employ ee ?--That is the Question, 33 N.Y.U. INSTITUTE ON FEDERAL TAXATION 578 ( 1975). 2 JOINT COMMITTEE ON TAXATION, ISSUES IN THE CLASSIFICATION OF INDIVIDUALS AS EMPLOYEES OR INDEPENDENT CONTRACTORS 7 (1979). ]GENERAL ACCOUNTING OFFICE, TAX TRE A TMENT OF EMPLOYEES AND SELF-EMPLOYED PERSONS BY THE INTERNAL REVENUE SERVICE: PROBLEMS AND SOLUTIONS 5 ( I 9 77) 4 s. REP. No. 1255, 80th Cong., 2d S e ss., p. 7 as cit e d in M.F.A. Mutual Insurance Co. v. U.S., 314 F. Supp. 595 (W.D. Mo. 1970). 5 National Labor Relations Board v. Hearst Publications, 322U.S. 120,121 (1944).

PAGE 25

6 see the list of cases in Marmoll, 391 TAX MNGM'T (BNA), 'EMPLOYEE' DEFINED, pp. c-6 thru c-10 (1979). 7 GENERAL ACCOUNTING OFFICE, supra note 3, at 13. 8 Id. at 23. 9 H.R. 3245, 96th Cong., 1st Sess., 125 CONG. REC. E 1342 (daily ed. March 27, 1979) (remarks of Rep. Gephardt). IOGENERAL ACCOUNTING OFFICE, supra note 3. 11 Hearings on H.R. 3245 Before the Subcorrnnittee on Select Revenue Measures, 96th Cong., 1st Sess. 5 ( 1979) (statement of Donald C. Lubick). 12 I.R,C. Section 3102(a). 13 I.R.C. Section 3402(a). 14 I.R.C. Section 3lll(a). IS I.R.C. Section 3301. 16 I.R.C. Section 604l(a). 17 I.R.C. Section 1401. 18 I.R.C. Section 6015. 19 I. R. C Sections 14 0 I 3 IO I and 3 I l I 16 20 I.R.C. Section 6205(a)( I) and Treas. Reg. Section 31.6205l(a)(J)(iii). 21 I.R.C. Section 620S(a)( I) and Treas. Reg Section 31.6205l(a)(J)(i). 22 I.R.C, Section 6157(a). 23 I.R.C. Sections 6601,6651, and 6653. 24 I.R.C. Section 3402(d). 25 I.R.C. Section 6521(a). 26 I.R.C. Section 7442 and Shaw v. U.S. 331 F. 2d. 493 (CA 9th 1964). 27 see Marv e l v. U.S., 548 F. 2d. 295 (10th Cir. 1977) as cited in JOINT COMMITTEE ON TAXATION, supra note 2, at 8.

PAGE 26

17 28 North, The Employment Tax Morass, 11 CREIGHTON LAI~ REVIEW 797 ( 1978). 29 I.R.C. Sections 6321, 6331, and 6862. 30 JOINT COMMITTEE ON TAXATION, supra note 2, at 9, 31 311 F. Supp. 715 (W.D. Pa. 1970). 32 I.R.C. Section 404(a)(S). 33 I.R.C. Section 106 34 I.R.C. Section 127. 35 I.R.C. Section 79. 36 I.R.C. Section I 19. 37 I.R.C. Sections 40, SOA and SOB. 38 I.R.C. Section SI. 39 Treas. Reg. Section 31.3121(d)-l(c)(3). 40 Lexis computer data bank is simply a hu g e data ba se con sistin g of several libraries, one of which is the Tax Library. In this library, amon g other things, all r e ported cas es since 1960 for the District Courts and all cases since 1942 for the Court of Claims are stor e d in its memory. Throu g h a determination of words and phrases common to a selected subject, th e researcher can obtain a listin g of all cas e s relevant to that issue. For th e years not contained in the L ex is system a manual search was performed to identify the relevant cases. 41 u.s. v. Silk, 331 U.S. 704 ( 1947) and Bartels v. Birm ingham, 332 U,S. 126 (1947). 42 Marmoll, supra note 6. 43 INTERNAL REVENUE SERVICE, INTERNAL REVENUE MANUAL CCH, ch. 4600, pp. 8463-1 thru 8463-5. 44 BI1DP BIOMEDICAL COMPUTER PROGRAMS P-SERIES 7 I I (W. Dixon & M Brown e d I 9 7 9 ) 45 Lach e nbruch, On Exp e ct e d Probabiliti es of Misclassifi cation in Discriminant Analysis, N e cessary Sample Si ze and a Re lation with the Multiple Correlation Coeffici e nts, 24 BIOMETRICS 823-834 (1968). 46 l . L 1 M. Ncrlove & 11. Pr e ss, Un1.var1.ate and Mu t1.var1.at e o 1.near and Logistic Models, RAND CORPORATION TECHNICAL REPORT R-1306-EDA NIH (1973).

PAGE 27

18 4 7 BMDP 44 5 supra note at 17. 48 chow, Tests of E ualit between Sets of Coefficients 1n Two Linear Regressions, 28 ECONOMETRICA 591 1960).

PAGE 28

CHAPTER II HISTORICAL BACKGROUND Development of Legislative History Social Security Act of 1935 With the devastating effects of the great depression on unemployment, attention was keenly focused on establishing the means by which such a calamity might be avoided in the future. 1 On January 28, 1934, a n Executive Order of the Presid e nt of the United States created the Committee on Economic Security. 2 On January 15, 1935, the Committee submitted its report which Congress promptly used as the substance for the drafting of le g islation. 3 4 The Social Security Act of 1935 came after long deliberation by both the President and the Congress on "the evil of the burdens that rest upon large numbers of our people because of the insecur ities of modern life, particularly old age and unemployment. 5 The original Social Security Act included eleven "Titles." Of principal concern for this study is Title VIII which created provisions dealing with old age benefits, Title IX which created provisions dealing with unemploym e nt compensation, and Title XI which provided the definitions to be used in the other ten Titles. The principal solution adopted by Titles VIII and IX was to provide for periodic payments similar to annuities to the elderly and to provide 19

PAGE 29

20 for compensation to workers durin g sustained periods of unemployment. The enactment of employment taxes was the method a dopted b y Congress to fund payments to the elderly and the unemployed. The primary purpose of this study is to determine the parties to whom the empl oyme nt taxes apply. Title V III of the original Social Security Act, provided for a tax on both the employee and employer of percent of the employee's "wages," Wages were defined in Section 811 of the Social Security A c t as .. all remuneration for emp loym e nt, in c ludin g the cash value of all remuneration paid in a n y m edi um other th a n cash; except th a t s uch t erm shall n o t include that part of the r emuneratio n which, afte r remuneration eq u a l to $3,000 has been paid to a n individu al by a n employe r with respect to employment during any calendar ye ar, is paid to such individual by suc h emplo ye r with respect to employment during such calendar year. (Emphasis added) __ One of the key t erms used in this definiti o n of wa ges was employ ment" and Section 811 d ef ined employment as "an y service, of what ever nature, performed within the United States by an employee or his employer." The unemplo y m e nt provisions of Title IX basicall y provided for an excise tax to be levied on employees of 1 percent of the employees' wages if the employer had eight or more employees The determination of whether an individual qualifies as an employee is critical in the application of the Title IX provisions. Title XI simply defines an employee as "an officer of a corporation ." The absence of an elabor.:ite definiti o n of emp loye e in either titlL' seems tL1 imply ; 1 ''I.Jillingness of Congress to allow the courts to Jctcnnin c t o whom the h ld 1 ,,6 statutes ou app y. Regulations were issued in 1936 by the Treasury that expanded the employee definition for both the old-age and the unemployment

PAGE 30

21 tax provisions. 7 The Regulations provide in part that within the meaning of the Ac t, any person is an employee if he is in the service of one or more employers .. for compensation An individual is in the service of an employer if he is subject to the continuing authori ty of the employer to supervise and direct the manner in which he renders services for compensation. It is not n ecessary that the employer actually direct or control the manner in which the services a r e performed; it is sufficient if the employer has the right to do so. The ri g ht of an employer to discharge an individual is also an important factor indicating that the individual is an employee. Other factors indicating th a t an indivi dual is an employee are th e furnishing of tools and the furnishing of a place to work b y the employer to the individu a l who p e rforms th e services. In ge n e ral, if a n individual is subjec t to the control or direction of an employer merely as to th e r es ult to be accompl ished by the work and not as to the means and methods for acco plishing the re s ult, he is a n independent contractor. An individual performing services as an ind epende nt contr acto r is not, as to such se rvices, an employee within the meaning of the Act 8 A comparison of this Regulation to the "definition" found in leading cases and legal authorities will disclose that the Regulation is substantially a brief statement of the common law definition of h d 1 h' 9 t e master an servant re ations ip. Common Law Definition of Master-Servant Relationship Maitland provides the following as a reasonable definition of the common law: The Common L aw (called also the unwritten law, or from its mode of development, th e customary law) emb rac es those rules of civil conduct which originat e d in th e common wisdom and experience of society, became in time established customs, and finally r eceive d judicial sanction a nd affirmance in the decisions of the courts of last resort. IO This system of common law was originally establish e d through the decisions reached by the English common law courts of Exchequer, 11 Common Pleas, and the King 's Bench as early as the year 1300.

PAGE 31

22 What helped promote the advancement of the common law was the increase in the recording of court decisions and the resulting increased availability of admissible legal precedent. 12 At first, court proceedings were circulated primarily only to judges and 13 select lawyers. But as the availability of court decisions became more widespread, citation of precedent also became more 14 prevalent by lawyers and more accepted by the courts. From this beginning, "the common law system was refined, worker over, modified, improved, and increased as the centuries passed until there developed an entire body of law. 1115 Even though the acceptance of the English system of law by the early American colonists was not immediate, the English system 11 d d h l . 1 f Am 1 16 eventua y 1 triump as t1e pr1nc1pa source o erican aw. With England's control over the original colonies, legal matters generally dealt with England and the English system. 17 This also resulted in the training received by the early American lawyers to 18 be heavily concentrated in English law. In addition to England's influence on the colonies, another significant reason for the acceptance of the English system was the fact that English was the primary language of Americans and the majority of early le g al writing was done in English. 19 For whatever reasons, the English system of law (including its common law) was the basis of the early American system of law. The common law is a "comprehensive body of law" providing a "comprehensive system of remedies. 112 Contained within this broad system of law is a body of case law and legal thought dealing with the master and servant relationship. The determination of whether

PAGE 32

23 or not a master-servant relationship existed was of primary impor tance under English common law in assessing the "liability of a master for a tort committed by a servant who was acting within the f h ,,21 scope o is employment. The question of a master-servant relationship also has become important in cases involving workmen 1 s compensation, contracts, and of primary importance to this study, the area of employment taxes. The 1936 Treasury Regulations dealing with employment taxes were promulgated pursuant to the common law definition of the master-servant relationship as defined by the American judicial system. An early example of the application of the common law definition of the master-servant relationship was in an 1889 Supreme Court case dealing with whether a corporation which had hired a commission salesman was liable for a negli ge nt act of that 22 same salesman. The salesman was furnished with a w ago n which was to be used exclusively in canvassing for s a les. 23 The salesman was convicted of driving the corporation's horse and wagon care lessly which resulted in personal injury to Katie Rohn as she 24 crossed a local street. The Supreme Court held that the salesman was a servant of the corporation and that such corporation was bl f K R h . 25 respons1 e or atie on s 1nJur1es. In arriving at this decision, the Supreme Court cited Railroad Co. v. H a nning as authority for th e definiti o n of .::i m ; 1 s t c rsc rv.::int r e lati o n s hip that follows: ... the relati o n of ma s t e r a nd servant ex ists when ever th e employer retains the right to dire c t the manner in which the business shall be done, as well as the result to be accomplished, or, in other words, "not onl y what shall be done, but how it shall be done. 1126

PAGE 33

24 In a later case, th e Su p reme Court u sed this sa me d ef inition in a situation dealing specifically with emplo ye e s a nd independent contractors. 27 In concluding that the work e rs involv e d in thi s case were independent contractors, th e Supreme Court held th a t the parties had the "liberty of action which excludes the idea of that control or ri gh t to control b y the e m p lo y er which cha r a ct e ri zes the rel a tion of employer a nd emplo yee and differentiates the emplo y ee or servant from the indep e ndent contractor 11 28 From this evolution of the common l aw comes th e co n cep t of the m as ter-s e rvant r e lati o n s hip that is cont a ined in the two Supreme Court decisi o n s di s cus se d previou s l y Even th ough th e term common law emplo yee is not specificall y mentioned in th e 1936 employment t ax Treasury Re g ulati o n s th e definition a ttri buted to an emplo yee b y th e Tre as ur y Departm e nt i s s imilar to th a t espoused b y th e tw o decisions di sc us sed abo v e These t wo d ec i s ions are mer e l y r ep r ese nt a ti ve of man y earl y deci si ons th a t u se t h i s same common l aw definition 29 Con se qu e ntl v a t t he ver y incep tion of the so cial sec urit y provi sio n s, th e n o ti o n of a common law emplo yee assumed a position o f im po rt a n ce as far as de t ermi nin g to whom the provisi o n s applied. Social S ec urit y Amendments of 1 939 The socia l sec urit y p r ov i sion s were ree x.il'li n e d fou r years after th e e na c tm en t of th e or i gi n al Soc i a l Securi t y A c t and pr i o r to the c o dificati o n of th e ex istin g ta x law int o th e 1939 Code The Social Securit y Hoard r ec omm e nded th a t th e coverage o f th e social s ec urit y pr ovisio n s be expa nd ed .J O Arthur J. A ltme ye r,

PAGE 34

25 Chairman of the Social Security Board, in hearings before the House Ways and Means Cormnittee recormnended that the social security provisions ... be expanded to the extent feasible to cover more of the persons who furnish primarily personal service. The intention of such an amendment would be to cover persons who are for all practical purposes employees, 31 but whose legal status may not be that of an employee. Using essentially the same language, the House adopted the Board's recommendation for the old-age insurance portion of the l . 32 socia security provisions. However, for some unexplained reason the House failed to include this expanded definition for the unem1 . 33 p oyment tax provisions. When the House proposal went to the Senate Finance Committee, the expanded definition of employee was deleted from the bill because it was "believ e d inexpedient to I I . 1 h. I 1. 1 34 c1ange t1e existing aw w ic1 imits coverag e to emp oyees. Therefore, the employee definition that was codifi e d in the 1939 Code was substantially unchanged from the original Social Security Act of 1935. Title VIII was codified into the 1939 Code as Sec35 tions 1400 through 1432, referred to as the Federal Income Contributions Act (FICA). The unemployment tax provisions were codi fied into the 1939 Code as Sections 1600 through 1611, referred 36 to as the Federal Unemployment Tax Act (FUTA). The 1939 FICA Treasury Regulations were also substantially unchanged from the language contained in the original 1936 employm e nt tax Treasury R 1 37 egu ations.

PAGE 35

26 Income Tax Withholding Provisions In 1942, in an attempt to raise additional funds to support the ff t C d h t k h V T 38 ware or, ongress enacte w a was nown as t e ictory ax This tax was a temporary income tax that was levied in addition to the normal tax and surtax The Victor y tax amounted to 5 percent of "Victory Tax net income," but was limited in as much as it could not result in a total tax greater than "90 percent of the individ 1 II 39 ua net income. Because of the substantial tax rates th a t existed during World War II, the Treasury suggested for the first time th a t incom e taxes 40 be withheld at the source. The Treasury recommended this procedure as a method of aiding taxpa yers in meeting their tax payments with a minimum strain. Greater equi t y and fairness is ga ined fo r th e g re a t body of income-tax payers by more uniform a ppl ica ti o n of the l aw, and th e administrative problem of collec ti on es pe cially in the case of taxp ayers who move from o ne localit y to another, will be mad e easier 41 The House adopted th e recommendation of the Treasur y and introduced legislation th at would require withholding of income taxes at 42 the source on di vide nds, bond interest, and wages As with the FICA and FUTA provisions, the question of whether the withholding requirements applied to a gi ven individual had to be answered. The House expressed the view th a t the common law rules which dev e loped largel y in the field of tort l aw w o uld be "unn ecessarily restri c tive if they stood a lone as the tests for withholding tax purposes. 1143 The definition of employee is acco rdin gly drawn to cover all common l .'.lw se rvc1nts, and in addi tion su c h o th er indi viduals performin g se rvi ces as are not in r e ality inde pend e nt businessmen or independent practitioners of a

PAGE 36

27 profession, whether or not they ar e "ind e p e nd e nt contrac tors" at common law. Thus under this definition an indi vidual may be an employee even though he is not subject to control. 1144 The Senate for reasons of "simplicity" and "ease of administra tion" required that collection at the source of payment be limited to sal aries, wages and other forms of compensation for personal services. 1145 The more s urprisin g result, and of greater importance to the tracing of th e common law definition of an employee was the fact that in th e Conference Committee the scope of the withholding provisions was limited to those individuals coming 46 under the common law definition of an employee. A year later, the definition of employee that became incor porated in the Victory Tax provisions was extended to the regular 47 income tax area by the Current Tax Payment Act of 1943. By so doing, Congress expanded the application of the common law employee definition from the social security tax area to the income tax area while at the same time resisting pressure to expand the definition of employee beyond that encompassed by the common law rules. Therefore, it was not until 1947 and the introduction of the 2conomic reality test that the common law definition of employee was again challenged. Judicial Interpretation Econ o mic Re<1litv Te s t. In 194 7, th e Supreme Court cons-id e rccl three federal employment tax cases which gave rise to what has been 48 termed the economic reality test." Of these three cases, the Supreme Court decided both the Silk and Greyvan Lines cases together on June 16 l 9 4 7.

PAGE 37

The Albert Silk Coal Company was in the business of selling coal at retail in the city of Topeka, Kansas. The question being considered by the Supreme Court in this case was whether workmen engaged in unloading railroad cars and truck drivers who made retail deliveries of the coal were employees for purposes of the employment tax provisions. The unloaders were paid an agreed price per ton to unload coal from the railroad cars. The men provided their own tools, worked when they desired, and were free to work for other employers. 49 When an unloader came to the yard 28 he was assigned a car to unload and told where it was to be unloaded. Some of the unloaders were quite regular workers while others were "fl d I d 1 . l 50 oaters an came to t1e yar son y intermittent y. Since Silk owned no trucks of his own, he contracted with individuals who owned their own trucks to deliver the coal at an agreed upon price per ton. The truckers were able to come and go as they pleased, wer e able to haul for others if the y d es ired, paid all their own operating expenses, furnished their own helpers when needed in delivering the coal, and were able to refuse to make d 1 ff l 51 certain e iveries without su ering any pena ties. The truckers were not instructed in how to perform their services, but simply were given a delivery ticket and told whether the purchase was cash 52 or charge. No record was kept of the truckdrivers' time and they settled up with Silk Co;-il Comp;-iny b y the trip, bv the day, or bv the week accorJing t1) the truckdr.i.v c c's wishL 'S Greyvan Lin es Inc., was a household-furnitur e truck e r operating in thirty-eight states and part of Canada. 54 Similar to Silk, the

PAGE 38

29 question that was decided by the Supreme Court was whether the truckmen who hauled the household furniture were em ployees under the employment tax provisions. These truckers w e re required to furnish their own trucks, haul exclusively for Greyvan Lines, pro vide any assistants needed to deliver the furniture, pay all oper ating expenses, carry the specified insurance coverage, drive their own trucks, and follow any regulations and instructions issued by G L 55 reyvan 1.nes. These instructions covered directions as to where and when to load and deli ve r the furniture. Compensation was d e termined on a commission basis with a bonus for successful completion of a job. The Suprem e Court held that the unloaders in the Silk case 56 were employees. The Court explained its decision as follows: (1) the tools the unloaders provided were only picks and shovels, (2) the y had no opportunity for gain or loss except through their own manual labor, (3) th e fact that many did not work r eg ularl y was insignificant, a nd (4) since th e unlo a d e rs did work in th e course of the employer s r eg ular tr ade or business, the y should d h l . 57 come un er t e emp oyment tax prov1.s1.ons. On the other hand, the Supreme Court held that the truckers b h d d 58 1.n ot cases were 1.n epen ent contractors. Even though on the surface, the description of the relationship of the truckers to their r e spec ti ve employers seemed quit e different, the Sup r eme Court's opinion did not di s tingui s h b e tw ee n the tw o cases with respect to the trucker's verdict. The Court indicated that eve n though the Greyvan Lin e truckers hauled just for one firm. this

PAGE 39

30 distinction was not controlling in the rend e rin g of its d e cision. 59 The Court held that the tru ckers were small businessmen who owned their own trucks and that "it i s th e tot a l situation, including the risk undertak e n, the control exercised, the opportunity for profit, from sound management, th a t m a rks th e se driver-owners a s independent contra c tors. 1160 The Silk and G r eyva n decisions themselve s a re not th e m os t significant factors o f th ese two cases. Despite the Con g res s i o n a l mandate that the common law rul es wer e to d e termin e if the e mplo y ment relation s hip existed i n other th a n the statutorily d efi n e d cases, 61 the Supreme Court in Silk opted for an expanded d ef ini tion of employee. As the fede r a l social secu rit y l eg isl a ti o n is a n a tt ac k on r ecognize d ev ils in our n a ti ona l eco n omy, a co s trict ed int erp r e tation of th e phr as ing by the co urts would not comport with its purpo se .... H e r e th e District Courts a nd th e Ci r c uit Courts of Appea l s d e t e mined th e ca~es lar gely i f n o t ind eed exclusive l y by app l y in g the so ca ll e d "c ommon law co ntr o l" t es t as th e c rit er i on This w<:1s cl ea rly wrong in v iew of th e Court's present ruling. 62 (Empha s is added) The Supr e me Court in Silk a l so li s t e d th e f ollow in g facto rs as import an t in d e t ermini n g whether e mpl oyme nt s t a tu s ac tu a ll y is present: (1) d eg r ee of co ntrol, ( 2 ) o pp o rtunities for profit or loss, (3) investm e nt in faci liti es (4) permanenc y of relation, and (5) skill requ i r e ct. 63 With r es pe c t to th ese fac tors, the Court s tat es th a t "n o one [f ac t or li s co ntr o llin g n o r is th e li s t complete. 1164 The S upr eme Court s op ini o n definitel y speaks of the d eg ree of control as simply o ne of several factors to be applied a nd n o t as the overall t es t.

PAGE 40

The Suprem e Court uph e ld both the Silk and Greyvan Lin es opini o ns a week later in a nother e mpl oy m e nt status case Bart e l s B . h 65 v. 1rm1ng am. This case deals with orchestras hired to pl ay for limited en gage m e nts; usually one-night stands The question being decided b y th e Supreme Court in this ca se is whether th e band leaders and other m e mbers of th e b a nd were the employees of the dance establishm e nts. These bands are "built a round a lead e r whose n a m e and distinctive style in th e presentation a nd r e ndition of d a nce music, is intend e d to g ive ea ch b a nd a m arke d individu ali t y 11 66 The l ea der hires a nd fires the musician s fixes their salaries, tells them what a nd h ow to play, a nd pays a ll of th e ir tr a nspor. d 67 tation an oper a tin g expe n ses It is also the b a nd l ea der who contracts with th e danc e es tabli s hm e nt s to play at a specified 68 pric e 31 The Suprem e Court held that th e b and leaders w e re the emp loy e rs of the individu a l mu sic i a ns and in turn th e band l eade r s were independent contractors with respe c t t o the dance estab lishm e nts. 69 In explaining its o pini o n, the Supreme Court stated th a t a b a nd leader organizes a nd tr ai ns the b a nd, selec ts the members, an d bears the risk of los ses that mi g ht occur. 7 o The r ela tionship between the b a nd l eade r a nd his orchestra is r e l a ti ve l y p e rm a n e nt, while th e relation s hLp b e tw ee n th e band a nd th e dan ce es t.1blish71 m e nt s i s o nl y t e mp orary In issuing th e Bartels o pini o n, th e Supreme Court cited th e Silks case and r es tated the notion th a t the c ommon l aw employee definition for employment tax purposes was "not to be d e t e rmin e d

PAGE 41

32 solely by the idea of control which an alleged employer m ay or could exercise over the details of the service rendered to his 72 business by the worker or workers." The Court goes on to suggest that determination of an employer-employee relationship is a matter f II 1 1173 A h s C h .. o economic rea ity. gain, t e upreme ourt was emp asizing that it is the total situation that controls, not simply the the common law control test. These three 1947 cases were not the fi r st instances in which the economic reality test was espoused by the Supreme Court Before the test was introduced into the employment tax area, the Supreme Court had already applied the economic reality doctrine N 1 L b R 1 B d IJ P bl 74 in ationa a or e ations oar v. 1earst u ications. In Hearst Publications, the Supreme Court held that n e wsbo y s were 75 employees under the National Labor Relations Act In interpreting the Act, the court stated that "Congress had in mind a wider field than the narrow technical legal relation of master and servant" and because of the broad language in the Act's definitions, there is "no doubt that its applicability is to be determined broadly, in doubtful situations, by underlying economic facts rather than technically and exclusively by previously established legal classif . ,,76 ications. After the economic reality test was introduced in the labor relations area, the Supreme Court held in the Silk case that "application of the social s e curity l e gi s lation s hould f o llow the sam e rules that we applied to the N : itional Labor Relation s Act ,, 7 7 in the Hearst case This then completes a bri e f history of the advent of the economic reality test and its integration into th e federal employment tax area

PAGE 42

Status Quo Amendment. With the success of the Supreme Court in altering the common law employee definition first in the labor relations area and then in the federal employment tax area, the Treasury Department quickly attempted to alter its employee defi nition by issuing Proposed Regulations in 1947. 78 These Regulations attempted to substitute the economic reality test for the common law control test in determining employment status for FICA 79 and FUTA purposes. These Regulations would have reduced the "right to control the performance of servic es by an individu a l" 33 to simply one of several factors that should b e examined in d e termining employment status instead of the overall t e st as e nvi s ioned by Congress in the original enactment of the social security pro visions. SO Congress viewed the Proposed Regulation s as a n attempt by the Treasury Departm e nt to usurp the authority of Congress in making 81 law. The issue involved in th e proposed r eg ul a ti o n s is whether the scope of social-security co ve ra ge sh o uld b e d e t e mined by Congress or b y other branch e s of th e Government ... acco rdin g l y under the propos e d r eg ulati o ns, th e question of co v e rage will b e d e t e rmined, n o t by C o ngress, but b y the Social Security Agency, the Tre a sury, and the courts 82 By joint resolution, Congress responded to the Proposed Regulations by passing legislation that would maintain the "st a tus quo" of the current statutes a nd r eg ulations and reaffirm th e l eg i sla tiv e intent of the original Social Security Act of 193 5 111e intent w as th;-it the usual common law rules, realistic a lly applied, would continue to be used to determine employment status for social security pur poses.83 The accompanying Senate Report stated emphatically that

PAGE 43

34 the "end point detennination" is whether th ere is an absence or existence of control and that whatever factors are pertinent to this determination should be used. 84 But the final determination is whether or not control actually exists. To further emphasize their desire to h ave the common law control test be the primary test of employment status, Congress added the following language to the FICA and FUTA provisions: ... such t e rm does not include (1) any individual who, under the usual common law rul es applicable in determining th e employer-employee rel a tionship, h as th e st a tus of an independ e nt contractor or ( 2 ) any individual ( ex cept a n officer of a corpora tig) who is not an employee under such common l a w rules. This is the first time that the common law definition of employee was officially included in the statute. Previousl y support for its application had been obtained either from legislative history of the social security provisions or the employment tax Treasury Regulations. Social Security Amen dments of 1950 Congressional s up port for the Status Quo Amendment was strong enough that it was passed over the veto of President Truman. 86 Representative Gearhar t explained the actions of Congress as follows: If thi s Congress had not interf ered t e n s of thous a nd s of peopl e in \mL ric.:1 who n eve r dr e~ 1m c t.! th e, \.Jt t:-c employed bv an,bod, and n e ver for o ne moment th o u g ht th ey were covered by soci.11 securit y o r subject to p.1:,roll tax es would have found that they h .:1d be e n swept into the social security system by bur ea u cratic ukase. In othe r words, th ey would suddenly hav e found th a t they h a d more employers than a dog has fleas. So, to end this confusion, Congress acted promptly and, after thorough-going debate, by a

PAGE 44

vote of nearly 7 to 1, proceeded by legislation to put the matter in order once again by restoring the ancient doctrine of the common law defining the relation of master and servant, employer and employee. 87 Despite this strong Congressional reaction to the Treasury's JS Proposed Regulations, in the very next Congressional session the Social Security Administration recommended the repeal of the Status Q Am d d f h l 88 uo en ment an enactment o t e econom1c rea 1ty test. The House adopted the proposal and passed H.R. 6000 which would have incorporated the economic reality test as a general definition of an employee for social security purposes. 89 Also the bill stat utorily determined that the following seven occupational groups would be classified as employees for the withholding and payment of social security taxes: (1) an outside salesman in the manufac turing or wholesale trade, (2) a full-time life insurance salesman, (3) a driver-lessee or a taxicab, (4) a home worker, (5) a contract logger, (6) a mining lessee or licensee, and (7) a house-to-house salesman. 90 The Senate did not accept H.R. 6000 as it was passed by the House. Instead, the Senate deleted the economic reality test, reaffirmed the common law definition of employee, and reduced the number of occupational categories to certain agent or commission drivers and full-time insurance salesmen. 91 Therefore, the economic reality test was again specifically rejected by Con g ress for use in resolving the employm e nt status qu e stion. For FICA purposes, the same definition of employee that was enacted in 1950 92 has not been substantially changed to date. In the next section of this chapter, this definition is reproduced in its full text as Section 312l(d) of the Internal Revenue Code of 1954.

PAGE 45

36 Before analyzing the current legislative status of the common law definition of employee, it is appropriate to consider another relevant piece of legislation that was introduced by the Social Security Amendments of 1950. To this point, this entire histor ical background discussion has been concerned with the FICA and FUTA provisions as they relate to the overall social security package adopted by Congress. No mention has been m a de of the status of the self-employed worker who did not come under the social security provisions discussed thus far. Another significant aspect of the 1950 amendments was the extension of the social security provisions to self-employed individuals. 93 The self-employment tax provisions were codified in the 1939 Code as Sections 480-482. 94 Section 482(b) states th a t subchapter E (which contains Section 480-48~ may be cited as the Self Employ ment Contributions Act (SECA). The original SECA provi s ions pro"d d f lf 1 ,, 95 vie or a tax on se -emp oyment income. Se lf-emp loymen t income was defined as "net earnings from self-emplo yme nt derived by an individual" except (1) those net earnings from self-employment that are less than $400 and (2) those net earnings from self employment in excess of $3,600 (which first had to be reduced by any wages receiv ed that h ad been subjected to the FICA t axes und e r Section 1426(b) of the 1939 Code). 96 d b 1 d" "d 1 1 d d 97 ness carrie on y suc1 in ivi ua ess certain e uctions. The term trad e or business was defined to h;Jve the same m ea ning as it did in Section 23 of the 1939 Code which dealt with the

PAGE 46

37 ordinary and necessary expenses incurred in connection with a trade or business that could be deducted in arriving at net income. 98 Section 481(c) contained an except clause that applied to the Section 23 definition of a trade or business which resulted in the exempting of certain activities from the self-employment tax. The most important 48l(c) exception for this study was the "per formance of services by an individual as an employee. 1199 Section 48l(d) referred to the FICA provisions to define the term employee. The FICA definition of employee, as has been discussed in considerable detail, means a common law employee. With this link to the FICA provisions, Congress had effectively extended the coverage of the social security provisions beyond merely employees and provided the means by which self-employed persons would now be subject to employment taxes. Sections 1401-1403 are the 1954 Code counterparts of Sections 481-483 of the 1939 Code. In substance, the self-employment tax provisions have been altered slightly in the past 30 years, but 11 h . h 100 essentia y t eir import is t e same. The most salient aspect of the SECA provisions for this study is that because of their construction, they provide a clear distinction between how an employee and and independent contractor will be treated for employ101 ment tax purposes. It is this distinction that has contributed to the increased importance of being able to determine when a common law employer-employee relationship actually exists.

