Title: Evaluation of federal marketing orders for fruits and vegetables using time varying parameters /
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00099365/00001
 Material Information
Title: Evaluation of federal marketing orders for fruits and vegetables using time varying parameters /
Alternate Title: Marketing orders for fruits and vegetables
Physical Description: xiii, 274 leaves : ill. ; 28 cm.
Language: English
Creator: Smith, Ernest Binford, 1950-
Publication Date: 1982
Copyright Date: 1982
 Subjects
Subject: Fruit trade -- United States   ( lcsh )
Vegetable trade -- United States   ( lcsh )
Food and Resource Economics thesis Ph. D
Dissertations, Academic -- Food and Resource Economics -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
 Notes
Thesis: Thesis (Ph. D.)--University of Florida, 1982.
Bibliography: Bibliography: leaves 267-273.
Statement of Responsibility: by Ernest Binford Smith.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099365
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000316152
oclc - 08561736
notis - ABU2943

Downloads

This item has the following downloads:

evaluationoffede00smit ( PDF )


Full Text










EVALUATION OF FEDERAL MARKETING
ORDERS FOR FRUITS AND VEGETABLES
USING TIME VARYING PARAMETERS















BY

ERNEST BINFORD SMITH


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





































For

CAROLYN and CATHERINE














ACKNOWLEDGEMENTS


All the individuals who contributed to the successful completion

of this dissertation cannot be acknowledged. Dr. Ronald W. Ward,

committee chairman, provided continual guidance and counsel. Other

committee members, Drs. Max Langham, Glenn Zepp, Richard Kilmer, and

Jim McClave, provided excellent reviews and suggestions for improvement

of the manuscript.

The Economic Research Service, U.S.D.A., provided major funding

for the research. The Agricultural Marketing Service, U.S.D.A., pro-

vided data and assistance critical to the completion of the study.

Kathy Carroll typed the rough draft and the final typing was done

by Ramona Rochester.

Any excellence in the manuscript must be attributed to the above

individuals and agencies. However, the analysis and conclusions are

those of the author and do not necessarily reflect the opinion of the

Economic Research Service or Agricultural Marketing Service.

Professional, financial, and clerical support are necessary but

not sufficient for a dissertation. Those individuals who provided the

emotional support that completes the necessary and sufficient conditions

deserve special acknowledgement. However, since some might be omitted,

none will be named.

Finally, the author's family, wife, daughter, and mother, is

ultimately responsible for the completion of this study. It would not

have been possible without their encouragement, support, and sacrifice.














TABLE OF CONTENTS

PAGE

ACKNOWLEDGEMENTS. . . . . . . . . ... . .. iii

LIST OF TABLES . . . . . . . . . ... . vii

LIST OF FIGURES . . . . . . . . . .. . ix

ABSTRACT. . . . . . . . . ... ....... .xii

CHAPTER

1 INTRODUCTION . . . . . . . . . . 1

Marketing Order Functions. . . . . . 3
Marketing Order Participants . . . . . 8
Problem. . . . . . . . ... ... 10
Objectives . . . . . . . . ... .13
Methodology and Scope. . . . . . ... .14
Organization . . . . . . . ... .16

2 MARKETING ORDERS . . . . . . . .... .17

Overview . . . . . . . .... . 17
Early History . . . . . . . . 18
Current Legislation . . . . . .. 25
Current Orders. . . . . . . ... .27
Citrus Fruits. . . . . . . .. 31
Noncitrus Fruits . . . . .... .33
Vegetables . . . . . . .... .35
Potatoes . . . . . . .... .36
Tree Nuts .............. . 37
Dried Fruits . . . . . . ... .37
Literature Review. . . . . . . ... .38

3 A THEORETICAL LINKAGE OF MARKETING ORDERS AND DEMAND 42

A Demand Model .. . ...... . . . 44
Marketing Order Activities . . . . . 46
Quality Control . . . . . . . 47
Quantity Control Activities . . . .. 48
Flow-to-Market . . . . . .... .48
Volume Management. . . . . . ... 50
Market Support. . . . . . . ... .52
Reflection in Supply and Demand ...... .53











Indicators of Marketing Order Intensity and
Potential . . . . . . . . . 55
A Static Indicator of Potential ...... 55
Objective Indicators of Intensity . . 57
Measuring the Association Between Order Activi-
ties and Demand. . . . . . . . ... 59

4 DATA AND EMPIRICAL PROCEDURES. . . . . ... 65

Model Definition . . . . . . .... .65
Time Reference. . . . . . . ... 66
Variable Definitions--Demand. . . . ... 66
Variable Definitions--Marketing Orders. . 70
Variable Definitions--Randomness. . . ... 72
Data Sources. . . . . . . . ... 74
Statistical Methods. . . . . . . ... 74
Error Transformation. . . . . . ... 75
Time Varying Paremeters . . . .... .76
Kendall Correlations. . . . . . 84

5 VARYING PARAMETER DEMAND ESTIMATES FOR SELECTED
ORDER AND NONORDER COMMODITIES . . . . .. 86

Oranges and Grapefruit . . . . . .. 88
Model Definition--Oranges and Grapefruit. . 88
Parameter Estimates--Oranges and Grapefruit 89
Time Varying Parameter Estimates--Oranges and
Grapefruit. . . . . . . . ... 96
Lemons and Limes . . . . . . .... .115
Model Definition--Lemons and Limes. .... . 115
Parameter Estimates--Lemons and Limes . 116
Potatoes . . . . . . . .... . 119
Model Definition--Potatoes. . . . ... 120
Parameter Estimates--Potatoes . . ... 120
Time Varying Parameter Estimates--Potatoes. 125
Marketing Order Vegetables . . . ... 125
Model Definition--Marketing Order Vegetables. 140
Parameter Estimates--Marketing Order Vege-
tables. . . . . . . . . ... 140
Nonmarketing Order Vegetables. . . . ... 143
Summary. . . . . . . . . . 145

6 MARKETING ORDER EFFECTIVENESS. . . . . ... 151

Direct Effects . . . . . . . . 151
Indirect Effects . . . . . . . . 156
Permanent Variation . . . . . . 163
Random Variation. . . . . . . . 167
Parameter deviation randomness . . .. 168
T-value deviation randomness . . .. 172
Parameter deviation vs t-value deviation 174
Summary . . . . . . . . ... .177


CHAPTER


PAGE








CHAPTER PAGE

7 SUMMARY . . . . . . . . . . 179

Marketing Orders . . . . . . . . 180
Marketing Order Activity and Demand--Theoretical
Expectation ............... . 182
Marketing Order Activity and Demand--Empirical
Results. ................. . 187
Implications for Future Research . . ... 190

APPENDICES

A THE AGRICULTURAL MARKETING AGREEMENT ACT OF 1937 193

B DATA . . . . . . . . ... . . .211

C TVP PROGRAM. . . . . . . . .... . 257

D ANNUAL PARAMETER ESTIMATES . . . . . . 261

BIBLIOGRAPHY. . . . . . . . . . . . . 267

BIOGRAPHICAL SKETCH . . . . . . . .... ..... 274















LIST OF TABLES


TABLE PAGE

1.1 Commodities Examined. . . . . . . . ... 15

2.1 Federal Marketing Orders Authorizing Quality Control
Regulations Only, as of May 1, 1979 . . . ... .28

2.2 Federal Marketing Orders Authorizing Intraseasonal
Control Regulations, as of May 1, 1979. . . .. 29

2.3 Federal Marketing Orders Authorizing Seasonal Control
Regulations, as of May 1, 1979. . . . . ... .30

4.1 Demand Variables. . . . . . . . . . 67

4.2 Marketing Order Variables . . . . . .. 71

5.1 Estimated Demand Parameters for Oranges ...... .90

5.2 Estimated Demand Parameters for Grapefruit. .... . 91

5.3 Estimated Flexibilities for Oranges . . . ... .93

5.4 Comparison of OLS and TVP Parameter Estimates for
Citrus. . . . . . . . ... ... .97

5.5 Estimated Demand Parameters for Lemons and Limes. .. 117

5.6 Estimated Flexibilities for Lemons and Limes ... 118

5.7 Estimated Demand Parameters for Potatoes. . . .. 121

5.8 Comparison of OLS and TVP Parameter Estimates for
Potatoes. . . . . . . . . .. . 126

5.9 Estimated Demand Parameters for Marketing Order
Vegetables. . . . . . . ... ..... 141

5.10 Estimated Demand Parameters for Selected Florida
Vegetables Without Marketing Orders . . . ... 144

5.11 Gamma Values and Percent Deviation of TVP Parameter
Estimates from OLS Parameter Estimates for Selected
Commodities . . . . . . . . ... . 146








TABLE


5.12 Percent Deviation of TVP T-Values from OLS T-Values
for Selected Commodities with Gamma=O. . . . .. 148

6.1 Summary of RMQ Coefficients and Fresh and Processed
Quantity Coefficients where the Regulated Quantity
is Associated with the Minimum Mean Squared Error
After Transformation . . . . . . .... 154

6.2 Years of Operation as of 1980 and Average Values of
Selected Discrete Measures of Order Activity for
Selected Commodities . . . . . . . . 159

6.3 Average, Minimum, and Maximum Values and Coefficients
of Variation for Total Annual Expenditures and
Selected Continuous Measures of Order Activity for
Selected Commodities . . . . . . . . 161

6.4 Kendall Correlation Coefficients Between Selected
Measures of Marketing Order Activity and the Degree
of Permanent Parameter Variation in Demand (gamma) . 165

6.5 Kendall Correlation Coefficients for Selected Measures
of Marketing Order Activity and Proportional Devi-
ations of TVP Parameter Estimates from OLS Parameter
Estimates Given Gamma = 0. . . . . . . ... 169

6.6 Kendall Correlation Coefficients for Selected Measures
of Marketing Order Activity and Proportional Devi-
ations of TVP t-Values from OLS t-Values Given Gamma =
0 . . . . . . . . . . . . 173

6.7 Summary Comparison of Kendall Correlation Results from
Alternative Measures of Randomness . . . ... 176














LIST OF FIGURES


FIGURE PAGE

3.1 A Simple Linkage of Marketing Orders with Supply and
Demand. . . . . . . . . ... ..... 43

5.1 Florida Oranges Adjustments in Intercept Parameter in
the TVP Model and Comparison with the OLS Model . . 98

5.2 Florida Oranges Adjustments in Fresh Quantity Co-
efficient in the TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . 99

5.3 Florida Oranges Adjustments in Processed Quantity Co-
efficient in the TVP Model and Comparison with the OLS
Model . . . . . . . . . . . . 100

5.4 Florida Oranges Adjustments in Fresh and Imported Sub-
stitutes Coefficient in the TVP Model and Comparison
with the OLS Model. . . . . . . . . ... 101

5.5 Florida Oranges Adjustments in Income Coefficient in
the TVP Model and Comparison with the OLS Model . . 102

5.6 Florida Grapefruit Adjustments in Intercept Parameter
in the TVP Model and Comparison with the OLS Model. . 103

5.7 Florida Grapefruit Adjustments in Fresh Quantity Co-
efficient in the TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . .104

5.8 Florida Grapefruit Adjustments in Processed Quantity
Coefficient in the TVP Model and Comparison with the
OLS Model . . . . . . . . ... . . 105

5.9 Florida Grapefruit Adjustments in Total Substitutes
Coefficient in TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . 106

5.10 Florida Grapefruit Adjustments in Income Coefficient in
Income Coefficient in the TVP Model and Comparison with
the OLS Model . . . . . . . . ... 107

5.11 Florida Grapefruit Adjustments in Export Ratio Co-
efficient in the TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . 108








FIGURE PAGE

5.12 Florida Grapefruit Adjustments in Regulated Ratio Co-
efficient in the TVP Model and Comparison with the
OLS Model . . . . . . . . ... ... .109

5.13 Texas Grapefruit Adjustments in Intercept Parameter in
the TVP Model and Comparison with the OLS Model . .. 110

5.14 Texas Grapefruit Adjustments in Fresh Quantity Coeffi-
cient in the TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . 111

5.15 Texas Grapefruit Adjustments in Processed Quantity Co-
efficient in the TVP Model and Comparison with the
OLS Model . . . . . . . . . . . 112

5.16 Texas Grapefruit Adjustments in Total Substitutes Co-
efficient in TVP Model and Comparison with the OLS
Model . . . . . . . . . . . . 113

5.17 Texas Grapefruit Adjustments in Income Coefficient in
the TVP Model and Comparison with the OLS Model . . 114

5.18 Idaho-Oregon Potatoes Adjustments in Intercept Param-
eter in the TVP Model and Comparison with the OLS Model 127

5.19 Idaho-Oregon Potateos Adjustments in Fresh Quantity
Coefficient in the TVP Model and Comparison with the
OLS Model . . . . . . . . . . . 128

5.20 Idaho-Oregon Potatoes Adjustments in Processed Quantity
Coefficient in the TVP Model and Comparison with the
OLS Model . . . . . . . .. . 129

5.21 Idaho-Oregon Potatoes Adjustments in Total Substitutes
Coefficient in TVP Model and Comparison with the OLS
Model . . . . . . . . . .. . . . 130

5.22 Idaho-Oregon Potatoes Adjustments in Income Coefficient
in the TVP Model and Comparison with the OLS Model. 131

5.23 Oregon-California Potatoes Adjustments in Intercept
Parameters in the TVP Model and Comparison with the OLS
Model . . . . . . . . 132

5.24 Oregon-California Potatoes Adjustments in Total Quan-
tity Coefficient in the TVP Model and Comparison with
the OLS Model . . . . . . . . .. . 133

5.25 Oregon-California Potatoes Adjustments in Total Substi-
tutes Coefficient in TVP Model and Comparison with the
OLS Model . . . . . . . . . . . 134









FIGURE


PAGE


5.26 Oregon-California Potatoes Adjustments in Income Co-
efficient in the TVP Model and Comparison with the
OLS Model . . . . . . . . . . . 135

5.27 Maine Potatoes Adjustments in Intercept Parameter in
the TVP Model and Comparison with the OLS Model . . 136

5.28 Maine Potatoes Adjustments in Total Quantity Coeffi-
cient in the TVP Model and Comparison with the OLS
Model . . . . . . . . . . . 137

5.29 Maine Potatoes Adjustments in Total Substitutes Co-
efficient in TVP Model and Comparison with the OLS
Model . . . . . . . . ... . . . 138

5.30 Maine Potatoes Adjustments in Income Coefficient in the
TVP Model and Comparison with the OLS Model . . .. 139














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


EVALUATION OF FEDERAL MARKETING ORDERS FOR
FRUITS AND VEGETABLES USING TIME VARYING PARAMETERS

By

Ernest Binford Smith

May 1982


Chairman: Ronald W. Ward
Major Department: Food and Resource Economics

Selected federal marketing orders for fruits and vegetables are

examined in order to evaluate the relationship between marketing order

activity and demand. Marketing orders have constituted the primary

policy instrument applicable to fresh fruits and vegetables, dried

fruits, and tree nuts for more than 40 years. This study is an attempt

to empirically measure selected results of marketing orders.

The general objective is to evaluate the relationship between order

activity and demand. A current and historical perspective of marketing

order use and evolution is presented. A theoretical specification

linking order activity and demand is developed and an empirical model

is developed and estimated. Finally, the empirical results are inter-

preted.

The 21 marketing orders applicable to citrus fruits, vegetables,

and potatoes as of October 1979 are examined. Most permitted order

activities are observed among this set of orders.








The reflection of order activity in demand is pursued from three

perspectives. The direct reflection, as measured by the response of

price to changes in the proportion marketed under regulation, is observed

in the demand for seven commodities. Two indirect reflections are also

pursued. Demand equations are estimated by the time varying parameter

method which allows the distinction between demand equations with random

coefficients which tend to fluctuate around estimated mean values and

demand equations with coefficients exhibiting permanent change, i.e.,

not tending to return to a mean value.