PAGE 47

Current Legislative Status For FICA purposes, the Internal Revenue Code of 1954 defines an employee substantially as it was defined after the 1950 amend ments. Section 312l(d) defines an employee for FICA purposes as: ( I) any officer of a corporation; or (2) any individual who, und e r the usual c o mmon law rules applicabl e in determinin g th e e mployer-employee rela tionship, has the status of an employe e ; or (3) any individual (other than an individual who is an employee under para g raph ( I) or (2)) who performs services for renumeration for any person-(A) as an a g ent-driver or commission-driver e n g a g ed in distributin g meat products, ve ge tabl e produ c ts, bakery products, bev e ra g es (other than milk), or laundry or dry cleaning services, for his principal; (B) as a full-time life insurance salesman; 38 (C) as a home worker p e rformin g work, accordin g to sp e cifications furnished by the person for whom the services are perform e d, on materials or g oods furnish e d by such a person which are r e quired to be returned to such person or a person designated by him; or (D) as a travelin g or city sa lesman, other than as an a ge nt-driver, e n g ag e d upon a full-tim e basis in the solcitation on b e half of, and th e transmissi o n to, his principal (exc e pt for side-line sales activities on behalf of som e other person) of ord e rs from whol e sal e rs, retail e rs, contractors, or operators of hot e ls, restaurants, or other similar e stablishments for merch a ndis e for resal e or suppli e s for use in th e ir business op e rations; if th e contract of service cont e mplated that substantially all of such servic e s are to b e p e rformed p e rsonally by such in dividual; e xc e pt that an individual shall.not be includ e d in th e term e mployee" under the provisions of this para g raph if such individual has a substantial investment in facilities used in conn e ction with the performanc e of such services (other than in facilities for transportation), or if th e services are in the nature of a sin g le transacti o n not part of a continuin g r e l a ti o nship with the person for whom the services are performed. (Emphasis added) It 1.s Section 3121(d)(2) dealin g with common law employe e s that 1.s of principal interest in this study. A common law employee is one of the classes of employees that is subj e ct to the FICA withholding and payment requirements. However, all six categories of employees defined 1.n paragraph (d) are subject to the if clause

PAGE 48

that follows the Section 3121(d)(3)(D) definition of a traveling or city salesman. Ignoring the language not pertinent to this study, paragraph (d) could be paraphrased as follows: an individual 1s treated as an employee for the FICA provisions if he is considered 39 an employee under the usual common law rules applicable in determining the employer-employee relationship if the contract of service contemplates that all the services would be performed personally by the colllIIlon law employee in question; except if such common law employee has a substantial investment in facilities used in con nection with the performance of such services or if the services are in the nature of a single transaction and not part of a con tinuing relationship with the person for whom the services are performed. This interpretation of the statute would mean that a common law employee is subject to the FICA withholding and payment re quirements only if he satisfies the "if clause" discussed above and also does not come within either of the situations described in the "except clause." Even though personal performance of the services, substantial investment, and a continuing relationship are all factors that are considered in applying the usual common law definition of employee, this statutory construction places these three factors as additional requirements that must be satisfied before an adjudged common law employee can be subjected to the FICA provisions. In the 139 cases included in this study, the argument never was presented to the courts that a common law employee should be excepted from the FICA withholding and payment requirements due to

PAGE 49

the statutory exceptions contained in para g raph (d) of Section 3121. Marmoll concludes that "there has b ee n a tendency in the authorities to de-emphasize--primarily through silence--the statu tory exceptions set forth in Section 312l(d). 11102 Additional dis cussion is devoted to these statutory exceptions in Chapter III when the factors are defined that are used in building the mathe matical models developed in this study. The additional discussion in Chapter III is included in the definitions of factor 3 (right to delegate) and fa c tor 4 (continuin g r e lationship). Treasury Re g ulation S e ction 31.3121(d)-l(c) further elaborates on the term common law employee for FICA purpos es (c) Common l aw em ploy ees ( I) Every individual i s an e mploy ee if und e r th e usual common l a w rul e s th e r e lationship b e twe e n him and th e person for whom h e p e rform s services is the le ga l r e lationship of em ploy e r a nd e mpl oyee (2) G e n e rally such r e lationship e xists wh e n th e p e rs o n for whom servic es are p e rform ed h as th e ri g ht to c o ntrol and dir e ct th e individual w ho perform s th e services, n o t only a s 40 to the r e sult t o b e accompli s h e d by th e work but also as to th e d e tails and means b y which that result is acc om plish e d. That is, an employee is s ubj e ct to the will and control of th e employer not onl y as to what shall be done but h ow it shall be done. In this co nn ec tion, it is not n e c essa ry that th e emp loyer actually dir ec t or control th e m a nn e r in which th e s e rvic es are perform e d; it is suffici e nt if h e has th e ri g ht to d o so. The ri gh t to dis c h arge is also an i mpo rtant factor indicating that the person poss e ssing th a t ri g ht is an em plo ye r. Other factors charact e ristic of an e mplo yer but not n ecessa rily pr e sent in e v er y case, are th e furnishin g of tools and th e furnishing of a plac e to work, to th e individual who per f or ms the services. In ge n e ral, i f a n individual is subject to th e control or direction of anoth e r merely as to the r es ult to be ac complish ed by th e work and n o t as to th e m eans and methods for accomplishin g the result, h e is an ind e p e nd e nt contra c tor. An individual performing servic es as a n ind e p en d e nt contractor is not as to suc h se rvic e s an e mpl oyee und e r th e usual common law rul es Individu a ls such a s physicians, l awye r s d e ntists, v e terinarian s construction contractors, public st e no g raphers, and auctioneers, engage d in th e pursuit of an ind e pend en t trad e busin ess or prof e ssion, in which they of fe r th e ir services to the public, are independent contractors and not employees.

PAGE 50

(3) Whether the relationship of e mploy e r and employee ex ists und er th e usual common law rul es will in doubtful cases b e determined upon an examination of the particular facts of each cas e (Emphasis add e d) 4 1 The FICA Regulation pertaining to the common law d e finition of an employee stat e s that an individu a l is an employee for FICA purposes if a le ga l relationship of employer and e mploy ee exists under the usual common law rules. Th e le ga l r e lationship of employer and employee generally exists when the following factors are pr e sent: ( I) the right to control and dir e ct the individual who performs the services, (2) the ri g ht to discharge, a nd (3) the furnishing of tools and a place to work by th e em ploy e r. Treasury Regulation S ec tion 3J.312l(d)-l(c) further stat es that an individual p erfo rmin g servic es as an ind epe nd e nt contractor is not considered a common law employee. Instead of de f ining factors that are characteristic o f an ind epende nt contractor, th e Regulation simply g ives examples of several prof essio ns that are in the pursuit of an independent tr a d e and who offer th e ir services to th e public; such a s, lawyers, physician s d e nti sts e tc. Finally, the Regulation states that d e terminin g an emp l oye rem pl oyee r e la tionship under th e usual common law rules is a qu es tion of fact and must be det e rmin ed for ea ch case individually. While Section 312l(d) and its corr e spondin g Tr eas ury Regulation provid e some elaboration on the d ef inition of a common law employee for FICA purposes, th e in com pl e t e n ess of th e statutory :1U thority is 103 evident by th e substantial amo unt of liti gat i on in this area. It has been l ef t lar ge ly to the courts to d ete rmin e wh e n a c omm on law employer-employee relationship exists within th e bro a d guidelines contained in S ec tion 3121(d) and Tr eas ury Regulation Section 31.3121 (d)-l(c).

PAGE 51

For FUTA purposes, Section 33O6(i) states that "the term employee has the meaning assigned to it by Section 3121(d), except that subparagraphs (B) and (C) of paragr?ph (3) shall not apply." This means that with the exception of certain typ e s of home workers and full-time insurance sales persons, the FIC A and FUTA definition of employees are identical. For the withholding of income tax e s, Section 34Ol(c) simply provides: 42 For purpos e s of this ch a pt e r, th e t e rm "employ e e" includes an offic e r, e mployee, or el e ct e d offici a l of the Unit e d States, a Stat e or any political subdivision ther e o f or th e District of Columbia, or an y a ge ncy or instru mentality of any one or mor e o f th e for eg oin g The term "employee" also includes an officer of a corporation. The Treasury R e gulations for S e ction 33O6(i) and 34O1(c) are substantially the same as the FICA Re g ulations alr e ady r e produced above and consequently ne e d not b e included. So e v e n though Section 34O1(c) does not m e ntion connnon law e mploy ee specifically, the Section 3401 re g ulation that defines employee is synonomous with the other r eg ulations that do define a common law employee for FICA purposes. While "employ ee has not be e n d e fin e d for incom e tax withholdin g purpos e s as any p a rticular t y p e of "employee," the authoriti e s have m e rely assum e d, sub s il e ntio in most cas e s, that the word e mploy ee a s used in the income tax withholdin g statut e s m ea ns "common law em ployee". B e caus e o f this historical tr e atm e nt o f th e incom e t a x withholdin g provisions and, furth e r, bec a us e th e statu tory s c h e m e would seem to support s uch an infer e nc e it m3y bL' inf e rr e d th.:1t .in 11 Pmpl o y e e" fLH inc o m<' tax wi t h o ldin g purpo se s m c.:i ns .1 com m o n l .1 w L'm ploy < ~ ", a s th.1t phrase is d e fin e d in the FICA rc g ulation s 70 ~ Consequently, under the curr e nt e mployment ta x provisions, when a taxpayer, an administrator, or a court is tryin g to d e termine if an employer-employe e relationship actually exi s ts, unless

PAGE 52

specifically exempted by the statutes, such a determination must be decided under the common law rules. In litigation, the court generally consolidates the FICA, FUTA, and income tax withholding issues into one question and that is whether or not the worker is an employee of the plaintiff under the co!IIlilon law rules (Treas. Reg. Section 31.3121(d)-(l)(c)(3)). Revenue Act of 1978 With the increased enforcement of the employment tax laws by the IRS, many taxpayers have complained that proposed reclas sifications of their workers represent a change in the IRS's . 105 position. Some taxpayers have relied on private letter rulings, technical advice memoranda, and results of prior audits for deter mining employment status but nevertheless have found themselves the subjects of reclassifications by the Service. 106 Congress found it appropriate to provide interim relief to taxpayers involved in employment tax status conflicts with the IRS until it had adequate l h 107 time to reso vet e issue. As a result, Section 530 of the Revenue Act of 1978 was enacted to provide relief for the 1979 108 tax year. Congress was unable to resolve the employment tax status issue during 1979, so H.R. 5505 extended the relief pro visions of Section 530 until the end of 1980. 109 In ge neral, Section 530 accomplishes two primary objectives. First, S e ction 530(a)( I) provides that if the taxpayer did not treat an individual as an employee for any period ending before January I, 1980, and all necessary Federal tax r e turns were cor rectly filed in 1979, then the individual will not be treated as 43

PAGE 53

an employee unless there is no reasonable basis for such treatment. Section 530(a)(2) stipulates that a reasonable basis exists for not treating a worker as an employee regardless of the circumstances if the taxpayer has relied on any of the followin g : ( I) judicial precedent, published rulin g s, technical advice with respect to the taxpayer, or a letter ruling to the taxpayer, (2) a past Internal R e venue Servic e audit of the tax payer in which there was no assessment attributable to the treatment (for employment tax purposes) of the individuals holdin g positions substantially similar to the position held by this individual, or 44 (3) lon g -standing the reco g ni ze d practic e of a si g nifi cant se g ment of the industry in which such individual was engaged. Second, Section 530(b) prohibits the Tr ea sury Department from publishing any Re g ulation or Revenue Rulin g after the enactment of the Revenue Act of 1978 (November 6, 1978) d e aling with the employment status of any individual for purposes of e mplo yme nt taxes. With the int e rim relief provid e d by S e ction 530 (and extended by H.R. 5505), Con g r e ss has had time to consid e r many su gg estions as to the remedies for the complex issues that beset the employment tax area. The next section of this study will briefly introduce the principal recommendations that have b ee n presented to Con g ress. Current Propos e d Solutions General Accounting Office The General Accounting Office (Gi\O), at the request of th e Joint Committee on Taxation, conducted a study of the problems with ad ministration of the social security tax and the definition of a 110 common law employee. In rev1ew1n g the problem of classification

PAGE 54

of workers as employees or independent contractors, the GAO con cluded that the principal problem involved uncertainty in the application of the common law control rules. Numerous examples were discovered where the IRS is inconsistently applying the common 111 law rules in different geographical areas. The GAO report 45 states that the "major cause of the employee/self-employed contro versy involves those cases in which the reclassified workers operate a business separate from the one that the IRS considers to be the 112 employer. The primary r e commendation of the GAO is that a cl e ar cut test needs to be devised to introduce a g reater d eg ree of certainty into the employment tax area for both the taxpayer and the IRS. The approach tak e n by the GAO is to us e four basic t e sts in d e ter minin g whether a tru e separate business e ntity actually exists. If a worker can satisfy the following four tests, th e IRS will be prohibited from reclassifying the worker as an em ployee. The individual must: ( I) Hav e a principal plac e of busin es s other than that furnished by the p e rson or p e rsons for whom h e or sh e p e rforms or furnishes s e rvices; (2) Maintain a separat e set of bo ok s and r e cords th a t r e fl ec t all items of income and expense of his or her trade or busin e ss; (3) B ea r th e risk of sufferin g a loss and the oppor tunity of making a profit; and (4) H o ld him se lf or h erse lf out in his o r h e r own na me as s e l f e mployed a nd/ o r makes his or h e r s e rvices generally available to the public. The report ar g ues that the four tests are "more precise, easi e r to understand, and can be applied more accurat e ly and consistently than the common law rules. 11113 How e ver, failure to m ee t all four

PAGE 55

criteria will not result in automatic employee status. Instead, if the worker satisfies three of the four criteria this will still be a possible indication of self-employment and such a situation will revert to the regular common law rules. If only two of the four criteria are satisfied, then the worker will be presumed to 46 be an employee. The GAO concludes that the number of cases that will require litigation to determine common law employee status will be greatly reduced by this tJpe of an approach. In addition to the tests, the GAO proposal prohibits retro active determinations (in absence of fraud) if a busin ess (I) obtains annually from the individual whom the business classifies as selfemployed a signed certificate stating that the individual meets all the separate business entity tests, and ( 2) annually provides the IRS with the name and employer identification number or social security number of each such certificate signer. To increase the effectiveness of these criteria the report also proposes that the taxpayer should sign the certificate under penalty of perjury. Another problem perceived in the GAO study is the problem of double collection of taxes on the same income as mentioned in the introduction of this study. The GAO report proposes that the IRS amend Section 6521 to allow an offset against the employees' por tion of the FICA tax by the appropriate portion of the SECA tax 114 already paid by the worker. H.R. 3245; Safe Harbor Remedy A similar approach to the GAO proposal was introduced by Representative Gephardt on March 27, 1979, as H.R. 3245. 115 The overriding purpose of this bill is to preserve the independent

PAGE 56

contractor status of workers in industries that have traditionally enjoyed such status and are currently being reclassified by the 116 IRS. This bill, like the GAO proposal, does not replace the connnon law rules. Those workers who can not satisfy the five "safe harbor" criteria for independent contractor status will then come under the common law rules. The five safe harbor rules are similar to several of those proposed by the GAO. 47 (I) An indep e nd e nt contractor must control th e hours he works. This rul e implies a tw o pronged test. The worker must control not only the aggregate number of hours actually worked, but h e also must control substantially all the scheduling of the hours worked. (2) Ind e pend e nt contractors provide th ei r own place of busin es s. Renting f rom the person for whom the worker provides s e rvices would not be prohibited as lon g as the rent is reasonable. (3) An ind epen d en t contractor ex p e riences investment or income fluctuations. His compensation is neither fixed nor guarenteed. Substantial inv es tm e nts in assets introduc e th e risk of busin ess failure and again this is a characteristic of self-employment. (4) Th e worker, as an independent contractor, must p erform services pursuant to a written contract that spell out the emp loym e nt status and the tax consequences that r es ult. This t es t is fairly mechanical, but it a lerts both parties to their relationship for both business and tax purposes. (5) The person for whom the services are provided must file the required tax r e turns. The worker must re c ei ve the nec e ssary information to fil e his tax returns and the IRS must r ece iv e the n e c essa ry information to permit collection of th e taxes due. Th e satisfaction of the five safe-harbor criteria by a work e r would result in automatic independent contractor status. If the criteria could not be satisfied then the work e r would fall under the usual counnon law rules. This notion is corroborated by Rep. Gephardt

PAGE 57

1n the following summary of the purpose and application of H.R. 3245. This bill creates a statutory scheme that promises to end the confusion and uncertainty created by IRS enforce ment practices that have characterized employment tax status determinations for nearly a decade. It is re cognized, however, that the tests cannot govern all cases. Therefore, if any one of the tests is not met, the common law test will be applied to determine the worker's status as an independent contractor or an employee. 1 17 Department of the Treasury The Treasury's principal arguments against the GAO and Gephardt proposals are that 48 ( I) the taxpayer will be able to manipulate the tests simply by changing the form of the business relation ship without changing the substance, (2) the tests are unclear and difficult to administer, (3) the proposals would continue the need for the use of the common law rules in many situations, and (4) the Service does not feel that th e proposals will d e ter the large number of independent contractors who willfully fail to comply with the employment tax provisions. 118 However, the solutions offered by the Treasury take an entirely different approach to the situation because they view the primary bl 119 pro em as one of noncompliance. This assessment of the problem is largely due to a recent study conducted by the IRS where they concluded that 47 percent of the workers treat e d as independent contractors failed to declare any of their compensation for income 120 tax purposes and 62 percent paid none of the social security tax due. Since the Treasury considers noncompliance of independent contractors as the most serious problem, their solutions have nothing

PAGE 58

to do with the definitional aspects of the common law rules. Instead, the Treasury supports a flat 10 percent withholding on compensation paid to independent contractors. This aids in the collection process for the IRS and also provides more information with which they can monitor independent contractors and the taxes they should be paying. In addition to the 10 percent withholding, the Treasury Department has proposed that instead of a one dollar penalty per information return not filed (with a maximum penalty of $1~00), that the penalty should be increased to 5 percent of ( $ 0 . ) 121 payments not reported with a 5 m1n1mum penalty An interestin g note to the Treasury proposal is that the taxpayer who is paying a worker must still d e t e rmine if he is an employee and subject to the graduated withholding requirements or if he is an independent contractor and subject to the flat 10 percent withholdin g requirements. Th e issue of noncompliance is obviously an important concern, but the fundamental problem that still remains is th e need to differentiate between an independent contractor and an employee. H.R. 5460 The fact that both noncompliance and the lack of certainty with the employee definition arc important problems is m anifest in the compromise bill, H.R. 5460, that was submitt e d t o the 122 House Ways and Means Committ ee shortly before the end of 1979. Briefly, the compromise bill incorporated the five safe harbor criteria that are contained in H.R. 3245 and the flat 10 perc e nt 49

PAGE 59

withholding from compensation paid to ind epe ndent contractors as proposed by the IRS. The only exceptions to the 10 percent withholding requirem en ts are that no withholdin g n ee d occur from a person who ( 1) performs similar services for five or more unrelated service-r e cipients; or (2) certifies that withholdin g would be excessive; or (3) makes direct sales for a supplier who provides no renumeration other than a volume sales bonus. 123 Therefore, unless a person qualifies under one of the above exce tions, regardless of a person's employment status some sort of withholding will be applied to any compensation received. Until this or similar legislation is enacted, th e reli ef provisions of Section 530 of the Revenue Act of 1978 will continue to be in force awaiting Congressional e nactment of permanent l eg islation in the employment tax area. Notes I J. HUGHES, THE FEDERAL SOCIAL SECURITY TAX I ( 1941). 2 Id. 3 R. COMPTON, THE SOCIAL SECURITY PAYROLL TAXES 4 ( 1940). 4 1 Socia Security Act of 1935, ch. 531, 49 Stat. 620 ( 1935). 5 u.s. v. Silk, 67 S.Ct. 1463 (1947). 6 R. COMPTON, supra note 3, at 129. 7 ~10 Ar e Employees, Title 26 C .F.R. S ec 40 1 .3 ( 1 939) and Employ e d Individuals, Title C.F.R. Sec. 400.205 ( 1939). 8 1d. 9 R. COMPTON, supra note 3, at 130. IOF. MAITLAND & F. MONTAGUE,/\ SKETCH OF ENGLISII LEGAL IIISTOKY 213 (1915). 50

PAGE 60

p. I IC. KINNANE, ANGLO-AMERICAN LAW 284 (2nd e dition, 1952). 12 Id. at 281. 131d. 141d. 15Id. l 6Id. at 4 7 I. I 7 Id. l 8Id. l 9Id. 201d. at 28 I. 2 If. P OLLOC K & F. MAITLAND, THE HISTOR Y OF ENGLISH LAW, vol.2, 528 ( 191 I). 22 sing e r Manu f acturing Co. v. R o hn, 132 U. S 518 (1899). 2 J Id. at 5 2 I 24 Id. at 518. 25 Id. at 524. 26 15 W all. 649, 656 as cit e d in Sin g e r M a nu fa ctu r in g C o ., supra not e 22, a t 523. 27 Me tcalf & Eddy v. M itch e ll, 269 U.S. 514 ( 1925). 28 rd. at 52 I. 29 see fn. 18 in J. HUGHES, supra not e I, at 48. 3O H e arin g s R e l a tiv e to th e Soci a l S e curity Ac t A m e ndm e nts of 1939 B e for e th e Committ ee on Ways and M ea n s 76th Con g ., 1st S e ss. 8, 11 ( 1939) (statem e nt of Arthur J. Altm e y e r). 3 IId. 32 H.R. NO. 728, 76th C o n g ., 1 s t S cs s. 61-62 ( 1939). 33 Id. at 74. 34 s.REP. NO. 734, 76th Con g ., 1st S e ss. 75 (1939). 51

PAGE 61

52 35 sections 1400-1432 (Subchapter A), Chapt e r 9, Int e rnal R e venue Code of 1939. 36 s e ctions 1600-1611 (Subchapt e r C), Chapt e r 9, Internal Revenue Code of 1939. 37 smith, Ind e p e nd e nt C ontract or o r E mpl o y ee ?--Th a t i s th e Q u e stion, 33 NEW YORK UNIVERSITY INSTITUTE ON FEDERAL TA XA TION 581 ( 1975). 38 s. R EP. NO. 1631, 77th Con g ., 2d S e ss. (1942). 39 Id. at 8. 40 H e arin g s on Rev e nu e R e vision of 1942 Be for e th e Co mmitt ee on Ways and Means, 77th Con g ., 2d Sess. 81 ( 1942) (stat e m e nt of Randolph Paul). 41 H.R. NO. 2333, 77th Con g ., 2d S e ss. 1 4 -15 (1942). 42 1d. at 125. 43 Id. at 127. 44 Id. at 127-128. 45 s.REP. NO. 1631, supra note 11, at 165. 46 H.R. NO. 2586, 77th Con g ., 2d Sess. 56 ( 1942). 47 H. R NO 510, 78th C o n g ., 1st S e ss. 30 ( 1943); Curr e nt Tax Payment Act of 1943, ch. 120, par. 2, 57 Stat. 126 ( 1943). 48 u. s v. Silk, 67 S.Ct. 1463 (19 4 7); H a rris o n v. Gr e yv a n Lin e s, Inc., 67 S.Ct. 1463 ( 1947); and B a rt e ls v. B irmin g ham, 6 7 S. Ct. 154 7 ( 194 7) 49 u.s. v. Silk, supra note 5, at 1465. 51Id. 52Id. 53Id. 54 Id. 55Id. at 1466. 56Id. at 14 70 5 7 Id.

PAGE 62

58 1d. at 1471. 591d. 601d. 61 s e e previous discussion dealin g with the Social Security Act Amendments of 1939 and the introduction to the income tax withholding provisions. 62 u.s. v. Silk, supra note 5, at 1467,1472. 63 rd. at 1469. 641d. 65 67 S.Ct. 1547 ( 1947). 66 1a. at 1548. 6 7 Id. 68Id. 69 rd. at 1551. 701d. 71 Id. 72 ra. at 1550. 73 rd. 74 322 U.S. Ill (1944). 75Id. 76 rd. at 124,129. 77 U S v. Si 1 k supra note 5 at 14 6 8 78 12 Fed. Reg. 7966-69 (1947) as cited in Smith, supra not e 3 7, at 584. 79I d sold. 81 H.R. NO. 1319, 80th Con g ., 2d S e ss. 3 (1948). 82 Id. at 304. 53

PAGE 63

468. 83 Act of June I, 1948, Pub. L. No. 80-642, ch. 469, 62 Stat. 84 s.REP. NO. 1255, 80th Cong., 2d. Sess. II (1948). 85 Act of June I, 1948, supra note 83, a t 438. 86 North, The Employment Tax Morass, I I CREIGHTON LAW REVIEW 784 ( 19 78). 87 see fn. 2 1n Bonney Motor Expr ess v. U.S., 10 AFTR 2d 5233 (E.D. Va. 1962). 88 ttearin gs on H.R. 2893 Before the Committ ee on Ways and Means, 81st Cong., 1st Sess. 1087-1088 (1949). 89 tt.R. NO. 1300, 81st Cong., 1st Sess. 14-15 ( 1949). 90 Id. at 81. 91 s.REP. NO. 1669. 81st Con g ., 2d Sess. 95-97 (1950). 9 2 s ocial Security A ct A mendm e nts of 1 950 P ub. L. No 81-734, ch. 809, Sections 104( a ) and 205(a), 64 Stat. 500,536 (1950). 93 Id., Sections 211-213, at 502. 94 26 U.S.C. S e ctions 480-482, pp. 3556-3558 ( 1952). 95 rd., Section 480, at 3557. 96 rd., Section 4 8 J(b), at 3557-3558. 97 Id., Section 48l(c), at 3558. 98 Id., S ec tion 23(a)(l)(A), at 3295. 99 Id., Section 481(c)(3), at 3558. l OO h 1 f 40 1403 f h 954 Compare t e anguage o Sections I Io t e I Code with Sections 480-482 of the 1939 Code. 54 IOloutsid e of t he em pl oyme nt t ax ar e a, th e notion of e mployee can b eco me e qu a lly confusin g For exa mpl e, th e b enefits of d e ferred ta xa ti o n of in come were ex t e nd ed to self-employed individuals throu g h the Keo g h or H.R. 10 pl a n s For purposes of qualified p e nsion, profit-sharin g o r stock b o nu s pl ans a shareho ld e e mploy ee of a Subchapt e r-S corporation is inter es tin g l y not treated as an em ployee, but rath e r is subj ec t to th e limit a ti o ns imp ose d on th e self-employed r etireme nt pl a ns Sec ti o n 401 (a)( 17) On th e other hand, p a rtner s who are tr ea t e d a s s elf emp l oyed p e r so ns for purpo ses of retirement b e n ef it s m a y b e tr e ated as em pl oyees for purposes of Section I 19 ( ex clusion from gross income f or meals

PAGE 64

and lod g ing provid e d to an em ploy ee for th e convenience of the employer) according to Armstrong v. Phinney, 394 F. 2d 661 (5th Cir. 1968). 102 11armoll, 'Employe e Defined', TAX MNGM'T (BNA) No. 391, p. A-3 ( 1979). 103 For an example of the num e rous cases liti ga t e d in this area see Marmoll, supra note 102, at C-6 through C-10. 104 11armoll, supra note 102, at A-5. 105 S.REP. NO. 1263, 95th Cong., 2d Sess. 210 (1978). l0 6 JOINT COMMITTEE ON TAXATION, ISSUES IN THE CLASSIFICATION OF INDIVIDUALS AS EMPLOYEES OR INDEPENDENT CONTRACTORS 8 ( 1979). 107 S.REP. NO. 1263, supra note 105. 108 Revenue Ac t of 1978, Pub. L. No. 95-600, Sec. 530, 92 Stat. 2885 ( 1978). 109 Act of December 29, 1979, Pub. L. No. 96-167, 93 Stat. 12 78. I IOGENERAL ACCOUNTING OFFICE, TAX TRE ATMENT OF EMPLOYEES AN D SELF-EMPLOYED PERSONS BY THE INTERNAL REVENUE SERVICE: PROBLEMS AND SOLUTIONS 2 ( 1977). l I I Id. at 9. l 12 Id. at 19. I 13 Id. at 2 1 114 Id. at 45. l 15 H,R. NO. 3245, 96th Con g ., 1st Sess., 125 CONG. REC. E 1342 (daily ed. March 27, 1979) (remarks of Rep. Gephardt). 116 125 CONG. REC. H2l40 (daily ed. April 9, 1979) (remarks of Rep. Gephardt). 117 Id. at H2142. 55 118 tt e arin g s on H.R. NO. 3245 il e fore th e Subcommitt e e on Select Revenue Measures, 96th Cong., 1st Sess. 5 (1979) (statement of Donald Lubick). 119 1d. at 7. 120 1d. at 14.

PAGE 65

56 12 I Id. at 12. 122 H.R. NO. 5460, 96th Cong., 1st S ess ., 125 CONG. REC. H8769 (daily ed. September 28, 1979) (remarks of Rep. Rostenkowski). 123 rd. at H8774.

PAGE 66

Variable Identification CHAPTER III RESEARCH METHODOLOGY Research Question The purpose of Research Question I is to identify thos e factors or variables that are relevant in determinin g wheth e r or not a worker is a common law employee. Once these factors are identified, the court cases can be examined to determine which factors w e re mentioned by judges in their written opinions. These data will th e n be used to build the models need e d to answer the remainin g three research questions. The first sourc e of relevant factors was contributed by the h I appropriate statutory aut ority. To supplement the limited statutory guidelines a lit e rature review was conducted. The most significant work in the common law employee area as far as identify ing relevant definitional factors was a BNA Tax Mana ge ment port folio.2 The Tax Management portfolio h e avily cit e d both the IRS 3 4 audit manual and the Restatement of Agency (hereinafter r eferred to as the Restatement). Consequently, th ese two sources were studied in detail and are r e lied on heavily in the subsequent ex planation and definition of the variabl e s. 57

PAGE 67

Operational Definition of Variables. The following is a list of the variables that were obtained from the above mentioned sources. Each of these eleven variables is discussed in detail in the pages that follow. I. Supervision. 2. Integration. 3. Right to delegate. 4. Continuing relationship. 5. Set hours of work. 6. Having control over the place of work. 7. Independent trade. 8. Method of payment. 9. Payment of business and/or travel expenses. 10. Furnishing of tools and materials. II. Realization of profit or loss. I. Supervision. Supervision as defined in this study, centers on the ability of a would be employer to control "how" 58 a worker accomplishes his assignment. Even thou g h it does not use the term supervision, Treasury Regulation Section 31.3121(d)-J(c)(2) states that a common law employee relationship generally "exists when the person for whom services are performed has the right to control and direct the individual who performs the services, not only as to the result to be accomplished by the work but .::ilso as to the details and means by which that result is accomplished'' (Emphasis added). The IRS audit manual discusses supervision in terms of three separate criteria. The first criterion is referred to as

PAGE 68

59 instructions and is defined as-A person who is required to comply with instructions about wh e n, where, and how he is to work is ordinarily an employee. Some employees may work without receiving instructions bec ause they are highly proficient and conscientious workers. Th e instructions which show how to reach the desired result may be oral or written.5 The IRS definition of instructions closely parallels the earlier definition of supervision taken from the Treasury Regulations. Both sources refer to the notion that an em ploy er lS concerned with more than simply the end results of a job. Instead,an employer generally desires to control how the job lS to be performed. The IRS definition also relates the need for supervision to the amount of skill the worker possesses. This same point is made in B e n v. U.S. where the District Court stated that "the element of freedom from control in the manner of the performance of the work, emphasized 1n some decisions, loses much of the significance when the skill of the worker 1s relied upon 1n the accomplishm e nt of the par ticular task. 117 The second criterion is referred to by th e IRS as training and 1s defined as-Trainin g a p e rson by an experienc e d e mploy e e workin g with him, by correspondenc e by r e quir e d attendance at m ee ings, and by other methods indicates that the e mploy e r wants the services perform e d in a particular m e thod or manner. This is e sp e cially tru e if the trainin g is g iven periodically or at frequent int e rvals. An independ e nt contractor ordinarily us es his o wn m e thods and rec e iv es no trainin g from the pur c h.::t sc r o f hi s ser vi ces In f act, it is usually his m e thods which brin g him to th e attention of the purchaser.8 If the employer requires an individual to att e nd meetings or work with an experience d employee, this is further evidence of a desire 6

PAGE 69

to control how the work is to be accomplished. 60 9 In Mccombs v. U.S., salaried supervisors were employed to observe workers who had contracted to apply aluminum siding to residential housing. The purpose of the supervisor was to "instruct the workers in certain techniques and procedures" and "train people who had no previous experience. The Court of Claims held that implicit in this find ing is the desire of the plaintiff to control not only the result b 1 h f h . .,10 ut a sot e manner o ac 1ev1ng it. The ability to "direct the ord e r or sequenc e in which the work must be done" is the third criterion used by the I R S that relates directly to the supervision factor. This criterion is defined in the IRS audit manual as-If a p e rson must p e rform se rvic e s in th e o rd e r or sequence set for him b y th e employ e r, it s how s that th e worker is n o t f re e to follow his own patt e rn of work but must foll o w th e e stablished routin e s a nd sch e dul es of th e e ploy e r. Oft e n, b ec aus e of the n a tur e of an o ccupati o n, the employer e ith e r does not s e t th e order of the s e rvices or s e ts th e m infr e quently. It is suffici e nt to show control, howev e r, if he r e tains th e right to do so.II Again, the reasonin g is that if a purchas e r of someon e 's services is able to control th e order or s e quence that must be follow e d by the worker, th e n such an individual is controlling th e d e tails of how the job is to be accomplished and 1s assuming an em ployer role. The home improv e ment industry is an exa mpl e of how the super vision factor has been applied in d e cidin g e mploye e v e rsus ind e pendent contractor cas e s. A firm hir es sal e sm e n t o solicit hom e repair jobs. As the jobs are acquired, an "applicator" is g iven the job of affixin g th e roofing or sidin g mat e rials. Th e qu e stion is the applicator's employment status. The home improv e ment

PAGE 70

61 company generally has supervisors to keep track of the applicators progress and make sure the customers are satisfied with the completed job. The court has the task of determinin g the degree of control these supervisors have exercised over the applicators. In certain situations the supervisor only appears at the job site "f d ff d "f . 1 d 12 to veri y progress an o er a vice i it is so ic1te In these cases, the courts have held that the lack of supervision d d d f" d. 13 tens to support an in epen ent contractor in 1ng. On the other hand, some courts have held that the supervisors had significant control over the applicators and through considerable training and instruction by the supervisors these workers become competent 1 14 app 1cators. In such cases, the courts h e ld that the presence f . d 1 f' d. IS o supervision supporte an emp oyee in in g Security Roofing and 16 d. .. Construction Co. v. U.S. iscounted the notion that supervision is an important factor when the work is largely routine. Although it is true that the applicator is normally unsupervised, h e is check e d on from time to time .. If he needs instruction, as, for instance, in how to conserve material, he receives it. After this, the work bein g lar ge ly routin e supervision is unn ece ssary. I do not re g ard its absence of controlling significance. One can b e an em lo ee e v e n thou gh h e su ervises himself as to the d e tails. Emphasis added The presence or absence of supervision is a determination of fact that must be made by the trier of fact and is often a difficult distinction to make. The Restatement defin es the overall test that should be applied in determining a master-servant relationship is whether "a person employed to perform services in th e affairs of another and who with respect to the physical conduct in the performance of the services I l 18 is subject to the others control or right to contro

PAGE 71

Howev e r, th e R e stat eme nt th e n list s "th e ex t e nt of co ntr o l which, by a g reem e nt, the mast e r may ex e rcis e ov e r th e d e tails of the work" as on e of the criteria to b e us e d in a pplyin g th e above describ e d 19 test. It appears that the Restatem e nt considers control of th e physic a l conduct of th e worker as th e ov e rall test to be us e d in determinin g employm e nt status whil e it also consid e rs control ov e r the details of th e work to simply be o n e o f th e factors to b e looked at in applyin g the overall test. This a ppar e nt dual rol e is somewhat confusin g and is reasonably similar to th e question h 1 20 tat surrounded th e e conomic rea ity t e st. Th e economic r ea lity 62 test att e mpted to r e le g at e control fr o m th e o v e rall t e st that should be appli e d in det e rminin g a common law e mpl oyee to a l e ss e r role as simply a sin g l e f actor out of many th a t would b e us e d in d e finin g e mploym e nt st a tus. 2. Int eg r a tion. Wh e th e r th e servic e s p e r fo rm e d b y a worker are an int eg r a l p a rt of th e busin e ss is cl ose ly r e l a t e d to th e independ e nt-trad e f ac tor, variabl e numb e r s e v e n, to b e discuss e d lat e r. How e v e r, th e I R S a udit m a nual tr e ats int eg r a ti o n as a uniqu e factor-Int eg rati o n of th e pers o n's s e rvic es int o th e busin es s op e rati o n s g en e rally shows th a t h e i s s u b j e c t to dir e ction and control. In a pplyin g th e int e g r a tion t e s t, first d e t e r m in e th e sco p e and fun c t io n of th e b u s in e ss a nd th e n wh e th e r th e servic e s of th e individu a l a r e m e r g e d int o it. Wh e n t h e suc c e ss o r co ntinu a ti o n of a bu si n e ss d e p e nd s t o a n a ppr e ci a bl e d eg r ee u pon th e p e r f o rm a n ce o f c e rt a in s e rvi c e s, th e p eo p l e wh o p e r fo r m th o s e s e rvi ces m ust n e c es sa rily b e s ubj ec t t o a ce rt a in a m o unt o f co nt r ol by the own e r o f the busin e ss. 21 A w o rk e r is con s id e r ed to be a int e g r a l p a rt of til e taxpa ye r's busin es s if th e s e rv i c e s p e rformed by th e work e r hav e "b e en mold e d ., 2 2 into th e taxpa y er's o v e rall busin e ss as on e int eg rat e d op e ration.

PAGE 72

63 A contra example to the notion of bein g int eg rated into the business 1s found in Aparacor v. u.s. 23 which is a case dealing with the employment status of commission salespersons or distributors. Aparacor was a lar ge corporation in the business of designin g and selling women's apparel which contracted with various indivi duals to distribute and sell its products at r e tail prices. The distributors furnished their own office facilities, hired th e ir own assistants, and were free to engage in other work or the sale of other products. They w ere neither reimbursed for expenses nor train e d by Aparacor. The Court further held that the retail distribution performed by th e salespersons was not part of Aparacor's "regular business of desi g ning and supplying merchandise. 1124 Aparacor was distin g uished from the cases reli e d on by the IRS because the court held that those cases involved "t y pical (albeit temporary) master servant r e lationships, that is, the p e rformance of relatively simple labor, in the course of and as an inte gra l part of the employer's business. 1125 Entertainers are an example of individuals that have b ee n held to be an inte g ral part of the business for which th ey perform their services. 26 One Court relied on the fact that b e for e certain belly dancers were featured as part of th e loun ge show entertainment, the 27 busin e ss had been losin g money. After the belly dancers and Greek music w ere added the es tablishm e nt b eca m e very profitable. In a similar case, another Co urt held that g ogo d.Jnccrs were .: in integral part of th e busincss. 28 The Court finds that the e videnc e do c s not sup p o rt the tax payer's contention that the entertainment provided by th ese individuals was m ere ly ancillary and subordinate to th e main busin ess On the contrary, we must conclude that the go-go dancers were an inte g ral part of the business.29

PAGE 73

In some situations, integration 1s consider e d an important factor 1n employment status cases. Howev er in th e 139 cases (since several of the cases had multiple verdicts, the total number of decisions rendered was 148) used in this study, inte gration was not explicitly mention ed in 120 of the decisions. In twenty-six decisions integration was specifically mentioned 1n favor of an employee finding and in only two decisions was the lack of integration mentioned in favor of an indep e nd e nt contractor finding. It appears that ge nerally the int eg ration factor either does not apply to the situation or it is a difficult factor to apply. A businessman is ge nerally not go in g to hir e a worker to perform services that are unnec essa ry to th e succ essf ul run ning of his business. To determin e if workers are an inte g ral part of the busin e ss is a fact jud gme nt that must b e made by the court. Only where a court explicitly mentions inte g ration 10 a written opinion is it considered as evidence supporting ei ther an employee or ind e pendent contractor findin g for purposes of this study. 3. Right to De l ega t e. If a worker is able to hir e pay, and fire his own help ers this evidences a certain d eg re e of freedom characteristic of an ind epe ndent contractcr. Gn th e other hand, an employer-employee relationship 1s ge nerally indicated if a worker is un a bl e to d e l egate work to another person, or if th e purchaser of his services controls the hirin g and firing of as sistants. The IRS audit manual defines the "right to dele gate in t e rms of the freedom to "employ assistants." The right to d e legate is 64

PAGE 74

defined in the manual as-If th e s e rvic e s must b e r e nder e d p e rson a ll y pr e sum a bly the e mplo ye r is inter e st e d in th e m e th o ds a s well a s the r e sults. H e is inter e sted in not only the result but also the worker.30 The ability of a worker to employ his own assistants is defined in the IRS audit manual as-65 Hirin g supervisin g and payin g assistants by the e mployer generally sh o ws control over the m e n on th e job. S om times one worker may hire, supervi se and pay the other workmen. He may do so as th e result o f a contract under which he is responsible for only th e a ttainm e nt of a r e sult. In this cas e h e is an ind e p e nd e nt contrac tor. On th e oth e r hand, if he hires, sup e rvis e s, a nd pays workm e n at the direction of th e e mploy e r, h e m a y b e an employ ee actin g in th e capacity of a for e man for or representativ e of the employ e r.31 The Internal Rev e nue Code is not sp e cific about what consti tutes a connnon law employee. For FICA and FUTA purposes, Section 3l21(d) states that an individual is an e mploye e und e r the usual common law rules if the contract of s e rvic e cont em plated that substantially all of such services are to be p e rform e d p e r so nally by such individual. This "if" clause s e ems to imply th a t it is not enough for a person to be adjud g ed a corrnnon law employe e but in addition the contract must contemplate that th e s e rvices be performed personally by th e worker in qu e stion. Thus, accordin g to this statutory contruction, the FICA withholdin g and payment requirements cannot b e applied to a coJJU11on law e mpl oyee unless the ri g ht-to-del eg ate factor is first sa tisfi e d. It is interestin g to note that ev e n thou g h this statutory construction e xists no argument was ever made in th e 139 cases includ e d in this study to avoid the FICA requir e m e nts due to the fact that the contract involved contemplated personal performanc e of th e s e rvices. Marmoll

PAGE 75

concludes that "there has been a t e nd e ncy in th e authorities to de-emphasize--primarily through sil e nce--th e statutory exceptions set forth in Section 3121(d). 1132 The presence or absence of th e right-to-dele g ate factor was a factor that was r e asonably easy to e licit from th e cas e s read. 66 Of the 148 judgements made by the District Courts and Court of Claims, ninety-three decisions incorporated the ability to dele g ate by the worker as a factor in favor of an ind e pendent contractor finding. An additional fifteen d e cisions d e termin e d that the worker did not have the fre e dom to dele g at e work or hire his own assis tants which tended to indicate an e mployer-employ ee relationship. In the remaining forty d e cisions th e ri g ht-to-dele g at e factor was not relevant, or its effect was not det e rminabl e by the courts. 4. Continuin g Relationship. In gen e ral, this factor supports an employer-employee findin g if the relationship betw ee n the prospective employer and the work e r is found to be of a p e rman e nt nature. Exactly what f a ctors must be pr e s e nt to constitute a permanent relationship is not easy to pinp o int fr o m the statutory law or cases. The Restatement's explanation of this factor is rath e r cursory. "If the time of e mploym e nt 1s short, the worker is l e ss apt to subject himself to control as to d e tails and the job is more likely to b e consid e r e d his job th a n th e j o b o f the o n e e mployin g h ,.33 1m. The probl e m with this definition is that it do e s not g iv e adequat e g uidance as to what should be consid e r e d p e r m anent e ployment. If a work e r perf o rms s e rvices on a sporadic job-by-job basis with no on e job lasting over a few days, is this then

PAGE 76

considered a non-permanent relationship? If this sporadic type of relationship continu e s for several years does this then indi cate a continuing relationship? If the relationship is contractual by nature, how long must the contract extend for a permanent rela tionship to exist? These are all qu e stions that can arise wh e n attempting to use the continuing-relationship factor to aid in determining an employee versus independent contractor cas e The IRS audit manual defines this factor as-A continuin g r e lationship betwe e n an individual and the person for whom he performs s e rvic e s is a factor which indicates that an employer-employee r e lationship exists. Continuin g services may includ e work p erfo rm e d at fre quently recurring thou g h som e what irre g ular int e rvals either on call of the em ployer or whenever the work is availabl e If the arran ge m e nt contempl a t e s continuing or r e currin g work, th e relationship is consid e red p e man e nt e ven if the services are part-time, seasonal,or of short duration.34 The IRS is fairly clear in its interpretation of how to handle sporadic employment that lasts for a reasonable len g th of time. h . h ld . I )S h h D. T eir view is up e in Tapager v. Birrning1am were t e istrict Court held that even thou g h certain salesp e rsons were part-tim e workers because their relationship was regularly recurr e nt this was an indication of an employer-employee r e lationship. The Court defined the continuin g -relationship factor as-The relationship of an independ e nt contractor g enerally contemplates th e ob t ainin g of an ag r ee d e nd and usu a lly cont e mplates the obtaining of that end within a stipu lated p er iod of tim e Th e r e l a tionship of em ploym e nt ge n e rally contempla t es a continuous and ind ef init e rcndE'rin g of servic e s which rcl.:iti.on s hip is t erm in a ble either at the op tion o[ the employer or employee without contractual liability.36 This definition implies that if a definiteness as to the tim e of the termination of the relationship exists, then a non-continuing 67

PAGE 77

relationship exists and we have an independent contractor finding. On the other hand, a continuing relationship or employee finding is indicated by an indefiniteness as to the time of the termination of the relationship. A somewhat contrary finding to Tapag e r was indicated in 37 Silver v. U.S. This case involved the home improvement industry and the classification of an applicator of roofin g and sidin g materials as an employee or independent contractor. The applicator in question had performed several jobs during the three year period under question. The Court held that "permanency of relationship can hardly be said to exist or be a wei g hty element when each obligation was of comparatively short duration and the worker was 68 f h ff f 1 bl. .,JS ree to accept or reJect t e o er o a new or simi ar o igation. Therefore, when services are performed sporadically, but over a reasonable len g th of time, it is not completely clear how a court will interpret this situation in terms of the continuin g -relationship factor. It is interestin g to note that while the continuin g -relationship factor is mentioned in various sources (Restatement, IRS audit manual, and court cases), that none of these sources refer to any statutory support for this factor. The continuing-relationship factor is listed in Section 312l(d) as an exception to the application of the FICA provisions to a common law e mploy ee Para graph (d) states that if a person is adjudged a common law employee and satisfies the "if clause" discussed in conjunction with factor three (right to delegate), then such a common law employee will be considered an employee for the FICA provisions "except if the

PAGE 78

services are 1n the nature of a single transaction not part of a continuing relationship with the person for whom the services are performed" (Emphasis added). The statute quite clearly provides that a worker who 1s classified as a common law employee, but whose services are not part of a continuing relationship is excepted from the FICA withholding and paym e nt requirements which for all intents and purposes treats the worker as an independent contractor. As was mentioned previously, the exceptions of Section 3121 have not been argued in the cases included in this study and appear to be substantially ignored in th e application of the common law employee rules for FICA purposes. For purposes of this study, the continuin g -relationship factor was fairly prevalent in the written opinions of the cases read. In forty-seven decisions the absence of a continuing rela tionship was indicated which favored an ind e p e ndent contractor findin g In an additional sixty-ei g ht decisions the presence of a continuing relationship was found to exist which favored an employee findin g In the rema1n1ng thirty-thr ee d e cisions the courts did not mention the continuing-relationship factor in their written opinions. S. Set Hours of Wo rk. The ability of a worker to set the aggregate number of hours he will work and also determine when he will work those hours is usually an indication of th e independ e nce possessed by a selfem ployed person. Wheth e r a pers o n can control the hours he works is ge nerally a readily d e terminable fact. This is somewhat evidenced by the fact that of the 148 decisions con sidered in this study, only eight times did the court fail to 69

PAGE 79

consider setting the hours of work as a relevant factor in their written opinion. In the other 140 decisions ninety-nine times the ability to set one's hours of work was mentioned in favor of an independent contractor finding while in the other forty-one decisions the ability of the prospective employer to control the hours worked was considered to favor a finding of employee status. In regards to the ability to control the hours an individual works, the IRS audit manual provides that-The establishment of set hours of work by the employer is a factor indicating control. This condition bars the worker from being master of his own time, which is the right of the independent contractor. If the nature of the occupation makes fixed hours impractical, a require ment that the worker work at certain times is an element of control.39 70 40 In H.R. 3245 as proposed by Representative Gephardt, the first of the "safe-harbor" provisions is controlling the hours worked. In hearings before the subcommittee on Select Revenue Measures, Representative Gephardt proposed that this factor contain a two41 prong attack. First, in order to satisfy the safe-harbor provisions, a worker must be able to control the aggregate hours he works. Second, the worker must also be able to control the schedule upon which he works these hours. If a worker meets these two requirements then a reasonable degree of freedom appears to exist favoring an independent contractor finding. The real crux of this factor that has been mentioned by both the IRS audit manual and 11.R. 3245 is that: if a work L~ r c.:in control how much, when, and how hard he works, then this is a strong indi cation of self-employment. If on the other hand, a person has to work a 9 to 5 job due to his employer's desire, then the worker is not master of his own time.

PAGE 80

6. Having Control Over th e Place of Work. The general premis e of this factor is that independent contractors gene rally provide their own place of business while an employee is provided 7 I a place to work by his employer. This general distinction, however, is somewhat of an oversimplification. There are c e rtain kinds of work that must be done on the employer's premises, but such work need not be done by an employee. For example, an e l ec trician must come to the employer's place of business, but by no means necessitates that the electrician is an employee. The IRS audit manual defines controlling-the-place-of-work factor as-Doing the work on the emp lo ye r's pr em ises 1n its e lf is not control. However, it does imply that the emp l oyer has control, especia lly wh e n th e work is the kind that could b e done e l sewhere A p erson workin g in the emp loy e r's plac e of bu s in ess is physically within the emp lo ye r's dir ectio n and s up e rv1s1on. The us e of d es k space and t e l e phon e a nd steno g raphic ser vic es provided by a n em ploy e r plac es th e worker within th e employer's direction and supervision.42 According to this definition, the Service's position seems to be that a fairly strong indication of an employ e r-employee relationship exists if a worker p erf orms his s e rvic es on an individual's business premises. Howev e r, Marmoll suggests that the Service's current position is that "furnishin g a plac e of work and of supplies--while factors to be consid ered in determining an employer employee relationship--ar e not alone sufficient to estab lish the 1 . ,.4 3 r e at1on s h1p. Controlling the plac e of work is a factor includ e d in both the GAO proposal and H.R. 3245. In the GAO proposal, th e second of th e four basic tests which must be met to b e g uarant ee d

PAGE 81

self-employment status requires the worker to have a "principal place of business other than that furnished by the persons for h h h f f h 11 44 w om e ors e per orms or urnis es services. As stated, this test seems to eliminate anyone being automatically considered an independent contractor if they happen to work on the employer's premises. However, in further explanation of this test, the GAO proposal states that the principal place of business can be rented from the would be employer if the rent if fair and reaonable. 45 While the ability to rent from the individual to whom the services are rendered seems to provide more flexibility, the GAO proposal states clearly that using a person's own residence for the princi pal place of work will not satisfy this test. 46 Providing one's own place of business is also the second requirement of the safe-harbor criteria suggested in H.R. 3245. However, the bill does recognize that certain individuals, by the very nature of the services they perform, do not have places of 72 b 1 h' 47 business and should not e he d tot 1s requirement. For example, insurance agents, door-to-door salesmen, real estate agents, truckers, . 1 1 48 and loggers do not perform their services at a singe ocation. The notion of salesmen and truckers introduces another element into controlling the place of work. It is possible that a salesman may be required to work in a certain area and that truckers may be required to drive certain routes. Accordin g to the IRS audit manual, this type of situation is also evidence of controlling the place of work. 49 The IRS audit manual states that-Control over the place of work is indicated when the employer has the right to compel a person to travel a designated route, to canvass a territory within a certain

PAGE 82

time or to work at specific plac e s as required. In some occupations services must be performed away from the premises of the employer, for example, ernplo 0 ees of construction contractors or taxicab drivers. 5 In most situations, the place-of-work factor is fairly easy to determine. This partially explains why it was chosen by both the GAO and Representative Gephardt in draftin g proposals that 73 would hopefully add certainty to the common law employee definition. In the cases included in this study, the employer controlling the place of work was mentioned seventy-seven times in favor of an employee finding. In another forty-nine decisions, the worker was determined to control the place of work which favored an inde pendent contractor findin g In only twenty-six decisions did the courts consider the place-of-work factor irrelevant. 7. Independent Trade. Whether or not a worker is engaged in an independent trade or business from th e person for whom he is providing services is a very fundamental e lement of the common law control test used in defining employment status. The Restate ment lists as one of the facts to be determined in establishing whether a worker is an independent contractor or employee "whether or not the one employed is engaged in a distinct occupation or business. 1151 In discussing common law employees, the Tr e asury Regulations state that individuals performing services as independ e nt contrac52 tors are not common law employees. /\s an e xample of these types of individuals, the Regulation lists "physicians, lawyers, dentists, veterinarians, construction contractors, public steno g raphers, and auctioneers, engaged in the pursuit of an independ e nt trade 7

PAGE 83

busin e s s or pro fess ion 1153 (Emphasis a dd e d). Th e entir e idea of an independent contractor is clos e ly ti e d to th e existence o f a separate or ind e pendent trade or business. The GAO, in th e ir report ~o the Joint Committee on Taxation discussed in Chapt e r II, directed th e empl o yment status test a t determining whether a tru e s e parate busin ess entity actually exists. 54 From this study, the GAO conclud e d that the major prob lem centered around those situations where the worker had a sep arate trade or busin e ss from the prosp e ctiv e employer, but the IRS 55 still reclassifi e d the worker as an e mploye e B e cau s e of the GAO's belief that many of these sep a rat e busin e sses should ind e ed 74 by treated as independent contractors, th e ir proposal was structured such that a trade or business not only h a d to e xist, but that trade or busin e ss had to satisfy four b a sic tests to be consid e red a s 1 56 a se e mploy e d entity. Th e IRS audit m a nual provides substantial detail in how to determine if a worker is e n g aed in a s e p a r a te trade or busin e ss. The manual divid e s the ind e pendent-trad e f ac tor into th e followin g three s e parat e cate g ori e s: (a) Can th e individu a l providin g the s e rvic e s work for a numb e r of firms at th e sam e time? (b) D oe s th e individual make his services availabl e to the g en e ral public? (c) Is t he individual r e quir e d t o d e vot e his f ull tim e to th e pers o n for whom h e p e rforms th e se rvic e s? The audit manual d ef in e s th e s e parat e cat eg ori e s as-\.Jor kin p fo r mo r e th a n o n e f irm at a tim e A p e r s on who works for a number of p e rs o ns or firms at th e sa me tim e is ge n e r a ll y an ind e p e ndent contractor b e cause he is usually free from control by any of the firms. It

PAGE 84

1s possibl e however, for a person to work for a number of people or firms and be an employee of one or all of them. Making service available to ge n er al public. Th e fact that a p e rson makes his servic e s available to the general public usually indicates an ind e pend e nt contrac tor relationship. An individual may hold his services out to th e public in a numb e r of ways; he may have his own office and assistants; he may han g out a "shin g le" in front of his home or office; h e may hold business licens e s; he may be list e d in business directories or maintain business listin gs in tel e phon e dir e ct o ries; or he may advertise in newpapers, trad e journals, ma ga zines, etc Full tim e r e quir e d. If the work e r must devote his full time to the business of the em ploy e r, th e e mployer has control ov e r the amount of tim e th e worker spends workin g and impliedly restricts him from doin g oth e r gainf ul work. An independ e nt contr ac tor, on th e other hand, is fr ee to work wh e n and for whom he chooses. Full tim e do es not necess ar ily m ea n an 8-hour day or a 75 5or 6-day w ee k. Its meanin g may var y with the int e nt of the parti es the natur e of th e occupation, and customs 10 the locality. These conditions should b e considered 10 definin g "full time." Full -tim e services may be required e ven thou g h not specifi e d in w ritin g or orally. For example, to produce a required minimum volum e of busin es s may co m p e l a p e rson to d e vote all of his workin g time to that business; or he may not be p e rmitted to work for anyone else, and to earn a livin g h e necessarily must work full time. 58 It is very difficult to id entify the distinction b e tween working for more than one firm at a time and havin g to devote one's full time to one em ployer. Simply stated, if a worker works for other ernp loyers and is not required to answer to just one boss, a certain degree of independence is shown which is consistent with bein g self-employed. For exam ple, this type of indep e ndenc e is characteristic of a worker known as a gy psy cha s e r." ,\ g ypsy chas e r is an individu a l who contracts with a truck driv e r to unload his truck which may contain furnit ur e, fruit e tc. Gypsy chasers generally frequent truck stops and attempt to solicit jobs as

PAGE 85

truckers enter the stop. Generally, a gypsy chaser is engaged by many different truck lines in a sin g le day and is not considered to work full time for any one firm. 59 The courts have generally interpreted this as significant evidence in favor of an independ f d. 60 ent contractor in ing The occupation of a telephone solicitor is an examp le of a different situation where a worker performs services for one person and is required to work a set number of hours per day. In Lieb v. U.S., telephone solicitiors who worked for an exterminator company were classified as employees by th e IRS. Am on g other fac tors discussed by the Court was the fact that the solicitiors were required to work th e ir entire shi ft in the e mployer's office for 62 five or six days a week. Such a finding was evidence that the basic relationship between the tel e phone solicitors and Lieb was that of an employer-employee. The distinction between working for m ore that one firm at 76 a time, or offerin g s er vices to th e general public, is also v e ry slight. The nature of the services perform e d by a work e r apparently determines whether an individual can provid e his servic e s to th e public at large, or wheth e r he can provid e his services to a number of different firms, and thus whether an employer-employee relationship exists. For example, a steel consultant's services are of fairly limited application to the ge n e ral public and would there fore probably be limit e d to certain cn ~ inc c rin g or steel-related industrial firms. In contrast an insuranc e salesman's servic e s are of the nature that are more applicable to the general public.

PAGE 86

It is this type of service that the IRS audit manual is referring to when it states that a worker may advertise in newspapers, magazines, or be listed in the telephone directory. In reading 77 the court cases that comprise the data for this study, it was found that working for more than one firm and working for the general public measured the same basic factor, but simply applied in different factual situations. The independent-trade factor was considered present in the cases read for this study if it was manifest in any of the three separate categories as defined by the IRS. In the majority of the decisions (ninety-four), an independent trade or business was deter mined to exist In twenty-five decisions the courts held that the worker was not involved in a separate indep e ndent trade or bus iness. In the remainin g twenty-nine decisions, whether or not an independent trade existed was not a relevant factor according to the court's written opinion or if it was relevant it was inde terminable whether or not it favored an independent contractor versus an employee finding. 8. Method of Paym e nt. The method of payment is the seventh fact that the Restatement mentions should be considered in determining whether "one acting for another is a servant or an independent con tractor.1163 The Restatement fails to elaborate on what methods of payment distin g uish an independent contractor findin g from an em ployee finding. The IRS audit manual's definition, however, is quite helpful in adding some substance to this factor. The IRS audit manual defines the method of payment as -Paym e nt by the hour, week or month generally points to an employer-employee relationship, provided that this

PAGE 87

method of payment if not just a convenient way of paying a lump sum agre e d upon as the cost of doin g a job. The payment by a firm of re g ular amounts at stated intervals to a worker stron g ly indicated an employer-employee relationship. The firm assumes the hazard that the services of the worker will be proportionate to the re g ular payments. This action warrants the assumption that, to protect its investment, the firm has the ri g ht to direct and control the performance of the worker. It is also assumed in absence of evidence to the con trary that the worker, by acceptin g paym e nt upon such basis, has agreed that the firm shall hav e such ri g ht of control. Obviously, the firm expects th e worker to give a day's work for a day's pay. Generally, a person is an employee if he is guaranteed a minimum salary or is given a drawin g account of a specified amount at stat e d intervals and is not required to repay any excess drawn over COIIIl1lissions earned. Payment made by the job or on a strai g ht commission generally indicat e d that the p e rson is an independ e nt contractor. Payment by th e job includes a lump sum comput e d by the number of hours requir e d to do the job at a fixed rate p e r hour. Such a p a y me nt should not be confused with payment by the hour. 64 The explanation of the m e thod-of-payment factor provided by th e IRS indicates that the basic distinction is that e mploy ee s are 78 paid regular amounts at stated intervals regardless of the employee's output, whereas an independent contractor is paid on a lump-sum or per-job basis and his compensation does depend on his completion of certain requirements. This basic distinction makes sense when one assumes that if a worker is to be paid by the hour his employer is int e rested in the worker completin g the task in the short e st reasonable amount of time. Since the employ e r must pay such a work e r by the hour, he likely will provide mor e detail e d instructions and supervision to ensure the quickest possible completion of the job. On the other hand, if the worker is paid by the job 1n a lump sum payment, this tends to indicate that the purchaser of the services is only interested in the final results to be accomplished.