Permanent temporal parameter adjustments are indicated for selected

citrus and potato orders. Temporal parameter variation is estimated

to be entirely random for all vegetables considered.

Several order activities are significantly correlated with permanent

change. However, the lack of any discernable temporal pattern in the

parameters suggests that the ability to increase demand through order

activities is limited at best.

Comparison of two alternative measures of randomness reveals that

order activities are not significantly associated with randomness in the

majority of cases. However, where the association is significant, an

increase in order intensity is associated with a decrease in randomness,

i.e., a more stable parameter, in most cases. No order activity is

significantly associated with increased randomness under both measures.














CHAPTER 1


INTRODUCTION


Marketing orders and agreements are the primary policy instrument

available for regulating marketing in the fruit and vegetable in-

dustries. The basic tenet of marketing orders is that producers

acting in concert to regulate the marketing of a given crop can do so

in a manner mutually beneficial to producers and society in the best

instance and in a manner at least not harmful to society in the worst

instance. This tenet arose during the economic chaos of the 1930's

and was part of the prevailing attitude that applied to much of the

U.S. economy [Benedict and Stine, p. 369]. During the same period the

primary modern policy instruments applicable to the other sectors of

agriculture were also adopted. A major dichotomy developed and still

exists between the fruit and vegetable industry policy instrument and

those applicable to other sectors of the agricultural community.

Marketing orders allow producers to exert a certain amount of

control and influence in the marketing of commodities subject to orders.

This control and influence is voluntarily acquired in the sense that at

least a majority of producers, by volume and/or number, must approve a

proposed order. Similarly, the actions taken under a marketing order

are initiated by a committee which is generally predominantly composed

of growers within the order area. In contrast, policy instruments

applicable to other areas of agriculture are generally not subject to








producer approval and are oriented toward supply control and price

maintenance and/or subsidy. These programs are voluntary in that in-

dividual producers are generally not directly required to participate.

This divergence in method results from two causes. First, the

technique employed under marketing orders had been previously utilized

on a voluntary basis by several fruit and vegetable cooperatives. These

attempts by cooperatives to improve grower income through market control

were unsuccessful, or at least less successful than they might have been,

due to the voluntary nature of the arrangement. Therefore, a. historical

basis existed for this method and for the necessity of an arrangement

that would be binding on all market participants. Secondly, fruit and

vegetable products are generally quite perishable and their production

is usually limited to relatively small, well defined geographic areas.

Other agricultural products are generally less perishable and produced

over a wider geographic area.

These two characteristics, perishability and geographic dispersion,

contribute significantly to the dichotomy of methods. The perishability

of fruits and vegetables precludes price support programs such as that

administered by the Commodity Credit Corporation, i.e., fruits and

vegetables cannot be stored in anticipation of higher prices. Target

prices might be developed for fruits and vegetables but the number of

commodities would create enormous administrative difficulties.

Additionally, the target price concept is predicated on a world market

price criteria which does not exist for most fruits and vegetables. The

concentration of production in relatively small geographic areas results

in a much greater risk of severe reductions in supply due to climatic

and other environmental, i.e., uncontrollable, variables. This increased









risk for fruits and vegetables suggests that acreage controls could

result in supply reductions much greater than anticipated or desired.

Crops subject to acreage controls should be more widely distributed

geographically in order to insulate the total supply from catastrophic

reductions in specific geographic areas. The geographic concentration

of fruits and vegetables therefore suggests that acreage controls are

an inappropriate policy instrument for these commodities.

Thus, the historical experience of producer cooperatives, the

perishability of fruits and vegetables, and the geographic distribution

of production combined with the prevailing philosophy that collusive

activities overseen by the government could be in the public interest

led to the inception of marketing orders and agreements during the

1930's.


Marketing Order Functions


Marketing orders serve several functions. The primary legislated

function is ". .. to establish and maintain such orderly marketing

conditions . as will establish, as the prices to farmers, parity

prices . [and] provide . an orderly flow of [commodities] . .

to market . to avoid unreasonable fluctuations in supplies and

prices" [U.S. Department of Agriculture 1979a, p. 50]. The method

utilized in serving this function results in several supplemental

functions which include the establishment of an industry headquarters,

the collection and dissemination of marketing and production data, the

creation of a forum within which intra-industry conflicts and industry

problems can be identified and debated, and, most importantly, pro-

ducers, handlers, and other market participants have the opportunity to

gain an increased level of economic awareness [Jamison 1965].








The function of establishing and maintaining orderly marketing con-

ditions is approached in the following manner. If growers of a specified

commodity within a specified geographic area organize and vote in favor

of a marketing order, then an advisory committee is established. This

committee is generally predominantly composed of growers, but often

includes handlers and may include representation of other groups such

as processors and consumers. The marketing order committee is authorized

to recommend regulations which are binding on all handlers within the

marketing order area. These regulations may pertain to two primary

control factors--quality and quantity. The committee may also recommend

actions categorized as market support activities. All recommendations

by the committee are subject to approval by the Secretary of Agriculture

or his designate.

The establishment of a marketing order committee, which usually

includes a full time manager and clerical staff, provides a focal point

for industry information and representation. Since the committee must

vote on activities to be undertaken, it is necessary for growers and

handlers to recognize the interdependence among growers and handlers

and between growers and handlers. Through this process of recognition

and reconciliation, growers and handlers become more aware of the economic

interrelationships existing between and among themselves. They also gain

an increased knowledge of the external economic factors which affect them.

The major external factor they must recognize is the potential response

to their activities of producers of substitutes, where substitutes in-

clude commodities that consumers purchase in lieu of the order commodity

as well as production of the same commodity in geographic areas not

subject to the marketing order regulations. Similarly, marketing order








committees must understand what may be termed the internal or direct

economic factors, specifically the response of growers given various

quality and/or quantity regulations, and likewise the response of

handlers, including wholesalers, retailers and ultimately the consumer.

The activities permitted under marketing orders may be categorized

as quality control, quantity control, and market support. There are

specific activities permitted under each category. All marketing orders

do not allow all permitted activities. Each order initially specifies

the activities which may be pursued under that order. Most orders in-

clude some type of quality control. Quantity control activities are

less prevalent and more diverse than quality control activities. Market

support activities vary from one order to another.

There are essentially three types of quality control activities--

grade, size, and maturity. Quantity control activities may be further

categorized into three types--quality regulations, market flow regula-

tions, and volume management regulations. Quality regulations fall under

the quantity control category if their purpose is to influence quantity

rather than to stabilize and standardize the product. Quality regula-

tions which change from one season to the next, tending to be higher during

the periods of larger supply, are appropriately regarded as quantity

regulations, though indirect, rather than quality control regulations.

Conversely, quality regulations which tend to remain constant from one

season to the next may be regarded as setting minimum standards for the

industry and appropriately fall under the category of quality control.

Market flow regulations and volume management regulations differ in

their intended purpose. Market flow regulations are designed to control

the flow of product to market during a given season. In contrast








volume management regulations are intended to limit the total quantity

marketed during the season. Though market flow regulations are pri-

marily intended to control the movement of supplies rather than limit

the total supply marketed, they can under some circumstances influence

the total supply.

Market flow regulations include shipping holidays and handler pro-

rates. Shipping holidays are utilized to reduce the build up in whole-

sale supplies which tend to occur during holiday periods and usually

coincide with calendar holidays such as Thanksgiving and Christmas.

Handler prorates are intended to smooth the flow of commodities through-

out the season and essentially limit the amount of the regulated

commodity that can be shipped during a specified time period. The

perishability of some regulated commodities may result in diversion

of fresh supplies to alternative, i.e., unregulated, outlets under

handler prorate regulations. This effect is similar to that achieved

under the market allocation regulations of the volume management pro-

visions.

Volume management regulations include market allocations, reserve

pools, and producer allotments. Market allocation essentially limits

the total quantity which may be marketed in the primary, usually fresh,

market. The remaining quantity may be marketed for alternative uses

such as export, animal feed, and sometimes processed products. The

reserve pool regulations are similar to market allocation, except that

a reserve supply of the product is maintained. Sales from the reserve

pool are controlled by the administrative committee and may be to the

primary market or to alternative markets. Therefore, the reserve pool

represents a potential supply for the primary market and tends to lessen









somewhat the effect of reduced supplies from diversion to the reserve

pool as compared to market allocation regulations, where no potential

supply exists for the primary market above that supply initially set

by the administrative committee. Producer allotments limit the quantity

of product that a grower may market and are generally based on some

percentage of product historically marketed. The producer allotment

regulation is the only regulation which serves to limit entry of new

firms to some extent. Though provisions are included for granting

allotments to new growers, the conditions which must be met can be

rather stringent.

Market support activities include standardization of container

and pack, production research, marketing research, advertising, pro-

hibition of unfair trade practices, price posting, and the provision

of marketing information.

The quality control, quantity control, and market support

activities provide the opportunity for organized control of product

characteristics, intraseasonal marketing, and intermarket marketing.

In addition, product demand may be influenced by advertising and pro-

motion activities. These activities do not grant monopoly control

in the strictest sense. Total production cannot be controlled by a

marketing order. Attempts to maintain price levels above the competi-

tive equilibrium may be successful for a limited time. However, prices

consistently above the competitive equilibrium will result in increased

production within and/or outside the regulated area. Continued efforts

to maintain prices above the competitive equilibrium would require

restrictions so severe as to be unacceptable to producers.








Even the presence of production control does not insure a non-

competitive pricing mechanism. The Florida celery marketing order pro-

vides for a producer allotment program which does create a friction,

if not a barrier, to entry. In addition, the Florida Celery Exchange,

a marketing cooperative, oversees the marketing of about 95 percent of

all celery produced in the state. This scenario clearly suggests a

monopolistic pricing mechanism. However, Shonkwiler and Pagoulatos

found the market to be consistent with competitive pricing. This result

is likely due to the influence of uncontrolled California marketing.


Marketing Order Participants


Direct participants in the marketing order process are growers,

handlers, occasionally processors and/or consumers, and the Secretary

of Agriculture or his delegate. Growers and handlers initiate a

request for a marketing order. The Secretary of Agriculture then

conducts hearings to determine if an order is justified. If it is

determined that justification exists, then a producer referendum, and

for some orders a processor referendum, is held. Concommitantly,

handlers receive a marketing agreement containing the same terms and

conditions as the proposed order for their signature. A necessary

condition for issuance of the order is that a specified percentage

of growers voting, by volume or number, and also in some cases pro-

cessors by a similar percentage, approve the proposed order. If, in

addition, at least 50 percent of the handlers sign the corresponding

agreement, then the proposed order and agreement are issued. If less

than 50 percent of the handlers sign the agreement, then the Secretary

may issue an order if he determines that an order is the only practicable









means of achieving the desired goals of the order. Orders and agreements

are rarely issued separately. However, if an order is issued the

existence of an agreement is generally irrelevant since an agreement is

binding only upon the signers, whereas the terms and conditions of a

marketing order are binding on all handlers. Thus, the lack of a

corresponding agreement does not change the method of order operations,

but may imply a substantial conflict between growers and handlers.

Regulations issued under an order apply to all handlers of the order

commodity in the specified order area. In addition to being the regu-

lated participants, handlers provide the financial support for local

marketing order operations through assessments on a unit of product

handled basis. The associated local costs include management and

clerical expenses for all orders, and advertising and/or research ex-

penditures for some orders. In some cases a provision for reimburse-

ments to handlers for selected expenditures on activities complementary

to marketing order operations, such as advertising, is included.

The administrative committee is the focal point for marketing order

operations. The composition of the committee varies among orders. Some

committees include growers only; other committees include growers and

handlers or processors; and in some instances a consumer representative

may also be included. Members of the committee are nominated by the

industry and appointed by the Secretary of Agriculture. The primary

role of the committee is to recommend proposed regulations and

activities to the Secretary of Agriculture.

The Secretary of Agriculture is responsible for appraising the

appropriateness of these proposals and subsequently approval or dis-

approval. Thus, the Secretary of Agriculture has continuing









responsibilities with respect to marketing orders. First, it is his

responsibility to determine if an order should be issued. Second, it

is his responsibility to determine if specific regulations recommended

by the administrative committee should be instituted. Third, the

Secretary is responsible for approving the administrative committee's

annual budget and assessment rate. Finally, the Secretary is required

to terminate an order if a majority of growers, by number and volume,

request termination or if the order no longer serves the goals of the

legislation under which it is authorized. These responsibilities are

generally delegated by the Secretary to an agency of the Department of

Agriculture, presently the Agricultural Marketing Service (AMS).

In addition to fulfilling the primary appraisal responsibilities,

AMS provides technical assistance and information to growers in initiating

an order and in the continuing operation of an order.


Problem


Four general perspectives emerge with respect to the marketing

order problem. The first is characterized by the argument that although

marketing orders grant monopoly type powers, these powers are suffi-

ciently limited by legislative and economic constraints [National

Commission on Food Marketing]. The second argument is from a theoretical

perspective. Jesse presents a theoretic exposition on the welfare

implications of selected marketing order activities. Some activities

are shown to unequivocally reduce social welfare in a theoretic context,

whereas other activities are shown to have an indeterminant effect.

Allotment programs were the only activities shown to unequivocally

reduce social welfare. This conclusion, however, is based on the


I








assumption of complete monopoly control; a condition which does not exist

for either commodity, hops and Florida celery, currently utilizing this

provision. In summary, ". .. it is important to recognize these con-

clusions as hypotheses subject to empirical tests" [Jesse, p. 18].

The last two perspectives are more critical. Jamison [1972] pre-

sents a descriptive analysis of several perennial commodities subject to

state or federal orders. Based on historical observations for these

perennials Jamison perceives a pattern leading to chronic surpluses

through the following process. An industry is most likely to cooperate

during periods of severe economic distress and therefore likely to approve

an order. During the first years of operation the order contributes

to an improved economic condition. However, the improvement results in

increased capacity for the industry which in turn calls for more strin-

gent controls. Although the industry may recognize its surplus problem,

it retains the order, fearing the catastrophic conditions which might

prevail if the order were rescinded. Jamison essentially documents the

argument presented by Wellman and Waugh that orders may result in "...

large and unnecessary productive capacity" [Wellman and Waugh, p. 5].

The fourth and final perspective is associated with the Federal

Trade Commission and the Department of Justice. Jesse and Johnson

identify the following seven issues as characteristic of this perspec-

tive:

1. The prevailing price, given a marketing order, is greater

than that price which would prevail under purely competitive

conditions.

2. Marketing orders reduce competition by preventing more efficient

firms from offering their product at a lower price.








3. Marketing orders protect less efficient firms and thus

contribute to excess resources in these industries.

4. Benefits arising from marketing orders are transitory,

since they tend to be capitalized into land values. Con-

versely, higher consumer prices continue in perpetuity.

5. Benefits from marketing orders tend to accrue more to larger

growers and therefore are inequitably distributed among

growers.

6. The current relatively stable economic environment fails to

justify marketing order programs initiated in a depressed

and unstable environment.

7. Marketing order programs are administered with too little

consumer input and/or recognition.

A common characteristic of all four perspectives is limited

empirical evidence which quantitatively relates marketing order activity

and theoretical expectations. Each perspective is essentially theo-

retical and/or subjective in nature. This common characteristic defines

the problem addressed in this treatise. Hoos [1979, p. 281] suggests

the existence of this problem with the thought ". . that the cost

effectiveness of marketing order programs has yet to be determined with

an underpinning of fully adequate economic theory and tested by inte-

grated theoretical and empirical investigations." Although Hoos'

thought was specifically directed at the question of cost effectiveness,

it is appropriate with respect to the entire area of marketing orders

and economic implications. Thus, the problem is a lack of empirical

verification of the subjective and theoretical implications of marketing

orders.









The specific problem of this research is then to provide additional

insight into the economic effects of marketing orders as they are em-

ployed in the fruit and vegetable industries. While the effects of

orders are potentially numerous and multifaceted, the limits to the

present analysis are set forth below.

Most past studies have concentrated on how orders affect the supply

side. While this is essential for completeness, little work has been

forthcoming as to the effects of order activity on demand. Hence, the

analysis will address the problem of order effects on commodity demand.


Objectives


Though numerous results are attributable to marketing order activity,

it is only through the fundamental relationships of supply and demand

that these results can occur. Given this premise, the primary objective

of this study is to examine the relationship between order activity and

demand.

Specific objectives necessary to achieve the primary objective

are

1. Development of a current and historical perspective on

marketing order use and evolution.

2. Development of a theoretical framework linking marketing order

activity and demand.

3. Development of an empirical model to estimate the association

between marketing order activity and demand.

4. Interpretation of the empirical results in light of the

theoretical expectations.









Methodology and Scope


The methodology and scope of the study are predicated on the two

following conditions: (1) federal marketing orders are applicable to

a large and diverse group of commodities, and (2) activities permitted

and utilized are numerous and variable among orders. Commodities

covered may be categorized as fresh fruits and vegetables, dried fruits,

tree nuts, and other horticultural specialities. Empirical analysis

is limited to the 21 orders applicable to citrus fruits, fresh vegetables,

and potatoes as of October 1979. Selected vegetables without marketing

orders are also examined for comparison. Most types of order activity

are observed among the orders included in this study and the methodology

utilized allows interpolation of the results to orders not empirically

considered. The focus of the study is the general association between

order activity and demand as opposed to commodity specific associations.

The methodology may be summarized as follows. Demand or price

functions are estimated for each of the 24 commodities listed in Table

1.1 which are defined by type and location of production, e.g., Florida

oranges and Texas oranges are two such commodities. A time varying

parameter method is utilized which allows permanent and random adjust-

ments in the demand parameters. The presence and degree of parameter

instability are subsequently related to various measures of marketing

order activity. The premise is that order activities should be somehow

reflected in adjustments in the demand parameters over time.

This approach is pursued due to the multidimensional nature of

order activities and the potential reflection of order activity as

interseasonal changes in parameters on one hand and as interseasonal












Table 1.1.--Commodities Examined.


CITRUS FRUITS


Florida Oranges
Texas Oranges
California-Arizona
California-Arizona


Navel Oranges
Valencia Oranges


Florida Grapefruit
Texas Grapefruit
California-Arizona Grapefruit

California-Arizona Lemons
Florida Limes

POTATOES


Idaho-Oregon Fall
Washington Fall
Oregon-California Fall
Colorado Fall
Maine Fall
Virginia-North Carolina


Spring-Summer


VEGETABLES

Idaho-Oregon Summer Onions
Texas Spring Onions

Texas Fall-Winter-Spring Tomatoes
Florida Fall-Winter-Spring Tomatoes

Florida Fall-Winter-Spring Celery

Texas Winter Lettuce

Florida Winter-Spring Cabbage (No Order)
Florida Fall-Winter-Spring Sweet Corn (No Order)
Florida Fall-Winter-Spring Green Peppers( No Order)








stability in parameters on the other. The inability to objectively

define a unique measure of order activity precludes direct incorporation

of activity levels in the individual demand functions. Since order

activities are ignored in the estimation process, the model estimated

is potentially misspecified and consequently subject to parameter

variation, e.g., advertising activities may increase demand over time.

The converse argument of increased parameter stability is based on the

potential ability of marketing orders to stabilize product quality,

pack, and intraseasonal availability.


Organization


The objectives previously listed are addressed in the following

five chapters and a summary provided in the final chapter. Chapter 2

provides a general overview of marketing orders and a review of litera-

ture. Objective two is addressed in Chapter 3. A simple demand model

is presented. Major order activities are related to demand. Alternative

measures of order activity and a method of pursuing the linkage between

order activities and demand are developed. Objective three is addressed

in Chapter 4, where the basic empirical model, including definitions and

statistical methods, is presented. The results from the empirical model

are presented in Chapters 5 and 6. Chapter 5 is devoted to the demand

results and time variant characteristics of the estimated parameters.

The evaluation of marketing order activity and demand is presented in

Chapter 6. A summary of the study is provided in Chapter 7.














CHAPTER 2


MARKETING ORDERS


This chapter is composed of two major sections intended to provide

a historical and current perspective on federal marketing order use and

evolution. The first section chronicles the historical development of

federal legislation pertaining to marketing orders and concludes with a

discussion of currently authorized orders. A review of literature is

provided in the final section.


Overview


The evolution of marketing orders from the conceptual definition

provided by voluntary programs of cooperatives and clearinghouses to

the legislative definition provided by the Agricultural Marketing

Agreement Act of 1937 may be traced over a period of less than 20 years.

In fact, the major transformation occurred during the decade of 1925

to 1935. During the latter part of the 19th century and continuing

until the sharp decline in prices during 1920--22, the agricultural

sector of the U.S. economy enjoyed a period of increasing prices and

output made possible by increasing consumer incomes and population, and

improved transportation and distribution systems. Although the price

declines of 1920--22 mark the end of this period of rapid growth and

prosperity, the problems unique to the marketing of fruits and vege-

tables existed throughout the period and continue currently.








The seasonal nature of production, coupled with the perishability

of most fruits and vegetables, tends to result in a high volume move-

ment during short time-periods and consequently low prices. These high

peaks of movement strain marketing facilities and contribute to in-

efficient utilization and product losses. In addition to intraseasonal

fluctuations in supply, interseasonal variations in supply are compounded

by the geographic concentrations common to horticultural crops. Vari-

ations in grades and sizes are also more pronounced among horticultural

crops. The geographic concentration, perishability, and grade and size

variation lead to distant buyers and tenuous confidence on the part of

these buyers as to the nature of the product. Similarly, the physical

distance between buyers and sellers and the pressure to sell quickly

a perishable commodity results in the opportunity for buyers to engage

in practices viewed by sellers as unfair and discriminatory.


Early History

The rapid growth of cooperatives in the fruit and vegetable industry

during the 1890--1920 period stems at least in part from these problems.

It is in the activities of some of these cooperatives that the basic

concepts of marketing orders are first defined. The California Fruit

Exchange initiated a market allocation program for lemons in 1925 which

allocated lemons between the fresh and processed markets. The benefits

accruing to non-participants and the increasing burden placed on partici-

pants eventually resulted in the demise of this program. Similar

experiences of other cooperatives led to the formation of "clearinghouses"

to coordinate the marketing of products from participating handlers.

The Florida Citrus Growers' Clearinghouse Association, in 1928, had as








its objectives: "(1) standardizing of grade and pack, (2) prohibit move-

ment in interstate commerce of poor grades and sizes, (3) regulate weekly

shipments. ., (4) regulate shipments to auction markets. ., (5) ad-

vertise, and (6) establish minimum prices. . which would at least

return cost of production to efficient producers" [Hamilton, quoted in

Farrell 1966, p.7]. This clearinghouse was operative from 1928 until 1933.

These objectives are essentially parallel with selected provisions

authorized for marketing orders. Similar voluntary programs were initi-

ated by a number of cooperatives and clearinghouses. However, lasting

results were achieved by few, if any, due to the benefits accruing to

non-participants at the expense of participating members.

The legislative evolution of marketing orders begins with the

Dickinson Bill of 1926 which was incorporated into the third and fourth

McNary-Haugen bills of 1926 and 1928, both of which were vetoed by

President Coolidge. The major policy instrument of the McNary-Haugen

bills was an export equalization fee levied on the first commercial

purchaser to cover losses incurred on commodities exported. Exporta-

tion was to be encouraged in order to maintain domestic prices at

renumerative levels. The objectives of the McNary-Haugen bills were

to promote "orderly marketing, to stabilize markets against undue and

excessive fluctuations, to preserve advantageous domestic markets,

[and] to minimize speculation and waste in marketing" [Nourse, p. 4].

These objectives were to be achieved through the actions of cooperatives

or "other agencies" under the direction of a federal farm board. Thus,

while the method, i.e., the export equalization fee, did not follow

through to current marketing order legislation the basic objectives

and philosophy of grower cooperation and control under governmental









counsel were firmly established. The fourth McNary-Haugen Bill, 1928,

contained the first reference to "marketing agreements." The proposed

federal farm board would have had the authority to ". . arrange for

the marketing of any part of the commodity [determined to be in surplus]

by means of marketing agreements with cooperative associations engaged

in handling the commodity or corporations created and controlled by one

or more such cooperative associations" [Nourse, pp. 5--6].

The act would have been administered by a federal farm board. The

board would have been authorized to investigate supply and marketing

conditions on its own initiative and required to carry-out such in-

vestigations if requested to do so by (a) an advisory council repre-

sentative of producers (required under the act), (b) leading cooperative

associations, or (c) organizations of producers.

Although the McNary-Haugen bills were all vetoed, they established

three basic principles of marketing orders. First, they established

the objective of orderly marketing and avoidance of excessive fluctu-

ations in supplies and prices. Second, they established the principle

of producer determined action under governmental auspices. Finally,

they established the principles of market control, i.e., improved

efficiency in marketing and distribution of given supplies at favorable

prices, as opposed to production controls.

The Agricultural Marketing Act of 1929 created the Federal Farm

Board. Two basic responsibilities of the Federal Farm Board were

(1) to encourage and assist in the formation and operation of "national"

cooperatives which would collectively bargain in the growers' interest,

and (2) to provide low cost financing to these national organizations

sufficient to enable them to carryover supplies during unfavorable








marketing periods. Although the act placed equal importance on "orderly

production" and "orderly marketing," the major emphasis during the

1929--1932 history of the Farm Board was in marketing.

The experience of the Farm Board during its short history was not

encouraging. Faced with a drastically reduced demand brought about by

a worldwide depression, marketing control proved insufficient and the

necessity for production controls soon became evident.

A series of "domestic allotment" bills were proposed at least

partially in response to the failure of the Federal Farm Board. These

proposals culminated with passage of the Agricultural Adjustment Act

of 1933 (AAA). The AAA contained three major provisions: 1) production

control, 2) licensing, and 3) marketing agreements. Current marketing

order legislation evolved directly from the marketing agreement and

licensing provisions.

Mehren describes the marketing agreement and licensing sections of

the AAA as ". . an experiment in new fields of law and economics"

[1947, p. 1]. Similarly, Benedict and Stine describe ". . the period

1933--35 . as one of experimenting with a great variety of untried

methods for relieving the distress in agriculture" [p. 327]. The

latitude for experimentation was quite broad since the details of the

creation and operation of marketing agreements were essentially at the

discretion of the Secretary of Agriculture. The act simply stated that

". .. the Secretary of Agriculture shall have power . to enter into

marketing agreements with processors, associations of producers, and

others engaged in the handling, in the current of interstate or foreign

commerce of any agricultural commodity or product thereof .

[Nourse, p. 22].








The generality of the AAA allowed diverse interpretations of the

purpose and relative importance of the production control, marketing

agreement, and licensing provisions. George Peek, first administrator

of the Agricultural Adjustment Administration and major author and

proponent of the marketing agreement provision, placed primary emphasis

on the marketing agreements provision and licensing provisions [Nourse].

He regarded agreements as a tool for securing the most favorable returns

from a given supply, primarily through export diversion, and licenses as

a method of controlling marketing practices and producer-processor margins.

Mr. Peek was philosophically opposed to production controls. Secretary

Wallace, in turn, viewed licensing as a tool to insure compliance with

the intent of the marketing and production programs and considered pro-

duction control essential [Nourse]. Still others viewed the legislation

as a opportunity to bring about radical reforms and reorganizations in the

agricultural sector [Nourse].

In this environment of diverse expectations and interpretations,

with the concommitant sense of crisis and limited legal and practical

precedent, it is not surprising that the original act of 1933 was

significantly amended in 1935. Mr. Peek was administrator of the

Agricultural Adjustment Administration from May of 1933 until December

of that same year. Chester C. Davis became Administrator shortly after

Mr. Peek's resignation. Mr. Davis' philosophical bent was much more

in line with that of Secretary Wallace and the direction of the Adjustment

Administration shifted significantly toward a more aggressive production

control posture with limited emphasis on marketing agreements and

associated marketing controls and reforms.









The principal result of the original AAA experience with respect to

marketing agreements and licenses was establishment of the distinct

dichotomy in agricultural policy for fruits, vegetables, and specialty

crops as contrasted with other agricultural commodities. Marketing

controls were determined to be relatively ineffectual policy instruments

for most agricultural commodities. However, for milk, fruits, vege-

tables, and specialty crops, characterized by historical voluntary

marketing programs, perishability, and/or geographic concentrations

of production, market controls emerged as primary policy instruments.

The amendments of 1935 are characterized by their degree of

specificity, the replacement of licenses with orders, and the limita-

tion of eligible commodities. Although the generality of the original

AAA provided the latitude essential for experimentation and develop-

ment, this same generality contributed to significant enforcement

problems. The generality and apparent delegation of regulatory

authority to producers, processors, and cooperatives, was not viewed

favorably by the courts. Under these circumstances the licensing

provisions of the act were essentially unenforceable by May of 1935

[Benedict and Stine, p. 374]. Replacement of the licensing provisions

with a Secretary's order provision ameliorated these problems by

specifying actions which could be undertaken and placing the regulatory

authority clearly with the Secretary of Agriculture. The order pro-

vision was also less punitive than the license provision, since it

provided for fining violators rather than revoking their licenses to

handle commodities.

Marketing orders, unlike agreements which may apply to any agri-

cultural commodity, were limited to specifically enumerated commodities.








Although the definition of eligible commodities has changed somewhat

over the last 45 years, marketing orders remain primarily limited to

milk, fresh fruits and vegetables, and tree nuts. The enumeration

of eligible commodities was at the expressed desire of handlers and

processors. Marketing orders may, under certain circumstances,

be promulgated without handler or processor approval. Since marketing

order regulations are at the handler level, it is not surprising that

handlers would oppose such a regulatory mechanism. The most vocal

opposition came from the canning industry. The canners argued that

(1) the history of market controls for canning crops was not encouraging,

(2) production control and/or acreage allotment programs would be

necessary, (3) proration would require the use of current or historical

bases, either of which might be inequitable, (4) a one year planning

horizon and limited ability to predict production would make controls

ineffectual, (5) no adequate secondary outlet exists for diverted

product, (6) the problem of carryover in canned goods was insur-

mountable, (7) processors would have insufficient influence in the

formulation of programs, and (8) the necessity for harvesting in a very

short time period would preclude effective grade and size controls

[Mehren 1947, p. 11]. This opposition was responsible for the enumera-

tion of eligible commodities, and explicit exclusion of fruits, other

than olives, and vegetables, other than asparagus, for canning, in the

1935 amendments.

The evolution from a conceptual definition to a legislative defi-

nition is essentially complete with the amendments of 1935. The current

enabling legislation, however, is not the AAA but the Agricultural

Marketing Agreement Act of 1937 (AMAA). Portions of the AAA not relating









to marketing orders or agreements were declared unconstitutional. The

AMAA was passed in order to reaffirm the marketing order and agreement

provisions of the AAA and remove any questions of constitutionality

relating to this portion of the act.


Current Legislation

A detailed description and chronology of the Agricultural Marketing

Agreement Act of 1937 (AMAA) is'provided in Appendix A. A summary of

the essential character of the legislation is presented in Chapter 1

and a discussion of the primary permitted activities is included in

Chapter 3. The major changes in the act are chronicled below.

The AMAA has been amended numerous times; however, the essential

character and method of operation have been retained. The number of

commodities eligible for marketing orders has increased but eligibility

remains essentially limited to milk, fresh fruits and vegetables, nuts,

and horticultural specialties. Resistance from canning and freezing

interests has successfully limited applicability to a select number of

enumerated commodities for processing.