PAGE 88

The amount of time to complete the job is irr e levant in this type of situation as lon g as the results are those that both parties agreed upon. In fact, it would be incumbent upon the worker to determine the quickest method to complete a job because his com pensation is directly related to the number of jobs that he can successfully complete. Also, the fact that a worker may receive payment in several installments before the job is completed is not an indication of employer-employee relationship. According to the IRS audit manual, if the receipt of intermittent payments 1s merely a convenient method to disperse a lump sum, then such . f d. 65 an arran ge ment still favors an ind e pendent contractor 1n 1ng. The rule that is bein g applied throughout the entire analysis of this factor is wheth e r the worker's comp e nsation 1s contingent on his performance of certain requir e ments (meanin g no guarantee of profit) or simply th e number of hours h e works. In general, the determination of how a worker is paid is rather straightforward. In the cases included in this study, 128 times the jud ge determined that the work e r was b e in g paid 1n a method that indicated an independent contractor finding. In an additional seventeen cases, the jud ges felt the method of payment seemed to indicate an employer-employee relationship. In only three cases was the method-of-payment factor not considered in the judge's written opinion. From this it is e vid e nt that the m e thodof-payment factor has b ee n applied in most of the cases includ e d in this study and the information needed to apply this factor 1n actual court case situations is usually readily available. 9. Payment of Busin e ss and/or Traveling Exp e ns es The reimbursement of expenses is concerned in general with the 79

PAGE 89

method-of-payment factor as discussed above, but is unique enough to be considered a separate factor. Reimbursement of expenses has a similar advantage to the method-of-payment factor; namely, it is a reasonably simple factor to apply. Either a worker is being re imbursed for his business expenses or he is not. This is not particularly consistent with the fact that in fifty-seven of the decisions analyzed in this study, the court did not appear to consider reimbursement of expenses in their written opinions. This is most likely attributable to the fact that a reimbursement of expenses was not mentioned specifically in the Restatement or 66 1n one of the landmark cases as was the method-of-payment factor. This may result in the situation where reimburs e ment of expenses is simply considered jointly with the method-of-payment factor as 11 f 1. 6 7 does Marmo 1n the Tax Management port o 10. This study takes the same approach as the IRS audit manual by treating reimbursement of expenses as a separate factor. The audit manual defines reimbursement of expenses as-80 If the employer pays the person's business and/or traveling expenses, the person is ordinarily an employee. The employer, to be able to control expens e s, must retain the ri g ht to regulate and direct the person's business activities. Conversely, a person who is paid on a job basis and who has to take care of all incidental exp e ns e s is g enerally an independent contractor. Sinc e he is account able only to himself for his expens e s, h e is free to work accordin g to his own methods and means.68 The audit manual implies th.:it a workC'r who h;:is to dC'pl'nd on ;:in employer for reimbursement of expenses is mor e subject to that employer's control than is a worker who pays his own expenses. In other words, reimbursement of a worker's expenses is simply

PAGE 90

another method of controlling that worker's business activities. An employer-employee relationship tends to be indicated when a purchaser of services pays for the worker's business licences, meals, transportation, or allows the worker to use the company's charge account. The Service's position as discussed above is somewhat inflexible. Marmoll states that-The Service's position, as s e t forth in the manual, on the payment of business or trav e l expenses i g nores the possibility of a cost plus fixed fe contract as w e ll as the mod e rn day custom of char gi n g it e mized expenses to the customer or client in addition to the fee for the work done. The clients of lawy e rs and accountants, for example, are customarily charged for the attorney's or the accountant's travel expenses and, yet, the cli e nt does not control the business activities of th e attorney or accountant, the relationship cl ea rly being one of client and independent contractor.69 The lawyer or accountant example above appears to su gge st that the reimbursement-of-expenses factor needs to be considered in terms of the overall compensation agreement. If the reimburs eme nt of expenses is in addition to an agreed upon lump sum fee, then Mar moll seems to sug ge st that the lump sum payment which supports 81 an independent contractor finding somehow overrides the fact that expenses are reimbursed which of course favors an employee finding. While it is true that the method-of-payment factor may carry more weight than the reimbursement-of-expenses factor when they have conflicting results, the fact still r e mains that reimbursement of expenses by the e mployer still favors an employee findin g In the cases included in this study, judges were able to apply the reimbursement-of-expenses factor on its own merits in nin ety one decisions. In seventy-four of these decisions the factor

PAGE 91

favored an independent contractor finding while in the other seventeen decisions the factor favored an employ ee findin g 10. Furnishing of Tools and Mat e rial s The more substantial the investment in tools and materials by a worker, the more likely the worker is to be considered an independent contractor. If the employer furnishes the tools and materials necessary for the accomplishment of the worker's assi g nments, then this tends to indicat e an employer-employee relationship. Th e r e ar e certain kinds of jobs for which th e tools n ee ded to p e rform th e assi g n e d 82 task are inconsequential. 70 For ex a mpl e a g ypsy chas e r who unloads trucks needs practically no tools to p e rform his s e rvic e s. Since the wor k is primarily m a nual, tt e furnishin g of tools would not be a particularly si g nificant f actor in a c a s e involvin g th e employm e nt status of a g yp s y chas e r. On th e oth e r h a nd, an a ppli cator of sidin g a nd ro of in g mat e ri a ls must hav e a rath e r compl e t e s e t of hand tools, ladd e rs, sca f folding, etc. and th e f urnishin g of tools would be a much mor e si g nificant f actor in this typ e of case. Both Treasury Reg ulation Section 31.3121(d)-l(c ) (2) and th e Restat e ment include th e f urnishin g of tools a s a factor that should be includ e d in the d e finition of a common law e mplo yee Th e R e state ment further stat e s that i f the tools ar e pr o vid e d by th e e mploy e r and ar e o f sub s tanti a l v a lu e it i s und e r s t oo d that th e work e r will follow th e directi o n s o f th e own e r in t h e i r us e and this indicat e s 7 I the own e r is mast e r. Th e IRS audit manual provid e s a dditional detail as to how th e f urni s hin g -of-tools-and-m a terials factor

PAGE 92

h ld b d 72 sou e 1nterprete The audit manual describes this factor as-The fact that an employer furnishes tools, materials, etc., tends to show th e existence of an employe r-emp loyee relationship. Such an employer can determine which tools the person is to use and to some extent in what order and how th e y shall be us e d. 83 An independent contractor ordinarily furnishes his own tools. However, in some occupational fields, e.g., skilled workmen, workers customarily furnish their own tools. They are usually small hand tools. Such a prac tice does not necessarily indicate a lack of control over the services of the worker. 73 Application of the furnishing-of-tools-and-materials factor 1s relatively straightforward which is evidenced by the fact that 1n 131 out of 148 decisions included in this study the judges men tion ed this factor in their written opinions. In eighty of those decisions, the courts held that the worker provided his own tools and/or materials which favored an independ e nt contractor finding. In another fifty-one decisions the courts h e ld that th e employer provided the tools and/or materials which fa vored an em ploy ee finding. In only seventeen decisions did the judges fail to discuss the fur nishing-of-tools-and-materials factor as part of their written opinion. I 1. Realization of Profit or Loss. The worker who can realize a profit or sustain a loss is generally an independent contractor and the worker who is prevented from earning a profit in excess of his stated compen sa tion or is insulated a ga inst sustainin g lossess, is ge n e rally considered an emp l oyee For example, in I 74 Nevin Inc. v. Rot1 e ns1cs the court held that a lessee of a drug store was an independent contractor. In this case, the parties involved e stablish e d a separate (though not a fully independent) business where the lessee had a considerable personal financial

PAGE 93

84 stake 1n the business. H e could make or lose m o n e y in th e busin e ss. The pos sibility o f profits could b e enh a n c e d by eff ici e n c y, economy and skill. He t o ok th e risk o f consid e rable per sonal loss in c as e th e busin e ss w as n o t s ucc e ssful a risk which involved a g ood d ea l mor e than the m e r e possi b ility of loss of position. li e obt a in e d compl e t e freedo m in conn e ction with his o w n ti me working hours, etc., and an opportunity to incr e ase his inc o m e not only by sales r e cords (as in th e c a s e of a corrnnission sal e sman) but by busin e ss judgm e nt in most of the matt e rs which are in the province of any indep e ndent retailer.75 Consequently, the l e ssee could aff e ct the amount of profits he was able to earn throu g h his own mana ge rial e fficiency and skill. The IRS audit manual also discusses the importance of mana ge rial skill in the realization of profits and avoidance of losses. 76 The manual states that "wh e ther a profit i s r e aliz e d or loss suffered generally d e p e nds upon mana g ement d e cisions; that is, the one responsible for a profit o r loss c a n us e his own in g enuity, initiative, and jud g m e nt in conductin g his busin e ss or e nt e rpris e 1177 Another important component of the re a lization-of-profit-or loss factor is d e terminin g wheth e r th e wor ke r has a si g ni f icant capital investment in th e business. The IRS audit manual provid e s that f or a si g nificant inv e stment to e xist, th e investm e nt must 78 be real, essential, a nd ad e quate. Such an inv e stm e nt is mad e generally in equipm e nt, furnitur e or real prop e rty n e c e ssary to run the business. Th e question of w h e th e r an inv es tm e nt in e q u i pme nt 1s r e al seems to b e conc e rn e d with wh e th e r th e work e r r ea ll y ha s a n e quity investm e nt, or is th e wor ke r simply buyin g th e e quipm e nt on ti me from the person for whom h e is providin g s e rvic e s. Th e audit m a nual illustrates this point in the followin g illustration.

PAGE 94

85 Little w e ight can be accorded to a worker's investment if the worker purchases equipment from his employer on a time basis but the employer r e tains title to the equipment, has the option of retainin g l eg al ownership by paying the worker the amount of his equity in the equipm e nt at any time before the equipment is fully paid for, requires its exclusive use in the operation of his busin e ss, and directs the worker in its use. Such investments are not "real".79 To be essential the investment must be in equipment or facil ities necessary for the performance of the worker's services. The example given in the IRS audit manual concerns a model who has a large investment in her wardrobe, but poses for a photographer that ordinarily provides the wardrobe and strictly as a matt e r of indul ge nc e the photographer allows the model to use her own clothes. Such an investment in clothes is not essential to the services that she performs. "The photographer hires her only for her photogenic qualities and her ability to pose; it is not required 80 that she furnish her own wardrobe." The critical test sug g ested by the IRS audit manual in deter mining if an investment is adequate is to compare the value of the investment of the worker to the total value of all the facilities needed to perform the work. An investm e nt in faciliti e s is not adequat e if the worker must rely appreciably on the facilities of others to perform the services. For instanc e an individual who is en g a g ed to perform a machine op e ration on his own premises and who furnishes his own equipm e nt of sub stantial value may be a s e lfe mployed contractor instead of an employ e e o f the manufacturer SI The realization-or-profit-or-loss factor is not only a maJor part of the IRS audit g uid elines, but it is also a factor included in both the GAO proposal and H.R. 3245. In the GAO proposal, "the risk of suffering a loss and opportunity of making a profit"

PAGE 95

is one of the four basic tests us e d to determine if an independent contractor engaged 1n a trade or business should be considered 82 self-employed. The only explanation of the factor provided by the proposal is that there should be a "real possibility that expenses directly related to the business will exceed business 1183 income. 86 On the other hand, H.R. 3245 provides consid e rable detail in describing its third safe-harbor criteria which it calls "invest ment or income fluctuation. 1184 The bill proposes that a worker who has a substantial investment or who risks income fluctuations should f h. f h b . SS sat1s y t 1s sa ear or cr1ter1a. It treats investment and income fluctuations as two separate components of an overall test which can be satisfied by the presence of either a substantial investment or risk of income fluctuation. The income fluctuation test is satisfied if the income level 1s neither fixed nor g uarenteed, but rather directly depends on the out-put of th e worker. This can be exemplified by a salesman who can make several sales presentations and not make a single sale. This is not the same as an employe e who is basically assured that if he works a certain number of hours at a reasonably acceptable level, he will receive a specified amount of compensation. 86 Application of the substantial-investment-in-assets test is similar to the incom e fluctuation te s t. The percent of the in vestment is not considered an important crit e ri o n d e spite the IRS audit manual, but rath e r th e worker must "inv e st a si g nificant amount of money or his time as in the income fluctuation test with no assurance of success. 1187

PAGE 96

With the importance placed on the realization-of-profit-or loss factor by the IRS and the inclusion of the factor in both 87 the GAO proposal and H.R. 3245, it is not surprising that the factor was discussed by judges in 119 of the 148 decisions included in this study. In only twenty-nine decisions did the courts fail to consider the realization-of-profit-or-loss factor in their written opinions. Before discussing the methods employed in the coding of the variables, those factors that were deleted from the study due to their infrequent occurrence in the written opinions will be pre sented. The followin g is a list of those s e ven factors discovered in the literature review, but not included in the analysis of this study. ( 1) Did the worker possess a special skill? (2) Was th e taxpayer's claim consistent with indu s try custom? (3) If required, who supplied special unifor ms? (4) What was the intent of the employer and worker? (5) Did th e worker receive any fringe b e n e fits? (6) Was th e worker required to submit written or oral r e ports? (7) Can th e w o rk e r t erm in a t e his relationship with the employer at any tim e h e wish e s without incurrin g liability? Since none of th e se factors were includ e d in m o r e than 10 p e rc e nt of the court cases, this was interpr e ted as e vid e nc e that th ese factors were not particularly relevant in d e terminin g what consti tutes a common law employee.

PAGE 97

88 The only oth e r factor that was discov e r e d 1n the search for relevant variables, but not used in this study was the "right to discharge." Treasury R e gulation S e ction 31.312l(d)-l(c)(2) mentions the right to discharge as "an important factor indicatin g that the person possessin g that right is an e mployer." The IRS audit manual discusses the ri g ht to dischar g e in the followin g manner-The ri g ht to dis c h a r ge i s an imp o rtant factor in indicatin g that th e p e rson possessin g the ri g ht is a n employer. He ex e rcis e s control through th e ev e r-pr e s e nt thr e at o f dis missal, which caus e s th e work e r to ob e y hi s instructi o ns. An ind e p e nd e nt c o ntr a ctor, on th e o th e r h a nd, cann o t be fired s o lon g a s h e produc e s a result which m e ets his contract specifications. 8 8 The IRS audit manual's d e finition, simply stat e d, is that an employ e e can b e fired anytim e but an ind e pendent contractor cannot b e dis charg e d exc e pt und e r th e t e rms of his contra c t. This rather sim plistic approa c h i g nor e s the pr e sence of uni o n c o ntr a cts that can often limit the pos s ibility of dischar g e of an e mpl o y ee This argument is support e d b y l1armoll as foll o ws-Sin ce th e l a w h as d e v e lop e d in su c h a m a nn e r t h at most e mplo yee s h a v e som e prot e ction a g a in s t a n arbitrary ri g ht of di s ch ar ge and sin ce m an y c o ntr ac tu a lly a g r e e with th e ir e mp l o ye rs to g iv e a c e rt a in a m o unt of n o tic e b ef or e t e rminatin g th e em ploym e nt r e l a ti o n s hip, the r e stat e m e nt's expl a nation of who is a "s e rv a nt" or e mployee" do e s not m e nti o n the ri g ht to dischar g e Mo reov e r, ther e a re courts which, whil e consid e rin g th e fact o r of ri g ht to di s ch a r g e o r ri g ht t o t e r m inat e hav e placed the factor in p e rspectiv e not al lo win g it t o g ov e rn th e cas e d es pit e the e xa g g e rat e d importanc e g iv e n to the factor in th e r e g ul a tions.89 On e District Court h as held th a t simply b cc .: rn sc til e taxp aye r had the ri g ht to dischar g e the work e r i f hi s s e rvic es w e r e n o t satisfactory 1n "no way" m e a nt that an e mployere mploye e rel a tionh . d 90 s 1p ex1ste

PAGE 98

This same ri g ht exists in every contractor who had employed a subcontractor wh e n the latt e r's performance falls below standard. An applicator who did not like a particular assignment was free to refuse it and in such a case the plaintiff would attempt to give him another job.91 This same position 1s taken again by another District Court in a similar case. While it 1s true that he could terminate an applicator's work b e fore the completion of a particular job if the desired result was not b e in g accomplished, this power was no different than that of a general contractor to declare a breach of his contract had been committed by a sub-contractor.92 What makes this factor even more difficult to understand 1n this type of study 1s that the written opinions of the judges ordinarily fail to provide enough information to adequately determine if this so-called right-to-discharge factor is really being applied to an independent contractor relationship or an employer-employee relationship. Because of the inability to ascertain when the ri g ht-to-dischar g e factor was present in the written opinions, it was decided to delete this factor from the study. With the selection of the eleven factors completed, the mathematical models necessary to answer Research Question 2 can be constructed. Before developing the methodology to be used in studyin g Research Question 2, it is first n ece ssary to discuss the coding procedures used in the analysis of th e data. Coding of th e Variables Each of the eleven variables previously defined was treated as a trichotomous random variable when collecting tl1e data from 89

PAGE 99

the court cases. There are alternate methods of coding trichotomous independent variables 10 building mathematical models. To obtain the maximum amount of information contained in this type of data, each variable should be represented by two dummy variables. For example factor 8, method of payment, would be coded as follows: (I) if the worker was paid by the job, variable x 1 would be coded as a I and O otherwise; (2) if the worker was paid by the hour, variable x 2 would be coded as a I and O other wise; and (3) if th e method-of-payment factor was not mentioned by the court, then both x 1 and x 2 would be equal to zero where This simple model could be expressed as (3. I) Y 1s th e probability of a worker bein g classified as an employee U is the error term. B's are the independent variable coefficients. Taking the expected values of this model results in E(Y I IX I = 0, x2 0) = BO + Bl E(Y I IX I = O, x2 = I ) = BO+ B2 E(Y I IX I = o, x2 = 0) = BO The interpretati o n of thes e coefficients would be that B 0 equals the "probability of an employee finding" (hereaft e r referred to as Y) if the method-of-payment factor was not mentioned in the case. If B 1 is added to B 0 this represents Y when a worker is 90

PAGE 100

paid by the job while B 0 + B 2 equals Y if the work is paid by the hour. The difference 1n Y according to whether a worker is paid by the hour or by the job 1s A different approach using a trichotomous independent variable 1n building a mathematical model would be to simply represent the method-of-payment factor with three values. For example 0 could represent the worker being paid by the job, I could re present the method-of-payment factor not being mentioned, and 2 could represent a worker being paid by the hour. This simple model could be represented as (3 .2) Taking the expected value would give E(Ylx 1 = 0) = BO, E(YIX1 = I ) = BO + BI, E(YjX 1 = 2) = BO+ 2B 2 At first glance the only difference in the expected values for model 3.2 and those calculated for model 3. I would be the expected values of Y given the worker is being paid by the hour. However, 91 a more important difference is that the expected values for model 3.2 imply that the difference in Y when the method-of-payment factor is not mentioned and wh e n the worker is paid by the job is while the difference 1n Y when the method-of-payment factor is not

PAGE 101

mentioned and when the worker is paid by the hour 1s That is, by representing the method-of-payment factor with just one variable with values 0, I, and 2, the difference 1n Y between the method-of-payment factor not being mentioned and a worker being paid by the job is forced to be equal to the difference in Y between the method-of-payment factor not being mentioned and a worker being paid by the hour. Therefore, the maximum amount of information contained 1n the data is not being utilized because the distance between each of the three cate go ries (O, I, 2) of the method-of-payment factor are forc e d to be equal when they may very well not be. In contrast, another consideration in selecting the method of codin g the variables is the fact that by usin g dummy variables, the mathematical models built in this study will increase from eleven independent variabl e s to tw e nty-two independent variables. Such an increase in the number of independent variables aff e cts the stability of the estimates, especially with the small number of data points available for use in this study. Therefore, even though representing the data with dummy variables theoretically utilizes more information, it has the disadvantage of doublin g the number of parameters to be estimated. Doublin g the number of p3ra meters to be estimated is an obvious trade-o ff to b e considered 1.n deciding how to code the data. To properly evaluate the effects of the different coding schemes, the initial analysis perform e d 1.n this study will be based on models built utilizin g both coding procedures. If it is discovered that the statistical results are 92

PAGE 102

not significantly improved by using the dummy variable coding scheme, the more parsamonious II-variable model will be used 1n the remainder of the analysis performed in this study. With this explanation of the coding procedures used in this study, the actual methodology used to analyze Research Question 2 can now be discussed. Research Question 2 It would be possible through a legal analysis of cases to 93 obtain a certain understanding of what factors espoused in Research Question I are relevant to judges in deciding an employee versus independent contractor case. However, this dissertation will attempt to go beyond that type of analysis and look at the question from a more quantitative approach. The ability to address tax research questions by the statis tical methodologies outlined in this study e nables the researcher to approach the question more objectively. The results of legal analysis are often biased by the researcher's own pre-conceived heuristics. 93 The statistical approach is an att e mpt to avoid as much subjectivity as possible in performing the study by al lowing the results to be the product of a systematic statistical estimation technique applied to the actual data. An added advanta g e of this type of r e s e arch approach is that a lar g e amount of d a ta can be effectively synth e siz e d : rnd used in analyzing the research qu e stions. It would b e difficult to ef ficiently reduce the results of 148 court decisions through a le g al analysis. However, by the aid of the computer and rather sophisticated

PAGE 103

software packages, the researcher is able to reduce this amount of data to a manageable size quite easily. The advanta g es of statistical analysis to certain types of tax questions have been illustrated in several recent tax dis sertations. Recently, multiple linear regression and discrimi nant analysis have become popular tools for tax researchers in analyzing the court's decision-makin g process. However, the use of such quantitative methods for modelin g courts' decision orocesses did not originate in the field of accounting. Researchers in the field of political science also have an interest in courts and their decision-makin g processes and th e se researchers have been using mathematical models and other techniques to predict judicial behavior. Mathematical Techniqu es 1n Tax Research Political Scienc e Re s e arch. The first published attempt at using a mathematical model to predict judicial behavior was by 94 Kort. This study uses pivotal factors (identified by the Supreme Court) to develop a model that would predict the outcome of "right to counsel" cases brou g ht before the Supreme Court. Kort used fourteen cases from the period 1932 to 1947 to estimate the model. Fourteen cases were then quantified, sunnned by case to ge t a composite value, and th e n by trial and error us e d to find a critical value that was able to correctly pr e dict the outcome of the last fourteen cases. In 1963, Kort improved on the above described wei g htin g scheme by using factor analysis to reduce the number of independent variables and then using this reduced set of independent variables to 94

PAGE 104

predict the case outcome (binary decision) using a multiple re. d 1 95 gression mo e Some possible problems contained in the early study were that the number of cases from which to obtain observa tions was relatively small in comparison with the independent variables and no consideration was given to the effect on the nor mality assumption of having a binary dependent variable in the model. An F-test was used to test for significance and this like wise is dependent on the normality assumption which appears to have been violated. Ulmer recognized the problem of binary dependent variables and in his study attempted to eliminate this problem by usin g multiple discriminant analysis. 96 Ulmer's study used six factors to formulate a discriminant function of Justice Felix Frankfurter's Supreme Court decisions dealing with civil liberties. Using data from cases involving the year 1959-1961 to estimate the function, the model was able to correctly classify six out of the seven decisions made by Frankfurter from 1961 to 1962. In addition to discriminant analysis and re g ression analysis, simulation techniques have been used to construct a model of Supreme Court decision-making dealing with civil liberty cases. 97 Also, straight correlation analysis has been used to predict future case outcomes in reapportionment cases. 98 In addition to political scientists, accountin g researchers have also employed statistical techniques to tax research. Recently, several dissertations have utilized both discriminant analysis and multiple re g ression to study certain tax questions. 95

PAGE 105

Tax Research in Accounting. En g lebrecht us e d a multiple regression analysis to study the valuation of closely held cor99 porations by the Tax Court for estate and gift tax purposes. The study used the Internal Revenue Code, Tr e asury R eg ulations, and Revenue Rulings to determine guidelines that the Tax Court should be using. He found that the g uidelines could be used to explain 86.5 percent of th e varianc e in the valuations in the Tax Court cases h e studied. Engl e br e cht also inv e stigated the question of whether or not the Tax Court act e d as an arbitrator between the IRS and th e taxpayer or wheth e r th e Tax Court's valua tions were unique. IOO Bromley disput e s th e fact that the regr e sion results d e scrib e d above ar e ad e quat e t o draw the conclusion that t h e Tax Court acts as a 10 I comprom1.sor. Broml e y attempt e d to replicate the study and e v e n thou g h similar re s ults were f ound 96 he interpreted th em quite differ e ntly. His disa g r e em e nt cent e r e d around the interpret a tion of the statistical r e sults a nd an apparent violation of a r e gr e ssion assu m ption. Engl e brecht count e red with his own arguments th a t obviously favor e d his findin g s a nd atte m pt e d 102 to disprov e Broml e y's results. One positiv e r e sult of the above d e bate is that it calls attention to th e probl e ms associat e d with this type of tax research and th e validity of th e conclusions that can be drawn, Bo y d us e d multipl e r eg r e ssion a n a l y sis t o d e t e rmin e if th e Tax C o urt u se d th e MJvs o n ~a nuf ac tur i n g Co g uid e lin es in d ec idin g bl 103 reasona e comp e n s ation cas e s, Dat a con s isted of s e v e nty-five cases from which coeffici e nts for the tw e lv e ind e p e nd e nt variabl e s were obtain e d. The results indicat e d that four variabl e s were

PAGE 106

consistently used by the Court and at that time, none of these was contained in the Regulations. A similar study that dealt with reasonable compensation was conducted usin g discriminant 1 h . 1 1 104 ana ysis as t e statistica too The results of th e study were supportive of Boyd's research in the s e nse that the Mayson Manufacturing Co. g uidelines were again eff e ctive in modelin g the Tax Court's decision-makin g process. Bond used a discriminant analysis in an att e mpt to discover the factors used by Tax Court jud ges in distin g uishin g between b 105 de t and equity. A sample of 101 Tax Court ca ses was us e d as data for thirty trichotomous variables. The r e sultant model con sist e d of two variabl es that wer e able to correctly classify ninety of nin e ty-six observations. Madeo us e d multiple discriminant analysis to study the ac1 d 106 cumu ate earnin g s tax. Usin g sixty-sev e n Tax Court decisions for data, th e study c o nclud e d that the mod e l b a s e d on I R S audit guidelines did bett e r at correctly classifyin g th e cas es than th e model based simply on th e Tr e asury R eg ulation s Taylor used a discriminant analysis to build a model that would distinguish b e tw ee n ordinary income and capital g ains in 1 107 rea e state transactions. Data for sev e nt ee n qu a litativ e variabl e s and thr ee quantitative variables were obtain e d from 108 Tax Court cas es Th e r es ults narrow e d th e ind e p e n de nt variabl e s to ten sta tistic a ll y si g nific. : rnt fac tors with th e maj o rity of the predictiv e ability contained in only four v aria bl e s. Kramer us es both multiple discriminant analysis and multiple regression analysis in examining th e variabl e s that affect th e 97

PAGE 107

108 valuation of large holdings of publicly traded stock. The discriminant analysis was used to build a model to predict whether courts will value the blocks of stock at market value or some other court determined value. The multiple regression analysis was used to predict the discount from market price. An analysis was also done in which a market model was built and compar e d with the court model to determine if judges diff er from the market as far as what factors they use to value large stock sales. This study intends to extend the scop e of this type of sta tistical tax research by introducin g a n ew es timation technique and also by usin g data from other than th e Tax Court. This will be the first tax study of this nature that d e p e nds e ntir e ly on federal tax cases obtained from outside th e Tax Court While the Tax Court has b e en successfully model e d in several recent dissertations, this study will att e mpt to e mpirically model the much more divers e judicial system of the combined Court of Claims d h . 109 an t e D1str1ct Courts In addition, this study will introduce logit analysis, a non-linear estimation technique, to help determine whether th e linear proc e dures applied in previous tax research studies are adequate Discriminant Analysis The depend e nt variable of int e rest in this study 1s wheth e r 98 a worker 1s an employee or an ind e pend e nt contractor Such a determination is strictly a binary decision. On e cannot choose 0.9 employee or O. I ind e p e ndent contractor, but the choice is betw ee n two discrete classifications which are repr ese ntable by ones or zeros

PAGE 108

Studies dealin g with a dichotomous dependent variable have often applied discriminant analysis in an attempt to distinguish 110 statistically between two groups of cases. To distinguish between the two groups, discriminant analysis attempts to develop from a set of predictors a single linear composite with values 111 that maximally distin g uish between members of the two groups. For purposes of this study, the computer pro g ram P7M of the 112 Biomedical Computer Program P Series (BMDP) was used to d e rive the discriminant function. The form of the discriminant function is D = (3.3) wher e D 1s th e score of the discriminant function, 99 Z's repres e nt th e elev e n discrimin a tin g v a ri a bles d e riv e d in Res e arch Qu e stion I th a t m e asur e char a ct e istics on which the groups are expect e d to differ, and d's ar e w e i g hts for the Z variables. The eleven variabl e s derived in Research Q uestion 1 were introduced into the analysis through a forward stepwise procedure. This type of procedure begins by selectin g the singl e most discrimi nating variable based on the minimization of Wilks' lambda. The variable that minimizes Wilks' lambda also maximizes the overall multivariate F-ratio which 1s a measur e of g roup discrimina tion. 113 The initial variable 1s then paired with each of th e re mainin g independent variables and the combination that produces the greatest discriminatin g power is included in the model. This procedure is continued until either all the indep e nd e nt variables

PAGE 109

are included in the model or the addition o f any of the remainin g ind e pendent variabl e s does not provide a pr e -sp e ci f i e d m1n1mum improv e m e nt 1n W ilks' lambda. Classification Functions. On e method of determining whether the variabl e s contain e d in the d e riv e d discriminant function are really able to distin g uish wheth e r a work e r is an e mployee or an independent contractor is to use the results of the discriminant analysis to predict th e outcome of th e 148 court cas e s included 100 in this study. Classification of th e court ca s es is achieved throu g h two classification functions (on e for e ach g roup) obtain e d from th e discriminant f unction. "Th e traditional classification equations are d e riv e d from the pooled withing roups covariance matrix and the c e ntroids for the di s criminatin g vari a bl e s. 1111 4 Th e classification e quation for on e g roup c o uld b e writt e n as where C. l ( 3.4) C. is th e cl as sificati o n scor e o f th e ith gro up, l C . 's ar e the cl a ssific a tion co e ffici e nts, lJ K. is th e constant, a nd l V's are th e observati o ns from th e individual cas e s. It is th e n possible to calculate a classification sc o r e for each case for both g r o ups ( e mploy e e v e r s us ind e p e nd e nt co ntr a ctor). The c ase is th e n cl ass i f i e d into th e g r o up in w hich i t h as th e hi g hest classification scor e Th e cl a ssific a tion scor e s can then be conv e rted to probabiliti e s of g roup m e mb e rship. In fact, classifying a cas e into a particular g roup is th e sam e as

PAGE 110

assigning the case into the group 1n which it has the greatest probability of membership. 115 Using classification functions to predict g roup membership for the same cases that were used to build the classification functions will obviously result in a high estimate of their clas sificational accuracy. The better procedure and one that would help overcome this bias would be to use a hold-out sample that 101 could be used to test the accuracy of the classification functions. However, because of the limited number of cas e s included in this study, there is not sufficient data to ge nerate a hold-out sample. The BMDP computer package has a technique referred to as a "jack knife-validation procedure" that was initially suggested by 116 Lachenbruch. This technique involves calculatin g a discriminant function based on all the cas e s but one. That discriminant function is then used to generate classification functions which assign the left out case to either the employee or independent contractor group This procedure is repeated leaving out a dif ferent case each time until all of the cases are classified. The r e sults of this procedure provide a less biased estimate of the accuracy of classification f unction s generated from the dis criminant analysis results. The number of misclassified cases 1s r e li e d upon heavily in determinin g the usefulness of the vari a bl e s includ ed 1n the dis criminant analysis m ode l. This 1s necessary b eca us e the discriminant analysis co ef ficients cannot be interpreted in the same manner as re g ression coefficients are. There are no t es ts of wheth e r a coe fficient 1s si g nificantly diff e rent from zero b e cause

PAGE 111

-----------------------102 d . 1 ff' . 117 1scr1m1nant ana ys1s coe 1c1ents are not unique. Therefore, classification accuracy becomes an important method of analyzing the discriminant analysis results. Another approach of deter mining the relative importance of the independent variables is to observe in what order they entered the discriminant function and what effect their entry had on the overall multivariate ratio. Even thou gh this type of analysis is helpful, it is not a completely reliable method since each variable is entered into the discriminant equation because of its effect on the discrimi natin g ability of th e function given the variables already included. Therefore, it is possible that a certain variable may be entered in the model not because of its own uniqu e discrimin a ting ability, but because of the improvem e nt it contributes in conjunction with the variables already contained in the mod e l. T e mporal Stability of the Model. Because the data used in this study involved a thirty-nine year span, it was considered necessary to test th e model for its stability over time. Th e data were therefore split in half with the first seventy-three cases being in the pre-1960 group and the other seventy-five cases being in the post-1959 group Because the coefficients are not unique, the building of a model for each time span and then comparing the discriminant function coefficients would not b e particularly il luminating. Instead, a model was built usin g th e pre-1960 cas es and the resultin g classification functions were us e d to assign both the pre-1960 and post-1959 cases to eit\1er the em plo yee or inde pendent contractor groups. The misclassification rat e s were then examined to determin e if any differ e nces in rates existed. A lar ge

PAGE 112

103 difference in the number of misclassified cases for the post-1959 group as compared to th e pre-1960 g roup would su gge st th a t th e model is not particularly stable over tim e Statistical Assumptions The us e of discriminant analysis imposes certain statistical requirements on th e distributional properties of th e independ e nt variables. First, th e independ e nt variables should h a ve a multivariate normal distribution. Since in this study all of the independent variables are discr e te var iables, this assumption do e s not hold. Oth e rs hav e ar g ued that discriminant analysis is still sufficiently robust to perform 11 . f . 1 f h 1 118 we in spite o this vio ation o t e norma 1ty assumption. Second, the disp e rsion matrices for each g roup sh o uld b e equal. This condition may also be violat e d in this study. Box's M t e st contain e d in th e SPSS software packa ge was us e d to test for th e 1 f d. . 119 equa ity o isp e rsion matrices. If th e matric e s ar e not e qual, the classification functions will tend to assi g n cas e s to th e g roup h h 1 . 1 b b 1 120 wit t e arger disp e rsion g iven e qua ex ante pro a i ities. If the assumption of equal disp e rsion m a tric e s i s violated, d . . 121 a qua ratic estimation proc e dure is appropriat e How e ver, Gilbert has shown that linear and quadratic methods yi e ld v e ry 122 similar results ev e n when the disp e rsion matric e s ar e not equal. In fact, in some cases a linear classification function provid e s better predictiv e p e r f ormance than does th e quadratic classification f 123 unction. With th e e vidence that su gge sts that th e diff e rences in lin e ar and quadratic m e thods ar e not si g nificant, only the linear classification results will be reported in this study.