Terms and conditions permitted under marketing orders have been

expanded but this expansion has been limited somewhat by commodity

specificity. In addition, activities of a more supportive as opposed

to regulatory nature have been added. The original permitted activities

were regulatory, allowing limitation of the total quantity marketed,

the allotment of quantities handlers may purchase and/or sale, the

disposal of surpluses, and the creation of reserve pools. Provisions

for inspection requirements were adopted in 1947. Pack and container

regulations, market research, and development activities were authorized








in 1954. Paid advertising and market promotion activities were authorized

for cherries in 1962 and extended to a significant number of enumerated

commodities since then. Production research was authorized in 1970.

Imports of certain enumerated commodities subject to marketing

orders were required to meet the equivalent grade, size, quality, and

maturity requirements with a 1954 amendment. As with paid advertising

and promotion, the list of enumerated commodities has increased since

1954.

In addition to changes in commodity eligibility, permitted activi-

ties, and the extension of selected regulations to imports of certain

enumerated commodities, the policy objectives of the act have also

evolved. The original policy stated that parity prices were to be

achieved through orderly marketing and that the interests of consumers

were to be protected by a gradual adjustment in prices and the pro-

hibition of actions intended to maintain prices above parity. Subse-

quent amendments have tempered the parity price objective and associated

contraint. Orderly marketing evolved from a method of 1937 to a goal

in 1961 with the expressed authority to continue any regulation until

the end of the season if discontinuance would disrupt the orderly

marketing of the commodity.

Marketing orders are voluntary in that producers must request an

order. If the Secretary of Agriculture determines through hearings

that an order is appropriate then a producer referendum is held. Handlers

concurrently indicate their approval of a proposed order by signing an

equivalent agreement. If the required proportion, by number or volume,

of producers voting approve the adoption of an order and the required

proportion of handlers sign the corresponding agreement the order is








promulgated. The Secretary of Agriculture may institute an order without

handler approval under certain circumstances.

Regulations issued under the order are binding on all handlers

(in contrast, marketing agreements are binding only upon signers).

An administrative committee, usually predominantly growers, is nominated

by the industry members and appointed by the Secretary of Agriculture.

This committee recommends regulations to the Secretary of Agriculture

or his delegate. Regulations may be issued only with the Secretary of

Agriculture's approval. Funding for marketing orders is provided through

handler assessments.

More than 40 marketing orders are currently applicable to more than

30 fruits, vegetables, nuts, and horticultural specialties. The 21

marketing orders applicable to citrus fruits, vegetables, and potatoes

as of October 1979 are examined in this study. Currently (October 1979)

active orders are discussed in the following section.


Current Orders

Currently active orders are examined in this section. A review of

authorized provisions and categorization of orders by authorized pro-

visions is initially presented. Discussions of orders by commodity

groups complete this section.

Authorized provisions include quality control, quantity control,

and market support activities. Most marketing orders authorize quality

control activities. Quantity control regulations and market support

activities are less common.

Marketing orders authorizing only quality control and selected

market support activities are listed in Table 2.1. Marketing orders

authorizing quantity control activities are listed in Tables 2.2 and 2.3.










Table 2.1.--Federal Marketing Orders Authorizing Quality Control Regu-
lations Only, as of May 1, 1979.


Order Area and Other Authorized Provisions
No. Commodity Pack and Research and Adver-
Container Development tising

906 Texas Oranges and Grapefruit X X X
909 Calif.-Ariz. Grapefruit X
916 Calif. Nectarines X X X
917 Calif. Pears, Plums and Peaches X X Plums only
918 Georgia Peaches -
919 Colorado Peaches X -
921 Washington Peaches X X -
922 Washington Apricots X X -
923 Washington Cherries X X -
924 Wash.-Oreg. Fresh Prunes X X -
925 Idaho-Oreg. Fresh Prunes X X -
927 Oreg.-Wash.-Calif. Winter Pears X -
928 Hawaii Papayas X X X
931 Wash.-Oreg. Bartlett Pears X X
932 Calif. Olives X X
945 Idaho-E.Oreg. Potatoes X
946 Washington Potatoes Pack
947 Oreg.-Calif. Potatoes Pack X
948 Colorado Potatoes X X
950 Maine Potatoesa X
953 Va.-N. Car. Potatoes -
965 Texas Valley Tomatoes X X
966 Florida Tomatoes X X
979 So. Texas Melons X X

Note: All quality regulations are grade and size.
Inactive.








29





























I I I I IXXXXXX
I >< >< I <><> >







XX XXX I IXXXXXX




I II>







IX XXXX XIXI I XX








XI I 1 X IXXXXXX







am
ICD.d
u ) 4-L 1

r) U 0 0) L 0
'-' o F 0 S- a

+M 1 LO S O- 33
) L OOL4C-

S> S- C t-La ) a)
L - SO f --j


-> N.c0) 0> 0 -0l o 0v

L * C * j s- u u > *1 C -M I -I O .C - }--
*a- *- S a -- a C -) 0

0)1-.0) -J -01-- I-
0 0oa 0 00 3 00






CDr o OOD CDO


S0


E
O
0






0
sZf
0


















I I I I


C-i
S_





L























C/i
_0
Ci D











CE







C i
10
cu
L


















>r1
OU












0 -
i- -



CD
0 +


) Ci
O





i 00
I 0


'0
) +












QJ Q



















(-'


u
Q_
0ir


Ou
Ci0 oi
0. c,


ci .0 X 0 X XXX IX


X I X


I I I X I


I I I X








X X I I I I I> X<


'A C I O
0 C '4- O


it'm 'r m u 4J ro I E
'AC rI 'A C 4 -' = C

WS.AW 0 i W I -- 0 0 -3 A W i ..
C- (uo aC' Ci> ^ .. Io 4 i -' Q a0M



CO LL t 0) o I o tou 0 a_





m0 0O rI. -- CN 1 r, 0- m
CU C1 0o CO CO o coC 0',
0 o) 0m 0O 0M m 0(n o 0 m O


X I X X


S-



0








5-

141
a)





O
Li
0

4-'






C
ro


o


id



















Ci
ra
0

















s-

Ci
0
0
Ci




CI












r<
0

o o


C i
Ci






10
C: C


00>,





0 0








S.0 E 0
0L





-- a C
0. 0 0 '
0. 0 C
CiZCi C
rci .0 0








These orders are segregated by types of quantity control regulations

authorized. Orders authorizing intraseasonal regulations are listed

in Table 2.2. The intraseasonal, or flow-to-market, regulations include

shipping holidays and handler prorates. Orders authorizing seasonal,

or volume management, regulations are listed in Table 2.3. These orders

authorize reserve pool, market allocation, and/or producer allotment

activities in addition to selected quality control and market support

activities. The orders and activities are discussed by commodity groups

in the following subsections.

Citrus Fruits. There are nine currently (October 1, 1979) active

federal marketing orders which regulate to some extent the marketing of

fresh citrus produced in four states--Arizona, California, Florida, and

Texas. Four orders apply jointly to citrus fruits produced in California

and Arizona; four orders are applicable to citrus fruits produced in

Florida; and one order is applicable to citrus fruits produced in Texas.

The types and/or varieties of citrus regulated vary among orders.

The four California-Arizona orders include two orders applicable to

different varieties of oranges, Navels and Valencias, and two orders

applicable to specific types of citrus, grapefruit and lemons. The

Texas order is applicable to oranges and grapefruit. The Florida

orders include three orders applicable to a single type of citrus

and one order applicable to "citrus fruit," which includes oranges,

grapefruits, tangerines, tangelos, and certain other citrus fruits

generally included in these categories, but does not include limes.

A separate Florida order is applicable to limes. The two other type-

specific Florida orders both apply to grapefruit but differ in the


geographic area regulated.





32


All of the citrus fruits regulated may be subject to size require-

ments and, with the exception of the California-Arizona orange and lemon

orders, to grade requirements. Similarly, with the exception of Texas

oranges and grapefruit and California-Arizona grapefruit, all may be

subject to flow-to-market regulations, i.e., weekly volume regulations

and/or shipping holidays. Among those orders authorizing flow-to-market

regulations only the Florida citrus fruit order does not authorize weekly

volume regulation. Research and development activities are authorized for

all orders except the Florida grapefruit and citrus fruit orders. Pack

and container regulations and advertising activities are authorized for

two orders, Texas oranges and grapefruit and Florida limes.

The citrus orders are characterized by the fact that six out of the

nine orders provide for weekly volume regulation. Only three noncitrus

orders contain similar provisions, California Tokay grapes, Florida

celery, and South Texas lettuce, none of which have utilized this pro-

vision. Weekly volume regulations have been utilized every season

during the 1960--61 through 1978--79 period by the California-Arizona

orange and lemon orders. Weekly volume regulations for Florida Indian

River grapefruit were first authorized during the 1961--62 season and

with the exception of the 1963--64 season were utilized each season

through the 1974--75 season. The Florida Interior grapefruit order

obtained this provision during the 1965--66 season and utilized it through

the 1975--76 season. The Florida lime order most recently obtained this

provision, during the 1971--72 season, and utilized it through the 1976--

77 season. The latest season considered in this analysis is the 1978--79

season. Thus, the Florida Indian River grapefruit order has not utilized

weekly volume regulations for the last four seasons; Florida Interior








grapefruit for the last three seasons; and Florida limes for the last two

seasons.

Noncitrus Fruits. There are 16 currently (October 1, 1979) active

federal marketing orders applicable to 13 noncitrus fruits produced in

20 states. Seven of the noncitrus fruit orders are applicable to a

single commodity produced in a single state. These single state--single

commodity orders are (1) Florida avocados, (2) California nectarines,

(3) Washington apricots, (4) Washington sweet cherries, (5) California

olives, (6) California Tokay grapes, and (7) Hawaii papayas.

Plums produced in California are the only domestically produced

plums subject to federal marketing order regulation; however, unlike the

seven commodities above, California plums are included in the only multi-

commodity noncitrus fruit order, the California Tree Fruits (pears,

plums, and peaches) order. Peaches are regulated under three orders in

addition to the California Tree Fruits order. Each of these orders is

applicable to a single state--Georgia, Colorado, and Washington.

Pears are covered under two orders in addition to the California

Tree Fruits order. The three orders applicable to pears include two

multi-state orders and apply to differing varieties. The fresh pears

regulated under the California Tree Fruit order include all varieties

except Beurre Hardy, Beurre D'Anjou, Bosc, Winter Nelis, Doyenne du

Comia, Buerre Easter, and Buerre Clairgeau. The preceding excepted

varieties, except Buerre Hardy, are regulated under a multistate order

covering Oregon, Washington, and California. A third order provides

regulatory authority for Bartlett pears grown in Oregon and Washington.

Three commodities are covered under single commodity--multistate

orders. Fresh prunes produced in Washington and Oregon are currently








regulated under a single order. A recently (1979) terminated order

covered fresh prunes produced in Idaho and Oregon. The two remaining

single commodity-multistate orders apply to tart cherries and cranberries.

The cranberry order applies to ten states--Massachusetts, Rhode Island,

Connecticut, New Jersey, Wisconsin, Michigan, Minnesota, Oregon,

Washington, and New York; and the tart cherry order to eight states--

Michigan, New York, Wisconsin, Pennsylvania, Ohio, Virginia, West Virginia,

and Maryland.

The cranberry, tart cherry, and California olive orders are the only

orders applicable to commodities for canning and/or freezing. Although

selected commodities for processing may be regulated under the marketing

order legislation, only these three have utilized this provision. In

addition to their uniqueness with respect to processing, the tart cherry

and cranberry orders are clearly the most geographically dispersed orders.

Activities authorized under the noncitrus fruit orders are relatively

similar among orders. All of these orders include authority for grade

and size regulations. Pack and container regulations are authorized for

10 of the 16 orders. Georgia and Colorado peaches, Oregon-Washington-

California winter pears, and the three processed fruit orders--tart

cherries, cranberries, and olives--are the only orders which do not

authorize pack and container regulations. Research and development activi-

ties are authorized for all but two orders--Georgia peaches, and tart

cherries. Advertising is authorized under six orders--avocados, necta-

rines, California Tree fruits, grapes, papayas, and olives.

Only four noncitrus fruit orders authorize quantity control activi-

ties and the type of authorized activity is different for each of these

four orders. The Florida avocado order permits shipping holidays; the









the California Tokay grape order premits weekly volume regulation; the

cranberry order permits producer allotments; and the tart cherry order

permits reserve pool activities.

Vegetables. There are seven currently (October 1, 1979) active

federal marketing orders applicable to selected fresh market vegetables

and melons produced in four states--Idaho, Oregon, Texas and Florida.

Only one vegetable order, Idaho-Oregon onions, is applicable to pro-

duction from more than one state. Two orders, for tomatoes and celery,

are active in Florida. The remaining four orders, applicable to onions,

tomatoes, lettuce and melons (except watermelons), all apply to Texas

fresh market production. Regulations have been issued during recent

seasons under all of these orders except the order covering fresh market

tomatoes produced in Texas. The administrative committee of the Texas

tomato order continues to be active and during 1979 initiated a limited

research project addressing the level of market acceptance for various

types of tomatoes.

Authorized activities are quite similar among the vegetable orders.

Grade, size, pack and container, and research and development activities

are authorized for each order. Intraseasonal volume regulation is

authorized for all but the two tomato orders. This volume regulation is

limited to shipping holidays for onions and melons. The celery and lettuce

orders authorize weekly volume regulations. This authority has never

been exercised for celery and was exercised to a limited extent only

during the lettuce order's first season of operation, 1960--61. Adver-

tising activities are authorized for only two orders, Idaho-Oregon onions

and Florida celery. Both orders have utilized this provision during

recent seasons.








The final authorized provision under the vegetable orders is producer

allotments. The Florida celery order is the only vegetable order, and one

of only two fruit and vegetable orders, which may utilize producer allot-

ments. This provision has been utilized each season since the order was

issued in 1965. A marketable quantity of celery is determined each

season. This quantity is apportioned among growers on the basis of each

individual's base quantity. Although actual marketing have historically

been less than the marketable quantity, the allocation of base quantities

to new producers and expansion of base quantities to existing producers

are at the discretion of the administrative committee. Thus, entry and

internal expansion are controllable to some extent.

Potatoes. There are five currently (October 1, 1979) active

marketing orders applicable to fresh market potatoes produced in six

states--Idaho, Oregon, Washington, California, Colorado, Virginia and

North Carolina. An order for Maine potatoes is currently inactive. Two

active orders apply to single states--Washington and Colorado. The re-

maining three active orders each apply to two states--Idaho and East

Oregon, Oregon and California, and Virginia and North Carolina.

Authorized provisions under the potato orders are quite similar.

Grade and size regulations are authorized for each order and pack and/or

container regulations are authorized for all but the Virginia-North

Carolina order. Research and development activities are authorized for

only two potato orders--Washington and Oregon-California.

Neither advertising activities nor quantity control regulations are

authorized for any of the potato orders. State programs for advertising

and promotion and the activities of the National Potato Promotion Board,

established under the Potato Research and Promotion Act, reduce the








likelihood of and need of advertising by individual potato marketing

orders.

Tree Nuts. There are three currently (October 1, 1979) active

orders applicable to three tree nuts--almonds, filberts, and walnuts--

produced in three states--California, Oregon, and Washington. Each

order applies to a single type of nut. The walnut and almond orders

are applicable to California production and the filbert order is

applicable to Washington and Oregon production.

Activities authorized under the tree nut orders are somewhat similar.

All of these orders authorize grade regulations and market allocation

activities. The market allocation provision is unique to the tree nut

and dried fruit orders. The California almond order is the only tree

nut order which authorizes advertising activities. It is also the only

tree nut order which does not include provisions for size and pack re-

quirements. The walnut order is the only tree nut order and one of

only four federal fruit and vegetable orders which authorizes the use of

a reserve pool. Research and development activities are authorized

under the almond and walnut orders, but not under the filbert order.

Dried Fruits. There are three currently (October 1, 1979) active

orders applicable to three dried fruits--dates, raisins, and prunes.

Each of the dried fruit orders is California specific. Grade and size

regulations and research and development activities are authorized under

each order. Advertising activities are authorized for dates and pack

regulations for prunes. Neither pack nor container regulations are

authorized for raisins. Market allocation activities are authorized for

dates and reserve pool activities are authorized for prunes. The raisin

order is the only dried fruit order authorizing both market allocation








and reserve pool activities. The only other federal fruit and vegetable

order with this dual authority is the walnut order.


Literature Review


The most recent general treatises on Marketing Orders and Agree-

ments are the report of the National Commission on Food Marketing (NCFM)

and a supplement by Farrell, both in 1966. The NCFM report is an exami-

nation and evaluation of then active State and Federal Marketing Orders.

The supplement by Farrell provides a history of the development and

evolution of Marketing Order legislation, as well as more detailed infor-

mation on State and Federal orders. The above studies and others docu-

mented here pertain almost exclusively to non-milk orders. Marketing

orders for milk are permitted under the same enabling legislation as

non-milk orders but differ substantially due to the unique price setting

authority provided under milk orders. This unique authority and the

widespread application of milk orders have contributed to the development

of an extensive body of literature not directly considered in the study.

In addition to the NCFM and Farrell studies, several general de-

scriptive and theoretical studies and articles have been published [See

Mehren 1947 and 1968; Smith; Hedlund; Benedict and Stine; Hoos 1957a,

1964, and 1979; Jamison 1965, 1966 and 1972; and Jesse]. Most of these

works are qualitative, theoretical, and/or descriptive in nature. Many

provide excellent evaluations, often based on years of study and experi-

ence, but they all recognize the difficulties of measuring order effec-

tiveness.

Jesse and Johnson provide the most general quantitative analysis of

marketing orders to date. Their analysis of marketing orders focuses on









the level and stability of order commodity prices compared to prices of

similar commodities without orders. They also categorize orders with

respect to their potential ability to raise and/or stabilize grower prices.

Their results indicate that grower prices for order commodities have been

neither higher nor more stable than prices for similar nonorder commodi-

ties and that price patterns for stronger orders do not differ from those

for weaker orders. However, it is recognized that commodities with

marketing orders may be subject to extreme price variability which the

orders only partially dampen.

The above studies may be supplemented by several studies describing

Federal and State orders applicable to California [see Hoos 1956 and

1957b; Foytik; Farrell and Wood; Jamison 1964; Garoyan and Youde; and

French, Tamimi, and Nuckton]. Another category of literature includes

commodity specific studies. Some individual commodities which have been

the subject of analysis are carrots [Shafer], dates [Dennis], raisins

[Pritchard, and Wohlgenant], cling peaches [Jamison 1966], tomatoes

[Brooker and Pearson], citrus [Thor, Thor and Jesse, Nicolatus, Hamilton,

Clodius, and Nelson and Robinson].

Thor and Jesse provide a recent example of a commodity specific

evaluation of marketing orders in their study of the California-Arizona

orange orders. This study simulates industry performance with and without

controls over historical and projected time periods. An econometric

model is utilized which allows the evaluation of selected performance

measures such as average returns, shipping season lengths, and bearing

acreage.

Thor and Jesse estimate the direct effect of marketing order controls

on intraseasonal allocation of marketing and season length. In








simulating performance without controls, the intraseasonal allocation

and season length are unconstrained. In addition, the influence of

order controls on the allocation of oranges between fresh and processed

markets is reconciled on the basis of either fresh and processed price

equivalence or maximum fresh marketing. Price equivalence is a net

concept which recognizes cost differences between the two markets.

Maximum fresh marketing is binding when the quantity of fruit suitable

for the fresh market is insufficient for price equivalence. Thus, the

model focuses on the effects of order controls on intraseasonal and

intermarket allocations. These allocations consequently influence

average returns which in turn influence bearing acreage or supply.

A final category of marketing order literature may be defined as

official criticism and/or critique. Official statements with respect

to Marketing Orders are found in U.S. Department of Agriculture 1975,

Federal Trade Commission 1977a and 1977b, Lipson and Batterton, Mason

1975, and National Commission for the Review of Antitrust Laws.

Additionally, personal views of Federal Trade Commission and Department

of Justice staff members are found in Masson 1976 and Masson, Masson and

Harris. The U.S. Department of Agriculture study appraises the potential

price impact of marketing order programs. Federal Trade Commission 1977a

and 1977b are official comments and exceptions to proposed regulations

under the cranberry and celery marketing orders, respectively. Both

statements are with respect to the use of the allotment provisions

allowed under these orders. Lipson and Batterton and Masson 1975 examine

the relationship between cooperatives and Marketing Orders. Masson 1976

comments on citrus marketing orders and Masson, Masson, and Harris con-

sider Marketing Orders and cooperatives.





41


The historical and current perspective on marketing orders presented

in this chapter provides the base for development of the theoretical

linkage between marketing order activity and demand. A more general

examination of order activity is accomplished in the following chapter

and a method developed to evaluate the reflection of order activity in

demand.















CHAPTER 3


A THEORETICAL LINKAGE OF MARKETING ORDERS AND DEMAND


The theoretical and conceptual foundation for evaluating the effects

of marketing orders is developed in support of the basic proposition that

marketing order activities are partially reflected in demand relation-

ships. The exposition begins with the development of a simple model of

demand. The major policy instruments of marketing orders are reviewed

and the potential reflection of their use in demand is examined. Multi-

dimensional measures of the degree of marketing order activity are

developed and a method utilizing these measures to identify the relation-

ship between marketing order activities and demand is presented. While

it is recognized that supplies may also be influenced by orders over

extended periods of time, this issue is not addressed in the analysis.

Figure 3.1 illustrates in the most simplistic terms the overall

framework for this chapter and subsequent analyses. Marketing order

effectiveness is transmitted through quantity controls, quality controls,

and market support functions. Each of these functions is implemented

through a multitude of market order activities occurring in varying

degrees of intensity. These activities will be discussed in later

sections.

In total, orders should affect supplies to some degree over time.

In fact, as noted in the literature review, most studies have concentrated

on the economic effect of orders on supplies. A second dimension of




















































Figure 3.1. A Simple Linkage of Marketing Orders with Supply
and Demand.








order effectiveness deals with the influence of orders on the demand for

the order commodities. As noted in Chapter 1, the scope of the current

analysis is limited to this dimension of the order issue. Little research

on this dimension of order activity is published. In this chapter, the

theoretical linkage between orders and adjustments in demand will be

developed. The regulatory characteristics will be detailed and how

they theoretically impact demand explained.


A Demand Model


The number of commodities examined in this study precludes individual

industry analyses of a very complex nature. An inherently simple model

which retains the basic elements of demand and captures at least some

of the most influential elements unique to individual commodities is

appropriate in this circumstance. While a rigorous theoretical derivation

of demand relationships may be developed and empirically applied in a

complete systems context, the transition from a theoretical to empirical

model in a partial systems context is neither as direct nor as mathe-

matically rigorous. A partial analysis draws upon the basic theoretical

postulates to define the critical factors, to assist in maintaining logical

consistency in model specification, and, finally to provide one basis for

evaluating the logical consistency of the empirical results.

The basic theoretical postulate of demand is that consumers act in

a manner consistent with the maximization of utility. Then demand is

directly predicated on an income constraint and relative prices. The

consumer seeks to purchase that bundle of goods which will maximize

utility subject to available income. Thus, on an individual basis, prices

of available goods, income, and the nature of the utility function are

the major factors which define demand.








The postulate of utility maximization subject to an income constraint

and the assumption of fixed supplies of agricultural commodities provide

a sufficient foundation for the specification of a relatively simple

demand relationship. Theoretical arguments imply that the prevailing

price for a given quantity of product is predicated on consideration of

the prevailing price and/or quantity of all available goods and services.

However, an ad hoc and intuitive appeal to the consumer budgeting approach

[see George and King, and Bieri and deJanvry] allows specification of the

less comprehensive general relationship:

P = f(Q,S,Y,Z) where

P = price,

Q = quantity,

S = a set of substitute goods,

Y = income, and

Z = a set of commodity specific variables.

The variable set composed of Q, S, and Y is the classical set of

demand variables. The sets S and Z are subject to interpretation. Sub-

stitutes, though a member of the classical group of demand determinants,

are not uniquely defined in a partial analysis. The substitute definitions

considered in this study diverge somewhat from the classical definitions

in all but one case. The commodities examined are spatially specific,

e.g., Florida oranges. Thus, the obvious substitute is available supplies

of the same commodity from other areas. This is in contrast to the more

usual definition of substitutes as similar but different commodities such

as beef and chicken for pork. A usual definition is pursued, however,

with respect to lemons and limes where each is examined as a potential

substitute for the other. The specific substitutes examined, as well as

other variables, are defined in the following chapter.








Two variables are included in the Z set, the proportion of order area

fresh production marketed under regulation and the proportion of total

fresh production exported in the case of citrus. The proportion marketed

under regulation is examined in an attempt to discern the direct effect

of marketing order allocation activities. These allocation activities

may be direct as in the determination of the total seasonal quantity of

Navel and Valencia oranges to be marketed in selected outlets or indirect

as is potentially observed with quality regulations. The interpretation

of the proportion regulated parameter is provided in the final section

of this chapter.

The second variable included in the commodity specific set, the

proportion of total production exported, is unique to the citrus equations.

The inclusion of this variable is in recognition of the recent variability

in citrus exports [U.S. Department of Agriculture 1980a]. Exports of

potatoes and fresh vegetables examined are minor and thus not considered

in the analysis.


Marketing Order Activities


Marketing order activities may be broadly categorized as quality

control, quantity control, and market support activities. Those activi-

ties most likely to have direct price impacts fall into the quality

control and quantity control categories and include grade and size,

flow-to-market, and volume management activities [U.S. Department of

Agriculture, 1975]. Market support activities include container and

pack standardization, research and development, prohibition of unfair

trade practices, price posting, and the provision of market information.








Wood suggests a fourth category which might be labeled as interaction

and facilitation activities. The interaction and facilitation activities

are not as clearly defineable as some other order activities. They in-

clude increased interaction and understanding between and among handlers

and producers, the discussion and potential resolution of industry problems,

and the combined effect of setting the "rules of the game" for the market

participants.

A discussion of each category follows. The discussions focus on

demand side relationships. Though the supply side is not explicitly

considered, a brief evaluation of potential supply effects is provided.


Quality Control

Grade, size, and maturity restrictions are not directly authorized

in the AMAA. The five permissible terms and conditions specified may be

applicable to any grade, size or quality of an eligible commodity. The

first permissible activity defined is the limitation of the total quantity

marketed in or transported to any or all markets. Limiting this quantity

to zero, making the restriction applicable to selected grades, sizes,

and/or qualities and defining markets by form, results in the grade,

size, and maturity regulations typically utilized.

Quality controls may be utilized to pursue two objectives. The

first objective is, in essence, to change or stabilize the commodity

from a form perspective, recall the time, space, and form dimensions

of utility. In the absence of quality control measures the form vari-

ability is potentially greater. Assuming that the buyer assesses the

product on the basis of perceived average quality, this variability

results in a lower perceived average and consequently a reduced demand

level. The second objective is to indirectly effect a reduction in








quantity marketed. The impact of quality controls on long run supply is

not obvious. If production adjustments can be made which increase the

quantity of product meeting the required quality standards without

significantly affecting production costs, the long run effect on supply

may be minimal if relatively constant quality standards are maintained.

If production adjustments cannot be made as described, a reduction in

long run supply could be expected.


Quantity Control Activities

Quantity control activities consist of two general types, flow-to-

market and volume management. These activities regulate the flow of

products over time and among markets. Flow-to-market activities are

uniquely temporal in character with indirect form implications. Con-

versely, volume management activities are primarily form oriented but

may have important and direct temporal implications.

Flow-to-Market. Prorates and shipping holidays are specific flow-

to-market regulations. Prorates limit the total quantity marketed during

specific time intervals. Shipping holidays prohibit marketing during

selected periods.

Prorates are operationalized in several different manners. The

essential characteristic of all prorate regulations is that handlers

are limited to a defined volume of trade during selected time periods.

Prior to each period, the volume to be marketed during the regulated

period is set by the administrative committee. This volume is sub-

sequently apportioned among handlers, i.e., prorated, on the basis of

each handler's proportionate share of the total volume available for

marketing. Other characteristics vary among individual marketing








orders. Mechanisms typically exist which allow handlers to overship,

undership, carry-over, and loan or transfer prorates. The prorate period

is generally a week, and may be unlimited, effective throughout the

season, or limited to a specific number of periods during the season

and/or a specified time interval within the season.

Prorates provide an opportunity for price discrimination under the

assumption of temporally variant demand. They also provide an opportunity

to stabilize commodity flows in what might otherwise be a glut or a

shortfall in the market. Two potential benefits arise from the more

stable market. First, buyers are more likely to patronize a market

characterized by consistently available supplies. Second, although

not empirically verified, it has been suggested that the psychology of

the market is such that the existence of very low prices will tend to

result in lower prices in the future than would have otherwise prevailed

[Nourse].

Although prorate regulations may necessitate financial adjustments

due to cash flow constraints, the direct cost to producers is not clearly

changed and consequently prorate regulations would not be expected to

affect the annual supply relationship. If buyers favor a more stable

market and prorates are utilized in this manner, an increase in demand

would be expected. Similiarly, if the market psychology is such that

currently low prices tend to result in future prices lower than other-

wise, prorates would be expected to generate a higher annual demand

relationship. Prorates utilized in a price discriminating manner should

not directly affect demand. However, if prorate activities increase

total revenue through price discrimination, then a higher average annual

price would be expected with a given annual quantity than would be








expected in the absence of price discriminating prorates. Thus, while

inherent demand may not have been changed, the observed average annual

price-quantity relationship may be at a higher level than otherwise in

the presence of effective price discriminating prorates.

Shipping holidays are periods during which all shipments are pro-

hibited. Generally coinciding with traditional holidays, primarily

Thanksgiving and Christmas, they are intended to avoid a build up in

supplies during periods of supposed lower wholesale demand. Since

shipping holidays are not nationally recognized, the prohibition of

shipments from a single geographic area provides an incentive for buyers

to purchase from alternative areas where shipments are not prohibited.

Buyers may tend to continue purchasing from the alternative supplier,

and thus a lowered demand for product from the regulated area may follow

a shipping holiday. Countering this argument is the tenet that temporally

related demands are such that low prices during a given period result

in a lower demand during subsequent periods. Gunter, Lee, and Fairchild

address this question with respect to Florida oranges and grapefruit.

Their results are somewhat inconclusive, however, due to severe data

limitations. Although shipping holidays may result in some loss of

product, the direct effect on producer costs appears minimal, and con-

sequently should not directly affect the supply relationships.

Volume Management. Market allocation, reserve pool, and producer

allotments are specific volume management activities. Each of these

activities is intended to limit the total quantity marketed in the pri-

mary market. The flexibility of each activity varies from relatively

fixed for producer allotments to relatively flexible for reserve pools.

Producer allotments are known prior to planting and essentially fix an









upper bound on the primary market quantity in any given season. Market

allocation, usually not known at planting, fixes the percentage of the

total quantity which may be marketed in the primary outlet. This

percentage may change as the season progresses, however, and does not

strictly bound seasonal supplies to the primary market. Reserve pools

may be partially, or totally, released to the primary market and

similarly do not fix strict primary market bounds.

Producer allotments fix the quantity that handlers may purchase from

individual producers. A base quantity, or base allotment, is associated

with each producer. The administrative committee recommends the per-

centage of the base allotment that is marketable in the primary outlet.