PAGE 113

104 Logit Analysis In considering a binary dependent variable such as is contained in this study, it is helpful to understand how such a variable should be interpreted. For example, the following simple re gression model is represented as where Y. l = (3 .s) Y. is a binary ( represented as O, I) dependent variable and 1 U. is the random disturbance term. 1 In this situation the expected value of Y. has a special inter1 pretation. Under the usual assumptions of the classical normal linear regression mod e l, the expected valu e of the disturbance term, U., is zero. This results 1n 1 E(Y.) 1 (3 .6) If Y. is considered to be a dichotomous discrete random variable, 1 the probability distribution can be stated as follows: Thus, P. 1 that \' = 1 Combining Y. Probability 1 0 P(Y. 0) = I P. 1 l P(Y. = I) = P. 1 1 1S the probability that Y. = .ind l o. The expected value of Y. must l E (Y.) = O(q.) + I (P. ) = p . l l l l equations (3 .6) and ( 3. 7) results P. 1 (3.8) = q. 1 q. 1S th e 1 then be (3. 7) 1n probability

PAGE 114

105 Therefore, the dichotomous d e pendent variable, Y., is simply l a representation of the probability of Y. bein g e qual to I, J. In conjunction with this study, if an independent contractor 1s represented as Y. = 0 and an employee is repres e nt e d as Y. = I, 1 l then the estimat e s of Y. can simply be discuss e d in terms of the 1 probability of a worker being classified as an employee. Unfortunately, binary dependent variables also cause special probl e ms when using linear regression to model an underlying relationship, First, an assumption of the classical normal linear multiple regression model is that the error terms are normally distributed. The error term when the d e pendent variable is dichotomous can take on only two valu e s. Second, another assumption of the classical model is that the variance of th e error terms is constant. If th e d e p e ndent variable is binary, th e variance of the error terms is a function of the indep e ndent variable and there fore is heterosk e dastic. Ordinary least squares (OLS) e stimat e s are no lon g er effici e nt. Third, since th e dependent variable has only two possible values (0 or I), any pr e dictions that do not fall within the (0, I) interval make littl e s e nse. Procedur e s could be d e veloped to overcome the above mentioned 124 problems if certain limiting assumptions would be made. However, such an attempt must be based on the assumption that the model is correctly specifi e d. If this specification assumption is n o t true, any such pr o c e dur e s d e v e lop e d w o uld be mi s l ea din g and would 125 not solv e the und e rlyin g problem. Corr e ct Functional F o rm. Wh e n trying to explain the relation ship between a continuous variable (Y) and a set of exo g enous

PAGE 115

106 independent variables (Xt), it is logical to begin with a linear . l 126 additive mode However, for the variable of interest in this study, such a functional form may not be so reasonable. If the model is written as Pt= Prob(Yt=I) = F(Xt), then the interest centers on how the independent variables affect the event in question and the proper functional form for F(Xt). Both theoretical and empirical considerations suggest that when the dependent variable is a binary variable, the shape of the response function will f 1 b 127 requent y e curv1 1near. Since F(Xt) must fall between 0 and I, the relationship must be nonlinear at least at the boundar1es. Perhaps an even more important consideration for this study 1.s that a linear application is simply not very realistic. For example, if the presence of the "supervision" variable is so strong that in an employment status tax case th e jud g e is almost certain to rule that an individual 1s an employe e th e n incr ea s e s in the other variables would have little or no effect on the probability of finding employee status. However, if the presence of the "supervision" variable was not so pronounced, the same increase 1n the other variables would take on greater importance. Such behavior indicates an interaction effect that cannot be captur e d 1n a linear probability model. Instead, such a r e lationship may clos e r approximat e a tilt e d S-shape where the independent variabl e \13s its greate s t impact un the probability of choosing a particular classification at the mid. f I d. b 128 point o t1e 1str1 ut1on. The rather low slopes at the endpoints

PAGE 116

relate to the previous example where the change 1n an independent variable had little or no impact on the judge's decision. Ordinary least squares regression 1s not limited to linear relationships because several common transformations exist that can be used to estimate nonlinear equations. However, because of 107 the dichotomous nature of the dependent variable, most of these models are ineffective since transformations involving dichotomous d d bl d d' h 1 129 epen ent var1a es st1 pro uce a 1c otomous resu t. To introduce the (0, I) bounds to the model, a transformation involving the dependent variable is required. Logit Transformation. One transformation often used to esti mate an S-shaped function is called the "logit transformation" because it is bas e d on the cumulative logistic probability function: where P. = F ( XB) 1 I + e -(XB) (3. 9) P. 1 represents the probability that the jud g e will deter mine a worker is an independent contractor given knowledge of X, X represents independent variables, B represents the coefficients of the independent vari ables, and e is the base of natural logarithms (e approx. = 2.718). One of the obvious advantages of this transformation is that the dependent variable now is constrained within the desired (0, I) boundaries. Solving equation (3.9) gives the following: XB e = P. 1 P. 1

PAGE 117

---~ -----------------------------and taking the log of both sides, XB = log P. i P. i 108 (3. 11) The right hand side of equation (3. I 1) is often referred to as the "log-odds ratio." The log-odds ratio 1.s a linear function of the f o o o A 130 X sand can be estimated by generalized least squares using P . i One important appeal of the logit model 1.s that it transforms the problem of predictine probabilities with a (0, I) interval to the problem of predicting the odds of an event occurring within the f h b 1 131 range o t e ent1.re num er ine. Once the model is estimated, the arithmetic of the logit trans formation can be reversed so as to obtain a predicted probability of a verdict of employee versus a verdict of independent contractor. This is necessary since the logit differences do not directly relate to the probability differences. Instead, the probabilities depend upon the values of the other variables. In other words, the model is additive in the lo g its, but nonlinear in terms of the probabil. 132 it1.es. The logit model will be estimated using both the Nerlov e and P 133 d h BMDP 1 . 134 ress an t e stepw1.se ogistic regression programs which generate maximum liklihood (ML) estimates using the individual (versus grouped) data. ML estimates are frequently used in statistics because they are usually known to be asymptotically efficient cstimators. Maximum liklihood estimates are based on the criterion of what underlying parameters would be "most likely" to have produced the

PAGE 118

observed data. The ML criterion is especially appealing in this study because its desirable statistical properties hold even with 135 small samples. 109 One assumption of the ML estimation technique is that the obser vations are independently obtained. Because of the nature of the judiciary system which relies on precedent of prior cases, this assumption will be violated to some degree. In this type of tax research, data collected from court cases is always subject to this bias, and any results obtained will have to be evaluated 1n terms of this limitation. 01S Results. Despite the theoretical limitations of usin g OLS on a binary dependent variable, for comparison purposes the model was first estimated using the GLM procedure available in the SAS software computer package. Since the data collected for this study represent all the cases that could be found that dealt with employ ment status for federal tax purposes, the data really r e present a population rather than a sample. For this reason, it would be incorrect to test for the significance of the OLS coefficients because such tests are designed to be able to make inferences about a population. Since the data used in this study represent the population, the coefficients were analyzed with respect to their impact on the probability estimates of Y., the indep e ndent variable 1 representing employee or independent contractor findin g Probability estimates for each case w e re obtain e d using the OLS coefficients and the cases were classified into groups using a probability of greater than 50 percent to repres e nt an employee finding and a probability of less than or equal to 50 percent to

PAGE 119

110 represent an independent contractor finding. The misclassification rates were then compared to the discriminant analysis and the logit analysis results to determine if the method of prediction makes any significant difference. As in the discriminant analysis section, the data were divided into pre-1960 and post-1959 groups with a separate model being estimated for both. 136 The Chow test was used to test the equality of the sets of coefficients between the two linear regressions. Logit Results. The resulting ML estimates were analyzed and compared with the discriminant analysis and 01S results to determine if the different methods selected the same variables to be included 1n the models. The ML estimates were then used to generate a log odds ratio for each case which was converted into a probabilty esti mate. Using a simple cutoff point of 50 percent, the cases were then classified as either belonging to the employee group or inde pendent contractor group. The misclassification rates obtained from the discriminant analysis, the 01S, and the logit analysis were then compared to determine which method was best able to classify the cases. In addition to determining which method was best able to classify the cases, the discriminant analysis, 01S, and logit models were examined to determine which variables were most important in all three methods. If certain variables are consid e red important no matter which method is used, then this is a g ood indi cation that such variables really have been useful to the courts in determining employment status for federal tax purposes.

PAGE 120

111 Research Question 3 Since the Tax Court does not have jurisdiction over the employment tax area, cases from the District Courts and Court of Claims will make up the data set for this study. Host previous tax dissertations of this nature have dealt primarily with the Tax Court. Because the Tax Court is composed of sixteen judges and decides only tax cases, it appears logical that there should be a certain degree of consistency in decisions. Howev e r, the Dis trict Courts and Court of Claims in total include a lar ge r numb e r of judges and hear many different typ e s of cases other than tax cases. Therefore, merely because of th e n a tur e of th e jurisdiction of employment tax e s, this study will consider whether or not tax research methodolo g i e s that have successfully d e t e rmined consensus in the Tax Court can b e extended into lar ge r and more divers e forums, a combination of the District Courts and the Court of Claims. Th e federal judicial system of the United Stat e s is ge o g raphically divided into e l e ven circuits, each havin g a Unit e d Stat e s Court of Appeals. Th e se circuits are th e n divid e d into districts, h . 137 wit over ninety Unit e d Stat e s District Courts. In ea ch district there may be many jud ge s sitting 1n various divisions. The Court of Claims was creat e d by a Congressional act on Febru a ry 24, 1855. 138 The Court of Claims r e sides in Washin g ton, D. C., and consists of five appointed jud ges with o ne b e in g d e si g n a ted as th e C hi ef Jud ge It is easy to sec that a combinati o n of the Distri c t Co urts and Court of Claims consists of a much more div e rse set of jud ges than do e s the Tax C o urt.

PAGE 121

I 12 Of particular interest to tax researchers is whe~her or not the choice of forum in a tax case is a critical factor for taxpayers to consider. Because of the nature of the jurisdiction of employ ment taxes, the data to be used in this study should provide an opportunity to determine if the District Court and Court of Claims differ in their decision-making processes. In dealing with the employee versus independent contractor question, it has been suggested that the Court of Claims would be the better forum. 139 In a recent dissertation, it was found that in the valuation of blocks of stock, a model based on the Tax Court cases was different from a model built from a combination of District Court and Court of Claims 140 cases. Therefore, the court that each case was tried in will be included as a variable 1n the model and analysis of this variable will determine the answer to Research Question J. To determine if the Court of Claims and District Courts have a different decision-makin g process, a discriminant analysis was used to build a model based solely on the District Court cases. The classification functions obtained were then used to assi g n both the District Court cases and the Court of Claims cases to either the employee group or the independent contractor group. A sub stantial increase in the misclassification rates for the Court of Claims cases would indicate that the legal forum may make a difference in an employee v e rsus indep e ndent contractor case for federal tax purposes. In an attempt to corroborate the discriminant analysis results, two additional OLS regression models were estimated. The first regression was based on the District Court cases and the second

PAGE 122

regression was based on the Court of Claims cases. The Chow 141 test was used to determine if the two equations came from the same population. These two tests comprise the analysis performed to answer the question concerning th e importance of the legal forum used in an employment status tax case. Research Question 4 The preceding research questions were designed to discover which variables are significant to the courts in deciding the employee versus independent contractor qu es tion. The results are based on the facts of the cases as discuss e d by the respective judges. The logical extension of this research is to determine if the variables found to be significant in Research Question 2 113 are related to the criteria contained in the proposals before Congress. If the criteria are quite differ e nt from those determined in this study, tax policy makers may want to reconsid e r those variables that hav e be e n su gge sted in ord e r to provide a more certain definition of an ind e pendent contractor. The analysis performed to answer this research question is not statistical in nature, but simply a comparison of the results obtained in this study to the four criteria proposed by the GAO and the five safe-harbor criteria contained in H.R. 5460 pres e ntly being considered by Con g ress. Assuming that the principal concern of Con gre ss in simplifyin g the common law employee definition is to add certainty to this area of the tax l a w, the add e d support of empirical findin g s of a study such as this, should result in added credibility to the factors selected for codification into the Internal Revenue Code.

PAGE 123

Notes 1 The Code and Regulation s ectio ns that were studi e d for relevant factors are the same sections discussed in det a il 10 Chapter II in the Current Legislative Status subs ec tion. 2 Marmoll, 'Employee' Defined, TAX MNGM'T (BNA) No. 391, pp. A-5 thru A-16 (1979). 3 INTERNAL REVENUE SERVICE, INTERNAL REVENUE MANUAL CCH, ch. 4600, pp. 8463-1 thru 8463-5 (1976). 4 RESTATEMENT (SECOND) OF AGENCY, Section 220, pp. 485-492 (1957). 5 INTERNAL REVENUE SERVICE, supra note 3, at 8463-1. 6 Ben v. U.S., 139 F. Supp. 883 (N.D. N.Y. 1956). 7 Id. at 886, 8 1NTERNAL REVENUE SERVICE, supra note 3, at 8463-1. 9 436 F. 2d 979 (Ct. Cl. 1971). IOid. at 981-982. I 1 INTERNAL REVENUE SE RV ICE, supra note 3, at 84 632. I 2 Ben v. U.S., 139 F. Supp. 883 (N.D. N .Y. 1956); Fa rm and Home Modernization Co rp. v. U.S., 138 F. Supp. 423 (N.D. N .Y. 1956); Fleeman v. U.S., 175 F. Supp. 336 (N.D. Ohio 1959); Bonded Insulation & Construction Co. v. U.S., 131 F. Supp. 635 (D.C. N.J. 1955). 13 Id. 14 McCombs v. U.S., 436 F. 2d 979 (Ct. Cl. 1971) a nd Alsco Storm Windows Inc. v. U.S., 311 F 2d 341 (9th Cir. 1962). 15Id. 16 163 F. Supp. 794 (D.C. Mass. 1958). 17 Id. at 795-796. 18 RESTATEMENT, supra note 4, at 485. 20 ror a mor e detailed discussion of the economic reality test, see Chapter II, pp. 27-32. 114

PAGE 124

21 rNTERNAL REVENUE SERVICE, supra note 3, at 8463-1. 22 Rin lin Bros. Barnum & Bail e Com. Shows v. Hi ins, 189 F. 2d 870 2nd Cir. 1951 as cited in Marmoll, supra note 2. 23 77-1 USTC 1 1 :)422 ( Ct. Cl. 197 7), 24 Id. at 87,072. 25 Id. at 87,073. 26 l. . ( ) F1 1p1d1s v. U.S., 27 AFTR 2d 71-595 O.C. Md. 1970 and T. L. Squared v. U.S., 33 AFTR 2d 74-991 (S.D. Ohio 1974). 27 .d. US 26 71 597 1 1.p1 1s v. . supra note at 28 r. L. Squar e d v. U.S., supra note 26, at 74-994. 29Id. 30 rNTERNAL REVENUE SERVICE, supra note 3, at 8463-1. 31Id. 32 Marmoll, supra note 2, at A-3. 33 RESTATEMENT, supra note 4, at 490, 34 INTERNAL REVENUE SERVICE, supra note 3, at 8463-1. 35 75 F. Supp. 386 (N.D. Ia. 1948). 36Id. 37 131 F. Supp. 212 (N.D. N.Y. 1954). 38Id. 39 INTERNAL REVENUE SERVICE, supra note 3, at 8463-2. 4 For a detailed discussion of H.R. No. 3245, see Chapter II, pp. 46-48. 41 11earin g on H. R 3245 B e fore th e Subc o mmitt ee on Select Revenue M e asur e s, 96th Con g ., 1st S e ss. 412 (1979) (statement of Rep. G e phardt). 42 INTERNAL REVENUE SERVICE, supra note 3, at 8463-2 43 Rev. Rul. 73-592, 1973 C.B. 338 as cited by Marmoll, supra note 2, at A-14. 115

PAGE 125

116 44 GENERAL ACCOUNTING OFFICE, TAX TREATMENT OF EMPLOYEES AND SELF EMPLOYED PERSONS BY THE INTERNAL REVENUE SERVICE: PROBLEMS AND SOLUTIONS 20 ( 1977). 45 Id. at 21. 461d. 47 H.R. 3245, 96th Con g ., 1st Se ss., 125 CONG REC El342 (daily ed. March 27, 1979) (remarks of Rep. G e phardt). 481d. 49 1NTERNAL REVENUE SERVICE, supra note 3, at 8463-2. sold. 51 RESTATEMENT, Supra note 4, at 485. 52 Treas. Reg. Sec. 31.3121(d)-l(c)(2). 53Id. 54 GENERAL ACCOUNTING OFFICE, supra not e 44, at 19. 55Id. 56 1d. at 20. 57 rNTERNAL REVENUE SERVICE, supra note 3, at 8463-2 and 8463-4. 58Id. 59 B o nney Motor Express v. U.S., 10 AFTR 2d 5231 (E.D. Va. 1962); Rahier Truckin g Inc v. U.S., 65-1 USTC ~ 9166 (E.D. Wash. 1965); and May Fre i ght S e rvice, Inc. v. U.S., 78-2 USTC ~9707 (D.C, Ut. 1978). 60Id. 61 77-1 USTC ~9356 (E.D. Okla. 1977). 621d. 63 RESTATEHENT, supra note 4, at 486. 64 rNTERNAL REVENUE SERVICE, supra note 3, at 8463-2. 651d. 66 11linois Tri-S e al Products, Inc. v. U.S., 353 F. 2d 216 Ct. Cl. 1965) mentions method of payment specifically as a factor that should be considered in defining a common law employee (at page 228).

PAGE 126

117 67 Marmoll, supra note 2, at A-12. 68 INTERNAL REVENUE SERVICE, supra note 3, at 4863-3. 69 Marmoll, supra note 2, at A-12. 7 For a more complete discussion of gy psy chasers see dis cussion of the independent trade factor supra. 71 Treas. Reg. Sec. 31.312l(d)-l(c)(2) and RESTATEMENT, supra note 4, at 485. 72 INTERNAL REVENUE SERVICE, supra note 3, at 4863-3. 73 rd. 74 58 F. Supp. 460 (E.D. Pa. 1945) as cited in Levin e Current factors that distinguish between "employ ee and ind e p e ndent con tractor, 76 rNTERNAL REVENUE SERVICE, supra note 3, 4864-4. 77 Id. 781d. 791d. sold. 8 lld. 82 GENERAL ACCOUNTING OFFICE, supra note 44. 83 Id. at 21. 84 H.R. 3245, supra note 47, at El342. 851d. 86Id. 87 ld. SSINTERNAL REVENUE SERVICE, supra note 3, at 8463-4. 89 Marmoll, supra note 2, at A-10, 90 Am e rican H ome s of New England, Inc. v. U.S., 173 F. Supp. 85 9 ( D. C. Mass 19 5 9)

PAGE 127

91Id. 92 Jagolinz e r v. U.S., 150 F. Supp. 492 (D.C. R.I. 1957). 93 Even a statistical m e thodology does not always r e sult 1n a completely objectiv e study. For a mor e complete discussion of this idea, see the Limitations section of Chapter I. 94 L O 11 Ko rt, Pred1ct1ng Supreme Court Decisions Matnemat1ca y: A Quantitative Analysis of the 'Right to Counsel' Cases, 51 THE AMERICAN POLITICAL SCIENCE REVIEW I ( 195 7). 95 1 . 1 . d 1 f Kort, Cont e nt A na ys1s of Jud1c1a Op1n1ons an Ru es o Law, in JUDICIAL DECISION-MAKING 133 (Schubert ed. 1963). 96 u1 h . . d Th 1 C mer, T e D1scr1rn1nant Function an a eore t1ca ontext for Its Use in Estimating the Votes of Judg es in FRONTIE RS OF JUDICIAL RESEARCH 335 (J. Grossman & J, Tanenhaus, 2d. 1969), 97 Greenbaum, Analytical and Sim ul a ti on Models for Explaining Judicial D e cision-Making, in FRONTIERS OF JUDICIAL RESEARCH 307 (J. Grossman & J. Tanenhaus, 2d. 1969). 98 Nagel, A Corr e lation Anal sis to Case Prediction, 42 TEXAS LAW REVIEW 1006 1964). 99 1 b h . 1 . I 1 Enge rec t, An Emp1r1ca Inv es tigation into t1e Va uat1on of Clos e ly Held Corporations by th e Tax Court for Estate a nd Gift Tax Purposes, Ph.D. dissertation, University of South Carolina (1976). 118 100 T. En g l e br e cht & D. Davison, A Statistical L ook at Ta x Co urt Com r om 1s e in Estate and Gif t Tax Valuation of Closel Held Stock, 55 TAXES 395 1977). 10 I Bromley, A Closer Statistical Look at Tax Court Co mpromis e 57 TAXES 325 ( 1979). 102 Englebrecht, A Reply, Analysis, and Extension of a Clos e r Statistical Look at Tax Court Compromise, 57 TAXES 607 (1979). 103 B o yd, Valid a tion of Guid e lin es for Dete rminin g Reaso nabl e Compen sa tion in Closely Held Corporations, Ph.D. dissertation, University of South Carolina ( 1977). 104 R. Kozub and IL Lath e n, V e rifi ca ti o n of the Tax Co urt Use of th e Mayson M.:inufacturing Co Guidelines i.n Dctermining Excessive Compensation, PROCEEDINGS OF THE SOUTHWESTERN AAA MEETING 245 ( 1979). I 05 . 1 . f d b Bond, ,\n Emp1r1c.:i Inv0st1gat1on o Cou rt Dcterunne De tEquity Attributes for Federal Incom e Tax Purpos e s, Ph.D. dissertation, University of South Carolina (1977).

PAGE 128

119 l0 6 d h 1 d . 1 Ma eo, T e Accurnu ate Earnings Tax: A n Ernpirica Analysis of the Tax Court 1 s Impl e mentation of Congressional Intent, Ph.D. dissertation, North Texas State University ( 1977). 107 1 f' . 1 . 1 Tay or, De inin g Capita Gains in Rea Estate Transactions by Applying Discriminant Analysis to D ef initional Elements, Ph.D. dissertation, University of Tennessee (1978). 108 Kramer, An Examination of th e Variabl es Affecting th e Val uation for Tax Pur oses o f Lare Holdin s of Publicl Traded Stock, Ph.D. dissertation, University of Texas at Austin 1979). 109 1 . b h . For a more compete discussion a out t e District Courts and Court of Claims judicial system, see the section of Chapter III pertaining to Research Question 3. l l O d 1 d B d I 05 See stu ies previous y mentione : on s upra note ; Madeo, supra note 106; and Kramer, supra note 108. I I 1 P. GREEN, ANALYZING MULTIVARIATE DATA 143 ( 1978). 112 BMDP BIOMEDICAL COMPUTER PROGRAMS P-SERIES--1979 (2d ed. W. Dixon & M. Brown 1979). I 13 N. NIE, STATISTICAL PACKAGE FOR THE SOCIAL SCIENCES (2d ed. 1975). 114 Id. at 445. I 15Id. 116 Lachenbruch, On Exp e cted Probabiliti e s of Misclassification in Discriminant Anal sis, Necessar Sam l e Siz e and a R elation with the Multiple Correlation Coefficient, 24 BI0!1ETRICS 823 ( 1968 117 b . f 11 h 1 f . Eisen eis, Pit a s int e App ic a tion o Discri m inant Analysis in Busin e s s Finance, and Economics, 32 JOURNAL OF FINANCE 883 ( 1977). 118 'lb O . . U Q 1 V bl Gi ert, n Discrimination sin g u a it a tiv e aria es JOURNAL OF THE AMERIC AN STATISTICAL ASSOCIATION 1399 (December, 1968) and Krzanowski, Discrimination Using Both Binary and Continuous Variables, JOURNAL OF THE Al1ERICAN STATISTICAL ASSOCIATION 782 (December, 1975) as cited in Madeo, supra not e 106. l 19 Id. at 229. 120 w. COOLEY & P. LOHNES, MULTIVARIATE DATA ANALYSIS 267 ( 1971). 121 Gilbert, Th e Effect of Unequal Varianc e -Covariance Matrices on Fisher's Linear Discriminant Function, 25 BIOMETRICS 505 ( 1969). 122 Id. at 512.

PAGE 129

120 123 Id. at 508. 124 unequal error variances can be handled by weighted least squares. The dependent variable responses can arbitrarily be forced to fall between O and I. Even though non-normal error terms exist, OLS gives unbiased estimators if the sample size is lar ge enou g h to have asymptotic properties. However, each of th e se solutions have tradeoffs and will place limitations on how the results can be interpreted. 125 E. HANUSHEK & J. JACKSON, STATISTICAL METHODS FOR SOCIAL SCIENTISTS 182 ( 1977) (hereafter cited as HANUSHEK). 126 Id.at 183. 127 J. NETER & W. WASSERMAN, APPLIED LINEAR STATISTICAL MODELS 329 ( 1974). 128 Id. at 32 9. 129 HANUSHEK, supra note 125, at 185. 130 1d. at 191. 131 R. PINDYCK & D. RUBINFELD, ECONOMETRIC MODELS AND ECONOMIC FORECASTS 248 ( 1976). 132 HANUSHEK, supra note 125, at 195. 133 The Nerlove and Press pro g ram was modified by th e Economics Departm e nt of North Carolina State University. This program is available as a SAS subroutine at the University of Florida. 134 BMDP, supra note 112, at 517.1. 135 HANUSHEK, supra note 125, at 203. 136 chow, Test of E ualit betw ee n Sets of Coefficients in Two Linear Regressions, 28 ECONOMETRICA 591 1960 137 tt. BICKFORD, SUCCESSFUL TAX PRACTICE 333 (1967). 138 1d. at 352. 139 williams,How to Alleviate Tax Burdens wh e n IRS Claims that Indep e ndent Contractors are Employees, 6 TAXATION FOR LA\,~ERS 364 (1978). 140 Kramer, supra note I 08, at 20 I. 14 ICh ow, supra note 136.

PAGE 130

CHAPTER IV EMPIRICAL RESULTS Research Question 2 Discriminant Analysis Results The stepwise 1 discriminant analysis pro g ram of th e BMDP computer software packa ge was used to derive a mod e l based on th e eleven independent variabl e s discussed in co njunction with Research Question I in Chapter III, Variables wer e entered into the model that maximized the difference betw ee n g roup m ea ns. To maximiz e the dif fe rence b et w ee n gro up means, th e variable that maximized this overall F-ratio (which also minimiz es Wilks' lambda) was entered into th e model. Tabl e 4. I presents the vari ables and the order in which they w e re entered into th e model alon g with their corresponding Wilks' lambda and their F-values when entered into the model. Usin g th e size of the F-ratio and Wilks' lambda as an indication of th e impact of certain variables on the model, it appears that supervision is the predominant factor 1n t e rms of imp ac t on d eter minin g a work e r's emp loym e nt status. The profit-or-l oss variabl e and th e independent-trade variable also provide sub s tantial decreases in Wilks' l ambda The r ema ining three variables are a lso statistically significant (at the 0.05 level) and do add discriminating ability to th e model e ven thou g h 12 I

PAGE 131

122 TABLE 4.1 BNDP 6-VARIABLE ~IODEL Order of Wilks' F Entry Variable Lambda Ratio 1 Supervision 0.411566 208.742 2 Profit or Loss 0.292841 58.787 3 Independent Trade 0.248914 25 .41 2 4 Permanent Relationship 0. 2 342 21 8. 971 5 Hiring of Assistants 0.224208 6. 342 6 Integration 0.216542 4.991

PAGE 132

123 their effe cts are n ot n e arly as g r e at as the fi rst thr ee vari a bl es entered into the model. The SPSS stepwise discriminant analysis pro g ram enters vari ables into the model until their F-values-to-enter are less than one. The BMDP stepwise discriminant analysis pro g ram will not enter a variable into the model unless th e F-value-to-enter is four or greater. Thus, the SPSS discriminant analysis program results in a model with nine variables instead of six. Table 4.2 presents the variabl es that e ntered th e mod e l with their corresponding entry-level F-ratios and Wilks' lambda. Comparing Table 4. I and 4.2 it can be seen that the two programs enter the same variables into the model up to and including th e sixth vari able, then the SPSS pro g ram enters an additional thr ee variables: (I) controlling the place of work, (2) reimbursement of business expenses, and (3) method of paym e nt. To d e t e rmin e i f the addition of these three variables is si g nificant, the followin g test su gges t e d 2 by Rao was used: where m n I Q = n p m equals the total sampl e size, n equals the number of variabl e s 1n th e full mod e l, p equals th e numb e r of variables in th e r es tricted (or BMDP) model, o 2 d eno t es the .MahaLinobis o 2 m e.:is ur e and C equals rn m 2 /m(m 2) where m 1 eq uals th e ac tu .:il number al ind e p e nd e nt contractor findin g s in the data and m 2 equals the number of employee findin gs in the data.