The issuance of new allotments and increases in existing allotments are

also recommended by the administrative committee. This control potentially

fixes an upper bound on total production through denial of entry to new

producers or increases to existing producers. The producer allotment is

unique in this ability to potentially control production. Although

recommendations on entry and increases are not strictly arbitrary, the

conditions necessary to gain entry or an increased allotment may be

quite stringent.

Market allocation activities fix the percentage of estimated total

supplies marketable in the primary outlet, usually the fresh domestic

market. The objective under market allocation is essentially price

discrimination, or the allocation of supplies in such a manner as to

take advantage of markets having dissimilar demands. In the absence

of allocation, prices, less delivery costs, would tend to equilibrate

in all markets. Under the assumption of a more inelastic primary market

demand, it is to the producers' benefit to limit primary outlet









marketing and thus increase secondary outlet marketing. The result is

a higher price in the primary market and a lower price in the secondary

market. Given a more inelastic primary market demand, the net effect

is to raise the average price over all markets. Thus, as in the case

of prorates, inherent demands may not be affected but the observed

average price, given a fixed total quantity, is higher in the case of

successful market allocation. Since market allocation activities do not

appear to directly affect producers' costs, the supply relationship should

remain unchanged.

Reserve pool operations allow both temporal and form types of price

discrimination. Reserve pools may be totally or partially released to

secondary markets or the primary market in the current or following

seasons. Thus, reserve pools potentially provide the ability to take

advantage of interseasonal, intraseasonal, and intermarket variations in

demand. Again, as with market allocation and prorate activities, inherent

demands are not necessarily affected by reserve pool operations but the

observed average price, given a fixed quantity, may be higher. Similarly,

to the extent that reserve pool operations leave producers' costs un-

changed, supply relationships should not be directly affected.


Market Support

Pack and container standardization, marketing and production research,

and market development are the primary market support activities. With

the exception of production research, these activities should impact

primarily on demand. Pack and container regulations serve to improve

buyer confidence. Market development, including paid advertising, is

a direct effort to expand demand. Marketing research may have several








objectives, one of which may be increased knowledge of demand and improved

ability to take advantage of demand variations. Other objectives may

include improved advertising or promotion programs, improved cost

efficiencies in marketing, or improved quality at the wholesale and/or

retail level. Production research, while primarily affecting supply, may

have as its objective the development of a variety or type with character-

istics preferred by the consumer.


Reflection in Supply and Demand

Marketing order activities may have several direct short run, i.e.,

single season, effects. Total production, total marketing, intraseasonal

allocation, interseasonal allocation, and intermarket allocation are

potentially subject to the direct influence of order activities. Total

production may be directly affected only through producer allotment

activities. Total marketing, in a given season, may be directly affected

by quality control regulations and reserve pool operations. Reserve

pool operations directly affecting total marketing are actually inter-

seasonal allocation activities. While quality control measures, e.g.,

maturity regulations, may effect a change in intraseasonal allocations,

flow-to-market regulations have the primary direct effect. Some influence

on intraseasonal allocations may also be exerted through market allocation

and reserve pool operations. Intermarket allocation is directly affected

by market allocation and reserve pool activities. Flow-to-market,

quality control, and producer allotments may also effect some change in

intermarket allocation, but this result is not a primary objective of

these activities.

Marketing order activities may effect a change in inherent demands

and/or apparent demand. Changes in apparent demand occur as a result of








intraseasonal, interseasonal, and intermarket allocations. Although

inherent demands, i.e., time and market dependent demands, may remain

unchanged, allocation activities may result in average prices different

from those which would otherwise prevail for a given seasonal quantity.

Changes in inherent demands may occur as a result of intraseaonal alloca-

tion activities, quality control measures, and selected market support

activities, e.g., advertising and pack and container standardization.

Although marketing orders may influence total production and total

marketing to some extent, the major direct effect of order activities

during a single season is on apparent demand, through intraseaonal and

intermarket allocation activities. The reflection of order activity in

supply and inherent demand is a potentially long term result. Advertising

programs, intraseasonal allocation activities, quality control regulations,

and similar activities intended to improve demand are not immediately

effective. A learning period is necessary to attain some potentially

maximum effect. Similarly, this result can be maintained only through a

continuation of the original activities.

Although not addressed in this study, order activities may be

reflected in supply through two primary mechanisms, anticipated prices

and perceived risk. If order activities result in a changed effective

demand, i.e., apparent or inherent, a movement along the existing supply

curve will be accomplished through an associated change in expected

prices. If order activities result in a change in perceived risk, then

a shift in the supply relationship will occur. A tertiary mechanism is

production research and marketing research resulting in changed production

costs and an associated shift in the supply relationships.









Intraseasonal and interseasonal price variability are obvious factors

affecting risk perception. However, the result of those activities or

functions considered by Wood and categorized here as interaction and

facilitation activities may also influence risk perception. The group

identity, the increased communication between and among producers and

handlers, the delineation of "the rules of the game," and the existence

of possible future gains, if not previously observed gains, may result

in producer optimism which reveals itself in a diminished perceived

risk. Thus, while order activities may not individually affect the

supply relationship, the marketing order as an entity may serve to

reduce the level of perceived risk, and consequently effect an outward

shift in the supply curve.


Indicators of Marketing Order Intensity and Potential


Marketing orders may affect demand as a result of the utilization

of individual order provisions. The degree of this influence is predi-

cated on internal intensity, i.e., the availability and utilization

levels of specific provisions, and tempered by the environment within

which the order operates. Therefore, in order to ultimately assess the

effects of marketing orders, some measure of this intensity must be

derived.


A Static Indicator of Potential

A composite index of potential order effectiveness is developed in

an unpublished paper by Jesse and Armbruster. This index, although

static and oriented toward overt quantity control measures, provides

a check list of internal and environmental factors and serves as a









basis for development of a dynamic index encompassing internal factors

beyond direct quantity control efforts.

The index developed by Jesse and Armbruster results from a subjective

scaling of selected internal and environmental factors and subsequent

summation of these scales. Order provisions and applicability of

regulations to imports are the internal factors considered. Primarily

environmental factors considered are the relative share of U.S. supply

subject to regulation, the number of orders applicable to the commodity,

the availability of alternative, i.e., unregulated, domestic markets,

the influence of cooperatives, and changes in per capital consumption.

Specific order provisions are assigned a scale value ranging from

zero to three. All order provisions other than direct quantity control

measures are valued at zero. Shipping holidays and limited handler

prorates are scaled equivalently at one. Market allocation and reserve

pool operations are similarly judged equivalent, but potentially of

greater impact with a scale value of two. Unlimited handler prorates

and producer allotments are judged to have the greatest potential impact

and are assigned a scale value of three. The applicability of quality

regulations to imports receives a scale value of one.

The relative share of total U.S. supply subject to regulation is

scaled from zero to three and varies in unit increments as the relative

share ranges from less than 10 percent, 10 to 50 percent, 50 to 75

percent and above 75 percent. Although average annual supplies provide

the primary basis, seasonal supplies are utilized where appropriate.

The remaining environmental factors are assigned zero/one scale

values. The existence of multiple orders is assigned a scale value of

zero versus a scale value of one for a single order. Multiple orders








are not coordinated and tend to reduce the potential impact associated

with a single control instrument. Similarly, the availability of unregu-

lated outlets accounting for at least 50 percent of regional marketing

is assigned a scale value of zero versus a scale value of one where regu-

lated marketing are in excess of 50 percent. A strong cooperative in-

fluence provides an additional degree of coordination and cohesion

contributing to increased potential. Thus, a scale value of one is

assigned when a single cooperative processes, packs, and/or sells more

than 50 percent of production under an order.

Order commodities which have experienced a relative change of a

positive nature in per capital consumption compared to the relative change

in per capital consumption of appropriate aggregate commodity groups are

assigned a scale value of one. Relatively positive changes in per capital

consumption may be a result of order activities and are therefore not

strictly indicative of increased potential. However, it might be assumed

that the primary determinants of per capital consumption are unrelated

to order activities. Thus, relatively positive changes in per capital

consumption may enhance the effectiveness of order activities. Con-

versely, a declining consumption per capital, possibly indicative of

a declining demand, may dampen the result of order activities.


Objective Indicators of Intensity

The index developed by Jesse and Armbruster is a subjective indi-

cator of order potential, based on environmental factors and the

availability of selected order provisions, primarily quantity control

provisions. An objective indicator of order intensity, based on en-

vironmental factors and the utilization of a comprehensive set of

order provisions,is more congruent with the objectives of this analysis.








A single objective indicator of intensity, however, is not defineable

through a mapping procedure such as that utilized by Jesse and Armbruster.

Such a procedure is inherently subjective. Even quantification of the

intensity of individual order activities is not feasible in light of

current information and analysis constraints. The quantity effect of

quality control activities, for example, is essentially unknown [U.S.

Department of Agriculture 1975, p. 36], and quantification of allocation

activities, while potentially possible, implies a degree of analysis

feasible only in the study of a limited number of commodities. Thus,

objectivity and empirical tractability constraints preclude the

development of a unidimensional measure of intensity and limit the

analysis to uncomplicated objectively measurable dimensions.

Although most order provisions can be observed only to the extent

that they are utilized or not, several quantifiable dimensions associated

with marketing orders may be defined. Total expenditures, total supplies

potentially subject to order regulations, and total supplies marketed

through regulated and unregulated outlets may be objectively measured.

In addition, specific order provisions may be categorized with respect

to their anticipated impact on the real and/or perceived nature of the

product, intraseasonal allocation, interseasonal allocation, and inter-

market allocation. The environmental dimensions of alternative markets

and commodity coverage are accounted for in selected quantifiable

dimensions enumerated above. Cooperative influence may be directly

quantified in terms of the quantity cooperatively marketed. Multiple

orders may be reflected in the quantity potentially subject to regulation

under each order relative to the total supply.









Measuring the Association Between
Order Activities and Demand


The theoretical development of a basic demand model has been pre-

sented with limited reference to order activities. Similarly, the

potential impact of order activities and methods of quantifying associ-

ated dimensions is presented with only general implications for demand.

The linkage is developed here. The obvious linkage would appear to be

discernable by direct incorporation of the quantifiable dimensions in

the demand model. This approach is not currently feasible, however.

The multidimensional nature of order activities, the relative

temporal stability for a given commodity of many order dimensions, and

a limited observational period preclude direct incorporation of the

quantifiable order dimensions. The number of quantifiable dimensions

is such that an expanded observational period or cross-commodity pooling

would be necessary to achieve a reasonable number of observations

relative to the number of independent variables. Similarly, the rela-

tive temporal stability of many dimensions would require cross-commodity

pooling in order to achieve sufficient dimensional variability. While

cross-commodity pooling is an attractive alternative, an appropriate

pooling procedure is not evident since the problem relates to commodities

that are not within the same product group.

However, as previously indicated, one order related variable is

directly incorporated in the demand function, the proportion marketed

under regulation. Marketing orders may indirectly allocate products

between markets through quality regulations or directly through allo-

cation activities. Some markets are exempt from regulations. Thus,

more restrictive regulations tend to force products into these unregulated








outlets. The effect of this allocation on the average price for all

products is dependent on the relative price flexibilities. If the

unregulated market is less flexible than the regulated market, then

the allocation of additional product to the unregulated market will

result in a higher average price. The proportion marketed under regula-

tion would be negatively related to average price in this circumstance.

Conversely, if the unregulated market is more flexible, a positive re-

lationship between price and the proportion regulated would prevail.

If regulations have been effected consistent with producers' interests,

a negative relationship would be expected.

The question of marketing order impacts on demand parameters is

obviously not addressed by the above procedure. In order to address

this question in the absence of direct incorporation of order activities,

it is enlightening to consider the potential causes and types of parameter

variation. Ward and Meyers suggest three general causes: (1) structural

change, (2) model misspecification, and (3) aggregation. Rosenberg

identifies two types of parameter variation, systematic and stochastic.

Each cause and type is potentially relevant to the problem addressed

here. The primary causes of interest are model misspecification and

aggregation. Structural change is of secondary interest given that the

marketing orders considered have generally been in existence throughout

the period studied.

Model misspecification and aggregation suggest systematic and

stochastic parameter variation, respectively. Order activities which

are excluded from the model may result in parameter variation over time,

e.g., advertising expenditures may result in an outward shift of the

demand curve. The aggregation effect is less obvious than the








misspecification effect and potentially yields an opposite effect. Price

is estimated as a function of quantity and other variables. Quantity

is a macro variable in the sense that various degrees of quality exist

and quantity is measured as the sum of the weights or units in each

quality category. Instability in the quality mix, while not evident

in the quantity data, may be reflected in a manner tantamount to random

or stochastic parameter variations. Thus, marketing order regulations

which serve to stabilize this quality mix might be reflected as more

stable demand parameters compared to those observed in the presence of

a less stable quality mix. A similar argument may be made where the

intraseasonal availability of supplies is made more stable or consistent

from season to season.

Marketing orders influence the demand relationship in two direct

ways. If the orders change the characteristics of the quantity shipped,

then the product is being directed toward a different set of preference

functions. These preferences may be for both old and new consumers.

Given that the quantity data do not reflect such characteristic changes,

then one would expect the demand parameters to vary because of the

aggregation problem over a narrowing (or widening) mix of qualities,

sizes, maturities, etc. This change in orders alone could lead to the

time varying parameters. The second dimension relates to the case where

orders can influence the information flow about the product. In this

situation, orders may directly alter the preference function for order

commodities. No doubt that this is precisely what is intended with

advertising and labelling programs. If this were the only dimension

to orders, then the time varying demand parameters could be measured

for homogeneous commodities. A change in preferences would be measured

with the time varying parameters due to omission of variables.








In actuality both information controls and changes in the product

mix occur simultaneously with orders and there is no clear way to

separate the two. What will be measured in the subsequent analysis is

the composite effect of both. The composite effect is more important

to the study since interest is primarily in the impact of orders and not

the various components. While demand changes are referred to throughout

the analysis, it is clearly recognized that both information changes and

product mix are elements of the change. Therefore, all references to

demand changes in the remaining discussion will entail two elements:

(1) there may be adjustments in the preference functions, and (2) the

estimated demand is an average over demand for different sets of product

characteristics. The final results, again, show the total effect of

changes in the estimated average demand.

The expected reflection of order activities in demand may thus be

identified as interseasonal changes in the demand parameters in one

instance and interseasonal stability in the demand parameters in the

other. Conventional estimation methods based on the assumption of fixed

parameters are insufficient in these circumstances. Several methods

are available, however, for dealing with the varying parameters problem.

Maddala provides an overview of selected methods and a brief synopsis

is provided in the following chapter of this study. The most general

of these methods is the Cooley-Prescott varying parameter regression

model. A detailed description of this method is provided in the

following chapter.

It is sufficient at this point to recognize only the critical

characteristics of the Cooley-Prescott varying parameter regression (TVP)

method. This method incorporates both types of parameter variation,








systematic (permanent) and stochastic (transitory). The relative degree

of permanent parameter variation is indicated by gamma, an estimated

statistic. As gamma varies from 0 to 1 the degree of permanent change

similarly varies from 0 to 100 percent. If gamma is zero the parameter

variation is completely transitory. This is interpreted to indicate

that though parameters may vary from some initial value they tend to

return to an estimated mean value. If gamma is greater than zero there

is no tendency to return to a mean value and unique parameter estimates

exist for each time period. Thus, the varying parameter method provides

a measure of the degree of permanent parameter change in gamma. However,

if gamma is less than one there remains a question with respect to the

degree of randomness. This is particularly true if gamma is zero.

If gamma is zero the Cooley-Prescott (TVP) model reduces to a

generalized least squares model with a heteroscedastic error structure.

Maddala, Kmenta, and other general econometric texts provide extensive

coverage of heteroscedastic models. The question of relative randomness

among parameters, however, is not addressed. In the absence of a

traditional and rigorously defined method, two primarily intuitive

measures of randomness are examined. The first measure is the pro-

portional deviations of the TVP parameter estimates from the ordinary

least squares (OLS) parameter estimates. The second measure is the

proportional deviations of TVP parameter t-Values from OLS parameter

t-Values. These measures are discussed in the following chapter.

Given the measures of permanent parameter variation and randomness

it remains necessary to identify their association with marketing order

activities. This is accomplished through the use of a non-parametric

test of the hypotheses that order activities and gamma are independent





64


and that order activities and randomness are independent. The test

utilized is the Kendall K described in the following chapter. An

associated statistic, the Kendall correlation coefficient, provides

an indication of whether the association is positive or negative. This

statistic is also described in the following chapter.

In summary, the procedure to be used is to estimate average demand

functions for each commodity allowing for both permanent and random

adjustments in the parameters. The presence and degree of parameter

instability is then related to various measurements of marketing order

activity.














CHAPTER 4


DATA AND EMPIRICAL PROCEDURES


The theoretical model enumerated in Chapter 3 is operationalized

in this chapter. The chapter is composed of two major sections. The

first section is primarily one of model specification, variable defi-

nitions, and associated data. The second section deals exclusively

with statistical methodology and is composed of three subsections. The

first subsection provides a description of an error transformation pro-

cedure utilized to allow comparison of linear and logarithmic models.

The second subsection provides a comparison of ordinary least squares

methodology and time varying parameter methodology with emphasis on the

time varying parameter or variable coefficients method. The third and

final subsection provides a discussion of the Kendall correlation co-

efficient and test used to indicate the association between selected

measures of marketing order activity and the degree of permanent and

transitory, or random, variation observed in estimated demand parameters.


Model Definition


The simple demand model developed in Chapter 3 may be stated as

FP = f(FQ, SUB, PCE, RMQ, RTEXPT), where FP is a measure of fresh price,

FQ is a measure of fresh quantity, SUB a measure of substitutes, PCE a

measure of income, RMQ a measure of regulated quantity, and RTEXPT a

measure of exports. Relevant temporal definitions, variable definitions








and associated data sources are provided in this section. Variable

definitions are found in Tables 4.1 and 4.2. A listing of the data

values is provided in Appendix B.


Time Reference

The appropriate time references are similar in that observations

for 19 seasons, the first ending in 1961 and the last ending in 1979,

are utilized in the estimation process for each commodity. Seasonal

definitions vary, however, among commodities. The seasonal definition

for citrus commodities is generally from October through September.

Potatoes considered in this analysis are fall potatoes in all but one

case. Virginia-North Carolina potatoes are spring and summer potatoes.

The spring season is considered for Texas onions and the summer season

for Idaho-Oregon onions. Both tomato orders, Texas and Florida, apply

to fall, winter, and spring production. Similarly, Florida celery,

green peppers, and sweet corn are fall, winter, and spring commodities.

The relevant season for Florida cabbage is winter and spring and the

winter season for Texas lettuce.


Variable Definitions--Demand

Summary definitions of demand variables are found in Table 4.1.

Fresh price is real average seasonal price for fresh citrus and vege-

tables. Although FP is used to designate the price variable for potatoes,

the price in this case is actually an average price for total production,

including both fresh and processed utilization. All prices are deflated

by the consumer price index (1967 = 100). The unit of measurement for

potato and vegetable prices is dollars per hundredweight. Citrus prices

are on a dollar per container basis and are on-tree equivalent prices.








Table 4.1.--Demand Variables.

Symbol Definition

FPa Fresh price--Season average farm level price of order area pro-
duction for fresh utilization deflated by the Consumer Price
Index (1967=100) in dollars per hundredweight for potatoes and
vegetables and dollars per box for citrus.

FQ Fresh quantity--Order area production for fresh utilization in
pounds per capital.

PQ Processed quantity--Order area production for processing in
pounds per capital.

TQ Total quantity--Total order area production in pounds per capital.

SUB Substitutes--Generic reference to the substitutes QSF, QST, IMP,
QSFIMP, and QSTIMP.

QSF Fresh substitutes--Domestic production outside the order area
for fresh utilization in pounds per capital.

QSTc Total substitutes--Total domestic production outside the order
area in pounds per capital.

IMP Imports in pounds per capital.

QSFIMP Fresh substitutes plus imports in pounds per capital.

QSTIMP Total substitutes plus imports in pounds per capital.

PCE Income--Annual Personal Consumption Expenditures deflated by the
Personal Consumption Expenditure Deflator (1972=100) in $1000
per capital.

RTEXP Relative exports--Proportion of total fresh domestic production
exported.

RMQ Relative regulated quantity--Proportion of order area production
for fresh utilization reported as marketed in regulated outlets.

aFP refers to the season average price for total potato production, in-
cluding fresh and processed.
bBox weights are 90 Ibs. for Florida oranges, 85 Ibs. for Texas oranges
and Florida grapefruit, 80 Ibs. for Texas grapefruit and Florida limes,
76 Ibs. for California-Arizona lemons, 75 Ibs. for California-Arizona
Valencia and Navel oranges, and 64 Ibs. for California-Arizona grapefruit.
CQST is production for fresh utilization only for vegetables and fresh
plus processed for citrus fruits and potatoes. QST includes imports for
tomatoes and green peppers.








Container weights are 90 pounds for Florida oranges, 85 pounds for Texas

oranges and Florida grapefruit, 80 pounds for Texas grapefruit and Florida

limes, 75 pounds for California-Arizona Valencia and Navel oranges, 64

pounds for California-Arizona grapefruit, and 76 pounds for California-

Arizona lemons.

Fresh quantity is defined as per capital fresh seasonal utilization

specific to order areas and is measured in pounds per capital for all

commodities. Total quantity, measured in pounds per capital and specific

to order areas, is examined in the case of potatoes where a comparable

measure of fresh utilization by areas is not available. An inferred

fresh quantity estimate for potatoes is examined. This estimate of

fresh quantity is inferred from order committee reports of regulated

quantity and, like other measures of quantity, is on a pounds per capital

basis.

Substitutes, though defined in varying ways for specific commodi-

ties, are generally fresh, and in some cases processed, quantities of

the same product available domestically from areas outside of the order

area. In addition, order area processed quantity of citrus and the

inferred order area processed quantity of potatoes are examined. The

only traditional substitutes examined are lemons for limes and limes for

lemons. Substitutes, like fresh quantity, are measured as pounds per

capital.

Six variables defineable as substitutes are examined. Processed

quantity (PQ), fresh substitutes (QSF), total substitutes (QST), imports

(IMP), fresh plus imports (QSFIMP), and total plus imports (QSTIMP) are

the variables falling into this category. All variables in this category

are not relevant to each commodity. Most are relevant to citrus only.








Estimated demand equations for vegetables include only fresh substitutes,

which include imports in the case of tomatoes and green peppers. The

inferred processed quantity is examined for potatoes, but fresh substi-

tutes are not. Total substitutes (QST), i.e., fresh and processed

utilization of production outside of the respective order area, is,

however, examined for potatoes.

Processed quantity (PQ) is order area production utilized for pro-

cessing. This variable is included in each of the demand equations for

citrus and is examined for potatoes. Fresh substitutes (QSF) are fresh

utilization of domestic citrus and potato production outside of the

respective order areas. In the case of limes, QSF is fresh lemon utili-

zation and in the case of lemons it is fresh lime utilization. Total

substitutes (QST) are combined fresh and processed utilization of

domestic citrus and potato production outside of the respective order

areas and fresh utilization of domestic vegetable production outside of

the respective order area. In the case of tomatoes and green peppers,

QST includes imports. Imports (IMP) are considered separately only for

lemons and limes. Fresh substitutes plus imports (QSFIMP) and total

substitutes plus imports (QSTIMP) are considered for oranges and

grapefruits.

The variety of substitute definitions for citrus, i.e., QSF, QST,

QSFIMP, and QSTIMP, were examined separately in the process utilized to

determine the preferred demand equation. The order area processed

quantity of citrus was considered of sufficient theoretical importance

to preclude its exclusion. Thus processed quantity and one of the

substitute definitions above was included in each demand equation for

citrus.








The measurements of regulated quantity (RMQ) and exports (RTEXP) are

ratio measures. Regulated quantity is defined as the proportion of order

area fresh quantity (total quantity for potatoes) reported as regulated.

The export measurement is relevant for citrus only and is defined as

the proportion of total fresh production of specific commodities exported.

The income measure (PCE) is personal consumption expenditures de-

flated by the personal consumption expenditures price deflator (1972=100)

and is in $1000 per capital. This measure is utilized rather than

personal disposable income or other income measures under the premise

that income allocated to personal consumption is more relevant to expendi-

tures on food commodities than the more gross allocation associated with

personal disposable income and other income measures.


Variable Definitions--Marketing Orders

A single measure of marketing order activity or intensity, discussed

in Chapter 3, is not objectively defineable. Thus, several measures,

none of which may be categorized as "the" measure, are examined.

Thirteen measures are considered. Summary definitions are provided in

Table 4.2. Four may be categorized as general measures. The remaining

nine are regulation specific. The general measures include years of

operations (NYRS), coverage (COVRG), expenditures relative to order area

sales (EXPM), and expenditures relative to total sales (EXPT). Regula-

tion specific measures include grade (G), size (S), shipping holiday

(SH), container (C), pack (P), weekly volume (WV), maturity (M),

packing holiday (PH), and producer allotment (PA) regulations. All

order measures utilized, with the exception of NYRS, are simple

averages over seasons of the seasonal measures defined below. Years of














Table 4.2.--Marketing Order Variables.


Symbol


NYRS Years of operation as of 1980.


COVRG Coverage--Average ratio of order area production for fresh
utilization to the selected substitute quantity (SUB) plus
order area production for fresh utilization (FQ), i.e.,
COVRG=FQ/(FQ+SUB) where SUB is either QSF, QST, QSFIMP, or
QSTIMP plus PQ for citrus. SUB, . ., PQ are defined in
Table 4.1.

EXPM Average annual expenditure of Marketing Order Committee as a
percent of order area sales 1960-61 through 1978-79.

EXPT Average annual expenditure of Marketing Order Committee as a
percent of total sales 1960-61 through 1978-79.


Grade regulation
Size regulation
Shipping holiday
Container regulation
Pack regulation
Weekly volume regulation
Maturity regulation
Packing holiday
Producer allotment


The variables G, . ., PA are
interpreted as the proportion of
seasons during which these re-
spective regulations were utilized
during the 19 seasons of 1960-61
through 1978-79.


Definition








operation (NYRS) is defined as of 1980, e.g., the current Texas lettuce

order was approved in 1960 and therefore NYRS is 20. A complete listing

of data is found in Appendix B.

The general measures, other than NYRS, are all relative and continuous

values, whereas the regulation specific measures for any year are discrete,

taking a value of 1 if the regulation is utilized and 0 otherwise. The

average values for regulation specific measures specify the proportion of

seasons that the specific regulation was utilized during the 19 seasons

examined. Coverage (COVRG) is a measure of the potential degree to which

apparent price influencing quantities may be regulated by an order. The

operational definition is FQ/(FQ + SUB), where FQ and SUB are as pre-

viously defined. The process by which SUB, the appropriate substitute,

is selected is described in the statistical methods section of this

chapter. Expenditures relative to order area sales (EXPM) are defined as

the ratio of order committee expenditures to order area sales, where

order area sales are defined as the order area price times reported regu-

lated quantity. Expenditures relative to total sales (EXPT) are defined

as the ratio of expenditures to total sales, where total sales are de-

fined as the order area price times all fresh quantities domestically

produced plus imports where appropriate.


Variable Definitions--Randomness

The final definitional unit is the measurement of randomness in

parameter estimates. Demand parameters are estimated under ordinary

least squares (OLS) and time varying parameter (TVP) methods, discussed

in the statistical methods section of this chapter. A basic assumption

of OLS is that parameters are constant, i.e., not random. Conversely,

parameters are assumed random with a potential permanent component under








TVP. The degree of permanent change and random change is measured by

gamma, a statistic of TVP. Gamma is discussed in a later section of

this chapter. Completely random parameter variation may be indicated

under TVP. In this circumstance a measure of randomness other than

gamma is appropriate.

Two measures of randomness are examined. The first measure is the

deviation of TVP parameter estimates from OLS parameter estimates. The

second measure is the deviation of parameter t-values under TVP from

parameter t-values under OLS. Given the assumption of constant

parameters under OLS it may be argued that similar TVP and OLS parameter

estimates indicate more nearly constant parameters and dissimilar esti-

mates indicate more random parameters. The two measures are examined

in recognition of the weakness inherent in the parameter test and an in-

ability to rigorously demonstrate the appropriateness of the t-value

test. The weakness inherent in the parameter test is that equality of

the OLS parameter estimate and the TVP parameter estimate does not insure

the existence of a nonrandom error structure. While the relationship

between t-values under OLS and TVP may be mathematically calculated,

the complexity of the relationship precludes rigorous specification of

how these values behave as the error structure becomes more random. A

monte carlo-like experiment was performed in order to address the question

of t-value comparisons. T-values under OLS and TVP were calculated for

specified degrees of randomness while holding gamma constant at zero,

i.e., assuming no permanent parameter variation. In 119 out of 120 cases

the OLS and TVP t-values consistently diverged as randomness increased.

The single exception clearly suggests that further work in this area is

needed. However, the monte carol results provide sufficient evidence

in support of the t-value comparison for this analysis.








Data Sources

The data base utilized in this study was developed from six general

U.S. Department of Agriculture sources and direct agency cooperation.

Annual and historical summaries of Citrus Fruits [U.S. Department of Agri-

culture 1980b, 1980c, 1976b, 1972a, 1967a], Potatoes and Sweet Potatoes

[U.S. Department of Agriculture 1980d, 1980e, 1977c, 1972b, 1967b], Vege-

tables [U.S. Department of Agriculture 1980f, 1981, 1977d, 1972c, 1967c]

provided the domestic quantity measures and prices for potatoes and vege-

tables. Fresh prices for citrus were taken from selected monthly issues

of Agricultural Prices [U.S. Department of Agriculture 1961--1980] and

citrus exports and imports were taken from selected annual issues of Agri-

cultural Statistics [U.S. Department of Agriculture 1980a, 1977a]. Per-

sonal consumption expenditures, population, the consumer price index,

and the personal consumption expenditure deflator, all initially reported

by the United States Department of Commerce, were taken from Working Data

for Demand Analysis [U.S. Department of Agriculture, 1980g]. Information

specific to marketing order activities was provided by the Agricultural

Marketing Service, United States Department of Agriculture.


Statistical Methods


As indicated in Chapter 3, three questions are quantitatively ad-

dressed. The first is model specification, including functional form and

selected variable definitions. The second is identification of permanent

and transitory, or random, parameter variation. The third is the associ-

ation between parameter variation and selected measures of marketing

order activity. Each question is uniquely associated with certain sta-

tistical procedures. Certain aspects of model specification are based

on the minimum mean squared error among selected OLS models. One such








aspect is functional form. Direct comparison of mean squared errors from

linear and logarithmic specifications is inappropriate. However, a trans-

formation suggested by Rao and Miller allows this comparison. The iden-

tification of permanent and transitory parameter variation is accomplished

through application of the time varying parameter method. Finally, the

association between parameter variation and selected measures of marketing

order activity is addressed through the Kendall correlation coefficient.


Error Transformation

Several alternative definitions of substitutes, the empirical

appropriateness of the proportion regulated (RMQ) and the proportion of

oranges and grapefruit exported (RTEXP), and two alternative functional

forms, linear and logarithmic, are examined. The alternative variables

and functional forms are evaluated in a manner akin to "all possible

regressions." This procedure results in a pretest estimator so the

traditional test of significance has no unambiguous interpretation.

Therefore, tests of significance are not presented.

The criterion of minimum mean squared error requires that either

the dependent variable is the same in each equation or that alternative

dependent variable definitions are such that the variance does not

change with changes in the unit of measurement. Consider the alternative

definitions of Y and log Y. The variance of log Y is unaffected by a

change in the units of measurement but the variance of Y is affected.