PAGE 133

124 TABLE 4.2 SPSS 9-VARIABLE MODEL Order of Wilks' F Entry Variable Lambda Ratio Supervision 0.411566 208.742 2 Profit or Loss 0. 29284 1 58.787 3 Independent Trade 0.248914 25.412 4 Permanent Relationship 0.234221 8.971 5 Hiring of Assistants 0.224208 6.342 6 Inte g ration 0.216542 4.991 7 Controlling Place of Work 0.212136 2.908 8 Reimbursement of Expenses 0.208831 2. 120 9 Method of Payment 0.205552 2.202

PAGE 134

125 The statistic Q has a F( I) distribution. Rao's n p, m n test is extremely useful in stepwise discrimination procedures as used here to determine if a set of additional factors contribute significantly to the discriminant function. When this test is applied to the addition of the three variables by the SPSS model, Q equals 1.22977 which is not significant at the 10 percent level. Another indication of whether the additional three SPSS variables contribute to the model is how well the 9-variable model does in its ability to classify the cases correctly as compared to the 6-variable model. From Table 4.3 it is evident that even though the actual classifications are not identical, the overall classification accuracy is unchanged for the two models. Because of the similarity in classification accuracy and the results of Rao's test, the more parsamonious 6-variable model will be used in preference to the 9-variable mod e l in the remainder of the discriminant analysis discussion. As discussed in Chapter III, the discriminant function coef ficients do not have tests of significance as th e re g ression coef ficients. The BMDP program provides standardized discriminant coefficients which have been calculated such that when applied to the data in standard form, the resulting discriminant score will have a zero mean and standard deviation of one. Therefore, ''the standardized dis c riminant coefficients arc of ~rca t an~lytical importance. When the si g n is ignored, each coefficiL'nt represents the relative contribution of its associated variable to that function .. analogous to the interpretation of beta weights in multiple regression. 113 Table 4.4 provides both the standardized

PAGE 135

TABLE 4.3 COMPARISON OF CLASSIFICATION ACCURACY Classification Results Independent Contractor c as es Correctly Classified Independent Contractor cases Incorrectly Classified Employee cases Correctly Classified Emplo yee cases Incorrectly Classified Percenta ge of Correct Classifications 6-Variable Model 95 2 48 3 96.6 126 9-Variable H odel 94 3 49 2 96 .6

PAGE 136

127 TABLE 4.4 DISCRIMINANT & CLASSIFICATION COEFFICIENTS FOR THE 6-VARIABLE MODEL Discriminant Classification Coefficients Coefficients Variable Standardized Unstandardized Emplo y ee Independent Contracto r Supervision -0.64582 -1. 1448 2 7.95209 3.40105 Profit or Loss -0.46002 -0. 75563 18.44061 16.04216 Independent Trade -0.36226 -0.59897 4.16722 2.66930 Permanent Relationship -0.30185 -0.377 84 5.32950 3.82747 Hirin g of Assistants -0. 22049 -0.37681 4.63323 1. 62934 Integration -0. 21458 -0.60334 4.63323 1.62934 Constant 6.624402 -55.2111 84 -26.45 177

PAGE 137

128 and unstandardized 4 discriminant coefficients as well as the classi fication function coefficients for the 6-variable model. The unstan dardized coefficients are also useful for our purposes because they can be applied to the data in their raw form and generate the same discriminant scores as do the standardized coefficients applied to the standardized data. By examining the standardized coefficients, it is again the supervision variable that provides the greatest relative contribution to the discriminant function. The profit-or-loss variable, the independent-trade variable, and the permanent-relationship variable provide l es s discriminating power that the supervision variable, but the three variables are substantially higher than either the hirin g -of-assistants or the integration variables that make up the rest of the model. Another measure of the success of the discriminant functions is how well the discriminant results can classify th e cases con tained in the study. Table 4.3 presents the classification co e ficients that are applied to the raw data to determine to which group the case should be assigned. Table 4.5 presents a classi fication matrix based on the classification coefficients of Table 4.3 and then the less biased results of the Lachenbruch jackknife classification proc e dure. The ov e rall classification accuracy of the BMDP 6-variable model is 96.9 percent while the jackknife procedure was slightly l e ss accurate at 95,3 perc e nt. Either method provides a hi g h perc e nta ge of corr ec t classifications which indicate that the variables included in the model are ef fective in discrimi nating between an employee versus independent contractor finding.

PAGE 138

6 -Variable Model Classification Matrix Jackkn ife Classification Matrix TABLE 4.5 CLASSIFICATION ACCU RAC Y MATRICES A c tual P e r ce nt G roup Corr e ct Ind e p e nd e nt C o ntractor -97. 9 Emplo ye e -94. 1 Total 96 6 In de p e nd e nt Contractor-7 6.9 Empl oye e -92. 2 Total 95.3 129 Pred ic ted Ind e p e nd e nt Contractor Employee 95 2 3 48 98 50 9 4 3 4 47 98 50

PAGE 139

130 Representing Data as Dummy Variables. In Chapter III the fact was discussed that treating e ach of th e variables as two dummy variables used more information contain e d in the data than simply representing them as sin g le vari a bl e s with thr ee valu es On the other hand, doublin g the number of indep ende nt variables r ed uces the stability of th e r e sults and makes the analysis much more complex. To determine if usin g dummy variabl es was superior to th e simpler II-variable model, the discriminant analysis procedure was run again using the twenty-two dummy variables. Table 4.6 presents the variables select e d to be in the m o d e l with th e ir corresponding Wilks' lambdas and F-va lu e s. The standardized and unstandardi ze d discriminant co e f f ici e nts are also pr ese nt e d. Each main effect variable is split into two dummy variabl es by assigning the vanable a or a 2. The I d e notes the absence of th e variable which favors an ind e p en d e nt contractor finding and a 2 d eno t es th e pr e s e nce of the variable which favors an em ploy ee finding. It is interesting to not e that th e supervision! dummy variable, profit-or-loss) and 2 dunnny variabl e s, and ind e pendent-trad e I and 2 dummy variabl e s were the first fiv e variables entered into th e model much as in the ori g inal comput e r run. An additional method of comparison for th e two mod e ls is to examine how th e models compare with respect to their ability to corrl'Ctly cl.:1ssify th e cases T.:1blc 4.7 shows that th e r.hld<.' l ob t a in ed irl)m thL' dunuuv v ; iri 1blL :-; cl:1s:-:ifics t \.JO cases correctly th a t the 6-variable model misclassifi e d. ll owc v c r, an increase in the co rrect classification r ate of approximately 1.3 percent was not considered material wh e n contrasted with the

PAGE 140

13 I TABLE 4.6 MODEL BASED ON 22 DUHHY VARIABLES Discriminant Coefficients Wilks f Variables Lambd a Ratio S t a ndardi zed U nst a ndardized (1) Supervisionl 0 39895 62 150 0.66549 -2.35639 (2) Profit or Lossl 0 32359 23 902 0.63683 -1. 58444 (3) Profit or Loss2 0.27593 7.545 0.37350 0.94744 ( 4) Independent Tradel 0.24764 22.085 -0.699 89 -1.70474 (5) Independent Trade2 0 23092 12.552 -0.5 3054 -1. 33532 (6) Permanent Relationship! 0.22053 13. 156 0 40 1 87 -0. 89088 ( 7) Integration2 0.20995 6 370 -0. 25728 -0.72993 (8) Furni shi n g of Tools2 0.20424 5 .1 42 -0. 23522 0.75258 (9) Supervision 2 0.19940 2 26 7 0. 155 30 1. 35367 (10) Place of Workl 0.19599 4.950 -0.237 48 -0.53146 (11) Place of Work2 0.19246 3. 496 0.20650 -0. 54496 (12) Method of Pa yme nt! 0.18932 3.358 0 20104 0.62371 (13) Hiring of Assistants2 0. 18703 8 206 -0. 46330 -1. 05150 (14) Hirin g of Assistants l 0 177 36 7.888 -0.47545 -1. 073 16 (15) Reimbursem e nt of Expenses! 0. 1751 6 1. 688 -0.1 3449 -0.2 8088 (16) Reimbursem e nt of Expenses2 0.17369 1. 105 -0.118 33 0.28930

PAGE 141

132 TABLE 4. 7 CLASSIFICATION ACCURACY MATRIX Pr ed i c t e d Actual P e r ce nt Ind e p e nd e nt Group Correct Contra c t o r Employee Indep e ndent Contractor 98.9 96 1 Emplo y ee 96 1 2 49 Total 9 7. 9 9 8 50

PAGE 142

133 fact that th e model contains an additional ten vari a bl es Since the variables contain e d in the two mod e ls are almost id e ntical coupl e d with the fact that the classification accuracy is increased only 1.3 percent by addin g the ten variables to the model, the discrim inant model that r es ulted from treatin g ea ch variable as a single trichotomous variabl e was us e d. This will re s ult in increased stability of the estimates obtained as well as simplifyin g th e inter pretation of the results b e cause of th e fewer numb e r of variables contained 1n the analysis. Prior Probabilities. The r e sults obtained thus far have been dependent on th e fact that the discriminant analysis assumed that no prior knowledge of the probabilities of g roup membership existed. Howev e r, in this study th e population fr e qu e ncies of em ploy ee and independent contractor d e cisions are known. The population fre quency of an employee findin g is 34 p e rcent and the population frequency of an ind e p e ndent contractor finding is 66 p e rcent. To determin e what the impact on classification accuracy would be if the population probabiliti e s were us e d in the Bayesian rev1s1on of probabilities performed by th e discriminant pro g ram, the model was re-estimated and the cases r e -classified bas e d on the ex ante population probabilities. The overall classification rate for the jackknife procedure was 95.3 perc e nt. When compared with Table 4.5, it can b e seen that this is id e ntical to th e r es ults obtain e d wh e n the prior prob a biliti es w e r e assumed to b e equal. The classification accuracy of th e r eg ular 6-variabl e mod e l applied to the ori g inal data is increas e d by one additional correctly classified case. The improvement 10 classific a tion accuracy was

PAGE 143

134 considered insi g nificant; therefore no furth e r analysis was perform e d usin g the population probabiliti es T em poral Stability of the Model. To test for the stability of the discriminant model over time, the model was es timat ed using onl y pre-1960 cases. The results w e re th e n us e d to classify not only the pre-1960 cases, but also the post-1959 cases. Table 4.8 presents the results of the re g ular classification matrix applied to the original data and the jackknife classification matrix. The overall classification accuracy of the re g ular classification procedure for th e pre-1960 cases is 95,9 p e rcent as compared with the overall classification accuracy for the post-1959 cases of 94.7 percent. The jackknif e r esults are e ven closer with both pre-1960 and post-1959 p e riods misclassifying four cases. From the results of the classific a tion matrices, it is apparent that the tim e in which a case was tried is not a si g nificant factor in det e rminin g th e e mploym e nt status of a worker Because of the simplicity of the above t em poral stability test, a similar test will be perform e d in the OLS section in an attempt to corroborate these discriminant analysis results. T es t of Equal Dispersion Matrices Th e equality of the disper sion matrices was tested usin g Box's 11 test. 5 Box d e fines the test crit e rion Mas where N represents the total sample size, g r e pres e nts th e number of g roups, D is the pool e d-withing roups covariance matrix, and w Dk is the covariance matrix of th e kth gro up.

PAGE 144

6 Variable Hodel Classification Matrix Jackknife Classification Matrix TABLE 4.8 TEr~ORAL STABILITY TEST Of CLASSIFICATION ACCURACY Actua l P e rc e n t Grou p Correct Pre -1 960 Indep e nd e nt Co ntractor-100.0 Pr e -1 960 Employee-85. 7 T o tal 95.9 Post -1 959 In d e p e n d e nt Contractor 95.6 Post1 959 Employee -93 3 Total 94. 7 Pre -1 960 Independent Contractor -100. 0 Pre -1 960 Employee-8 1. 0 Total 94 5 r o st -1 959 Ind ( pe n dc nt Co ntr a c t o r-95.6 P os t-1959 Employee -93.3 Total 94.7 135 Predicted In dep end e nt Contractor Em?loyee 52 0 3 18 55 18 43 2 2 28 45 30 52 0 4 17 56 17 43 2 2 28 45 JO

PAGE 145

136 The SPSS disc~iminant analysis pro g ram estimated the Box M t es t statistic to be 239.7. The Box M statistic has an F-distribution with twenty-on e de g rees of freedom in the numerator and 39,434 degrees of freedom in the denominator. The resulting F ratio is 10.839 which is si g nificant even at the one p e r ce nt l e v e l. Therefore, with the F-ratio bein g hi g hly si g nificant it is r e as o able to assume that the dispersion matric e s of the two g roups are not equal. As discussed in Chapter III, this may r e sult in more cases bein g classified into the g roup with the lar ge r dis persion. This possibility has not materializ e d in this study because the classification accuracy is extremely hi gh for both the employ ee and indep e ndent g roups. 0LS Regression Results Even thou g h usin g 0LS to mod e l a dichotom o us d e p e ndent var iable results in error t e rms that are h e t e rosk e dastic and n o normal as well as estimat e s that are not constrained to th e (0, I) interval for which a dichotomous depend e nt variable is d e fin e d, the 0LS results w e r e very similar to the discriminant analysis results previously discussed. Tabl e 4.9 pr e sents the results of OLS applied to th e eleven variabl es that r e sult e d from Research Question I. Even thou g h statistical si g nificanc e is not really appropriate b e cause we are dealin g with a p op ulation a nd not a sampl e the lev e ls of significance as pr ese nt e d in Tabl e 4.9 give some indication of tl1 e more important v a riabl es The three most significant variabl es a r e th e same thr ee variabl e s resultin g from the discriminant analysis; i. e .,

PAGE 146

13 7 TABLE 4.9 OLS REGRESSION RESULTS Regression Significance Variable Coefficients Level(%) Supervision 0.21197195 0.0001 Integration 0. 1044468 2 0.0001 Hirin g of Assistants 0.08306364 0.0535 Perman e nt Relationship 0.07638216 0.0107 Set Hours of Work 0.00438250 0.0016 Controlling the Place of Work 0.03915166 0.8696 Independent Trade o. 13592924 o. 1042 Method of Payment -0.05033621 0.0001 Reimbursement of Expenses 0. 04 5504 39 o. 1406 Furnishing of Tools -0.00302913 o. 1380 Profit or Loss 0. 14 105293 0.9013 Intercept -l .03689 2 .'.i9 0 0001 1/ o. 794524

PAGE 147

138 supervision, profit or loss, and independent tr a d e The oth er three variables that would be statistically significant at th e 10 p e rcent level are p e rman e nt relationship, hirin g of assistants, and integration. When these r es ults are compared with Table 4. I it is observed that the six variables produced by the Bl1DP step wise discriminant analysis are id e ntical to those variabl es pro duced by the OLS r eg ression analysis and are statistically sig nificiant at the 10 percent lev e l. To further compare th e discriminant analysis and OLS results, a stepwise multiple regression m o d e l was estimated. Table 4. 10 lists the variables that were entered into the mod e l, the order in which they were entered, th e re g r ess ion coefficients, and the . 2 correspondin g improv eme nt 1n R In comparing Tabl e 4. 10 with the discriminant results present e d 1n T a ble 4. I, th e only diff e ence 1n the mod e ls is that the stepwise re g r essio n model adds a seventh variable; i.e., controllin g th e plac e of work. The other six variables are the same in both stepwise models e v e n with r es p e ct to the order in which th ey were e~tered into the anal ys is. Th ere fore, whether a r eg ular r eg ression or a stepwise procedure is used, the variabl e s which prov e significant ar e identical (with one exception) to those variables ge nerat e d in the stepwise discrimi nant analysis. Th e coefficients of th e ind e pend e nt variabl e s can be inter pret ed as the impact of .1 p.1rticular variabll' on the prob ; 1bility of an employee finding. The way th e dat a wer e coded, a pr e diction of an emp loye e finding should be close to one and a pr e diction of an independ e nt contractor findin g should be close to zero.

PAGE 148

139 TABLE 4. 10 OLS STEPWISE REGRESSION RESULTS Order o f Regression IJTJpr oveme nt Total Entry Variables Coefficients in R 2 R2 (1) Supervision 0. 2 1636 0.513843 0 513843 (2) Profit or Loss 0.14957 0.11873 0.70715 (3) Ind epe ndent Trade 0.11762 0.04393 o. 75109 (4) Perm a n e nt Relationship 0.07249 0.01469 0 76578 (5) Hirin g of Assistants 0.07513 0.01001 0. 77579 (6) Int eg ration 0. 10408 0.00767 0.78346 (7) Controlling Place of Work 0.03821 0.00441 0 78786 Constant -0 79700

PAGE 149

140 If all eleven variabl e s w e re cit e d in a cas e in favor of an in dependent contractor findin g the OLS model would pr e dict a score of -0.2483701 which is less than zero and really is difficult to interpret sinc e th e ind e p e nd e nt variable is d ef in e d only between zero and one. If all th e variabl es w e re e ith e r not mentioned or deemed to b e unimportant, this mod e l would pr e dict a score of 0.5368539 which would b e interpr e t e d as a 53.7 p e rc e nt probability of an em ployee findin g A ss umin g that a lack of a ny important factors in an employment status c a se would result in approximately a SO-percent chanc e of an e mploy ee finding and a SO-p e rc e nt chanc e of an independ e nt contractor findin g this r es ult is not unr ea listic. If all eleven variables w e r e m e ntion e d in favor of an em ploy ee findin g the pr e dict e d score would b e 1.3 253 767 which is in the right direction, but it is again o ut o f th e d ef in e d r a n ge for our d e pendent vari a bl e In other words it makes no sense to conclude that th e re i s a 1 32 per ce nt chanc e of an em ploy ee finding The pr ec edin g discussion points out o n e of th e pr o bl e ms of usin g OLS. H owe v e r, to illustr a t e th e problem, tw o p o lar cases w ere sel e cted in which all the variables were e ither in favor of an employee findin g or all variabl es were in favor of an ind e pend e nt contractor finding Wh e n usin g ac tual cases th e results are more meanin gf ul. To determin e how th e OLS m o d e l does in predictin g th e 148 decisions us e d in this study, a pr e dict e d score was ge n e r a t e d for each c ase Usin g a simple cutoff p o int of 50 p ercen t (below 50 perc e nt indi ca tin g an independent contractor finding and SO percent or abov e indicatin g an e mploy ee finding), th e OLS mod e l was

PAGE 150

14 I able to correctly classify 141 cases for an overall classification accuracy of 95.3 p e rcent. These classification rates are within 2 perc en t of the discriminant analysis classification rates. Both discriminant analysis and OLS re gressio n are able to pr e dict a high percentage of cas es corr ec tly, with neither model doin g significantly better than the other. The caution must again be ex pr essed that classifying tl1e same cas es from which the estimates were obtained is not a completely valid test of a model's si g ni ficance, but it do e s give some indication of whether or not the model is capturin g the basic mathematical relationship The OLS regression results were also used to test the t em poral stability of the model. The data were divided into two time periods, pre-1960 and post-1959. The mod e l was then es timat ed for each . 6 tim e p e riod and th e Chow test was used to test for th e equality of re g ression coefficients between the two eq uations. The fo lowin g statistic was calculated : (SEETOTAL SEEPRE SEEPOST) / k where SEE represents the residual sum of squar es k ref e rs to the numb e r of coefficients in each equation n e quals th e total number of observations, TOT,\L r efe 1 -s t o th e combined eq uati on PRE r efer s to the pr e -1960 re g ression eq u a tion, a nd POST ref ers to the post1959 re g re ss ion equat ion. This statistic is distributed according to an F-distribution with k d egrees of freedom 1n th e num e r ato r and n 2k d eg rees of freedom

PAGE 151

14 2 1n the denominator. Th e valu e of the statistic 1n this cas e is 0.58708 and when compar e d with th e critical F-value of 1.75 (this is at the 5 perc e nt si g ni f icanc e l e vel). Th e null h y poth e sis that the two re g ressions come from the same population clearly can not be rejected. Th e refore, th e r e sults of th e Chow t e st cor roborate the discriminant analysis r e sult by concludin g that the re g ressions estimat e d for the two separat e time p e riods are not si g nificantly differ e nt from each oth e r. The conclusion from these two separate tests is that the model appears to be stable over time. Logit Analysis R e sults Two different approaches were tak e n in obt a inin g ML estimat e s from the logit analysis. First, the Nerlove and Pr e ss pro g ram as contained in the S A S computer software packa ge w a s us e d to estimate th e e ntire I I-variable mod e l. S e cond, the Bl 1D P stepwis e logistic re g ression pro g ram was us e d to obt a in a mor e parsamon1ous model. If comparisons of the full model with th e mor e parsa m onious model prove that a smaller number of variabl e s can e stimate th e model just as w e ll, the mor e concis e mod e l will b e us e d for com parison purposes with the discriminant analysis and OLS re g ression results. Table 4. I I pr e s e nts th e r e sults o f the SA S e sti ma tion o f th e full I I-vari a ble model. The coeffici e nts b y th emse lv e s ar c dif ficult to int e rpr e t b e cau se they ar e not dir e ctl y pr o p o rti o nal to the probability estimat e s. This is due to th e f a ct that th e probability estim a t e s d e p e nd on th e valu e s of th e oth e r variabl e s

PAGE 152

TABLE 4. 11 COEFF ICIENTS AND CLASSIFICATION MATRIX FOR FULL 1 I-V ARIABLE MODE L Variabl e Supervision Integration Hiring of Assistants Permanent Relationship Set Hours of Work Controlling th e Place of Work Independent Trad e Method of Pa y m e nt Reim bur sement of Expense Furn is hing of Tools Profit or Loss Constant Classification Accuracy M a trix Actual P e rc e nt Ind e p e ndent Group Correct Contractor Ind e p e ndent Contractor-96 ,9 94 Employee 9 4. I 3 Total 95,9 97 143 Coefficients -3.51 07 -3. 9233 0.0867 -2.8738 -1.6459 -0.2561 -3.9 850 2 16 13 -1. 6685 -0.2168 I.010 I 33.5703 Predicted Employee 3 48 51

PAGE 153

1n the model. Usin g th e co e ffici e nts applied to th e raw d a ta, a lo g -odds ratio w a s calculated for e ach cas e To conv e rt this ratio into a probability th e followin g transformation was perform e d: wher e p = e l 1 I + e P r e pr e s e nts th e prob a bility es tim a t e and 1 e r e pr ese nts th e anti-l og o f th e lo g -odds ratio. Usin g a simpl e c uto ff p o int of 50 p e rc e nt (b e low 50 p e rc e nt r e pres e nts an ind e p e nd e nt contractor and 50 p e rc e nt and ab o v e r e pres e nts an e mplo yee ) th e ca se s we r e c l ass ifi e d into e mploy ee or 144 ind e p e ndent contra c tor g roups. T a bl e 4. I I sh o w s that thr ee e mploy ee plus three ind e p e nd e nt contractor ca se s w e r e m isclas s i f i e d for an overall classi f i c ation a ccur ac y o f 95.9 p e r ce nt. Table 4. 1 2 pr e s e nt s th e results of th e B l 1D P st e p w i se l og i s tic r eg ression r e sults. Onl y f iv e vari a bl es we r e a bl e to s urviv e th e entry requirem e nts o f th e st e pwis e pr og ram. Th e sa m e pr o c e dur e was p e rformed for th e 5-variable mod e l with r e sp ec t to calculatin g log-odds ratios and prob a bility pr e di c ti o n s as wa s pr e viou s l y described in conn e cti o n w ith th e full I I-v a ri a bl e mod e l. Thr e e employ ee cas e s a nd on e ind e p e nd e nt c o ntr a ct o r c ase we r e mi sc l as sifi e d which r es ult e d in an ov e rall c l a s sif ic a ti o n acc ur acy of 97.3 p e rcent. Th e fact th a t th e 5-v a ri a bl c BMD P m o d e l i s b e tt e r abl e t o pr e dict than th e I I-v a ri a ble S AS mod e l is di ff icult to ex pl a in. To d e t e rmine if the di f f e r e nc e is s impl y du e to th e di ffe r e nt

PAGE 154

Order of Entry ( I) (2) (3) (4) (5) Actual Group Independent TABLE 4. 12 COEFFICIENTS AND CLASSIFICATION MATRIX FOR THE BMDP 5-VARIABLE MODEL Variable Profit or Loss Supervision Independent Trade Permanent Relationship Integration Constant Classification Accuracy Matrix 145 Coefficients -1. 528 -2.826 -I. 761 -1.977 -2. 726 2 I. 155 Predicted Percent Ind e p e ndent Correct Contractor Employee Contractor-98.9 96 Employee-94. I 3 48 Total 97.3 99 49

PAGE 155

146 computer pro g rams, the same S-variable mod e l was estimated by the SAS N e rlove and Press pro g ram. The ind e pendent variable coef ficients and the probability estimates obtained from the SAS program were identical with those obtain e d from the BMDP stepwise procedure. Therefore, the increased classification accuracy was not du e to the diff e rence in comput e r programs. Inst ea d, the more parsamonious 5-variable model is simply better able to classify the cases than the full I I-variable mod e l. For classification purposes, the addition of the extra six variabl e s appears to obscure the underlyin g r e lationship. Consequently, the 5-variable model will be used in comparin g the r e sults of the discriminant, the OLS regression, and lo g it analys e s. Comparisons of D iscrimin a nt, OLS Regr e ssion, a nd L ogi t A n a lys es In each of th e thr ee t e chniques us e d to model t he e mploy e e versus indep e nd e nt contractor decision, c e rtain vari a bl e s have repeatedly surfac e d as th e factors that hav e th e most impact on an employment status cas e Table 4. 13 lists the variables that were included in th e steowise orocedur es us e d for ea ch of these analvses. The discriminant analysis and OLS r eg ression results are id e ntical except that the OLS analysis included controllin g the place of work as an additional variabl e Th e le g it analysis re sults ar e al s o v e ry s imilar to th e two lin ea r es ti ma tion techniqu es Th e m.Jin diffL r c nc c 111 tht' lo l-! it tL 'S ults is lhL rL'VL r s in 1 ~ L)l tlw entry p os iti o n s o f th e s up e rvi sio n and pr o fit(, r-lo ss variabh's and th e e xclusion of th e hirin g -of-assi s t a nt s vari ab l e All three m e thods includ e (I) supervision, ( 2 ) profit or l oss

PAGE 156

14 7 TABLE 4. 13 COMPARISON OF STEPWISE RESULTS Ord e r of Entr y Discrimin ;: mt OLS Lo g it Variable Analysis Re g ression Analysis Supervision 1 1 2 Profit or Loss 3 2 Indep e ndent Trade 2 3 3 Permanent Relationship 4 4 4 Hirin g of Assistants 5 5 Integration 6 6 5 Controllin g th e Place of Work 7

PAGE 157

148 (3) independent trade, (4) permanent relationship, and (5) in tegration as variables that are strongly related to th e employment status question. The hiring-of-assistants variable was included in two of the analyses and th e controlling-the-place-of-work variable was includ e d in only the OLS regression results. As far as which of the thr ee techniqu es was "best" for this particular study is difficult to determine. The th eo r e tic super iority of the lo g it analysis has already b ee n w e ll documented in Chapter III. Howev e r, if the classification r es ults of the three methods are compar ed th e theoretic superiority of th e lo g it analysis is not matched by its superior pr e dictive ability The discriminant model misclassified five cas e s, th e OLS re g ression model misclassified s e v e n cases, and the logit mod e l misclassified four cases. In lookin g at th e misclassified cases it is apparent th a t some of the jud g es d ec ision-makin g models are not r e pr e sented by the composite court mod e l construct e d 1n this study In total, seven different cases were misclassified by th e thr ee est imation techniques. Of those sev e n cases, four were clearly borderline cases and the probabiliti e s generated by th e m a thematical models were still reasonable. However, in th e oth e r three cases 7 the probability predictions w e re substantially diff e r e nt from th e actual verdicts rendered. In examining till cases individu.11ly, there i s no common tllrcad bL LWL'L'n tli c c ~ l SL S tli;:it would indic ~ 1l c why th e jud ge s ee m e d to discuss several f~ctors in fa vor of o ne findin g but then hold for the opposite findin g A District Court judge in one of th e misclassified cas es ex plain ed

PAGE 158

the difficulty 1n decidin g some employment status cas e s as follows: 149 C e rtain of th e factors r e co g ni ze d as t e sts to d e t e rmine wheth e r an employer-employee r e l a ti o nship exists t e nd to show such a relationship whil e other of th e t es ts tend to show th a t th e swamp e rs were ind e p e nd e nt con tractors. At th e same tim e many of th e reco g niz e d tests are not h e lpful in th ese cas e s because the nature of the work done leaves th e m m e anin g less 8 As is pointed out by this District Court jud g e, th e se types of taxpayers generally are not clearly e mpl oyee s or ind e p e ndent contractors. Th e jud g e must attach a certain amount of weight to each factor and th e n make his decision. In th e thr ee cas e s 1n which the math e matical models w e re unabl e to clo se ly approxi mate the actual v e rdict, the jud ge simply did not attach much weight to th e variables that constitute th e math e matic a l models, For example in all thr ee cases th e sup e rvisi o n variabl e was discuss ed in favor of a finding opposite to th e verdict. In two of th e three cas e s, th e indep e nd e nt-tr a d e v a riabl e wa s dis cuss e d in favor of a findin g opposit e to th e v e rdict. Ho we v e r, other variables which th e judge obviously we i g h e d v e ry h ea vily were found to support a conclusion diff ere nt from that of th e supervision and ind e p e nd e nt-trad e variabl es Th e obvious c o clusion that is illustrat e d by th e se misclassified cases is that the composite m a th e matical models obtained in this study arc not going to accurat e ly r e pres e nt each of th e individu a l f c d c r.:il t.:ix judg es decision-makin g b e havior. Ev e n th o u g h th e l og it mod e l miscl.:i ss ifi e d th e few e st nu m b e r of cases, th e amount of improv e m e nt w~s minim a l. C la ss ification accuracy is not nec e ssarily the only valid method of comparin g

PAGE 159

150 the three analyses p e rformed in this study. It does indicate, however, that for the employment status question, either the linear or non-linear estimation techniques seem to do equally well This situation may not necessarily occur 1n other studies dealing with a dichotomous dep e ndent variabl e and th e r e fore the results of this study should not be taken as an indication that the non linear (legit) analysis will nev e r produce superior results when modeling a curvilinear relationship. For example, e ven in this study the le g it r es ults mad e more sense in t e rms of the actual estimated probabiliti es because all of th e pr e dictions were at least constrained b e twe e n zero and one On the other hand fifty-one (34.5 perc e nt) of the OLS estimated probabilities f e ll outside of the (0, I) interval. For classification purpos es th e y were simply assumed to b e e ither zero (if th e estimated probability was ne g ative) or one (if the estimated probability was g r ea ter than one), but when a lit e ral int e rpr e tation is d es ir e d, th e OLS estimates are not always relevant. In ge neral, the discriminant, OLS r eg ression, and lo g it analyses agreed on which factors were most relevant in determinin g what constitutes a coannon law employee. All three analyses were able t o classify in excess of 95 percent of the cases correctly. With these results from Research Question 2, the diff e rent l eg islativ e propo sa ls c.rn n ow be inv es tigated t o ascertain if those prop osa ls are suppLHt d by L'mpi.ri c. : 11 f'Vltkncc. lkf l ) r f' the' le g is lative proposals are a n a ly zed and compared in R esearch Question 4, the importance of th e l ega l forum 1n em plo yme nt tax cases will be consider e d in Research Question J.

PAGE 160

151 Research Qu e stion 3 To determine wh e ther the le g al forum (District Court versus Court of Claims) makes a differenc e in a common law e mploy ment tax case, a similar approach to that u sed in answerin g th e temporal stability question was impl eme nt e d. F irst, th e BHDP discriminant analysis pro g ram was used to es timat e the mod e l with only th e 127 District Court cas es This model was th e n us ed to classify the cas es from th e District Courts and Court of Claims. Tabl e 4. 14 pres e nts the classification matrix for th e jackknife technique and the r eg ular classification matrix obtain e d by apply ing th e 6-variable model classification f uncti o ns t o th e ori g inal data (which is dif fe r en t from th e jackknife pr o c e dur e wh ich with holds one ca se e stimates the classification functions, and classifies the left out case). Using eith e r classification techniqu e r es ults in substan tially no diff ere nc e in the classificati o n accuracy b e tw ee n th e two diff e r e nt courts. For e xample, usin g th e jackknife t ec hniqu e results in 93,7 perc e nt of the District Court cases being cor rectly classified a nd 95,2 p e rc e nt of th e Court of Claims cas es bein g correctly classifi e d. This initi a l test seems to disa g re e with the supposition that it mak e s a diff e r e n ce in which court an empl oyme nt tax case is tri e d 1n. S e c o nd, the OLS r eg r es sion m ode l was es ti mate d u s in g th e followin g thre e diff e r e nt s e ts of obs e rvations: ( I) th e e ntire 148 c ases (2) only th e 128 District Cour t cas es a nd (3) th e rema1n1n g tw e ntyo ne Co urt of Cl a ims cases. T h e C h ow t e st as discussed in th e t e mporal stabi lity qu es tion was th e n appli e d

PAGE 161

6-Variable Model Classificati o n Matrix Jackknife Classifi ca ti o n Matrix TABLE 4.14 CLASSIFICATION ACCURACY FOR CHOICE OF FORUM TEST A c tual Per ce nt r. r o up Co rr ec t Distri c t CourtInd e p e nd e nt Co ntr ac tor-97.6 Dis tri c t CourtEmp l oyee -93 0 T o t a l 96 .1 Court o f Claims Ind epe nd e nt Contr ac t o r 100.0 Cou rt of Claims Emp l oyee -87 5 Total 95.2 Di s tri c t CourtInd epe ndent Co ntra c tor -96 4 Di s t r i c t Court Emp l oyee 88 .4 T o t a l 9 3. 7 C o urt o f Claims I nrl e p e nckn t Co ntr ac tor100.0 Co urt o f C laim s Emp l o v ee 8 7. 5 T o t a l 95 2 1 52 Predicted Ind epe n de nt Cont r acto r Em pl oyee 8 2 2 3 40 85 42 13 0 1 7 14 7 8 1 3 5 38 86 41 I J 0 1 7 Jt, 7

PAGE 162

153 to d e termine if the two r eg ressions com e from the sa me population. The Chow test statistic in this case has an F-di s tribution with twelve degrees of fr e edom 1n the num e rator and 1 2 4 de g r e es of freedom in th e d e nominator. Th e critic a l F-valu e (0.05 si g ni f canc e l e vel) is 1,75 which is g r e ater than th e C how test statistic which has a value of I. 144. Th e r e fore, w e f a il t o r e j e ct th e null hypothesis that the two r eg r e ssions com e f r o m th e s a me popu lation. Th e conclu s ion of th e Chow t e st supp o rts th e res ults of the classi f ication a ccur a cy t es t in su gge s tin g th a t th e r e i s no si g ni f icant diff e r e nc e in d e terminin g co m m on la w em plo yee status due to the l eg al forum th a t the c a s e w as tri e d 1n. T his conclusion does not necessaril y c on f l i ct with th e r es ults fo und 1n a r e c e nt study which con c lud e d th a t a combin a ti o n o f Di s t r ict Co urt and Court o f Claims d ec isions w as sys t e m a tic a ll y di ffe r e nt f rom similar decisi o ns d ec id e d in th e Tax C o urt. 9 In th e pr e s e nt study it is simply prop o s e d th a t f o r on e sp e cific qu e s ti o n; i. e ., w h a t constitut e s a co mm on la w e mploy ee f or f e d e r a l t ax pu r po se s, th e District Courts a nd C ourt of Cl a im s do n o t a pp e ar t o dif f er. Resea rch Q u e sti o n 4 In comparin g th e r es ult s of this stud y w ith th e c u r r e nt le g isl a tive pr 0 p o s a l s it i s fir s t 1 1L'Cc>ss .:ir y t0 di s t i n g uish wh ;-i t crit e ri a should m o s t l o g i c ally be includ cJ in l e g i s l a ti o n r c l:it i n g to th e common l a w e m ploy ee d e finiti o n f o r fe d e r a l em pl oym e nt tax purpos e s. Th e m os t s i g nific a nt a r g um e nt f o r in c lu s i o n in le g isl a tion of th e fa c t o r s that e v o lv e d f r o m a n a l yz i ng Resea r c h

PAGE 163

Question 2 is that these factors have proved to b e most relevant to federal jud g es in deciding common law employee tax cases. However, this is not the only necessary characteristic of 154 a factor that is to be recommended for in clusion in legislation. Another essential characteristic of s u ch a factor would be simplicity of application. In other words, if a factor 1s to add certainty to the corrnnon law employee definition, it must be able to b e objectively determined. For example, the method-of -p ayment factor can be objectively applied Eith e r a worke r was paid by the j ob or he was paid by the hour. This typ e of factor could be easily draft ed into the Internal R e v e nu e Code and its application should be quite cl e ar. To aid in comparing the results of this study with the current legislative proposals, Table 4. 1 5 was d e si g ned to list the factors that were deemed most rel e vant in analyzin g R e search Qu e stion 2 as well as thos e factors included in tl1e GAO proposal and House Bill H.R. 5460. l O Th e only two factors that were included in all three of these proposals was the profit-or-loss variable and the controlling-the-place-of-work vari able The independent-trad e variable was also included in the results of Research Question 2 and the GAO proposal. Beside these three factors, the rest of the variables listed in Table 4. 15 have no overlap am o n g th e three proposals being considered in this r e search qu e sti o n Tl1e most important variable in the mathematical mod e ls was sup e rvisi o n and neither of the l eg isl a tiv e proposals includ e d this variable in their safe-harbor crit e ri a To determin e how th e three variables that w e r e includ e d 1n at least two of the proposals would do 1n e xplainin g the tax cases

PAGE 164

155 TABLE 4. 15 FACTORS CONTAINED IN ALL THREE PROPOSALS H.R. G !\O !1athematical Variable 5460 Proposal Models Supervision X Profit or Loss X X X Independent Trade X X Permanent Relationship X Hiring of Assistants X Integration X Controllin g the Place of Work X X X Set Hours of Ho rk X Separate Set of Books X

PAGE 165

156 included in this model; a discriminant model, OLS re g ression model, and legit model were estimated using only the profit-or-loss variable, the independent-trade variable, and controlling -th e place-of-work variable. The respective models were then used to predict probabilities with which each cas e was classified into either an independent contractor g roup or employee g roup. As shown in Table 4 16, the three models are very close in their classification accuracy results with the overall percentage of correctly classified cases bein g approximately 90 percent. This is a drop in accuracy of about 5 to 8 percent from the mathematical models calculated in Research Question 2. Howev e r, even thou g h a drop of 5 to 8 perc e nt may be significant, an overall correct classification rate of 90 percent is still quite respectable This is especially note-worthy when it is realized th e these three variables have the support of th e GAO, a Con g ressional subcom mittee, and the empirical results of this study. Supervision is th e one variable that prov ed to be quit e significant in the results of this study, but was not includ e d 1n the 3-variable model. To determine what improvem e nt would result from adding the supervision variable, each of the models was re estimated using all four variables. Each of the thr ee methods misclassified seven cases for an overall classification accuracy of 95.3 perc e nt. By includin g the supervision variabl e an ad ditional 5 percent of th e cases were classified correctly Even though a 5 p e rc e nt increase may not b e v e ry lar g e, in this case it represents 50 p e rcent of th e cases misclassifi e d by the 3-variable model. In that r es pect, the classification accuracy added by the supervi s ion variable is quite substantial.

PAGE 166

Actual Group Ind e pendent Contractor-Employee-Total Independent ContractorEmployee-Total Ind e pendent Contractor-Employee-Total TABLE 4. 16 CLASSIFICATION ACCURACY MATRICES FOR THE 3-VARIABLE MODELS Discriminant Classification Matrix Pr e dicted Perc e nt Independent Correct Contractor 93.8 91 82.4 9 89.9 100 OLS Classification Matrix Logit 94. 8 82.4 90.5 Classification 94.8 82.4 90.5 92 9 99 Matrix 92 9 99 15 7 Employee 6 42 48 5 42 47 .'.:, 42 47

PAGE 167

158 The determination of wh e ther this 5 p e rcent increase in accuracy 1s si g nificant enough to b e included in the l eg islativ e proposal that follows is lar ge ly a subjective question. One factor a g ainst inclusion of the supervision variable in prospective legislation is that its presenc e is difficult to objectiv e ly d e t e rmine. There fore, for purposes of this study, the additional 5 p e rcent of clas sification accuracy g ained by inclusion of the supervision variable was not deemed si g nificant e nou g h to warrant inclusion in the followin g legislative proposal. In conclusion, th e r es ults of this study are not compl e tely analo go us with the current legislative proposals. How e ver by combinin g the 3-variable model with the two m ec hanical tests includ e d in H.R. 5460, a r eason able safe-harbor proposal would be as follows: A worker c a n not b e reclas s ifi e d as an em plo yee if the followin g five t es ts are satisfied: ( I) the individual has th e risk of sufferin g a loss and the opportunity of makin g a profit, (2) the individual hold s him se lf or h e rs elf out in his or h er own n ame as se lfemp loy e d and/or makes his servic e s ge n e rally available to the public, (3) th e individu a l has a prin c ip a l plac e of busin ess other than that furnished by th e p e rson or p e r s ons for whom he or she performs or furnishes services, (4) th e individual must p e rform ser vi ces pur sua nt to a writt e n contract th a t spells out th e em pl oyme nt status and th e tax consequences that r es ult, and (5) th e p e r so n for whom the ser vice s arc provided must file th e r e quir ed t ax returns. Such a proposal includ e s factors that incorp o rat e th e thinking and suggestions of the GAO, a Congressional subcommittee, and th e e mpirical results of this study. Hop ef ully, a consensus such as this would

PAGE 168

159 result 1n a piece of legislation that could be objectively applied to add the needed certainty to the common law e mployee d e finition for federal employment tax purposes. Notes 1 The stepwise results r e port e d in this study were based on a forward stepwis e proc e dure which starts with n o variabl e s in th e model and enters variables until they no lon ge r add sufficient discriminating power to th e model. For comparison purposes, the backward stepwise procedure was also used. The resultin g model was identical to the forward stepwise procedure. Consequent ly, only the forward stepwis e results are reported in this study. 2 P. GREEN, ANALYZING MULTIVARIATE DATA 1 82 (1978). 3 N N IE e t al ., STATISTICAL PACKAGE FOR THE SOCIAL SCIENCES 443 (2d ed. 1975). 4 Th e unstandardiz e d co e ffici e nts w e re obtain e d from the SPSS discriminant analysis r e su lt s b e cause the B!IDP program only prints the standardiz e d coefficients. 5 w. COOLEY & P. LOHNES, MULTIVARIATE DATA A NA LYSIS 229 ( 1 9 71). 6 chow, T e st s of Equ a lity b e tw ee n S e ts o f Co e ffici e nts 1n Two Lin e ar Regressions, 28 ECONOMETRICA 591 ( 1960). 7 G e nsler-L ee Inc. v. U S ., 70 F. Supp. 675 (N.O. C al. 1 946); Moore v. U .S., 68-2 USTC 11 9661 (E.D. T e x. 1 968); and Rahier Trucking Inc. v. U.S ., 65-1 USTC 119 166 (E.D. Wash. 1965). 8 h. k. 7 Ra 1er True 1ng, Inc. v. U.S., supra not e 9 Kramer, An Exa m ination of th e Variabl es A f fe cti ng th e Valuation fo r T a x Purp o s e s of Larg e Holdin gs of Publicly Trad e d Stock, Ph.D. diss ertation University of T e xas at Austin 201 (1979). 10 rwo of the factors from H.R. 5 460, J writt e n c o ntract and filin g of rc quir c d L1x r c lurns, : ir e strictly m cch;inic.1 1 c ritl'ria and not r e ally p a rt o f th e dcfinition.Jl asp ec t o f a c ommo n law employ e e. For this reason th e y were not includ e d in Tabl e 4 15.

PAGE 169

CHAPTER V SUMMARY, CONCLUSIONS, AND RECOMM ENDA TIONS In Chapter I four r es earch qu es ti o n s w e r e pre s ented dealin g with the determination of a coITU11o n law e mpl oyee f o r feder a l e mplo y m e nt tax purpo s es. Th e fir s t qu es tion pertain e d t o th e id e ntificati o n of the r e l e vant vari ab l es in d e finin g a commo n l aw e mplo yee R esea rch Que s ti o n 1: \.J h at f ac t ors o r va ri a bl e s h ave b ee n u sed in d e t e rminin g er~p lo y m e nt s t a tu s? The s eco nd res ea rch qu es tion was aimed at d e t e rminin g whic h of th e variabl es discu ssed in Resea r c h Que s ti o n 1 we r e a bl e to mo s t ef f ec tiv e ly Q o del th e fa c t ors u se d b y court s in d ec idin g an e mplo yme nt status case for fed e ral t ax purposes. Resea r c h Q u es t io n 2 : 1n co n s t ruc tin g a ma th ema t ica l moci e l o f th e court s de c i s i o n-m a kin g b e havi o r, w hich o f th e va ri ab l es espo used in Resea r c h Ques tion l p r ove t o b e s i g nific a nt ? Du e to th e fact th a t th e T ax Court lacks jurisdiction ove r e mplo y m e nt tax cas es the dat a u se d in this study are comp ri sed of Distri c t Court and Co urt of Claims cases Th e third r esea r c h qu es ti o n was concern e d with wheth e r th e c h oice of l ega l fo rum was important in liti g atin g an e mpl oyee v e r s u s ind epe nd e nt co ntr acto r cnse Re sea rch Q u es ti o n 3 : In d e t e rr:1i nin g th e e mp l oyme nt sta tu s o f th e c;:ises ex.c 1111in e d, d o t il e di ffcr e n t co urt s o f o ri gi n n l jurisdi c tion (Di st ri c t Co urt s v e r s us Cn urt of C l aims) u se simi l;:ir fa c t o r s in th e r e n d e rin g of d ec i s i o n s? Du e to the la ck of ce rtainty s urr o undin g th e common l aw emp l o yee d e finiti o n f o r e mpl oy m e nt t ax p11rpo se s, Con g r ess is c urr e ntly cons id e rin g 160

PAGE 170

16 I legislation that would hop ef ull y reduc e th e confusion that now exi sts in det e rmining a work e r's employm e nt status. Research Question 4: (a) Do th e vari ab l es d e t e rmin e d in Research Qu es tion 2 hav e an y s imilarit y with the crit e ria in th e c urrent prop osa ls before Congr e ss? (b) What po ss ibl e r ecomme nd a tion s might th ese r es ult s s u gg est? This chapter will surmnarize the r es ults of the pr ece ding res ea rch questions. After th e surmnary and co nclusion s the r eco mm e ndations coming o ut of this s tud y will b e presented alon g with ap propriate suggestions for futur e r esea r c h. S u(T!lTl a ry and Con c lu s i o n s Th e determination o f whether a worker i s an e mpl oyee or an independ e nt contract o r is a questi o n th a t has b eco me in c r eas in g l y important in the l as t deca d e In Chapt e r I, Table 1.1 vividly portrays the differenc es in th e e mployment tax es paid b y t axpaye r s according to wheth e r th ey ar e tr ea t e d as e mplo yees o r ind e p e nd e nt co ntra c tors. With th e substantial increases in e mplo y m e nt t a xes that hav e occurred since 197 8 and th e addi ti o n a l incr eases th a t h a v e been pr o j ec ted through 1987, the gap betw ee n the a mount o f tax es paid b y a n e mployee versus an independent contr ac t o r co ntinu es to wid e n. Th e IRS has also becom e m o r e active in m o nit o rin g th e em pl oyme nt t ax area. An in c r eas in g nu mbe r o f l o ngestn bl.i. s h e d vocations that have b ee n tre a ted as ind epende nt contractors ar e 11 0\-' b e ing r ec 1.:1 ssif iLd as employees b y th e IR S T h e esc;i lati. o n in IRS-t axr.1ye r co nfli c ts in thi s a r e a is du e to th e ambiguity th a t surro und s th e de finition of a common law employe e

PAGE 171

A com mo n law employee i s one of the s ix 1 s t a tutoril v m e ntion ed cate go ries of w o rk e rs that come und er th e p ay m e nt a nd withholdin g requir e m e nt s of th e FICA provision s Howev e r, n eit h e r the Internal Revenue Cod e n o r th e correspondin g R eg ulati o n s co tain an adequate definition of what constitute s a commo n l aw emp lo yee Because of th e lack of s tatutor y compl e t eness pertainin g to th e commo n l aw e mplo ye e definition, th e co urt s h a v e been l e ft larg ely o n their own t o d e termine whether a worker is a common law em p lo y ee or a n ind epe nd e nt contractor. Histori ca l Background of C o mmon L aw Employ ee Chapter II co nt ai n s a co nci se histor y o f th e c ommo n l aw a nd 162 how it ha s impa c ted o n til e c urr e nt s t a tu s of th e co:nmo n L1w emp lo yee d e finition. Th e fund:imc nt a l un der l y in g p rin c ip : 11 of th e cor.un on l aw emplo y ee d e f initi o n i s til e n o ti o n of co ntr ol. Bas i ca lly, if th e p e r son for \.Jhom th e se rvi ces a r e t o be p e rfor me d has t he ri gh t t o control n o t o nl y t he r es ults t o be accom pli s h ed but a l so the d e tails and me a n s by which th a t r es ult is accomplished th e n an emplo y er-employee r e l atio n s hip i s ge n eral l y co n s id e r ed t o ex i s t. It is this definiti o n of a commo n l a w emp l oyee th a t was espoused by Con g ress in th e ori g inal Social S e curit v Act of 1 935 Th e applic a tion of th e commo n law noti o n of emp lo yee to t he em plo y m e nt tax ar ea h as b ee n c h al l e n 3 2 d num erous tim es In 1 939 t l 1 e Social Sec urit y Board r e c omm e nd e d l egis l a ti o n th.:it w o uld d e l e t e til e commo n law d efin iti o n fr om til e soc ial sec urit y provi s i o n s Th e IRS attempted thr o u g h til e i ss uanc e of Emp l oym e nt Tax R eg ulation s to

PAGE 172

r e pl ace th e c omm o n l aw de finition with a br oa d e r co n cep t t e r me d the eco n o mi c r ea lit y t es t. Th e S u p r e m e Co urt a tt empted t o br o ad e n the n o ti o n of wha t co n s titut es a n e mpl oye e f o r fe d e r a l empl oy m e nt t a x purp oses in t wo o f i t s 1 9 47 d ec i s i o n s 2 H o w eve r, in ea ch o f th ese cases, Co n g r ess r e j ec t ed t he a tt emp t s t o a lt e r 163 the co mmo n la w em plo y ee d ef initi o n. In a fin a l r es p o n se C o n g r ess inclu ded in t he Soc i a l S ecu rit y Ame ndm e nt s of 1 950 la n g u age that s p ec ifi ca ll y in c lud ed a c omm o n l aw emp l oyee as one of th e c a te go r ies of w o rk e r s t ha t was s u bject t o t he FICA provis i o n s Vari ab l e Id e nti fica ti o n O n ce t he deba t e o v e r w h e th er th e commo n l aw emp l oyee d ef ini ti o n was appl icab l e in t he e m p lo y m e nt t ax a r ea was r eso l ve d, th e n ex t conce rn was h ow t o app l y s u c h a d e finiti o n. Chapte r I II di s cu sses th ose fac t o r s th a t h a v e bee n used b y t he co urt s in decidin g e mpl oyme nt t ax cases Til e fo ll ow in g i s a li s t o f th e vari ables pr ese nt in th e co urt cas es th a t w e r e in c lu ded i n t he mathem a ti ca l m ode l s tl1 a t w e r e e s tim a t e d i n thi s s tu d y 1. S u pe r v i s i o n. 2. Int eg r a ti o n. 3. Ri gh t t o d isc h a r ge 4. Co ntinuin g r e l a t io n s h i p. 5. Se t h o ur s o f wo rk. 6 H avi n g co n t r o l ove r t h e p l ace of \ JO r k. 7. ln d e pc n J c nt t r aclc 8 M e th o d o f p ayme nt.

PAGE 173

164 9. Pa yme nt of hu si n ess a nd/ or travel exp e nses. 10. Furnishing of tools ,1 nd materials. 11. Rea li za tion of profit o r loss. Empirical Result s There were 14 8 Di s trict Court a nd Court of Claims cases that comprised th e data us ed in thi s s tud y Each case was r ea d to d e termine th e pre se n ce o r absence of th e e l e ven variab les. Both discriminant a n alysis and OLS re g r ess i o 11 a n a l ysis were u se d to estimate th e model. In add iti o n to th ese two lin e ar es tim a ti on techniqu es a non-lin ea r l og it a n a l ys is procedure was also per formed. Chapter IV pres e nt s th e empi ric al r es ult s of the se thr ee estimation t ec hniqu es Th e forward s tep\vise discriminant t echn iqu e resulted in a 6 -v a ri able model which wa s ab l e t o co rr ec tl y clas sify 96 6 p ercent of th e cases into ei th e r t he emp l oyee o r in de pendent co ntractor g roups. The followin g is a li s t of th e s i x variables th a t make up th e discrimin a nt mod e l. 1. Supe rvi sio n. 2. Pro fit or l oss 3. Ind ependen t trade. 4. Perman e nt r e l a tion s hip. 5. Hirin g o f : 1 ss i s t.111ts. b. llltL'~Llt iun. Th e forw.:ird s t e p 1s is e OLS r eg r e s s ion tL 'c liniqu c r es ulted in a 7-varia ble m o d e l th at able tn co rr ec tl y classifv 95 J percent o( t!i e c a S L' S Th e first s i x v : 1riahl cs in th e OLS

PAGE 174

165 regres s ion model were id e ntical to the six variables in the discriminant mod e l. Controlling th e plac e of work was the addi tional variable added b y th e OLS regression model. The 7-variable OLS regression mod e l was able to ex plain approximatel y 79 perc e nt of the variance in the mod e l. The s t e pwise l og istic re g ression technique resulted in a 5 variable model that w a s able to cor rectl y classif y 97.3 percent of th e cases The 5 variables co tained in the lo g it m o del were al so part of the discriminant and OLS models. It is appar e nt from th e empirical r es ult s th a t th e differ e nc es obtained from u s in g th e thr ee estimation t ec hniqu es are nominal. With onl y minor excep ti o ns, the thr ee techni q u es id e ntified the same variables in d e rivin g their fin a l mod e l s Each technique was able to cla ss if y app r oxi matel y 95 t o 97 p e r ce nt of th e cases correctl y Ther efo r e eve n thou gh a n argument was made that t he logit analysis w as s up e rior in es tim a tin g dichotomous d epe nd e nt variables, thi s a r g u me nt was not s upport e d b y tl1e r es ult s of thi s stud y The failure o f th e l og it m o d e l to generate s up er ior result s was not due to th e f ac t that the lo g it a n a l ys i s did a poor job. On t he co ntrar y the lo gi t re s ult s were quite impre ss ive. It wa s the abilit y of th e linear es timation techniqu es that was surpri s in g Despite the vio lation o f certain s t a ti s ti ca l as sumptions, 3 both th e di sc ri m inant ; mal ys i s .: rnd th e OLS r eg r ess i o n techniqu es w e r e .-:iblc to prl1vide s ub s tanti a llv the s am e r es ults a s the lo g it ;: rn a l vs i s Tliis d ocs n o t r e fut e th e th co r c t icill st1pl'ri ority of the lo g it analysis. How eve r, tile th eo r e tic a l s up e riorit y of the l og it analysis did n o t r es ult .in any practic.-:il differences

PAGE 175

for purpo s es of th e spec ifi c r esearc h qu es ti o n s add r essed in thi s study. Tempor a l Stability of the Mode l 166 Due to the thirt y -nin e year span of the data, t es t s w e r e per formed to determine if th e model was stable over tim e Fir s t, a discriminant model was es timated u s in g only pre-1960 cases The resultin g model was u se d t o c las s if y not o nl y th e pre-1960 cases, but also the po s t-195 9 cases If th e miscla ss ific a ti o n rat es were substantiall y hi g h e r fo r th e post-1959 cases th a n for th e pre-1960 c ases, thi s would be an indication that th e model w as changing over time. H oweve r, th e r esu lt s of this t es t were that th e classification accuracy for both tim es periods was approximately 95 percent. Therefore, th e initial t es t indic.:ited that th e model w;is not c h ;i n g ing ove r tim e S eco nd, t wo O L S r egressio n s w e r e es tim ~ 1t ed The first r eg ressi o n was ba se d on th e pre -1 960 cases a nd th e seco nd r eg r es sio n was bas ed o n t he post-1959 cases Th e Chow t es t was u se d t o estimate whether t he two r eg ressi o ns came from t he same population. The Ch o w test s t a ti s ti c wa s not s i g nific a nt a t th e 5 perce nt significance le ve l. This f indin g s upports the di sc rimin a nt analysis te s t for t empo r a l s tabilit y by co n c lu di n g th a t th e tw o regr essio n s do come fr om til e ~.:1111c p o pulat Ll ll ; m e.I llil'r c' fore t ile mod e l is n o t ch a n g in g ove r tim e

PAGE 176

Current L eg i s lativ e Pr oposa l s Th e vari a bl es c ont ai n e d in th e m a th ematica l m odels of thi s stud y were compar e d wi th th e proposals currently being con s id e r e d b y Con g r ess Thr ee v a riabl es were found th a t we r e n o t onl y con tained in the math e m a ti ca l mo d e l s built in thi s st ud y but th ese three variables were a l so co nt ai n e d in various l egis l a ti ve p ro posal s Th es e thr ee variables wer e (1) profit or lo ss ( 2 ) indep e nd e nt tr a d e a nd (3) co ntr o llin g th e plac e o f wo rk. Thi s 3-variable mod e l is th e c omb ined result of th e e mpirical conclusions of thi s stud y a Congressional s ub commi tt ee an d th e GAO Con seq u e ntl y th ese tl1r ee variables alon g wi th t wo add iti o n a l mechanical-t y p e f ac t o r s we r e combined t o fo rm a pos s ibl e le g i s lativ e propo sa l. To d e t e rmin e h ow we ll s u c h a s m a ll numb e r of v ariab l es wou ld do in modelin g th e emp l oymen t s t a tu s ques t ion l eg it a n alysis discrimin a nt a n a l ys i s and OL S r eg r ess i on a n a l ys is we re all used to estimate th e 3-variable model. A ga in, th e r es ult s of th e 16 7 three different es tim a ti o n t ec hn iques we r e quite s imilar. E ac h procedur e wa s able to ge n era te estimates th a t correctly classified approxim a tel y 9 0 p e r ce nt of th e cases Hhil e thi s i s a d r o p of rou gh l y 5 percent in acc ur acy fr om th e l a r ge r ( 5 6 o r 7 va ri ab l e ) m o d e ls. 90 perc e nt clas~if i ca ti o n acc ur ncv i s sti ll qt 1 it c re spectable. Ch o ic e of L ega l Fo ru m Anoth e r qu es ti o n that was addre sse d in this s tu d y W.'.lS 1.;hether the ch oice of l ega l fo rum wa s imp o rt an t in a n emp l oym e nt s tatus

PAGE 177

case. The appro.1ches t o studying this qu es ti o n were b as i c .1ll v the s am e proc e dure s th a t w e r e u se d in the t empora l s t ab ilit y tests. First, a di sc riminant mod e l was estimated u s in g only the Di s trict Court cases. Thi s mod e l was u se d to classify n o t 168 4 only the Di s trict Court cases, but also th e Court of Claims cases. The mod e l correctly classified about 94 percent of the cases in both of the l ega l forum s This findin g was corroborated b y the results of the Ch ow te s t applied to two OLS r eg r ess i o n s Th e fir s t regression wa s ba se d o n th e Di s trict C o urt cases a nd the seco nd regre ss ion wa s based on th e Court of C l a im s cases The results of the Chow test were that the two r eg r ess i o ns came f r o m the same population. Therefore, both t es t s imply that the decision models used by the Di s trict Court and Court of Claims a re subs t a ntiall y the same when d ec idin g a common l a w e mpl oyee ve r s u s ind epe nd e nt contractor case Recomme nd a ti o n s The primar y r ecomme ndati o n f o rth co min g from thi s s tud y i s the possibl e l egis l a tive proposal contain e d in Chapter IV. The 5-variable propo s al co n sists of tw o m ec h a nic a l t es t s a lr e ady sug ges ted b y a Congressional s ub com mitt ee plus an add itional three d ef initi o n al vari a bl es The se thr ee varL:ibles were se l ec t e d s ional subcomm itt ee but al so wer l' part of th e mathemJt i cal mod e l s s u gges t e d in thi s s tu
PAGE 178

169 from th e empirical analysis of thi s s tud y be g iv e n con s id e r a tion. Supervision, permanent relationship, hirin g of a ss istants, a nd inte g r a tion were al l factors that contributed s i g nific an tl y in modelin g the emp l oyme nt s t a tu s issue and a t present a r e not part of th e sugges ted l eg islativ e proposal. Con se quentl y any of th ese four va ri ab l es would b e r easo nabl e addition s to the safe -har bo r criteria developed to defin e a common l aw emp lo yee If fact, two of these variables hav e th e added be nefit of being quite objectiv e l y determined variabl es The ab ilit y t o hir e assistants would be subs t a ntiall y an obj ec tive determination. To a l esse r extent, d e termining wh e th e r a n emp l oymen t relati o n s hip if of a permanent n a ture also could be quite objectivel y defin e d. The objectivit y of a fact o r is imp o rt a nt wh e n the principal purpose of th e l eg isl a ti o n i s to add certaint y to a t ax a rea. TI1is fits th e e raplo y ment s t at us is s ue be ca use th e principal co cern of the proposed le g i s lation is to add certaint v to the commo n law d efi nition of emp l oyee Su gges ti o n s for Futur e Research On e issue th a t thi s stud y did not addr ess wa s w h e th e r cases th a t were tri e d wi th a jur y see m e d to u se a di ffe r e nt d ec ision mod e l. It mi g ht prov e inter es tin g t o obtain th e full co urt tr a n s cript s for a sample of jur y -tri a l e m plo y m e nt s t a tus cases in orde r to determin e if tl 1e de c i s ion s w e re ba se d o n th e same fac t o rs as th e v e rdi c t s r e nd e r e d s tri c tlv hv jud g~s With the ge ographical division s th a t m a ke up the U nit e d State s Di s trict Court sys t em it would b e u s eful t o d e t e rmin e if

PAGE 179

170 there are differenc e s in court mod e ls bas e d on ge o gra phical location. This qu e stion could not b e addr esse d in this study because there was not enough cases 1n any on e district to analyz e it separately. How e ver, i f an ar e a of tax coul d be found that spawned sufficient litigation, this would be an important question to address. The question has oft e n been pos e d of wh e ther th e choice of legal forum is a r e levant factor for purpos e s of liti g atin g a federal tax case. To a limited d eg r ee that qu es tion was ad dressed in this study. To adequately d e t e rmin e whether th e Tax Cour the Court of Claims, or th e District Courts ex hibit dif ferent d e cision-makin g b e h a vior, an area of tax law would n ee d to be found that is broad enough to includ e substantial liti ga tion 1n each of the three forums. Most lik e ly, th e tax qu e stion could not be v e ry specific in nature. For exa mpl e s e v e r a l dissertations discuss e d earlier 1n this study hav e d ea lt with diff e r e nt asp e cts of valuation. If c e rtain ge n e ral factors could b e id e ntifi e d that applied to all valuation qu e stions, th e cas e law should be num erous enough to allow th e l eg al forum qu e stion to be adequately studied. Logit analysis w a s proposed in this study a s an alt e rnative technique to be used in tax studies that are d e alin g with a dichotomous d e p e nd e nt v a ri a ble. Even thou g h th e lo g it r e sults 1n this study were not vastly superior to th e lin ea r es ti m ation results, future studies o f this nature s h o uld continu e to pr omo t e the statistical t ec hniqu e s that b e st apply to th e typ es of res e arch qu e stions b e in g ask e d and th e data us e d in the r e sp ec tive analysis of those questions.

PAGE 180

Note s 1 For a li s t of th e s ix statutorily d e fin e d catego ri es of workers that come und e r th e pa y ment and withholdin g requir eme nt s of the FICA provisi o ns, see th e discu ss ion ab o ut Se c tion 312l(d) in Ch a p t e r II p 3 8 2 u. S. v. Silk, 67 S. Ct. 14 63 (194 7) and Harrison v. Greyva n Line s, Inc., 67 S.Ct. 1463 (1947). 17 I 3 Th e discriminant a nal ys is t e chniqu e r e quires that th e predictor variable s be multivariate normall y distribut e d and t hat the dispersion matrices of tl1e employee and indep e nd e nt co ntra c t or g roups be e qual. Both of the se assumptions arc violated in this stud y In the OLS regre ssio n analysis the e rror terms a re n e ith e r n ormally di st ribut e d nor constant whic h violate s two of th e lin e ar re g r ess ion assumptions. Also, th e estima t es from th e OLS m o d e l m ay fall ou t s id e th e (0,1) interv a l, whi ch is th e ran ge for which th e d e pend e nt variable in this study is defined. 4 Th e data are comprised of Di s trict Court and C o urt of Claims cases~ b eca u se the T ax Court doe s not have juri s diction over emplo y ment tax cases.

PAGE 181

APPENDIX A LIST or DISTRICT COURT CASES INVOLVIN~ THE E ~W LOYE E VERSUS INDEPENDE N T CONTRAC T O R QUESTION Decision Year 194 3 1972 Case Aberd ee n Aerie No 24 of Fr ate rn a l Order of Eagl es Air Termin a l Cab 1949 Allen & Co. 1970 American Co nsultin g Corp. 1959 1979 1977 1948 Am e ri ca n Ho mes of Ne w England American In s titut e of Famil y Relations Apaca Atlanti c Coas t Life Insur a nc e Co. 1973 Avis Rent A Car S y stem 1966 196 7 Azad Barrett 1945 Bartels 19 4~ B e, 1 vercL il c M c morL-11 1946 1956 r ;irk Beckwith B e n 172 Decisi o n Year 19 43 19 71 1955 Case Bilt g e n v. Re y nolds Bishop Bonded In s ul a ti o n & Const ru c ti o n Comp a n y 19 62 Bonney Moto r Express 1942 Burruss v. Early 1 945 195 0 1 963 19 7 8 Ca nn o n Valley ~i llin g Co. Capital Life & H e al th In s ur ance Co mpa n y Capi tal Trawlers, Inc. Chase ~anufacturing 1965 Club Hubba Hubba 1943 Combi n ed Metals Re duction 1959 1 948 1945 C o mpan y Co n so l ida ted Housecraft, Inc. Crosse tt Lum ber Co. Emard v. Sq uir e

PAGE 182

Decision Year 1946 Case Entner 1977 Erickson 196 7 1956 Ethert o n Farm & H ome nod erniza tion Corp. 1970 Filipidis 1959 Fleem a n 1946 Gaus e-Wa r e Service 1942 Insuran ce Co. Gen e ral Wayne Inn, Inc. v. Rothensies 1946 Gensler-Lee, In c 19 77 Gilmore 1964 1944 1947 19 74 1946 Gram e r cy Construction Comp a n y Gulf Oil Corp. Haines v. Kavangh Hamlin H ea rst Public a tions, Inc. 1978 Henr y 194 6 H e nrv Broderic k, Inc. 1944 Hir sc h v. Rothensies 1940 Indi a n Refining Co De c ision Year Case 173 1969 Int e rn a ti o n a l Exterminator Corporation 19 7 9 196 7 In ves t o rs He rita ge Life Insurance Co. Jack so n v. Phinne y 1957 J ago linz e r 1976 J effe r 1979 J o n e s, E C. 19 78 1941 Jon e s, M .H. Kentucky Cottage Industries 197 8 Klingler 1 978 Krobl in Ref ri ge r a ted Xpress 1968 Kurio 1946 1952 1976 1962 1948 Lakie L ev in v. Ma nnin g Li e b L oe b Lorin g 1944 L os A n ge les Athletic Club 1 967 : 1 1rtin Jq 73 ~b" Fre i g ht Service In c 1 960 ~ e rv is

PAGE 183

Decision Year Case 1954 Metropolitan Roofing & llodernizi.ng Company, Inc. 1970 M F.A. ~ !utual In s urance Company 1956 Millard's Inc. 1968 1968 1943 1945 1942 194 2 194 7 1942 Mladinich Moore Mutual Tru ck ing Co. Nevin, Inc. O Hara Vesse ls, Inc. v. Hass e t Oran ge State Oil Co. v. Fahs Part y Cab Co. Radi o City '.'lusic Hall Corporation 1965 Rahier Trucking, Inc 1974 Reser ve Natio nal Insur1966 1946 1958 ance Company R & H Corporation Schwing Se c urit y R oo fin g & Construction Comp;1ny 1949 Shr evepo rt Laundries, Inc. 1954 Silver 174 De c ision Year Case 1963 Smith, Paul 1975 Smith, Robert 1943 Spirella Co Inc 1978 Standard Chemical Mfg Co. 1975 Standard Life & Accident 1943 1963 19 77 1954 Insurance Co Standard Oil Co. v Glenn Star Fish & Oyster C o Star Oil Company Sterns v Clauson 1 943 Stone 1948 1941 1940 Tapager v. Birmingham T e n Ey ck Co., Inc. Texas Co v. Hi gg ins 1959 Thor Compan y 1945 Tidwell 1962 19 74 1946 1967 Titanium Ores Corp. T. L Squared Tomlin h' e bb Inc., W ~I 1943 h11alen v. Harrison 1952 Willard St0ra ge Battery Co. v. Ca r e y

PAGE 184

Decision Year Case 1944 Willard Sugar Co. v. Ge ntsh 1941 Williams 1956 Williams, E. F. 1959 Willi ams P a cki n g & Naviga t ion Co ., Inc. v. Enochs 1977 Wolfe 17 5 Decision Year Case 1949 Woolson Spice Co. 1976 World ~a rk e ts Centers In c. v I. R. S. 1944 Yean-1ood 19 62 Yellow Cab Corporation 1957 Zipley

PAGE 185

APPEHDI X B LIST OF COURT Of CLAIMS CASES INVOLVING THE E r ~LO YEE VERSUS INDEPENDENT CONTRACTOR OUESTION De cisio n Deci sio n Year Case Yea r Case 19 77 Aparacor, In c 19 7 l McCorn bs 1965 Arthur Venn e ri Compa ny 19 76 ~lcCo rmi ck 1963 Ben Construction Corp 1 977 Mor i sh 197 8 Be ye r 19 7 5 Pacific Fa r East L ine Inc. 1964 Cap e Sh o r e fish Co., Inc. 1 9 70 Powers 1960 Cutler 1 9 72 Ralls, In c 194 7 De-Raef Corporation 1 966 Ravh ill 195 8 Edw a rd s 1 952 Thornton 1965 Illin ois Tri-Seal 197 5 Tristate Deve l opers Inc. Product s 1 95 7 Wea ther g u a rd Corp. 1965 Kirkconn ell 176

PAGE 186

BIBLIOGRAPHY Articles Bromley, Robert G. "A Clos e r Sti:1tistical Lo ok at Tax Court Compr om i se ," Taxe s -The Tax l-lagaz in e 5 7 (r:ay 19 79) : 325-330. Chow, Gregory C. "T es t s of [q u ;:ility between S e t s of Coef ficients in Tw o Linear R eg re ssio n s, EconoP1etri ca 28 (Jul y 1 960): 591-605. Eisenbei s R obe rt A. "Pitfall s in th e Application of Discriminant An a l ys i s in Bu s in ess fin a nc e, and Econ o mics, The Journal of Finan ce 32 (Jun e 1977): 875-900. Englebrecht, Ted D. A R e pl v Anal vs i s a nd Ex t e n s i o n o f a Closer Statistical L oo k a t T ax Court Compromis e ," Tax es --The Tax ilagazine 57 (S e pt e mber 1979): 607-6 1 4. ----, and Davis o n, D a l e L. "A Statistical Lo o k a t Tax Court Compromi s e in E s t a t e and Gift Tax Valuation o f Clo s el y Held Stock," Taxe sThe Tax ~lagazine 55 (Jun e 1977): 395-400. Greenbaum, Werner F. Anal y tical : rnd Simulation : l ode .l s fo r Expl a inin g Judicfal Deci s ion-Making." In Fronti e r s of Judi c i a l Resea r c h, pp 307-334. Edit e d by Jo e l B. Grossman and Jo se ph T a nenhaus. New York: John Wiley and Sons, 1969. Gilbert, E. S. "On Discriminating Using Qualitative Variables, Journal of th e Am e rican Statistical As so ciation 63 (Decemb e r 1968): 1399-1412. "Th e [ffe c t of Un e quc1.l \'c1ri.:111 ce -C ovc1ria nc e ~b tri ces o n Fisher' s Lin e ar Discrimi11 1nt Fun c ti o n," Biom e tri cs 25 (September 1 969): 505-515. Kort, Fred. "Pr e dictin g Supr e me Court D e ci s i o n s t!a th em :itic:i.llv : A qu:rntitativ e Anal y sis of th e Ri ~ ht to Coun se l' Case s ," T he Am e ri ca n Politi c.:1 1 S c i e nc e Review 5 1 (:,!arch 1 95 7): 1-1 2 177

PAGE 187

178 "Cont e nt Anal ys i s o f Judi c ial Opini o n s a nd R ul es of Law." In Judicial D e ci s i o n-M .;:J king, pp. 133-197. Edited b y Schubert. N e w York: TI1e free Pr e ss of G l e ncoe, 1963. Kozub, Robert M., and L a th e n, Willi a m C. "V e rific a ti o n of th e Tax Court's U se of th e Ma yso n ~Ianufacturin g C o G uid e lin es in D e terminin g Ex ce s s iv e Compen sa tion," Pr ocee din gs o f the S o uth we st e rn Am e rican A cco unting A s soci a tion Me eting. San Antonio, Tex.: n.p., 1979. Krzanow s ki, I L J. O n Di sc rimin a ti o n and Cl ass ifi ca ti o n U s in g B o th Bin a r y and C o ntinu o u s Variabl es ," J o urn;:il o f th e A me rican Statistical Asso c iation 70 (December 1975): 782-790. Lachenbruch, Pet e r A O n Ex p e ct e d Pr o babiliti es of Mi sc la ss ifica ti o n in Di sc rimin a nt J\n 3 Jy s i s N ecess~ 1r y S;1111pl e Siz e and a R e lati o n \ J ith til e ~!ult ipl e C o rr e lati o n Co e fficient," Biom e tri cs 2 4 (D ece mber 1968): 8 2 38 3 4 Levine, Aaron. "Curr e nt F ac t o r s th a t Di s tin g ui s h bet1,L'e n Emrlo ye e and Ind e p e nd e nt C o ntractor," Journ.:il o f Ta xa tion 37 (S e ptember 1972): 18 8-193. Libb y Rob e rt. Acc o untin g Ra tio s and t he P r e di c ti o n of Failur e : Som e B e h av ioral E vid e n ce ," J o urn a l o f A cco unting Re s earch 13 (Sprin g 1 975) : 1 5 0-161 1'agel, S.S. 'Ap p l y in g C o rr e lati o n An a l ys i s to Case Pr ed icti o n," Texa s Law R e vi e w 4 2 (Octob e r 1 9 64 ): 1 006 -1 0 17. Nor th, John E. "The Em p l oyme nt T.1x Mora ss ," Cr e i g ht o n L aw R ev i ew 11 (1978): 7758 06 Smith, William A. "Ind e p e nd e nt C o ntr a ct o r o r E: mp lo yee ?--That I s th e Q u es ti o n," New York Un i v e r s it y In s titut e o n F e d e r a l Taxation 33 (1975): 577-599. St reer, Paul, and B oy d, J ose ph "Er.iplo yee o r Ind epe n cle n t Cont ra c tor? Pr opose d G uid e lin es Ma y L ess en th e C o ntr o v e r sy ," Taxes--Th e T a x Mag azin e 56 (Au g u s t 1978): 4 89-495 Ulmer, Sidne y S "Th e Di sc r i minant Function and a Th e or e tic a l Cont e xt for Its U se in E s timatin g th e V o t es o f Jud ges ." In Fronti e r s of Jl!dicial R ese ar c h, pp 3 3 5 -3 63 Edit e d b, J o L l !3 C n1 ssm .111 Z"ln L l J , s L pli Ti1t1 L' 11 li. 1u s '.'J L I, '\'l,rk: JL,llll h' ill : rnd SL1ns, l L1 ( C) Willi a m s J o hn 13. "!1 0 1, tP \l l cv iatL' Tnx Bllr dc n \ li e n 1 1\S C l.:ii ms that Ind' pe nd c nt Co 11tr ; 1 c tor s art? Em p !L W l 'C s," L 1x,1tio11 for L a \, 1 y er s 6 UL:l y -Jun c 1978): 3 6 -'l-3 6 8

PAGE 188

179 Books and Dissertations Bickford, Hugh C. Successful Tax Practice. Englewood Cliffs, N.J.: Prentice-Ha 11 Inc., 196 7. Bond, Jam e s B. "An Empirical Investigation of Court Determined Debt Equity Attributes for Federal Incom e Tax Purposes." Ph.D. dissertation, University of South Carolina, 1977. Boyd, Jos eph L. "Validation of Guid e lin es for Determining Reasonable Compensation in Closely Held Corporations." Ph.D. disserta tion, University of South Carolina, 1977. Compton, Ralph T. The Social Security Payroll Tax es Chicago: Commerce Clearing House, 1 940. Cooley, William W., and Lohn es Paul R. Multivariate Data Analysis. New York: John Wil e y & Sons, Inc., 1971. Dixon, W. J., e d. Biom e dical Computer Programs P-Series-1979. Berkeley: University of California Press, 1979. Englebrecht, Ted D. "An Empirical Inv es ti gat ion into th e Valuation of Closely Held Corporations by the Tax Court for Estate and Gift Tax Purpose ." Ph .D. diss e rtation, University of South Carolina, 1976. General Accounting Office. "Tax Tr ea tm e nt of Employees an d Self-Employed Persons by the Int er nal Revenue Service: Problems and Solutions." U.S. Government Printin g Office, Report to the Joint Committ ee on Taxation. November 1977. Green, Paul E. Analyzing l~ultivariat e Data Hinsdale, Illinois: The Dryden Press, 1978. Hanushek, E. A., and Jackson, J.E. Statistical M e thods for Social Sci e ntists. New York: Academic Pr e ss, 1977. Hughes, John E. Th e Federal Social Security Tax Chicago: The Flood Co., 194 I. Internal Revenue S er vic e Internal Revenue Manual New York: Commerce Clearing House, 1976. Joint Committ ee on Taxati on ll.S. Gover nm e nt Printin g Off ic e 1 ssucs in th e Cl.:issification of lndividu, :. ils as Employees or Ind e p e nd e nt Contractors. February 197 9. Kinnane, Charles H. Anglo American Law. Indianpolis: The Bobbs-M e rrill Company, Inc. 2d. 1952.

PAGE 189

180 Kramer, Sandra s. "i\n Examination o[ the Variables Affecting th e Val uation for T ax Purposes of Lar ge Holdin gs of Publicly Traded Stock." Ph.D. Dissertation, Univ e rsity of Texas at Austin, 1979. Madeo, S. A. "Th e Acc umul ated Earnings Tax: An Empirical Analysis of the Tax Court's Implementation of Congressional Int e nt." Ph.D. diss e rtation, North Texas State Univ e rsity, 1977. Maitland, Frederic W ., and Monta g ue, Francis C. A Sketch of English Le ga l History. New York: G. P. Putnam's Sons, 1915. Marmoll, Helen E. 391 Tax M a na geme nt Portfolio, Employee De fin e d. Washington, D.C.: Tax Management, Inc., 1979. Nerlov e M., and Press, M. J. "Univari ate and Multivariate L og -lin ea r and Logistic Models." Rand Corporation Technical Report R-1306-EDA/NIH, Rand Corporation, December 1973. Neter, John, and Wasserman William. App li ed Lin ea r Statis tical Models. Homewood, Il 1 inoi s: R ichard Irwin, Inc., 1974. Nie, Norman H.; Hull, C. Hadlai; J e nkins, Jean G.; Steinbrenner, Karin; and Bent, Dal e H. S tatistic a l Pa ckage fo r th e Socia l Sciences. 2d ed. New York: McGraw-Hill Book Company, 1975. Pindyck, Robert S., and Rubi nf e ld, Daniel L. Econometric Mod e ls an d Economic Forecasts. New York: McGraw-Hill, Inc., 1976. Pollock, Frederick, and Maitland, Frederic W Th e H i story of English Law. Cambrid ge : University Press, 1911. Taylor, Ronald L. "D e finin g Capital Gains in Real Estate Transactions by Applyin g Discriminant A n a l ys is to Definitional Elements." Ph. D. diss e rtation, Univ e rsity of Tennessee, 1978. Judicial D e cisions Aberdeen Aerie No. 24 of fraternal Order of Ea g les v. U.S., 50 F. Supp. 734 (W.D. Wash. 1943). Air Terminal Cab, Inc. v. U.S., 29 AFTR 2d 72-998 (E.D. Mo. 1972). Allen & Co. v. Hensl ee 86 f. Supp. 295 (1 1 .D. Tenn. 1949). American Consultin g Corporation v. U.S., 311 F. Su pp. 715 (W.D. Pa. 1970). American Homes of N ew England, Inc. v. U.S., 173 F. Supp. 857 (D.C. Ma s s I 9 5 9 ) American Institut e of Family Relations v. U.S., 44 AFTR 2d 79-5043 (C.D. Calif. 1979).

PAGE 190

Atlantic Coa s t Lif e In s urance Co. v. U.S., 76 F Supp. 627 (E.D. s.c. 1 948). Apaca v. U.S., 39 AFTR 2d 77-631 (D.C. N.l !. 1 977). Aparacor, Inc. v. U.S., 556 F. 2d 1004 (Ct. Cl. 1 977). Arthur Venneri Compan y v. U.S., 340 F. 2d 337 (Ct. Cl 1 96 5). 181 Avis Re nt A Car System v. U.S., 364 F. Supp 605 (E.D. N .Y. 1 973) Azad v. U.S., 277 F. Supp. 258 (D.C. Minn. 196 6) Barrett v. Phinne y 278 F. Sup p. 65 (S.D. Tex. 1 967). Bartels v. Birmingh am 59 F Supp. 84 (S.D. Ia. 1 945) Beaverd ale Memorial P.:irk Inc. v. U.S., 47 F Supp. 663 (D.C. Conn. 19 42) Beckwith v. U.S., 67 F. Supp. 902 (D.C. Mass. 1 946) Ben Construction Corporati o n v. U.S., 312 F. 2d 781 (Ct. Cl. 1 963) Ben v. U.S., 13 9 F. Supp 883 (N.D. N .Y. 1 956) Beyer v. U.S., 78 -1 USTC 'i9307 (Ct. Cl. 1 978). Biltgen v. Re y nolds, 58 F. S upp. 909 (D.C. Hin n. 1 943) Bishop v. U.S., 334 F. Supp. 415 (S.D. T ex 1 971) Bo nded In s u] ation & Constru c tion Company, Inc. v. U.S ., 131 F. Supp. 635 (D.C. N .J. 1 95 5). Bonney Motor Express v U.S., 206 F. Supp. 22 (E.D. Va. 1 962) Burruss v. Earl y 44 F. Supp. 21 (h'.D. V a 19 42). Cannon Valley nillin g Co v. U.S., 59 F. Supp. 735 (D.C. Minn 1945). Cape Shore Fi s h Co., Inc. v. U.S., 330 F. 2d 96 1 (Ct. Cl. 196 4) Capital Life & H ea lth In s urance Co v. Bow e rs, 90 F. Supp 6 00 (E.D. S.C. 1950). Cap i t .:i l Tr i1 w 1 c r s l 11 c v U S 2 1(, L S u pp ~ 4 0 ( D. C Mc I l) 6 J ) Chase l. : rnub c turin g v. U.S., 446 F. Supp. 693 (E.D. :1o 1 978) Club Hubba llubb a v. U.S., 239 F. Supp. 3211 (D.C. Hi. 1965).

PAGE 191

Combined Metals Reduction Co. v. U.S., 53 F. Supp. 739 (D.C. Ut. 1943). Consolidated Housecraft, Inc. v. U.S., 170 F. Supp. 842 (E.D. N.Y. 1959). Co-op Cab Co. v. Allen, 82 F. S upp. 842 (E.D. N.Y. 1959). Crawford Packing v. U.S., 228 F. Supp. 549 (S.D. Tex. 1962). Crossett Lumber Co. v. U.S., 79 F. Supp. 20 (W.D. Ark. 1948). Cutler v. U.S. 1 80 F. Supp. 360 (Ct. Cl. 1960). De-Raef Corpor a tion v. U.S., 70 F. Supp. 264 (Ct. Cl. 194 7). Edwards v. U.S., 168 F. Supp. 955 (Ct. Cl. 1958). Emard v. Squire, 58 F. Supp. 281 (W.D. Wash. 1 945) Entner v. U.S., 67 F. Supp. 684 (S.D. Ohio 1 946 ). Erickson v. U.S., 41 AFTR 2d 78-456 (W.D. W is. 1977). Etherton v. U.S., 278 F. Supp. 568 (E.D. Tenn. 1967). Farm & H ome Modernization Corpo r atio n v. U.S., 1 38 F. Supp. 423 (N.D. N.Y. 1956). Filipidis v. U.S., 27 AFTR 2d 71-595 (D.C. Md 1970). Fleeman v. U.S., 175 F. Supp. 336 (N.D. Ohio 1959). Gause-Ware Servic e Insurance Co. v. Thom as, 76 F. Supp. 626 ( N .D. Te x 19 46) General Wa y n e Inn, In c v. R o thensies, 47 F. Supp 391 (E.D. Pa. 1942). Gensler-Lee, Inc. v. U.S., 70 F. Supp. 675 (N.D. Cal. 1946). Gilmore v. U.S., 41 AFTR 2d 7 8 -808 (D.C. Md. 1977). Gramercy Construction Co. v. U.S., 64-1 USTC 9333 (S.D. N.Y. 196 4) Gulf Oil Corporation v U.S., 57 F. Supp 376 (W.D. Pa. 1 944) Haines v. Kavanagh 70 F. Supp. 705 (E.D. Mich. 1 947). H a mlin, 74-2 USTC 95 78 (D.C. Kan. 1974). Henry v. U.S., 42 AFTR 2d 78-5053 (E.D. Tenn. 1978). 182

PAGE 192

183 Henry Broderick, Inc. v. Squire, 69 F. Supp. 109 (W.D. Wash. 19 46) Hirsch v. Rothensies, 56 F. Supp. 92 (E.D. Pa. 1944). Illinois Tri-Seal Products, Inc. v. U.S., 353 F. 2 d 216 (Ct. Cl. 1 965). Indian Refining Co. v. Dallman, 31 F. Supp. 455 (S.D. Ill. 1940). International Extermin a tor Corporation v. U.S., 305 F. Supp. 676 (N.D. Tex 1969). Investors Herita ge Lif e Insurance Co. v. U.S., 79-1 USTC II 9246 (E.D. Ky. 1979). Jackson v. Phinne y 266 F. Supp 835 (W.D. Tex. 1967). Jagolinzer v. U.S., 150 F. Supp. 489 (D.C. R.I. 1957). Jeffer v. U.S., 37 AFTR 2d 76-1430 (D.C. N .J. 1976). E. C. Jones v. U.S., 79-1 USTC II 9120 (E.D. Tex. 197 8) M. H. Jon es v. U.S., 76-2 USTC II 9607 (N.D. Okla. 1 976). Kentucky Cotta g e Industries, Inc. v. Glenn, 39 F. Supp. 642 (W.D. Ky. 1941). Kirkconnell v. U.S., 347 F. 2d 260 (Ct. Cl. 1965). Klingler v. U.S., 78-2 USTC II 9668 (N.D. Ia. 1978). Kroblin Refriger a ted Xpress, Inc. v. U.S., 78 -2 USTC 9809 (N.D. Ia. 1978). Kurio v. U.S ., 68-1 USTC II 9382 (S.D. Tex. 1968). Lakie v. U.S., 70 F. Supp. 665 (E.D. Mich 1946). Levin v. Manning, 124 F. Supp. 192 (D.C. N.J. 1952). Lieb v. U.S., 77-1 USTC II 9356 ( E .D. Okla. 1976). Loeb v. U.S., 209 F. Sup p. 22 (E.D. La. 1 962 ). Loring v. U.S. 80 F. Supp. 781 (D. C. ~bss. 194 8 ). Los Angeles Athletic Cl ub v. U.S., 54 F. Supp. 702 (S. D. Cal. 1944). Martin v. U.S., 267 F. Supp. 268 (~1.D. FL:i. 1967). May freight Service, Inc. U.S., 78-2 USTC II 9707 (D.C. Ut. 1978).

PAGE 193

Mccombs Company, Inc. v. U.S., 436 F. 2d 979 (Ct. Cl. 1971). McCormick v. U.S., 76-1 USTC ft 9319 (Ct. Cl. 1976). Mervis v. U.S., 187 F. Supp. 248 (E.D. La. 1960). 184 Metropolitan Roofing & Moderniz ing Company, Inc. c. U.S., 125 F. Supp. 670 (D.C. Mass. 1954). M.F.A. Mutual Insur a nce Co. v. U.S., 314 F. Supp. 590 (W.D. Mo 19 70). Millard's Inc. v. U.S .. 146 F. Supp. 385 (D.C. N.J. 1956). Mladinich v. U.S., 379 F. Supp. 117 (S.D. Miss. 1974). Moore v. U.S., 68-2 USTC ft 9661 (E.O. Tex. 1 968) Morish v. U.S., 555 F. 2d 794 (Ct. Cl. 1977). Mutual Trucking Co. v. U.S., 51 F. Supp. 114 (N.O. Ohio 1943). Nevin, Inc. v. Rothensies, 58 F. Supp. 460 (E.D. Pa. 1945). O'Hara Vessels, Inc. v. Ha sse t, 60 f. Supp. 672 (D.C. ~lass. 1942). Orange State Oil Co. v. Fahs, 52 F. Supp. 509 (S.D. Fla. 194 2) Pacific Far East Line, Inc. v. U.S., 75-1 USTC ,i 9317 (Ct. Cl. 1975). Party Cab Co. v. U.S., 75 F. Supp. 307 (N.D. Ill. 19 47). Powers v. U.S., 424 F. 2d 593 (Ct. Cl. 1970). Radio Cit y Music Hall Corporation v. U.S., 50 F. Supp. 329 (S.D. N.Y. 1942). Rahier Truckinr,In c .v. U.S., 65-1 USTC ft 9166 (E.D. Wash. 1965). Ralls, Inc. v. U.S., 470 F. 2d 579 (Ct. Cl. 1972). Rayhill v. U.S., 364 F'. 2d 34 7 (Ct. Cl. 196fi). Reserve National Insuranct! Company v. U.S., 7 4 1 USTC ft 94S6 (1 -.' .D. Ok L-1. 19 7 4) R & H Corporation v. U.S., ~55 F. S upp. 870 (ILD. Penn. 196 6) Schwing v. U.S., 65 F. Supp. 227 (E.D. Pa. 1 946 ). Security Roofing & Construction Co. v. U.S., 1 6 3 F. Supp. 794 (D.C. Hass. 195 8).

PAGE 194

185 Shreveport L a undri es Inc. v. U.S., 84 F. Supp. 435 (W.D. La. 1949). Silver v. U.S., 131 F. Supp. 209 (N.D. N.Y. 1954). Paul Smith v. U.S., 64-2 USTC 11 9511 (N.D. Tex. 1963). Spirella Co., Inc. v. McGowan, 52 F. Supp 302 (W.D. N.Y. 1943). Standard Chemical Ma nu fac turing Company v. U.S., 42 AFTR 2d 78-5124 (D.C. Neb. 1978). Standard Life & Accident In s urance Compan y v. U.S., 75-1 USTC 11 9352 (W. D. Okla. 1975). Standard Oil Co. v. Glenn, 52 F. Supp. 755 (W'.D. Ky. 1943). Star Fish & Oyster Company v. U.S., 223 F. Supp. 402 (S.D. Ala. 1963). Star Oil Company v. U.S., 41 AFTR 2d 78-570 (W.D. Okla. 1977). Sterns v. Clauson, 122 F. Supp. 795 (D.C. Me. 1954). Stone v. U.S., 55 F. Supp. 230 (E.D. Pa. 1943). Tapager v. Birmingham, 75 f. Supp. 375 (N.D. Ll. 194 8). Ten Eyck Co., In c v. U.S., 4 1 F. Supp. 375 (N.D. N.Y. 1 94 1). Texas Co. v. Hig gi n s, 32 F. Supp 428 (S.D. N.Y. 1940). Thor Company v. U.S., 173 F. Supp. 65 (D.C. Mass 1959). Thorton v. U.S., 102 F. Supp. 554 (Ct. Cl. 195 2) Tidwell v. U.S., 63 F. Supp. 609 (W.D. Tenn. 1945). Titanium Ores Corporation v. U.S., 205 F. Supp. 606 (D.C. Nd. 1962). T. L. Squared v. U.S., 33 AFTR 2d 74-991 (S.D. Ohio 1974). Tomlin v. U.S., 70 F. Supp. 677 (N.D. Cal. 1 946) Tristate Developer s Inc. v. U.S., 752 USTC 11 9773 (Ct. Cl. 1975). 1-leather gua rd Corp. v. U.S., 50 AFTR 1 270 (Ct. Cl. 1957). Webb Inc., W. M. v. U.S., 271 F. Supp. 249 (E.D. La. 1967). Whalen v. Harrison, 51 F. Supp. 511 (N.D. Ill. 1943). Willard Storage Battery Co. v. Carey, 103 F. Supp. 7 (N.D. Ohio 1952).

PAGE 195

Willard Sugar Co. v. G e ntsch, 59 F. Supp. 82 (N.D. Ohio 1944). Williams v. U.S., 38 F Supp. 536 (N .D. Ill. 1941). Williams, E. F. v. U.S., 1 39 F. Supp. 875 (N.D. N.Y. 1956). Williams Packin g & Navigation Co., Inc. v. Enochs, 176 F. Supp. 168 (D.C. Hiss. 1959). Wolfe v. U.S., 39 AFTR 2d 77-1353 (D.C. N.D 1977). Woolson Spice Co. v. U.S., 85 F. Supp. 242 (N.D. Ohio 1949). World Market Centers, Inc. v. I. R. S., 77-1 USTC' 32 77 (N.D. Tex. 1 976) Yearwood v. U.S., 55 F. Supp. 295 (W.D. La. 1944). Yellow Cab Corpor a tion v. U.S., 210 F. Supp. 64 (D.C. N.J. 1962). Zipley v. U.S., 156 F. Supp. 141 (E.D. Pa 1957). 186

PAGE 196

BIOGRAPHIC A L S K ETCH Dav e Nelson St e wart was born in Alamo, N e vada, on June 21, 1952, the son of Ald e n Levi and Carma N e lson St e wart. Aft e r g raduation from Ed W. Clark High School, Las V eg as, N e vad~ in 1970, he entered Brigham Young University in Provo, Utah. H e r e c e iv e d his Bach e lor of Arts and Master of Accountancy in June, 1977. In S e ptemb e r, 1977, he enter e d the Graduat e School of the University of Florida in Gaines ville, Florida. In March 1 9 80, Mr. St e w a rt acc e pt e d e mployment as an Assistant Professor in th e Institut e of Pro fe ssional A ccountancy at Brigham Young University. In July 1974, h e marri e d Jan e F e rry of C o rinn e Ut a h. Th e y are the par e nts of thr ee childr e n: David L e vi, N a tali e Su z ann e and Lara. 1 8 7

PAGE 197

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope of quality, as a dissertation for the degree of Doctor of Philosophy. (___ _John L. Kram e r: airman Associate Professor of Accounting I certify that I have read this study and that in my op1n1on it conforms to acceptable standards of scholarly pr e s e ntation and is fully adequate, in scope of quality, as a dissertation for the degree of Doctor of Philosophy. William A Collins Assistant Professor of Accounting I certify that I have read this study and that in my opinio n it conform s to acceptable standards of scholarly pr e s e ntation and is fully adequate, in scop e of quality, as a dissertation for the d e gree of Doctor of Philosophy. -; i / ( r L \ D e nnis A. C a lf ee Associate Professor of Taxation

PAGE 198

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope of quality, as a dissertation for the degree of Doctor of Philosophy. John C. H e nretta Assistant Professor of Sociology This dissertation was submitted to th e Graduate Faculty of the Departme nt of Accounting in th e Colle g e of Busin e ss Administration and to the Graduate Council, and was accepted as partial fulfillment of the require ments for the degree of Doctor of Philosophy. June 1980 Dean, Graduate School

PAGE 199

U NI V ER SITY O F FL O RIDA II I II IIIIII Ill Ill llll l lllll II IIIIII IIII II IIIIII IIII II IIII IIIII I I 3 1262 08553 5077

PAGE 200

U NI V ER SITY O F FL O RIDA II I II IIIIII Ill Ill llll l lllll II IIIIII IIII II IIIIII IIII II IIII IIIII I I 3 1262 08553 5077