Thus, the mean squared error of Y may be increased or decreased by the

selection of appropriate units of measurement and comparison is not

meaningful.

Standardization of Y such that the variance is not related to the

unit of measurement will allow comparison of the resulting mean squared








errors. A transformation which results in the same "Jacobian" of trans-
*
formation for Y and log Y where Y is the standardized variable, will

yield a unit of measurement neutral variance [Rao and Miller, p. 108].

A transformation subsequently suggested by Rao and Miller is

utilized in this analysis. The transformation is defined as

Yt = c Yt
where
-z log Yt
c = T the inverse of the geometric mean of Y.

*
Direct estimation of Y and log Y is not necessary. The mean
2
squared error (MSE) associated with Y is c MSE(Y) and as indicated

previously the MSE(log Y) is not changed by this type of transformation,

i.e., MSE(log Y) = MSE(log Y). Thus, it is necessary only to calculate

c, the inverse of the geometric mean of price, and apply the transforma-

tion MSE(Y*) = c2 MSE(Y), which is then comparable with the MSE(log

Y ) = MSE(log Y).


Time Varying Parameters

Nonconstant parameters may arise from a number of sources. These

sources fall into three general categories of (1) structural changes,

(2) misspecification, and (3) aggregation. Marketing order activities

may result in parameter variation related to all three categories;

however, misspecification and aggregation are the primary sources

suggested in this study. While a marketing order constitutes a structural

change, the individual commodities examined in this study have for the

most part been either subject to marketing order activities throughout

the observation period or not subject to marketing order activities

throughout the period. Thus, while other structural changes may have








occurred, the introduction of a marketing order has not been a major

structural change for most of the commodities examined.

Misspecification includes a number of aspects. The selection of

relevant variables and appropriate functional form are the most common

aspects. In this analysis the examination of linear and logarithmic

models and various substitute definitions is an attempt to improve

specification. However, the exclusion of marketing order activity

measures potentially results in misspecification.

Aggregation across micro units will yield stable macro units only

if the micro units remain stable. Thus, if the micro units are not

stable over time the aggregate unit will not be stable. This potential

instability for fresh fruits and vegetables is manifested in the vari-

able quality from one season to the next. One would expect consumers

to respond differently to the same quantity when the implicit quality

differs. Thus, while a simple summation of quantity in terms of weights

or units provides a stable response to the aggregate when the average

product quality is constant, the response to the aggregate is not

stable in the presence of interseasonal variations in average quality.

If a marketing order contributes to a more stable product, i.e., micro

unit, then the resulting response to changes in the aggregate, i.e.,

total quantity, should be more stable. This suggests that selected

marketing order activities potentially contribute to more stable demand

parameters.

Estimation techniques for dealing with nonconstant parameters may

be broadly categorized as ". . (1) Random coefficients models, (2)

Systematic (non-random) variation models, and (3) Kalman-filter models"

[Belsey and Kuh, p. 375]. The technique utilized in this analysis is








a member of the Random-coefficients models group. A synopsis of the

two latter categories, a discussion of the random coefficients models

category, and a descriptive discussion of the technique utilized in this

analysis follows.

Systematic (non-random) variation models are characterized by B

variation dependent on identifiable factors. The B variation may be

of a discrete nature, for example varying by time sets, or cross-

sectional groupings [see Goldfeld and Quandt, and Maddala, pp. 394--396]

or B may be functionally related to a set of variables [see Belsey, and

Maddala, pp. 390--392]. This latter argument of B being a function of

a given set of variables is appropriate in the analysis of the effect

of measurable policy variables which enter an optimizing decision problem

[Maddala, p. 403] and therefore potentially applicable to the problem

addressed in this treatise. However, the relative stability of order

activities for a given commodity and a limited number of observations

preclude its application.

Kalman filter models have their origin in the engineering field of

optimal control theory. The essential characteristic of the Kalman

filter model is the existence of a transition matrix T which with an

associated error define the stochastic process associated with the

random coefficients. While the technique has obvious econometric appeal,

the transition matrix T and associated error structure is much more

difficult to specify for econometric problems as compared to engineering

problems [Cooper]. A discussion of this model is provided by Cooper.

Random coefficients models are characterized by the definition of

some or all of the parameters as random variates. Maddala provides a

summary of several models in this category, a model considered by








Hildreth and Houck, the adaptive regression model and varying parameter

regression model [Cooley and Prescott 1973], stochastically convergent

parameter models [Rosenberg], and the pure random coefficient models.

The pure random coefficient models are unique among the group summarized

due to the absence of any systematic variation in the parameters. The

most general of these models is the varying parameter regression model

considered by Cooley and Prescott, which allows systematic and/or non-

systematic variations as well. In the absence of systematic variation

the varying parameter model reduces to the pure random coefficient model

[Cooley and Prescott 1973, p. 467].

Selected applications of the TVP method included Cooley and DeCanio,

Ward and Myers, Ward and Tilley, and Shonkwiler. Additionally, Cooley

and Prescott [1973] provide examples of applicable problems and ongoing

research at that time. Cooley and DeCanio compare TVP estimated vari-

ation in supply response parameters with variations implied by a rational

expectations model. Ward and Myers address the question of advertising

and associated parameter variation for frozen concentrated orange juice.

Ward and Tilley examine the differences in parameter variation among

the three major orange juice products, frozen concentrate, chilled, and

canned single strength juice. Shonkwiler, while focusing on TVP estimates

for a recursive system and consequent updating by a Kalman filter method,

provides an empirical application to the problem of forecasting cattle

prices. Each of the above references provides an adequate exposition

on model definition. However, Cooley and Prescott [1976] provide a more

complete theoretical exposition including estimation methods and the

derivation of asymptotic properties.








The rudimentary elements of the TVP model are

(1) Yt = xt Bt, where

(2) Bt = 5p + ut, and

(3) +P = P1 t+

The permanent component, $, is subject to transitory, or random, vari-

ation indicated by the error ut. The time path of the permanent component

follows a moving average process with error vt. The model is rather

general, reducing to the pure random coefficient model if equation (3)

is excluded. If in addition the randomness is limited to the intercept

the model reduces to the conventional ordinary least squares model.

The error components ut and vt are identically and independently

distributed normal variates with mean 0 and covariance structures known

up to different scale factors. Specifically, the covariance structure is

(4) COV(ut) = (1-y)a2 u, and

(5) COV(vt) = yo2v.

Since u and z are known up to a scale factor an element of each may

be normalized to 1. If this normalization is with respect to the elements

of z and v corresponding to the intercept, i.e.,al = o~l = 1, then

the transitory change associated with the intercept is equivalent to the

additive disturbance term of the ordinary least squares model.

The value y (gamma) provides a measure of the degree of permanent

change from one period to another. As indicated previously, if gamma

is 0 the system reduces to the random coefficient model and although

the permanent component 8P is stationary it is subject to random change

about its mean value. This is in contrast to the ordinary least squares

model where the B are nonrandom. If gamma is greater than 0 then the








permanent component changes from one period to the next independent

of any central tendency among periods.

The non-stationary nature of the parameter generating process pre-

cludes specification of the likelihood function over all observations.

However, the likelihood function conditional upon the value of the

generating process at a specific point in time may be specified. If

the parameter set to be estimated is B P, where T is the number of
T+l, where T is the number of
observations available, then
T+1
(6) P v v P v
+(6) = + T+1 = + E v, and
s=t+1
T+1
(7) B= vs +
s=t+l

Given equation (7),the basic relationship of equation (1) may be re-

written as
T+1
(8) yt = xtB + xut x' E vs, or
s=t+l

(9) yt = xtfB, + wt' where
T+1
(10) wt = xtut + x Z vs
s=t+l
It can be shown that

(11) COV(w) = o2[(I-y)R+yQ] = a2 ().

The matrix R is diagonal with

(12) rii = (x- Eu xi).

The matrix Q is composed of the elements

qij = min(T-i+l, T-j+1) x xx.

If the parameters for some prior period, say z, are to be estimated the

elements of Q are modified such that







min (Iz-il, Iz-jl) xiv xj, if (z-i)(z-j) > 0
(13) qij =
0, if (z-i)(z-j) < 0

This modification is seen to be appropriate upon recognition of the
relationship
z
B = B v, if t < z
(.14) s=t+l
t
t = SB + E vs, if t > z.
s=z+l

This allows observation of the dynamic path which the parameters have

followed. The path may verify a priori expectations and/or suggest a

respecification of the model.

Given the specification of Q (y) the log likelihood function is

(15) L(Y, B, o2 X 2
(15) L(Y, 2, y, X) = In 2n In o2 In


(Y-Xe)- n(y)-l(Y-XB).
2

The log likelihood function may be partially maximized with respect to
B and a2 to yield the following estimators conditional on y.

(16) B(y) = [X'a(y)-1]-1 X -(y)-1 Y

(17) s2(y) = [(Y-Xe(y))- S(y)-1 (Y-XB(y))].

Substitution of (16) and (17) into (15) yields the concentrated likelihood

function conditional on y. Since y lies between 0 and 1 an interactive
process may be utilized to estimate the value of y which maximizes the
concentrated likelihood function. This value of y is consequently uti-

lized to estimate B and 02 from equations (16) and (17). A computa-

tionally convenient method is presented in Cooley and Prescott [1976]








and is incorporated in the SAS program utilized in this analysis. A copy

of the program is provided in Appendix C.

Utilization of this technique requires knowledge of tu and v up

to a scale factor. In the absence of any information to indicate other-

wise zu and v are assumed equal. Similarly, since there is no reason to

believe that random changes in parameters are correlated, zu and Ev are

assumed to be diagonal, i.e.,

1 0.. 0
0 022 . 0

u = vE


0 . k .
Given this configuration it is necessary only to specify the relative

variability of the parameters. The estimated parameter variances re-

sulting from the OLS estimates are utilized to specify the relative

variability and are normalized with respect to the estimated variance of

the OLS intercept, i.e.,

u v -OLS -OLS u v -OLS -OLS u v OLS -OLS
011 =11 =1 al1 022 = 022 = 2 1 . .' kk = akk Ok 1l

Cooley and DeCanio indicate that the TVP results are relatively

robust with respect to the specification of zu and Ev. However, Madalla

[p. 398] suggests that the alternatives considered may limit the degree

to which robustness may be inferred.

The asymptotic properties of this estimator are presented in Cooley

and Prescott [1976]. A procedure to test the difference between OLS

estimates and TVP estimates has not been developed [Ward and Myers, p.

9]. Similarly, a test for significant differences between interperiod

parameter estimates does not currently exist.








Kendall Correlations

Given the existence of permanent parameter variation indicated by

y > 0 or the inferred existence of random parameters given y < .98 and

BTVP f BOLS, the question of association between marketing order activity
and permanent and/or random parameter variations arises. A nonparametric

test for independence based on the Kendall correlation coefficient [see

Hollander and Wolfe, Chpt. 8] is used to address this question.

The Kendall correlation coefficient is based on a bivariate sample

of size n. In this study two types of bivariate samples are considered.

The first is marketing order activity and gamma. The second is marketing

order activity and the absolute proportional deviation of the TVP

parameter estimate or t-value from the OLS parameter estimate or t-value.

The proportional deviation provides a measure of the degree of randomness

and is relevant only in the case of y=O.

The Kendall correlation coefficient T is defined as

T = 2 P[(X1-X2)(Y-Y2) > 0] 1.

In this study X. corresponds to a selected measure of marketing order

activity and Yi corresponds to either the gamma value or proportional

parameter or t-value deviation for commodity i. If X and Y are indepen-

dent then T = 0. The null hypothesis H : T = 0 may be tested by the K

statistic which is defined as
n-l n
K = z E h(Xi, Xj, Yi Yj), where
i=l j=i+l

1 if (Xi-Xj)(Yi-Yj) > 0,

h(Xi,Xj,YiYj) 0 if (X.-Xj)(Yi-Yj) = 0,

-1 if (X.-Xj)(Yi-Yj) < 0.








An estimator for the Kendall correlation coefficient is given by

S 2K


The Kendall correlation coefficient is readily interpretable.

Consider,

T 0- P(X1-X2)(Y-Y2) > 0] > 1/2.

Thus if r is greater than zero then the probability that an increase

(decrease) in X is accompanied by an increase (decrease) in Y is greater

than 50 percent. Similarly, if T is less than zero the probability

that a change in X is accompanied by an opposite change in Y is greater

than 50 percent. Hence, by using the Kendall statistic one can test

the association of marketing order activity with demand adjustments,

whether permanent or transitory.

The multidimensional nature of order activities, the relative sta-

bility of order activities for a given commodity, and the potentially

lengthy time period necessary for order activities to be reflected in

demand preclude direct incorporation of order activities in the demand

model. The Kendall correlation coefficient and associated test provide

a partial solution to this problem. Although the association cannot

be quantified to the same degree, the Kendall procedure does allow

determination of the existence and direction of association.

The data and empirical procedures described in this chapter are

utilized to develop the results presented in the two following chapters.

Chapter 5 contains the estimated demand equations and discussion of the

presence and degree of parameter variation. The association between

order activity and variation in the demand parameters is addressed in

Chapter 6.














CHAPTER 5


VARYING PARAMETER DEMAND ESTIMATES FOR
SELECTED ORDER AND NONORDER COMMODITIES


Drawing from the demand theory in Chapter 3, demand equations were

first estimated by ordinary least squares (OLS) under the following

guidelines. Own quantity, substitutes, and income were deemed to be

of sufficient theoretical importance to be included in the final

equations selected. Two structural forms, linear and logarithmic,

were examined. Similarly, in the case of citrus several alternative

definitions of substitutes were examined and the influence of exports

was considered to be an empirical question. With the exception of Texas

winter lettuce, order committee reports (OCR) of fresh quantities regu-

lated are generally smaller than Crop Reporting Board (CRB) reports of

fresh market production or utilization. The influence of this informa-

tion on price was also considered to be an empirical question.

Thus, the final OLS equations selected include own quantity, a

measure of substitutes, and income. The functional form, the incorpora-

tion of information regarding order committee--Crop Reporting Board

differences, and in the case of citrus the substitute definition and the

incorporation of export information were determined on the basis of a

minimum mean squared error criterion. In order to allow the comparison

of mean squared errors from linear and logarithmic models a transforma-

tion suggested by Rao and Miller [pp. 107--111] was utilized.








After selecting the equation for each commodity which minimized the

mean squared error, given the guideline constraints, the covariance

matrix was utilized to specify the relative variability of the parameters

and consequently provide an estimate of two components of the error

structure necessary to estimate these same equations in a time varying

parameter (TVP) framework as outlined in detail in Chapter 4.

The five following categories, plus a closing overview section,

provide the organizational structure for this chapter: (1) Oranges and

Grapefruit, (2) Lemons and Limes, (3) Potatoes, (4) Marketing Order

Vegetables, and (5) Non-Marketing Order Vegetables. Variables considered

and methods of incorporating order committee and Crop Reporting Board

differences vary between categories and in some cases within categories.

Thus, in each category variables and methods are described, estimated

OLS and TVP parameters are presented, and commentary provided. The

pretest bias resulting from the minimum mean squared error selection

criterion is recognized and, although t-values are presented, parameter

estimates are not evaluated with respect to significance levels. In the

categorical discussions with respect to flexibilities, the random

parameter estimates, i.e., those where gamma is zero, are not explicitly

considered. Although differences exist between the random parameter

estimates and the constant parameter estimates, the inferred flexibilities

change only in degree and not in type, i.e., flexible versus inflexible.

If a flexible response is indicated under TVP where gamma is zero the

response under OLS is flexible also and vice versa. The difference

between the TVP and OLS parameter estimates are discussed in the closing

section of this chapter. The overall purpose of this chapter is to pro-

vide the empirical results relating to the theoretical model set forth




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs