Group Title: evolution of industrial land use with the Knoxville metropolitan region
Title: The evolution of industrial land use with the Knoxville metropolitan region
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102837/00001
 Material Information
Title: The evolution of industrial land use with the Knoxville metropolitan region
Physical Description: Book
Language: English
Creator: Honea, Robert B
Copyright Date: 1975
 Record Information
Bibliographic ID: UF00102837
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: ltuf - ADB3750
oclc - 14184163

Full Text












THE EVOLUTION OF INDUSTRIAL LAND USE WITHIN
THE KNOXVILLE METROPOLITAN REGION: AN ANALYSIS
USING AERIAL PHOTOGRAPHY AND HISTORICAL DATA
FOR THE PURPOSE OF LAND USE MODELING







By




ROBERT B. HONEA


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
1975



































DEDICATED TO:

Dr. (Jim) James C. Wilkinson










AC KNOWLEDGEMENTS


Of the many individuals who have contributed to this study,

foremost is Dr. Clark I. Cross, Chairman of my Supervisory Conurttee.

I am appreciative for his professional and personal assistance. His

periodic encouragement has been most opportune. Dr. Gary Shannon has

contributed both to this study and at other stages of my graduate

career. I am grateful for his help. Dr. James R. Anderson has been a

source of encouragement before and during my graduate tenure. To him

I extend my appreciation. To Dr. David L. Niddrie 1 extend my appre-

ciation for many engaging discussions. His intellectual vigor is a

source of professional inspiration. The lectures of Dr. Clayton

Curti. provided many ideas which arc :c uded in tniS study. To him

my sincere appreciation is extended. Gratitude is also extended to

Dr. Sh3nnon McCune, Chairman of 'he De'partment of Geography and other

menibers of the faculty for their assistance.



who contributed to the research effort: Mr. C. T. N. (Ted) Paludan at

Marshall Space Flight Center; Dr. C. W. (Pete) Craven, Mr. Richard C.

DurCee, anid Dr. Michael M. McCarthy at Oak' Ridge National La:boratory;

rand > i~mcT.'ily ,.'ho patiently occupied tlhe remainder -f ::y huro while i.

hibern ated in my office.










TABLE OF CONTENTS


Page


ACKNOWLEGEMENTS . . . . .

LIS" OF TABLES . . . . . .

LIST OF ILLUSTRATIONS. . . . .

ABSTRACT . . . . . .

CHAPTER

I. INTRODUCTION . . . . .


- . i-i


. . . . . . . vii.

. . . . . . ix


Problem Background ..
Statement of Problem . .
Problem Operationalization
Model Operation . . .
Motivation for Study . .
Organization of Study. .


II. THEORETICAL AND EMPIRICAL ST'IDIES RELEVANT 10 'THE
DEVELOPMENT OF INDUSTRIAL SITE-SELECTION ALGORITHM . .

Selected Theoretical Works . . .

'er IsolerLe Stat . . . . ..
lUber den Standort der Iridustr -cn .. . .....
The Location of Economiic Rehavior. ..
Imperfect Competition Ir.dI the Divlnonl., neb,) pr ..
'ihe Theories of David M. Smith aiu
Melvin L. Greenhut . . . . . .....

Empirical Studies . . . .

The Studies of Allen Prod and Richard Londale .
Historical Studies of Industrial Location Factors.

Land Use Modeling Studies. . . . . . . . .
4A icrtative List of Variables. ..

III. DESCRIPTION UF THE STUDY AREA AND ANALYSIS PROCEDURES. .

The Study Region . . . . . . .....

Industry within the Region . . . . . ..

Food products industries (SIC 20) . ..
Tobacco products industries (SIC 21) . .
Textile mill product-; industries (SIC 22).


19
22

5 32
35


39
S 39

. '1


. . . . . .
S. . . . . .









Page

Clothing and related products (SIC 23). .. . 58
Lumber and wood products except furniture
(SIC 24) .. . .. . . . . . 58
Furniture and fixtures industries (SIC 25) . 64
Paper and pulp products industries (SIC 26) . 68
Printing and publishing industries (SIC 27) . 68
Chemical products industries (SIC 28). . . . 6
Petroleum refining and paving and roofing
products (SIC 29). . . . . . . . 74
Rubber and plastic products industries (SIC 30). 74
Leather products industries (SIC 31) ..... 74
Stone, clay, and gland; products industries
(SIC 32) . . . . . . . . .. . 81
Primary metal industries (SIC 33). . . . . 81
Fabricated metal prodLcts industries (SIC 34). . 81
Machinery (except electrical) industries
(SIC 35) . .... . . . . . .. ... . 87
Electrical machinery, equipment, and supplies
industries (SIC 36). ... ... . . . 87
Transportation equipment industries (SIC 37) .. 94
Instrument and related products industries
(SIC 38) .. . ...... . ...... 94
Miscellaneous industries (SIC 39). ... .... 94

Data Collection. . . . . . ... . . . . 101
Statistical Procedures . . . . . . . 106

IV. EMPIRICAL ANALYSIS . . . . . . . . . . ll

Descriptive Statistic' . .. . .. . I.1

Description of Sample. . . . .. ... .. .. 113
Description of Variable Measurements .. ...... . 112

Results of the Factor Analysis .. .. .. ... 148

V. SUMMARY AND CONCLUSIONS . . . . . . .. 158

Summary. . ... . . . . . . . . . 158
Conclusions. .. . . . . . . . . . 159

Conclusions Regarding the Form of the Site-Selection
Algorithm . . . . . . ... . . .. 159
Conclusions Regarding the Proper Form of Model
Operation . . . . . . . . . . 160
Conclusions Regarding the Use of Aerial Photography
in Model Design. . . . . . . . . . 160
Conclusions Regarding Recommendations for Future
Research . . .. . . . . 161







vi


Page

RESEARCH BI LIOGRAPHY. . . . . . . . . 163

Periodicals ........ ................ 163
Books . . . . . . . . . . . . . . 168
Government Doci..,ients and Agency and Institute Research
Reports. . . . . . . . . . ..... 172
Unpublished Sources. . . . . . . . .... .177

APPENDICES

A. Sample Survey Forms, . . ... .. . . . .. . 179
B. Map Overlays ........ .. ............. .182

BIOGRAPHICAL SKETCH . . . . . . ..... .. 184










LIST OF TABLES


Table

1. Factors Most Influential in the Location Decisions
of Florida Industries, 1956-1957 . . . . . .

2. List of Possible Factors Influencing Industry Location
as Utilized in the Carrier and Schriver Survey . .

3. Ten Factors Most Frequently Mentioned by Tennessee
Firms as Affecting Location Decisions . . . .

4. Ten Location Factors with Highest Mean Number Points
Assigned by Tennessee Firms Interviewed. . . . .


Page


. 24


. 26


. 27


. 28


Location Factors Suggested by Bullington . .

Tentative List of Variables . . ...

Expansion of ETDD Industries Between 1952-1973

Distribution of Sample by SIC Category . .


9. Variable -

i0. Variable -

L1. Variable -

12. Variable -

!3. Variable -

14. Variable -

15. Variable -

16, Variable -
Half Mi.lcs

i7. Variable -

[8. Variable -

[9. Variable -

20. Variable -

21. Variable -


Slope of Land . . . . . . . . .

Drainage . . . . . . . . .

Clearing-Cover Condition . . . . .

Distance to Center of Town . . . . .

Distance to Major Throughfare . . . . .

Density of Land Use in Immnediate Vicinity . .

Rating of Price of Land . . . . . .

Proportion of Urban Area Within Two and a


Distance to Major Highway . . . . . .

Distance to Secondary Road . . . . .

Distance co Railway . . . . . . .

Distance to Waterway . . . ......

Distance to Airport . . . . . . .


31

37

53

313

114

115

117

118

119

120

122


123

12b

126

127

128

130







viii


Tab e Page

22. Variable Distance to Interstate. . . . . . . 131

23. Variable Oveiall Quality of Accessibility-Then . . 133

24. Variable Overall Quality of Accessibility-Now. ... . 134

25. Variable City Water Availability . . . . ... 135

26. Variable City Sewage Availability. . . . . ... 135

27. Variable Gas Availability. . . . . .. . . 136

28. Variable -- Did Community Have Zoning Then? . ... . .136

29. Variable Was Site Zoned for Industry?. . . . .. 137

30. Variable Was Industry Already in Area?. . . . 137

31. Variable Overall Rating of r'oltigous Land
Use Compatability .... . . . . . ... 159

.'. Vari.able Condition of NeighL ohood . ... . . 140

33 Yariable Density of Land Use . ... .. .. .. ... 141

34. Variable Nearby Community Service. . . . . . .143

55. Variable Was the Site in an Industrial Park? ... .. 144

6. c.'" ibje Was was o Qu .ality of the i'a-?. . . . 1J,4

37. Variable Proximity to Knoxville. . . . . . ... 46

33. Variable Amount of Other Industry Located Nearby . 147

39. Correlation Matrix ... .... .. . . . 149

A(j. Eigen Values ..... . . . .. . . .. 152

41, Unrotated Factor Matrix. .. . . .. .. 154

42. Varimax Rotated Factor Matrix. . . . . . . .. 155











LIST OF FIGURES


Figure

1.
i

2

3.

4.

5.

6.

7.

8.

9.

'10.

11.

12.

13.




15.
7l.






17.

18.

19.

20.

21.

22.

23.

24.


Food Products (Plant It!:.ti

- Food Products (Employment)

- Textile Products (Plant Loc

- Textile Products (Employmen


Urban Land-Use Location Scheme. .

Smith's Model with Comparisons. .

East Tennessee Development District

Physiography . . . . .

Towns and Communities . . .

Plant Location 1943 . .

Employment 1943 . . . . .

Plant Location 1953 . ....

Employment 1953 . . . ....

Plant Locations 1963 .. .

Employment 1963 . . .

Plant Locations 1973. . ..

Employment 1973 . . . . .


Apparel and TRelared 'roduci. (Plant L.oc:ti

- Apparel and Related Products (Employricent).

- Lumber and Wood Products (Plant Locacions)

-Lumber and Wood Products (Employment). .

- Furniture and Fixtures (Plant Locations) .

- Furniture and Fixtures (Employment). . .

- Paper and Pulp Products (Plant Locations).


ons)


61

62

63

65

66

67

69


Page

. . . . . 11

. . . . 17

. . . 40

. . . . . 42

. . . . . 43

. . . . . 45

. . . 46

S . . . . . 47

. . . . . 48




. . . 50

. . . . . 51

. . . .. 52

on,; . . . 56

. . . . 57

nations) . . . 59

t). . . . . 60


SIC

SIC


SIC


siC
SIC

SIC

SIC

SIC

SIC

SIC









Figure Page

25. SIC 26 Paper and Pulp Products (Employnent) . . .. 70

26. SIC 27 Printing and Publishing (Plant Locations). .. 71

2-. SIC 27 Printing and Publishing (Employment) . . .. 72

28. SIC 28 Chemical Products (Plant Locations). . . ... 73

29. SIC 28 Chemical Products employmentt) . . . ... '75

30. SIC 29 Petroleum Refining and Paving and Roofing
Products (Plant Locations) ...... . 76

31. SIC 29 Petroleum Refining and Paving and Roofing
Products (Employment). . . . . . ... 77

32. SIC 30 Rubber and Plastic Products (Plant Locations) 78

33. SIC 30 R bber and Plastic Products (Fmploymet) 79

34. SIC 31 Leather Products (Plant Locations) . ... 80

35. SIC 31 -Leather Products (Employn.ent) . .. . .... 81

36. SIC 32 Stone, Clay and Glass Products (Plant
Locations) . . . . . . . . . 83

37. SIC 32 Stone, Clay and Glass Products (Employment). 84

38. STC 33 Primary cetal T.c'ustcies (Plant I.ocation) . 85

39. SIC 33 Primary Metal Industries (Employment).. . . 86

40. SIC 34 Fabricated Metal Products (Plant Locations). 88

41. SIC 34 Fabricated Metal Products (Employrment) .. .. 89

47. IC .'S MAchi nery ,Pia.nt Locations) ....... .. 90

43. SIC 35 Machinery (Employ.ent) . . . . 91

44. SIC 36 Electrical Machinery, Equipment and
Supplies (Plant Locations) . ... . ..... 92

45. SIC 36 Electrical Machinery, Equipment and
Supplies (Employment) . . . . . . 93

46. SIC 37 Transportation Equipment (Plant Locations) . 95







xi


Figure Page

47. SIC 37 Transportation Equipment (Employment). ... . 96

48. SIC 38 Instruments and Related Products (Plant
Locations) . ... . . . . . . . 97

49. SIC 38 Instruments and Related Products (Employment). 98

50. SIC 39 Miscellaneous Industries (Plant Locations) . 99

51. SIC 39 Miscellaneous Industries (Employment). ... 100

52. 1958 Stereo Image of Beta-Tok Industrial Site .... .102

53. 1974 Stereo Image of Beta-'ek Industlial Site . .103









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

THE EVOLUTION OF INDUSTRIAL LAND USE WITHIN
THE KNOXVILLE METROPOLITAN REGION: AN ANALYSIS
USING AERIAL PHOTOGRAPHY AND HISTORICAL DATA
FOR T1HE PURPOSE OF LAND USE MODELING

By

Robert B. Honea

March, 1975

Chairman: Dr. Clark I. Cross
Major Department: Geography

This research effort represents a portion of a much larger

research goal, the identification of the determinants affecting land

use change. This study focuses, however, upon only one aspect of this

problem, the conversion of land to industrial use. The approach

assumes that the determinants of Lhe land use conversion process are

found in the "market place," where land transactions among buyers and

sellers occur. Only one side of the market transaction process is

studied, however, namely that of ite purchaser's desires (in this

case, the industrial developer) in securing an ideal or suitable site.

The problem was to identify the ideal qualities, quantities or attri-

butes desired in an industrial site and to formulate a general algo-

rithmic statement to identify potential industrial sites.

Research procedures involv'-!d ihe developing o7 a .ist of var:iabl es

previously noted in the literature to be related to industrial site

selection and streamlining the list to a set suitable for statistical

testing. A sample of 157 industrLes which have located (or relocated)

in the 16-county Knoxville metropolitan region since 1950 was selected

for analyst s. Ui current a&nd 1'slio-. al a:..i l. ,,hctrn.ra ch data






xiii


were collected on the location characteristics of each industrial

site. These data were then subjected to factor analysis to determine

the interrelations of variables, to minimize the list of variables

needed to describe the industrial site-selection process, and to

determine if the preconceived ideas concerning the factors affecting

the process were valid. Seven factors accounting for 72 percent of

the variance found in the original data were identified.

The 30 variables studied did not group as previously conceived

but nevertheless the factors present a logical arrangement of vari-

ables. Four factors are almost singularly identified by accessibility

characteristics. Reexamination of the accessibility factors indicates

they could be combined, thereby reducing the number of indexes to two

or three. The results suggest that a zoned or designated place for

industry to locate, open space suitable for industrial development,

and accessibility to the site acc the three most important considera-

tions in industrial site-selectio~i decisions. Any algorithm intended

to simulate industrial land use coriversion should be constructed with

these parameters as major components.















CHAPTER I

INTRODUCTION


Problem Background


Most urban and regional planners seek to answer the question,

"What would happen if . .?" The answer is not simply determined.

"Even though equipped with the best information system, no . .

decision maker, however adept, can anticipate or visualize the many

relationships which must be considered in a comprehensive, regional

plan."1 If, however, we could specify the correct development pro-

cesses and represent them symbolically as a set of equations, the

computer can approximate through simulation, years of regional

growth in a matter of minutes. Before we can accurately simulate

regional growth, the development processes must first be defined and

understood, and by observing the limited success of previous land use

modeling efforts we become acutely aware that we do not sufficiently

understand them.2




1Charles R. Meyers, Jr., "New Tools for Regional Planning," AIA
Journal, October, 1971, p. 1.

2Models should not be judged simply on a success-failure basis
but rather in terms of the models' ability to answer questions about
real world. In the past, some models have been more successful than
others. For a review of previous modeling efforts one should consult:
Ira S. Lowry, Seven Models of Urban Development: A Structural Com-
parison (Santa Monica, California: RAND Corporation, 1967); B. W.
Mar and IV. T. Newell, Assessment of Selected RANN Environmental Modeling
Efforts. A report prepared for the Environmental Systems and Resources









We would expect land use change or growth to be stimulated or

brought about by certain "triggering" events or mechanisms. Many

early land use modeling efforts assumed the "triggering" mechanism to

be brought about by the location of new industry. We now recognize

that other stimulating mechanisms such as improvements in transpor-

tation or expanding demand for personal services also may prompt

regional growth.3 This research problem, however, is predicated upon

the assumption that industry (specifically export industry) is the

major stimulus for urban and regional growth and thus deserves initial

consideration in any modeling effort.

Brown et al. recognized "the critical importance of industrial

location. .yet land-use modelers have devoted surprisingly little

effort to analyzing the determinants of industrial location."4 Only

a few research efforts have made serious attempts to model the basic

determinants of industrial location choices and "for the most part

these attempts have been quite limited and crude."5 The focus of




Division, National Science Foundation (Seattle: University of Washing-
ton, June, 1973); and 0. Stradel and B. G. Hutchinson, Notes for a
Short Course on Practical Applications of Regional Development Models
(Waterloo, Ontario: University of Waterloo, The Transport Group,
Department of Civil Engineering, November, 1971).

3Homer Hoyt was among the first urban economists to study the
growth and development of urban land. His primary axiom was that "no
city could grow by taking in its own wash." Homer Hoyt, The Structure
and Growth of Residential Neighborhoods in American Cities (Washington,
D. C.: U. S. Government Printing Office, 1939).

4james II. Brown and others, Empirical Models of Urban Land Use:
Suggestions on Research Objectives and Organizations, Exploratory
Report 6 (New York: Columbia University Press, 1972), p. 82.

Ibid.









this study is on the conversion of land from prior uses to industrial

use.

Statement of Problem


An initial requisite of any model purporting to replicate and

predict industrial land use development is the identification of those

land parcels having the greatest probability to change to industrial

land use. It is necessary, therefore, to construct an algorithm which

calculates the suitability of some land parcel for a specific indus-

trial land use. The primary objective of this research effort is to

develop a site-selection algorithm which identifies land parcels

having the highest probability of attracting industrial land use.

More specifically this research attempts to identify and measure the

importance of the intrinsic site characteristics which appear to

control the conversion of land to industrial use within the 16-county,

Knoxville metropolitan region and to formulate a simulation algorithm

to identify potential industrial sites.


Problem Operationalization


Intrinsic site characteristics refer to the qualities possessed

by some land parcel which identify its suitability for a specific land

use. In the case of industrial land use, examples of some of the

attributes which should be examined are: proximity to rail service;

access to nearby highways; site preparation costs; availability of

city services; or whether the site is within a developed industrial

park. The importance of these attributes may be measured and subse-

quently aggregated to form a combined measure or index.









The following is a list of indexes which are often identified as

affecting industrial land use choices:

Site Preparation Costs--The cost of converting rural or open
land to manufacturing use. Matters which might be taken
into account are foundation conditions, drainage conditions,
slope of the land, and type of vegetation cover. This
factor does not include the purchase price of the land.

Market Price of Land--The per-acre cost of land to the
manufacturer who wishes to build a new plant or expand his
existing plant. This would include property taxes but not
necessarily taxes on improvements to the property.

Proximity to Suitable Labor Force--The number of suitable
workers within easy access of the site.

Transportation Accessibility--The ease with which people or
materials can be moved from the plant site to roads, rail-
roads, airports, or waterways. In this factor, the concern
is with the linkage between the plant and major transporta-
tion facilities within the immediate area.

Utilities--The kinds and quality of utilities available at
the potential plant site. Some utilities which might be
considered are water, sewer, and gas.

Compatibility with Existing Land Uses--The compatibility of
general manufacturing activity with other existing land uses
adjacent to the potential site.

Neighborhood or Community Attractiveness and Amenities--The
condition and density of dwelling units and business estab-
lishments in the immediate area of the plant site, and
proximity to hospitals, schools, parks, and churches.

Industrial Park Space--The availability of suitable buildings
or land in an industrial park. The appearance of the park
and the quality of services produced are some of the con-
siderations to be examined.

This list was compiled from both a review of literature and

preliminary study. Each factor is a composite of several variables

collectively identified as an index. Particular combinations of

variables are not unique but simply represent logical groupings based

upon the literature search. It is convenient for the purpose of this









study to utilize this list as a beginning point to be refined after

further study.

Each index posited is considered to reflect an information unit

to be evaluted in the selection of industrial sites at the intra-urban

scale. By combining these indices it is possible to develop an

aggregate measure of the suitability of a site for industrial use.

Before doing so, however, a weight should be attached to each index to

reflect its relative importance in the location decision. A general

mathematical form for the site selection algorithm follows:

N
LUS. W. I.
i=l i, ,


where LUS = Attractiveness score for land use category 2
(in this case k = industrial land use),

N = Number of Indices,
th th
W. = Index weight for the i index and the th land
use category, and

.th th 6
I. = i index for the Z land use category.

-7
Model Operation


In modeling industrial land use development attractiveness scores

are utilized in the following manner: Projected industry growth




Oak Ridge National Laboratory, Regional Environmental System
Analysis, A Research Proposal Submitted to the National Science
Foundation, Research Applied to National Needs (RANN) Feb. 1, 1973,
p. 9.

Land use models are constantly evolvi;ig and thus one can only
speak of the framework of the model in a hypothetical manner. This
rity or may not become the structure of the nodel and represents only
a tentative view as to the operating manner of the iodel.









provided via a socioeconomic model is distributed initially to various

subregions through a subregional allocation algorithm. Industrial

expansion within the subregion is distributed to available land par-

cels on the basis of attractiveness scores calculated in the above

manner. The availability of various land parcels is determined from

knowledge of the probability that specific land parcels will be

offered for consumption.

The land use model as presently conceptualized is perhaps best

characterized as a stochastic (Monte Carlo) model. Projected growth

in land use will be awarded to various land parcels on the basis of

the land use scores (LUSk). Basically, however, the model is heuris-

tic in that as the land use conversion processes are better under-

stood, analytic-deterministic procedures will be substituted for

simulation-stochastic procedures.


Motivation for Study


It has been noted that "urban spatial organization is the outcome

of a process which allocates activities to sites. In our society,




8The assumption that all land is potentially available is not a
valid premise in that not all property owners are willing to sell
property. Cadastral data and ownership characteristics necessary to
develop a site-availability algorithm are not readily available and,
therefore, very few land use models have included this consideration.
The sociopolitical modeling team at Oak Ridge National Laboratory is
currently studying the methods whereby local opinion and knowledge of
owner characteristics may be utilized to identify land parcels which
may be available for various land use activities. See: Osbin L.
Ervin and Charles R. Meyers, Jr., The Utilization of Local Opinion
on Land Use Simulation Modeling: A Delphi Approach (Oak Ridge,
Tennessee: Oak Ridge National Laboratory, ORNL-NSF Environmental
Program, 1973). Olaf Helmar, The Delphi Method for Systemizing
Judgments about the Future (Los Angeles: University of California,
April, 1966).








the process is mainly one of transactions between owners of real

estate and those who wish to rent or purchase space for homes and

businesses." The market process of "transactions between willing

buyers and willing sellers determines the spatial organization of

urban activities ." and thus should dictate the methodological

structure of land use models.10

The market place however is not perfect. Individual land specu-

lation, family or corporate property gifts, over-building by contrac-

tors, land use planners over-estimating demand, governments exercising

property rights, the tendency for land uses to remain intact and the

perpetuation of mistakes rather than corrections, all spoil the simple

modeling of market transactions. However, after investigating numerous

approaches toward modeling land use development, the market place

appears to be the most viable way to approach the simulation of urban

activities.11 The site-selection algorithm seeks only to replicate

the buyer's considerations in selecting an adequate industrial site.

Seller considerations and market perturbations are not considered at

this time.

The results of this research effort are intended to integrate

with a "holistic" environmental model to be developed later. The

motivation which stimulated this research problem, however, was the




9John P. Crecine, Computer Simulation in Urban Research (Santa
Monica, California: The RAND Corporation, 1967), p. 2.
10
Ira S. Lowry, Seven Models of Urban Development: A Structural
Comparison (Santa Monica, California: The RAND Corporation, 1967),
p. 5.
Ibid.










desire to understand the land use development process. After some

preliminary analysis and review of previous research it became appar-

ent that understanding and modeling land use change contained problems

to occupy several dissertations. The decision was made, therefore, to

focus upon one aspect of the land use evolutionary process, the develop-

ment of industrial land use, in the hope that by expanding the under-

standing of this process, rigor could be added to the eventual

development of a comprehensive land use model.


Organization of Study


The organization of this study is: the present chapter attempts

to introduce the reader to the research problem and to rationalize the

need for the study. Chapter II discusses the literature from which a

tentative list of site-selection variables was developed. In Chapter

III, the test region is described and the research methodologies and

procedures explained. Chapter IV presents the results of an empirical

study conducted to identify and determine the relative importance of

various location factors in the 16-county Knoxville region. Finally,

Chapter V summarizes the conclusions derived from the empirical study

and presents recommendations for the construction of a site-selection

algorithm and recommendations for further research.















CHAPTER II

THEORETICAL AND EMPIRICAL STUDIES RELEVANT TO
THE DEVELOPMENT OF INDUSTRIAL SITE-SELECTION ALGORITHM


The purpose of this chapter is to relate to this research effort

ideas expressed within selected theoretical and empirical studies and

to discuss the evolution of the indexes listed in the previous chapter.

Discussion of some works may be abbreviated depending upon their

relevance to industrial site-selection processes. Theoretical works

involving classical location theory are discussed first, followed by

a discussion of empirical studies of industrial location factors.

The final sections discuss approaches utilized in other land use

models and the development of a tentative list of variables.


Selected Theoretical Works


The location of industry in an intrametropolitan area is the

result of a complex interaction of variables that can best be under-

stood only through careful examination of historical events. Classi-

cal location theory, however, abstracts from reality by use of the

principle ceteris paribus where one or two variables are permitted to

vary while all other variables are held constant. Von Thtinen's and

Weber's studies represent early contributions to location theory and,

thus, deserve fir.:t consideration in this study. More recent theo-

retical works follow.









Der Isolierte Staat1

Von ThUnen's work dealt with a hypothetical agricultural land use

system whereby transportation costs and economic rent were used to

explain the location of specific types of agricultural production as

they were concentrically displaced around a market center. In a

similar manner, one might expect industrial, commercial, and residen-

tial land uses to arrange concentrically around an urban center (Figure

1). For a specific land use the utility (accessibility) derived from

any location would decline with increasing distance from the CBD as in

(a). This, of course, assumes that the CBD is the most accessible

point within the urban area. The optimal price one would pay for

utility or accessibility is illustrated in (b). Cost incurred in

obtaining utility (Y) is represented by the area (XYZA) but the profit
2
to be derived at utility (Y) is (ABCZ). The profit is similar to the

land rent of von Thinen's agricultural model. Thus, rent for various

land uses can be expressed as a function of distance from the CBD as

in (c). Based upon the comparative bid-rent capabilities of each land

use, one would expect (RR1) to represent the rent function for

residential land use, (QQ1) for wholesaling and industry land use,




Johann Heinrich von ThUnen, Der Isolierte Staat in Beziehung
Auf Landwirtschaft Und Nationalikonomies (Berlin: Schumacher-Zarchin,
1875) in K. W. Kapp and L. L. Kapp, Eds., Readings in Economics (New
York: Barnes and Noble, 1949).

2Michael E. Eliot Hurst, A Geography of Economic Behavior: An
Introduction, Belmont, California: Duxbury Press, 1972, p. 231.






Page
Missing
or
Unavailable









and (PP1) for commercial and service activities. The intersections at

Y and Z would form the boundaries between the various land uses.

Commercial activities are the highest bidders for sites and

usually occupy the most accessible places within the city. But commer-

cial land uses are also found in the industrial and residential zones.

Similarly, residential uses may also be found in other zones. For

example, a four-story apartment building may yield several times more

rent as a one-story commercial operation on the same site and thus may

out-bid competitors for the property.4

Bid-rent functions are a very important real world phenomenon

encountered in explaining industrial site selections. Industries

seeking highly accessible sites upon which to build must compete

against other bidders for those same sites. Consequently cost per

unit acre, distance from CBD and proximity to transportation facili-

ties were included as variables to be examined in this study.


Uber den Standort der Industrien5

Alfred Weber was among the first economists to pose a general

theory of industrial plant location. The optimal location for an

industrial plant was seen to be a formation of three factors:




3Ibid.

Ronald Reed Boyce, The Bases of Economic Geography: An Essay
on the Spatial Characteristics of Man's Economic Activities (Atlanta:
Holt, Rinehart and Winston, Inc., 1974), p. 264.

5Alfred Weber, Uber den Standort der Industrien (Chicago: Univer-
sity of Chicago Press, 1928).









transportation costs, labor costs, and agglomerative forces. Weber

theorized that the optimal location would be found:

1. where total transportation costs per unit of output were at

a minimum or,

2. where transportation diseconomies were offset by savings

through agglomeration factors or access to labor.

Of particular interest to this study are Weber's ideas concerning-

agglomerative factors and proximity to labor. In general, the ideas

posited by Weber are of greater importance at the regional or sub-

regional levels of industrial location than at the site level. Some

of Weber's agglomerative factors, however, relate to industrial site

characteristics and, thus, are considered in this analysis.

Weber's agglomeration factors are:

(1) The joint development of industries which promotes the

attraction of auxiliary industries and increases the efficiency of

large scale production and utilization of special technical equipment.

(2) The development and growth of specialized labor due to the

greater opportunity for work in the area.

(3) The greater accessibility to raw material suppliers who can

provide material regularly and on short notice.

(4) The reduction in overhead costs, such as gas, water, elec-

tricity, roads, and communications.6




6Robert G. Turner, "General Theories of Plant Location: A
Survey," AIDC Journal, VI (October, 1971), pp. 25-26.










The Location of Economic Behavior7

The A.\nercan economist, Edtgar Iloover initially attempted to

improve i''eber's explanation of industrial location )by including the

consideration of such variables as the size of mIarket area and insti-

tutional forces in his theoretical analysis. Yet Hoover, although

critical of Weber's analysis, ultimately proceeded along the same

lines to present his location theory primarily in terms of costs,

Hoover, however, did expand Weber's agglomerative forces to include

the importance of banks, utilities, fire and police protection,
8
climate, property tax, and lower interest rates.

Hoover also developed some generalizations in the locational

habits of light industry versus heavy industry. Because of the

necessity of "handling of large quantities of goods either coming in

from elsewhere or being shipped out, heavier types of manufacturing,

warehousing, and wholesaling prefer locations in transshipment zones

along rail or waterways." "Manufactures, wholesalers and warehouses

of less bulky goods need not be located on railroads or waterfronts at

all, since they can be served by truck." They locate more in response

to "the attractions of labor supply, cheap land, and nearness to local

suppliers or customers. As a rule, they are found interspersed with

commercial and inferior residence uses."'




Edgar M. Hoover, The Location of Economic Behavior (New York:
McGraw-Hill, 1949).

Turner, op. cit., p. 27.

Hoover, op. cit., pp. 128-129.








The site considerations of industries of the 1930's and 40's have

changed in more recent times but several basic locational rules as

expressed by Hoover and Weber remain intact. Accordingly, several

variables such as neighborhood amenities, neighborhood compatibility,

proximity to labor force, availability of utilities, adjacent land

uses, and transportation accessibility have been included in this

analysis.


Imperfect Competition and the Duopoly Debate

A number of location theorists believed pure competition was not

a suitable theoretical structure for the study of plant locations, and

sought to explain locations in terms of the competition between two

firms attempting to capture the largest share of a market area.

Fetterl0 and Hotelling11 were among the earliest to expound on
12
duopoly location theory. They were followed by Lerner and Singer,2
13 14
Smithies,13 and Chamberlin4 who expanded the original concept.




10Frank A. Fetter, "The Economic Law of Market Areas," Quarterly
Journal of Economics, XXXVIII (May, 1924), pp. 520-529.

11Harold Hotelling, "Stability in Competition," Economic Journal,
XXXIV (March, 1929), pp. 41-57.

12A. P. Lerner and H. W. Singer, "Some Notes on Duopoly and
Spatial Competition," Journal of Political Economy, XLV (April,
1937), pp. 145-186.

13Arthur Smithies, "Optimal Location in Spatial Competition," The
Journal of Political Economy, XLIX (June, 1941), pp. 423-439.

1E. H. Chamberlin, The Theory of Monopolistic Competition
(Cambridge: Harvard University Press, 1936), pp. 194-196.









Devletogloul recently commented on the economic irrationalit) of

the approach and presented arguments against the theoretical base for

such location activities. Imrpo:rtant contributions, however, are still

to be noted. Duopoly theory sugg-sts consideration should be given to

the repelling or attracting properties of industries which depend upon

local markets.6


The Theories of David M. Smithl7 and Melvin L. Greenhut18

Smith's and Greenhut's contribution to the theory of industrial

location incorporated the minimum costs approach of Weber with the

maximum profit solutions of manufacturing location posited by August

Lisch.19 Smith calls this the maximin solution. The concept devel-
20
oped is illustrated in Figure 2.2 In (a), the costs of production

are permitted to vary over space (distance) and revenue obtained

(demand) is kept constant. This is essentially the Weber solution.

The basic concept of Lbsch's model is shown in (b), where revenue is




1Nicos E. Devletoglou, "A Dissenting View of Duopoly and Spatial
Competition," Economica (May, 1965), pp. 140-160.

1Turner, op. cit., p. 31.

1 DIvid M. Smith, "A Theoretical tFramework for Geographical
Studies oE Industrial Location," Economic Geography XLII (April,
1966), pp. 95-113.
18
18Melvin L. Greenhut, Plant Location in Theory and in Practice
(Chapel Hill: University of North Carolina Press, 1956).
19
William H. Weglom and Wolfgang F. Stalper (Translators), The
Economics of Location, by August Lbsch (New York: John Wiley and
Sons, 1957).
20Smith, op. cit., p. 96.
Smith, op. cit., p. 96.









COSTS
VARYING



/ LOSS
tossI


PROFIT


(a)





o
cr
aC
0
z
4
I-
u
0


DISTANCE -


-COSTS
VARYING


DISTANCE --a b


SMITH'S MODEL WITH COMPARISONS (AFTER SMITH, p. 96)
Figure 2.


WEBER


/


DISTANCE -p-



LOSCH


REVENUE
CONSTANT









permitted to vary over space and costs are held constant. Finally, in

(c) the combined solution suggested by Smith is offered with maximum

profit occurring at A, where costs are lowest and profit is highest.

Note that maximum revenue, however, is obtained at B.21

An interesting variation of Smith's model was the introduction of

noneconomic factors, in particular, the concept of psychic income.22

This innovation permitted social, psychological, or other personal

factors to be entered into the model, hence relaxing the assumption of

economic man. Such considerations according to Smith tend to divert

the location of a plant from the ideal site to locations closer to the

owner's home, a golf course, or perhaps a parochial school.

Smith suggests that stochastic procedures may ultimately have to

be used to simulate industrial location decisions as personal factors

cannot be accounted for by rigorous mathematical reasoning.23 This

research assumes that personal considerations may be accounted for by

noting neighborhood amenities near the potential site. The importance

of housing quality, proximity to churches, hospitals, schools, or

parks and personal services availability in the immediate vicinity of

the site are examined in this study.


Empirical Studies


The following empirical studies were significant resources in the

development of a tentative list of site-selection variables. Two




21Ibid., pp. 96-97.

22Ibid., p. 108.
23David M. Smith, Industrial Location: An Economic Geographical
Analysis (New York: John Wiley and Sons, Inc., 1971), pp. 269-273.









types of studies are presented: those studies directed primarily

toward the industrial developer who seeks new industry for a commu-

nity; and those studies which attempt to anAilyze industrial location

coraiderations by way of a large sample of new industries and to cate-

gorize the locational considerations by industry types. Only one

study is directed toward the development of a site-selection algorithm

for a land use model.

24 75
The Studies of Allen Pred24 and Richard Lonsdale2

Allen Pred has compiled a study of the history and present status

of industrial location decisions within a metropolitan region. These

patterns discussed pose interesting hypotheses for empirical analysis

but the interest of this research is directed primarily to the site

characteristics discussed by Pred. It should be noted that Pred's

analysis focused upon a single metropolitan area whereas this study

encompasses a region with a hierarchy of urban places.

In discussing location patterns, Pred identifies seven types:

1. Ubiquitous industries concentrated near the CBD The market

area of these industries is generally coincident with that of the

metropolis or city. Food processing industries, specifically bakery

goods, package foods, and fresh milk products, are Lome of the examples

of these types of industries.




24Allen R. Pred, "The Intrametropolitan Location of American
Manufacturing," Annals of the Association of American Geographers, LIV
(June, 1964), pp. 165-180.

25Richard E. Lonsdale, "Rural Labor as an Attraction for Indus-
try," AIDC Journal, IV (October, 1969), pp. 11-17.









2. Centrally located "communication-economy" industries Job-

printing industries, newspaper printing, and advertising printing

would be representative industries of this category.

3. Local market industries with local raw material sources -

These industries show a high degree of randomness in their locational

pattern but with some tendency toward CBD locations. Samples would be

ice plants, concrete brick and block industries or industries whose

raw materials are by-products of other large-scale industries such as

the pulp and paper products industry.

4. Non-local market industries with high value products Typi-

cally these industries provide a high value per unit weight product

and are insensitive to transport considerations within the local

region. The pattern is "at least superficially irrational." Computer

and related industries and chemical industries are typical examples.

5. Noncentrally located "communication-economy" industries The

subset of industries includes those which are not necessarily pulled

to any functional area of the city but rather tend to cluster together

in any suitable area primarily because of the necessity to "keep

abreast of the latest innovations or forthcoming contracts." Elec-

tronic, military equipment, and space age industries such as found in

Huntsville, Alabama, or Houston, Texas, are examples.

6. Non-local market industries on the waterfront Industries

where primary raw materials are imported by water or whose finished

product is often moved by water comprise this group. Petroleum

refining, coffee roasting, and sugar refining are prominent among

these types of industries.









7. Industries oriented toward nationLral markets These indus-

tries have extensive market areas and arc influenced by high transpor-

tation costs on a bulky fieiishcd product. A large percentages o-fr those

industries have a tendency to locate d:n t h side of the metrc olis

facing the most important market regii.. Pred identifies the Noway :

Lowland refining indust-ry anid th: i)e;. -c_ utomo(i.ie industry a.s
26
oyanp les.2

Richard Lonsdale in a study oi the locationa.1 habits of rural

industry notes that those industr~i-;. a .fected by transportation coAi.

tend to locate in urban areas whiJ.; ;.:o:;e highly affected by labor

cost, gravitate to rural areas. (t addition, Lonscda.ie note rural

firms tend to space themselves out; ij.t'. .irde- to assure a labor supply

industries wi-Lh tendencies toward i:rl 0oc.-tions are apiparc

food "rodC..ti's, textit, l J. umber :,d-w' .,*/ y r \, t;s,, p"p.cr 'p ourcG" ,;

chen.cal, and electrical machinery :L,-ecially routine assembly. I.o

profit margins, keen competition, and nigh petrc tageO2 of produce in

workers. are s5ome basicc character i.s;i.-. of thest :i J:Lus ries.

The patterns of in T.ustrial location detailed by Fred and Lon:',d:l

are rimari- ; identified with a spatl.. level slightAy higher tha-: j.11

;sit specific l e,'el which is the foc 'r:; this sty In relation i.

the ... jvelcpmen -. 1'e i. to-select i.o-. '.i rithim, '.a La ions in ::.

rerrn re signif. ic r it is obvi'.ort: sep.r t :tings wil

have -o be developeJd for each i ndi.sty type, but this is not a prjim',ry




6Pred, op, cit., pp. 175-1.78,

Lonsdale, :p. cit.









objective of this study. It was anticipated that after developing a

general site-selection algorithm additional research would permit the

matching of various types of manufacturing to specific types of sites.
28
Historical Studies of Industrial Location Factors8

Numerous empirical studies exist of the factors associated with

actual industrial location events. Although many of these studies

were undertaken by academicians, the viewpoint is that of the indus-

trial developer seeking to attract new industry or expand existing

industry within the community.

Characteristically, investigations of this type are not concerned

with a specified theoretical framework for approaching the problem




28Discussion within this section is based mostly upon a review of
the following articles: J. S. Bullington, "Utilization of State-Wide
Site Evaluation Committee to Aide in the Location or Relocation of
Plant Facilities," AIDC Journal, IV (October, 1969), pp. 27-42; James
E. Chapman and William H. Wells, "Factors in Industrial Location in
Atlanta, 1946-1955," Atlanta Economic Review, IX (September, 1959),
pp. 3-8; Ronald E. Carrier and William R. Schriver, "Location Theory:
An Empirical Model and Selected Findings," Land Economics, XLIC
(November, 1968), pp. 450-460, and a more complete explanation of the
study: Ronald E. Carrier and William R. Schriver, Plant Location
Analysis: An Investigation of Plant Location in Tennessee (Memphis:
Memphis State University, 1969); Melvin L. Greenhut and Marshall R.
Colberg, Factors in the Location of Florida Industry (Tallahassee:
The Florida State University, 1962); T. E. McMillan, "Why Manufacturers
Change Plant Location versus Determinants of Plant Location," Land
Economics, XLI (August, 1965), pp. 239-243; N. J. Stefaniak, Indus-
trial Location within the Urban Area: A Case Study of Locational
Characteristics of 950 Manufacturing Plants in Milwaukee County
(Milwaukee: Wisconsin Commerce Reports, 1962); Charles M. Tiebout,
"Location Theory, Empirical Evidence and Economic Evolution," Regional
Science Association, Papers, III (1957), pp. 74-86; U. S. Department
of Commerce, Industrial Location Determinants, 1971-1975 (Washington,
D. C.: U. S. Department of Commerce, Economic Development Administra-
tion, February, 1973); and D. C. Williams and Donnie L. Daniel,
"Industrial Sites for Small Communities," AIDC Journal, VI (April,
1971), pp. 33-39.









but rather to learn the reasons for the location decision from persons

acquainted with the location event. None of these studies were con-

cerned with the construction of an algorithm to simulate the process

of industrial land use development.

Greenhut and Colberg29 analyzed factors influencing the decision

of 400 manufacturers locating in the State of Florida between 1956 and

1957. Location considerations were divided into three groups: demand

(market) considerations, costs (assembly) considerations, and personal

(psychic) considerations. Access to markets and potential markets

(Table 1) rated the highest among the location factors with the

remaining factors surprisingly low. The study, however, was slanted

toward measuring regional and subregional factors and thus was of

limited value to this study.

The extensive study undertaken by Carrier and Schriver30 of

plant locations in Tennessee between 1955 and 1965 was conducted

within the framework of existing location theory and, in part, did

focus upon site-location factors. Many of the variables included

in this analysis are based upon the conclusions reached in this study.

Carrier and Schriver identified six classes of location factors

believed capable of affecting plant locations: (1) personal factors,

(2) procurement-cost factors, (3) processing-cost factors, (4) distri-

bution-cost factors, (5) location demand factors (including locational
31
interdependency considerations), and (6) certainty factors. (This




2Greenhut and Colberg, op. cit.

30Carrier and Schriver, op. cit.

Carrier and Schriver, op. cit., p. 451.




















Percentage of 400 Plant
Location Factors Listing as Primary Factor

Access to markets 51.9

Anticipation of market growth 12.8

Good labor relations 1.7

Lower wages 2.6

Ease of attracting out-of-state
personnel, including research 4.7

Low freight cost on obtaining raw
materials and components 7.7

Low cost on freight on shipping final
product 10.7

Climate as it affects operations 1.8

Community facilities (education, police,
medical, etc.) 2.9

All other factors 3.2


FACTORS MOST INFLUENTIAL IN THE LOCATION
DECISIONS OF FLORIDA INDUSTRIES, 1956-1957

TABLE 1









class was identified after the interviews.) Certainty factors were

defined as the confidence that the "prevailing and forecasted data used

to identify the site offering maximum profits would persist into the

future."32

Persons involved in the selection of sites for 307 manufacturing

plants were interviewed. Each respondent was asked to select six

factors from those listed in Table 2 and to distribute 100 points among

these six in order to indicate the relative importance each factor

contributed to the total plant location decision.

Of the 36 factors listed in Table 2, low cost and availability of

labor was mentioned most frequently as the primary factor affecting

the location decision (Table 3). Personal considerations without

economic advantages received the highest average number of points,

followed by low cost and availability of labor (Table 4).

On the basis of the interviews the authors grouped industries

according to the six factors previously listed:

(1) Personal Factors Miscellaneous manufacturing, furniture and

fixtures, and food and kindred products were highly sensitive to

personal factors with most of those firms being "home-grown."

(2) Procurement-Cost Factors Industries which need large

volumes of low-unit-value or perishable raw materials were character-

istically affected by this group of factors. Food and kindred prod-

ucts, stone, clay and glass products, and lumber and wood products

industries indicated greater sensitivity to these factors.



32bid.
Ibid.









1. Personal Factors;

Personal with economic
advantages
Personal without economic
advantages

2. Procurement-Cost Factors:

Better service from seller
of raw materials and
components
Low cost on raw materials
or components
Availability of low cost
raw materials

3. Processing-Cost Factors:

Low cost and availability
of labor
Low cost of fuel
Low cost of electric power
Low cost of financing project
through Area Redevelopment
Administration
Climate
Favorable labor-management
relations
Low cost of satisfactory
type of water
Adequate waste disposal
Low cost of building and
land
Low cost of financing plant
through revenue or
general obligation bonds
Favorable community and state
tax structure
Community concessions
Available existing plant
Available existing building
Particular characteristics
of building site


4. Distribution-Cost Factors:

Low freight cost,
finished product

5. Location-Demand Factors:

Greater demand in
the area
Greater demand poten-
tial in the area

6. Certainty Factors:

Nearness to metro-
politan city
Community facilities
Community planning
and zoning laws
Cultural qualities
of the town
Community leaders'
cooperation
Size of city
Data provided by
Chamber of Commerce,
community, etc.
Information provided
by local manufac-
turers
Recreation, a good
place to live, etc.
Nearness to corporate
headquarters
Local supporting
services
State administration
neutral in labor-
management relations
Progress in racial
adjustment
Data provided by the
state industrial
development agency


LIST OF POSSIBLE FACTORS INFLUENCING INDUSTRY
LOCATION AS UTILIZED IN THE CARRIER AND SCHRIVER SURVEY33

TABLE 2


33p. 453.
Ibid., p. 453.



















Percent
Factor of Firms Rank


Low cost and availability of labor 65.6 1

Low cost of electric power 36.0 2

Favorable labor management relations 35.7 3

Community leaders' cooperation 32.2 4

Low cost of building and land 19.8 5

Low freight cost, finished product 17.9 6

Available existing plant 17.5 7

Favorable community and state tax
structure 17.2 8

Low cost of financing plant through
revenue or general obligation bonds 16.9 9

Available existing building 16.6 10


TEN FACTORS MOST FREQUENTLY MENTIONED BY TENNESSEE
FIRMS AS AFFECTING THE LOCATION DECISION

TABLE 3



















Percent
Factor of Firms Rank


Personal without economic advantages 49.2 1

Low cost and availability of labor 38.0 2

Available existing plant 35.9 3

Personal with economic advantages 32.9 4

Availability of low cost raw materials 31.9 5

Greater demand in area 30.2 6

Greater demand potential in area 29.8 7

Low cost of financing project through
area Redevelopment Administration 29.6 8

Available existing building 27.6 9

Nearness to corporate headquarters 26.0 10


TEN LOCATION FACTORS WITH HIGHEST MEAN NUMBER
POINTS ASSIGNED BY TENNESSEE FIRMS INTERVIEWED

TABLE 4









(3) Processing-Cost Factors These factors are associated with

in-plant costs in assembling or processing the finished product,

e.g., labor, energy, external services, capital, land costs, etc.

Electrical machinery, apparel and related products, and textile mill

products industries were affected by these factors.

(4) Distribution-Cost Factors These factors reflect the costs

incurred in shipping the finished products to the buyer. Among the

most sensitive to these factors were food and kindred products,

miscellaneous manufacturing, and paper and allied products industries.

(5) Location-Demand Factors Industries affected by these

factors are highly sensitive to market-demand in terms of proximity.

Included in this category are paper and allied products, printing and

publishing, and primary metal industries.

(6) Certainty Factors The validity of existing and forecasted

data is considered to be highly important by industries affected by

these considerations. In other words, these industries want to know

the future stability of costs in production and the probable con-

tinuance of existing markets. Printing and publishing, leather and

leather products, and transportation industries were highly sensitive

to these factors.

It is obvious that the scope of the Carrier and Schriver study is

much broader than the objectives of this study. Its utility is

limited for this research purpose. The factors considered by Carrier

and Schriver span several spatial levels of locational decisions. The

result is that factors which may be very important at the site-

selection level are weighted low in comparison to the total list of

factors. Also, the disproportionate number of factors offered for









consideration under the six categories tends to skew the weightings.

Finally, the lack of a very large sample in specific SIC categories

tends to decrease the validity of the results of the weightings and,

therefore, the conclusions reached regarding the typical locational

patterns of specific industries.

Bullington34 offers a scheme to locate potential industrial

sites on a state-wide basis by suggesting the scoring of location

factors on an ordinal scale and aggregating them into an index to

determine the site potential for specific industries. Local and

state governments could then match the qualities of the industrial

sites available in the community to specific industries. The factors

which Bullington suggests are listed in Table 5.

The U. S. Department of Commerce recently published the partially

aggregated results of an extensive 5-digit industrial location survey

conducted by mail throughout the U.S. The purpose of the survey

was "to assist the nation's underdeveloped and declining areas in the

development of their economic resources and potentials."36 Only

manufacturing industries demonstrating "reasonable" growth between

1958 and 1967 were selected for inclusion in the survey. Survey

forms were mailed to a total of 2,950 companies in 254 different




34J. S. Bullington, "Utilization of a State-Wide Site Evalua-
tion Committee to Aide in the Location or Relocation of Plant
Facilities," AIDC Journal, IV (October, 1969), pp. 27-42.

35U. S. Department of Commerce, Industrial Location Determi-
nants, 1971-1975 (Washington, D. C.: U. S. Department of Commerce,
Economic Development Administration, February, 1973).

3Ibid, p. 1.





















Site Characteristics


a. Size of Parcel
b. Shape of Parcel
c. Topography
d. Drainage
e. Flood Record
f. Condition and
Appearance
g. Underground Water
h. Soil Bearing Capacity

Acceptability

(This referred to the
potential friction or
good-will prompted by
the location of industry)

Accessibility


a. Highway
b. Secondary Roads
c. Rail


d. City Water
e. City Sewer
f. Limitations of Site

Community Factors

a. Commercial Air Service
b. Water Transport
c. Location in State
d. Mileage Rate
e. Airport Facilities
f. Comprehensive Planning
and Zoning
g. Retail Accommodations
h. College
i. Community Appearance
a. retail
b. residential
j. Highways
k. Presentation of Facts
by Community
1. Sanitary Sewer and
Water Treatment
and Facilities


LOCATION FACTORS SUGGESTED BY BULLINGTON


TABLE 5









SIC categories. One form, Survey of Industrial Location Determiinants,

was to be completed by all companies to i.Jcntif'y the locational and

operating characteristics of existing plants.

Only the results of the Survey of Industrial Location Det-rmi-

nants have been published to date and its usefulness for this study

is limited because of the highly disaggregated form of the data.

Hopefully, these data have been digitized by now and can be utilized

for future analysis purposes.

The attributes measured or assessed in the survey are only

slightly coincident with those sought in this analysis. Also the

survey spans several spatial levels of locational decisions. Never-

theless, some of the variables utilized in the survey were included

in this analysis.


Land Use Modeling Studies


Site-selection algorithms are certainly not novel to land use

modeling methodologies. Numerous modeling efforts have utilized

various allocation systems to distribute projected change in land

use. Most of these approaches have been different from this approach

in framework and in terms of the spatial scale of allocation.

Previous land-use models usually operate at the census tract or

count) level simply because data (e.g., census materials) to calibrate

the models are more readily available at those levels. The following

studies, however, have approached the allocation problem similar to

this study and consequently are briefly discussed below.

Among the first land-use models to utilize a site-selection

algorithm to distribute projected land use change was one developed









at the University of North Carolina by Donelly, Chapin, and Weiss.37

The primary purpose of the model was to simulate the growth and spa-

tial spread of residential land use. Historical data (1948-1960) were

utilized to calculate an attractiveness probability to simulate the

conversion of rural land to urban residential use. Changes were

recorded by means of 1000-foot cells composed of nine 2.5-acre land

development units.

The model operated in the following manner:

1. All land within the city unsuitable for development is

eliminated from consideration at the beginning and the

supply of land remaining is identified as available for

residential use.

2. For each 1000 foot cell, a measure of relative value is

established, i.e., land value, as a measure of its attrac-

tiveness for residential development.

3. The effect that "priming" (expansion of municipal services,

commercial services, and industrial development) decisions

will have on modifying the value of the property is then

calculated for each cell. These are assumed to be exoge-

nously given but in this case the exact amount is known from

historical data from 1948 to 1960.

4. Land parcels are then "reassessed" to obtain a new attrac-

tiveness score.




37Thomas G. Donnelly, F. Stuart Chapin, and Shirley F. Weiss,
A Probabilistic Model for Residential Growth (Chapel ilill: University
of North Carolina, Institute for Research in Social Science, 1964).









5. Density conIstraints (numiibe.rs of units per acre/year) are

then introduced.

6. Finally, the known gro'H,;h of residential households between

1948 and 1960 is allocated on a probability basis.38

The drawbacks to the model are: (1) the need to have historical

data in order to develop the transitiGn probabilities; (2) the disre-

gard of the effect of changes in industrial location and employment;

and (3) the vast amount of programming needed to load and run the

model for a small region such as a single city. The concepts gene-

rated by this study, however, have become basic to many other modeling

efforts and were equally important in this study.

The Pittsburgh industrial location model, INIMP39 (Industrial

Impact Model), is similar to the hypothetical design of this modeling

approach; the major difference is that growth is distributed to

census tracts and consequently the variables are more aggregated than

those being considered in this study.

Four variables (attributes of census tracts) and one constraint

were adopted as being sufficiently discriminatory to determine site

locations. These are: weighted mean unit-assessed value of land;

weighted mean unit-assessed value of buildings; weighted mean struc-

tural density, and amount of industrial clustering. These measures

were determined by census tract. The constraint can either be imposed




3Ibid, p. 11.
39
3Steven H. Putman, "Intraurban Industrial Location Model Design
and Implementations," Regional Science Association, Papers, IXX
(1966), pp. 199-214.









artificially as in the case of zoning controls or non-existence of

services; or directly imposed by the model operator. On the basis of

the aggregated scores of the above indexes, the :nodel distributes a

portion of projected city-wide employment changes among exist ,ig

facilities and, upon reaching certain critical values of saturation,

switches to a separate routine to distribute new facilities to census

tracts having the highest scores for the remainder of the projected

industrial employment growth. The distribution to various census

tracts is accomplished by way of the maximum score. However, in the

event of a tie, the allocation algorithm switches to a Monte Carlo

routine. Similar but more elegant models of this type are being

developed at Harvard40 and the University of British Columbia.41


A Tentative List of Variables


Each theoretical study reviewed approached the problem of

explaining industrial location in terms of three components: demand,

cost, and personal factors. Carrier and Schriver subdivided the

process into six components: personal factors, procurement-cost

factors, processing-cost factors, distribution-cost factors, location-

demand factors, and certainty factors. Only processing costs, pro-

curement costs, distributing costs, personal costs, and certainty




40
4Carl Steintz and Peter Rogers. A System Analysis Model of
Urbanization and Change: An Experiment in Interdisciplinary Educa-
tion (Cambridge: M.I.T. Press, 1971).
41
4M. A. Goldberg, Quantitative Approaches to Land Management
(Vancouver, B. C.: University of British Columbia, The Resource
Science Center, 1970).









factors have any direct relationship to site location considerations.

Certainty factors may be considered a variation of location-demand

costs. At the site level, procurement costs and distribution costs

are sensitive to ,ne variable, accessibility. Considerations of

freight rates, transport modes, proximity to raw materials, supplies,

etc. are more related to locational considerations at the subregional

level.

In an effort to arrive at a tentative set of variables to be

utilized in this analysis, a list of variables mentioned in the

industrial survey literature as being related to the site selection

decision was compiled. This list is found in Table 6.

A number of variables may be eliminated as not related to the

site-selection process while others can be combined with other vari-

ables. After considerable study the following list of variables was

developed to be used in a survey of the site conditions of past

industrial location events. Selection was based upon the relevance of

the variable to industrial site-selection and the ability to obtain

measures of the variable. The variables are listed below:

I. Site Preparation Cost

a. Slope of land
b. Drainage
c. Clearing-Cover conditions

II. Market Price of Land

a. Distance to center of town
b. Distance to nearest major thoroughfare
c. Density of urban use in immediate vicinity
d. Overall rating of price of land from 1 to 10

III. Proximity to Work Force

a. Proportion of nearest city within 2-1/2 miles
b. Population of nearest community




















1. Industrial Park Space
2. Industrial Park Quality
3. Industry Nearby
4. Zoning and Building
Restrictions
5. Pollution Regulations
6. Land Costs
7. Site Preparation Costs
8. Topography
9. Drainage
10. Soil Conditions
11. Place to Dump Effluent
12. Processing Water
13. Utilities
14. Municipal Water
15. Sewage
16. Natural Gas Service
17. Proximity to Local
Markets
18. Proximity to Local
Raw Materials
19. Proximity to Sup-
porting Industry
20. Water Transport


21. Railroad Transport
22. Highway Transport
23. Commercial Airport
24. Distance to CBD
25. Community Transportation
26. Community Parking
27. Neighborhood Services
(Restaurants, Hospitals,
Parks, Gas Stations, etc.)
28. Nearby Housing
29. Nearby Labor
30. Community Cooperation
31. Community Stability
32. Community Wealth
33. Community Taxes
34. Community Progressiveness
35. Community Attractiveness
36. Community Labor Climate
37. Community University
38. Nearby Government and
Institutional Facilities
39. Community Wage Rates
40. Space for Expansion


TENTATIVE LIST OF VARIABLES


TABLE 6









IV. Transportation Accessibility

a. Distance to major highxi',-
1). Di stance to secoindaryl roa
c. Distance to rail
d. Distance to airport
e. Waterway service
f. Distance to nearest Interstate interchange
g. Overall quality of accessibility from 1 to 10

V. Utilities

a. Water available
b. Gas available
c. Sewerage available

VI. Compatibility with Existing Land Uses

a. Did community have zoning at time of location?
b. Was site zoned for industry?
c. Was zone changed to accept industry?
d. Was industry already in immediate area?
e. Overall rating of contiguous land use compatibility

VII. Neighborhood or Community Attractiveness and Amenities

a. Condition of neighborhood
b. Density of land use in immediate vicinity
c. Nearby community services

VIII. Industrial Park Space

a. Was the site in an industrial park?
b. Overall rating of the quality of park?

Additional Data Collected

a. Proximity of site to Knoxville
b. Amount of other industry located nearby at time
of event
c. Was building already there?

This list may omit variables which should be considered and,

therefore, should not be considered exhaustive. However, it is antic-

ipated that in measuring the importance of each variable some may be

eliminated thus simplifying the site-selection algorithm. The next

chapter explains the methodology and research procedures used to test

the importance of each variable.















CILI fER III

DESCRIPTION OF THE STUDY AREA
AND ANALYSIS PROCEDURES


This chapter describes the research methodologies utilized in

this study. The first part of the chapter describes the eastern

Tennessee study region, the second part the data collection proce-

dures, and the final part the analysis procedures.


The Study Region


The study region encompasses 16 counties surrounding and including

Knoxville, Tennessee, (Figure 3) and represents an administrative

entity called the East Tennessee Development District (ETDD). The

region spans 6,500 square miles and contains a population of approxi-

mately 750,000 people. Its selection for this study was based upon

the availability of data in the Oak Ridge National Laboratory (ORNL)

Data Base.1 ORNL selected the region on the basis of "the diversity

of the region, the availability of data, the presence of cooperative

and interested user groups, close proximity, etc."2




1Oak Ridge National Laboratory is currently developing a "holis-
tic" environmental model for the ETDD region. This research was
supported by this program. ORNL-NSF Environmental Program, Regional
Environmental Systems Analysis (A Research Proposal Submitted to the
National Science Foundation, February, 1972).

Ibid, p. 3.












The ETDD region is centered in the southern portion of the Ridge

and Valley Province (the "CGreat Vally"'), bordered to the northwest by

the Cumberland Plateau and Mountains, and to the southeast by the

Great Smoky Mountain complex (Figure 4). Both the Smokies and nhe

Cumberlands are characterized by steep slopes and forest cover, with

the Cumberlands distinguished by strip mining scars.

The area is drained by the Tennessee River system, the natural

flow of which has been vastly altered by the Tennessee Valley Author-

ity (TVA). By harnessing the power of the river system to produce low

cost electrical power and developing a navigable channel to Knoxville,

TVA has become the major development agency within the study region.

As a result much of the industrial development in the region has been

structured by TVA activities.

The largest urban center in the region is Knoxville which serves

as the major economic and transportation focus for the region. Sur-

rounding Knoxville are Oak Ridge, Maryville-Alcoa, and more distant

Morristown, each with between 20 to 35,000 in population. These

cities perform subregional functions. Remaining urban centers are

small in population and are mostly located within the confines of the

plateau and the Blue Ridge (Figure 5).


Industry within the Region

This analysis has concentrated upon secondary manufacturing (SIC

20 through 39) and excluded extraction industry (SIC 10 through 19).3

The location determinants of extraction (or primary) industry are




Industries are referred to by their Standard Industrial Code
(SIC) number throughout this study.












































.7~


/4


0


'I

























^1



I





to
b

















tr


t
,

k
^ I

fcl
,











u-1


b;

__


~-t









dictated more by the distribution of raw materials and consequently

the locational criteria of these industries will be different from

secondary industries. For this reason primary industry is not con-

sidered in this st-.dy.

Categorically one could state that most of the industry within

the region is concentrated in the Knoxville area. Of the 1002 indus-

tries in the region, approximately 46 percent are located in the

immediate vicinity of Knoxville. Knoxville also has the greatest

diversity of industry, whereas industry in other communities is

characterized by specific categories. (An example is the concen-

tration of furniture industry in Morristown.)

Figures 6 through 13 provide a visual overview of the dynamics of

industrial development in the region since 1943. Since 1953 a 48

percent turnover in industry has occurred. Of the 1002 industries

within the region in 1973, 488 have developed in the 16-county region

since 1953.

Table 7 provides a statistical summary of the new industry-new

employment expansion (or decline) from 1952 to 1973. The table does

not reflect the absolute employment expansion rate but rather the

amount of employment gained or lost by the birth or death of specific

two-digit SIC industries. A high average expansion rate indicates new

plants have developed in the region and have significantly increased

the total employment in that specific industry.

In the following discussion each industry type is briefly

described. Maps are included to show the regional distribution of

specific industries. Graphs illustrating the historical employment





















15



tr-- 4- -,-- 1-

i:i
*-- 1
.. *. .
I. *. *






















S





ti
I \ !---L~;~,-.- t































!




k
-4 1



00


PI


I I I









49























-Ii


4;


2--i

4;


+M











51


















r I4r


____ I








r. ~ .







*~ Ii








5' 'I
Lr 6



I I i i i i i






52










i iCC
- ,
*. a -o g g
__^ / .~ m oo -
*6 S-^ e SA a-
I *





I I
oA

4. 6








S.J

c
^-- ^r--- 1r






S_ S ** I <
--------------Y--- .*~-













UI
\ ^____i i ..
S
*:. .. ;' 7 :'




S-~
S j i- - -i- -- -Y -.
^_ ,_^ p -" lv ^ ; "i



,t~ P *!'-v;

S-- -^1)




"--
-~----^--*.^ry_.





a I 6 a i i











Average New
I industry Expansion
by Rate
SIC 1955 1958 1960 1963 1966 1969 1973 1952-1973


20
Food +9.4 +10.0 +0.5 +8.7 +6.5 -3.2 +11.8 +3.8

21
Tobacco

22


Textile


23
Apparel

24
Lumber

25
Furniture

26
Paper

27
Printing

28
Chemical

29
Petroleum

30
Rubber

31
Leather

32
Stone,
Glass, Etc.


Primary
Metals

34
Fabricated
Metals


-4.6 -2.7


+28.3


-40.4


+24.0


-4.7


+9.3


-.6


-8.2


-1.4


-9.7


-13.6


-3.2


+30.4


-16.4


-2.0


+13.4


+4.7


-11.6


-7.1


-7.5


+2.3


-2.2


+53.1


+12.3


+24.5


-4.3


+5.2


+9.8


+25.8


+1.4 +3.2 +7.6



-6.9 +10.1 +.5



+36.6 +12.0 +.8


+8.3 -3.9 +5.2


+20.1


+2.1


+13.8


+10.2


-4.1


+2.3


-41.9


+10.3


+13.1


+11.2


-0.3


+5.2


+4.2


+2.2


-14.7


+27.2


+79.5


+25.9


+40.3


-7.2


+5.9


+49.1


+21.0


-2.4


+43.0


+61.6


+62.6


-19.7


-12.4


+11.2


+7.8


+41.8


-3.3


-25.6


-5.0


-32.4


+63.5


+.3 -11.0 -4.4 +25.5


-22.8 -4.6 -6.9


+10.1


-2.4


-3.6


+10.9


-5.7


+20.0


+13.8


+6.8


+2.8


+3.6


+16.5


+24.6



+3.2



-4.7


+4.7 +36.8 +6.3 +15.3


EXPANSION OF ETDD INDUSTRIES BETWEEN 1952-1973


TABLE 7
















195' 1958 1960 1963


1966 1969 1973


Average New
Expansion
Rate
1952-1973


35
Machinery
Non-Electric

36
Electrical
Machinery

37
Transpor-
tation

38
Instrument

39
Miscellaneous


+6.7


-4/.7 -1.3 -6.9 4409.8 --21.1 -30.1 +19.1


+26.4 -17.2 +22.5


+37.3 -1.2 -3.9


-15.7 +2.6 -8.4 +


+3.7 +439.0 -13.2 +136.5


+,.5 -1.1 +9.5 -3.3


34.5 +4.4 +11.6 -26.1


+85.4


+3.3


+10.6


*This table is based upon ORNL industrial location data. Expans:
rates are determined on the basis of new employment prompted by
new plants (or death of plants) since the previous time period.
base year is 1952.


ion


The


TABLL 7 Contiiinued


Industry)
by
SIC


_ ~_I_


___ _I_____


_(__ ~_ _____~_ ___


-3.4 +1i.2 -8,6


,.3.1 -15., +55.3 +0.5









trends relative to total regional employment and the industry's share

of total regional manufacturing employment accompany each map.4


Food products industries (SIC 20)5

Of the 103 plants existing within the region, 66 are located in

the Knoxville region. Most of the remaining plants are located in the

southern half of the region bordering the Smoky Mountains near the

better crop producing areas along the margins of the Nolichucky,

Pigeon, Little Tennessee, and the French Broad Rivers (Figure 14).

Approximately 75 percent of these industries employ less than 100

persons. Employment growth has been gradual and in accord with total

regional growth. Proportionally, however, employment has been de-

clining (Figure 15).


Tobacco products industries (SIC 21)

No tobacco industries exist within the study region.


Textile mill products industries (SIC 22)

The textile industry is distributed broadly from northeast to

southwest through the center of the region with Knoxville having the

greatest concentration followed by Morristown and Sweetwater (Figure




4The maps and graphs utilized in this chapter were prepared by a
CALCOMP plotter driven by an IBM 360 computer. The data utilized were
originally collected by Osbin L. Ervin. For a complete presentation
of these data, see: C. R. Mcyers, Jr., 0. L. Ervin, D. L. Wilson, and
P. A. Lesslie, Spatial Distributions and Employment Trends of Manufac-
turing Industries in East Tennessee (19413-1973) (Oak Ridge, Tennessee,
Oak Ridge National Laboratory, June, 1974).

5Transparent overlays found in Appendix B may be superimposed
over each industry map for spatial referencing.





























E-4
16



















0
bb

















C,


H









I.
---?
i0
--*1















0




a)







M H
C )



/ (Dn 0 0 l





o





OS Sh Oh S 0
o ^ 4














a)



0
U-)


















C)





S0 4
\ -4














V










In -
















os ll l o h se oII I S sI
SO1 S 0t S L0t S zOt S -
un
^ t 'fl 2 z
I '2 M

















Sol s oi s cot 1 9O so S









16). Average employment in the industry is approxi-atcly 210 persons

per plant. Rockwood and Harriman have only four plants but share

nearly 50% of the total employment inl the region. The distribution

of plants reflects a tendency to gravitate to re1.Ltively small communi-

ties as suggested by Lonsdale6 and Pred.7 Textile industry has declined

in terms of the percent of total regional employment in recent years

perhaps resulting from the introduction of higher wage industry into

the region (Figure 17).


Clothing and related products (SIC 23)

Knoxville leads the area with the greatest concentration in both

plants and employment (Figure 18). Average employment is approxi-

mately 315 persons; however, the median is closer to 150 employees.

This type of industry though concentrated in Knoxville is scattered

throughout the region with numerous small communities having at least

one (sometimes two or three) small employment industry. In terms of

new industry expansion the clothing industry expanded at an average

rate of 10.9 percent between 1953 and 1973. Historically, however,

growth has been erratic (Figure 19).


Lumber and wood products except furniture (SIC 24)

This industry is widely scattered throughout the region mostly in

small communities (Figure 20). It .s a small employment industry

mainly utilizing local raw materials. In recent years, an overall




Richard E. Lonsdale, cited in Chapter 2.

Allen Pred, cited in Chapter 2.




















































cn

L)












rLr






L)L
1-4
'n






















































Io, s, ~ .... ,. .,... I I I... Ji
OS Sh Oh SE OE SZ O SI O S 2


I.......I .....iii 111111 I I


li 1 I 1 II I I I I - - - .. . . .
SO ?U UL e


111111 i I I


u






CD
z z







uE




HZ






O[-


Ht
C"-





E-4






n


EU


sOI s vUI








61


































00


















~C'4




-4



























H




00










H
l-n


OS Sh Oh SE OE SZ OZ SI O S



D(
12


I ~ .. i i I 1 1 1


I i li i


I I I I I I I il lI I I A I I i i I r


cOI s ZoI s


I


Ii n n


sOl s h01





































t



iI




















* .

S-* .


S.Y




S

S--


a a k
a a


CL.














ho
) a,




ca
I-
< 4x0




E-i

UT -































a \ Pa a


1.









decline in the industry has occurred partly due to a decline in small-

operated lumbering in general for the region. The average employment

per plant is only 32 persons with the median closer to 15 persons per

pl'-.t.

The growth prospects for this industry are rated good by the TVA

Regional Planning Office but it is expected to be in terms of large,

permanent sawmills utilizing local oak, hickory, and poplar forest

resources (Figure 21). Those species are being increasingly utilized

by the furniture manufacturers.8


Furniture and fixtures industries (SIC 25)

The furniture and related products industries (e.g. woodworking

industry) is heavily concentrated in Morristown (with 20 of the 46

plants) even though Knoxville has 17 plants (Figure 22). The average

number of employees per plant in Knoxville, however, is 161 persons;

whereas the average in Morristown is approximately 310 persons.

The furniture industry located in Morristown not because of the

local timber sources but because of cheap labor and good rail connec-

tions to large mills in the South and to market regions in the North.

With increasing reliance upon local timber, however, the growth

prospects for this industry are seen to be significant (Figure 23).




Communication with Bill Ogden, TVA Regional Planning Office,
Knoxville, Tennessee, May 3, 1973.

Ibid.

























































I ... I I I I I I I I I i I I I I I II I ... I I I... It r
OS S O S SE OE S2 0 SI 01 S


C


A s. I A . I I I . I 1 I 1 I I 1 I I I I I I I


Z









H


0

M Q)
et'















E-4


0)















u




1-
u


~


I,,,,,~ ~ ,


I I I I I I I I I


I, ,,


hOT s


EOl s 201 s




















2'E







I,,
-4
'-
tjD









67




0




CO







H-4
0o '










Z


NJH
Ln z








V a)
=r
0 0










E-4






C1


H
1 1







\ t





ta E-4
0)
iD C






en z







O






III I I I I I Jl ll ) I I I I l llI II I I l ill lu I I
sOl S hO I S 01 S O S 2s









Paper and pulp products industries (SIC 26)

The paper and pulp industry is concentrated in the Knoxville

area both in terms of employment and number of plants (Figure 24).

Intuitively, one suspects market-Kemand to be a strong locational

factor except for the plants in Morristown which are locationally

related to the by-products produced by the furniture industry.

Employment averages 78 persons with the largest plant employment only

190 persons. Total employment in the paper and pulp products industry

is relatively low compared to other industries, only 1377 employees

in 1973. The average growth rate, however, has been significant

(Figure 25).


Printing and publishing industries (SIC 27)

Printing and publishing includes local newspaper printing and

for obvious reasons is widely dispersed throughout the 16-county

region (Figure 26). Knoxville has the greatest concentration with

over 75 percent of the employment and 2/3 of the total number of

establishments. Total employment is low for the region (1670 employees

in 1970) with an average of 19 workers per plant. Growth has increased

along with population with no great expansion predicted (Figure 27).


Chemical products industries (SIC 28)

The shunning of large urban places by chemical plants noted in

Chapter II at first appearance is not substantiated by the regional

pattern of chemical industries within ETDD (Figure 28). Industries

are found in the other communities (e.g., Morristown and Oak Ridge)

but Knoxville is the leader in number of plants with all but six of

the 33 plants in the region. However, the Knoxville industries are









































04











1C1

1-4
n,


r
































































os S O S O E se oO S O s 01 s So


,T C II I I I |
sl 1I L


- -
z z







0
o o









o.
E-4


I


111 1 I I I I 1! i i 1


Il l I i i


I . .


IUL


EUL s


Su i




















r- -


0






I i l ;


-c------,













.. ^ _i ._-









+ i --_ --


I


(fl*
c=
.- a
C
=
a. --


6t,~







t2













E-4 r.I


P.4


H


4-L


6 b
E >


II






















































OS Sh Ot SE O0 SZ 0 S 1O SI 0


I I I I I I I I I


SI I I I I I I


II III I I I


SW z u


I l I I


u








z 3
Z H




HU




^^3


"


I


SUI S


201 s






73




















P P4




00
"il



cn-














C- ------ ----
T *~





tt
_ _ I _ __ ____ ___














_ _i I



I ft SI I









relatively small (the largest has 69 employees) wh;hle the ENKA plant

located at Lowland (six miles south of Mlorrito-..n) employees over

4500 persons. Large employment figures in Oak Ridge and M'orristown

(Lowland) skew the average to 308 persons per plant; therefore, the

median 10 per plant is more descriptive of the true employment

picture. Employment in the chemical industry has been erratic in the

last two decades (Figure 29) and future expansion is difficult to

predict.


Petroleum refining and paving and roofing products (SIC 291

Employment and plant number are relatively insignificant relative

to total industry within the region with all but one plant (seven

employees) located in Knoxville (Figure 30). Little change in employ-

ment has been noted in recent years (Figure 31).


Rubber and plastic products industries (SIC 30)

Ten of the 12 rubber and plastic products industries are found

in Knoxville presenting over 60 percent of the employment (Figure

32). The average number of employees is 68 people with 20 employees

as the median. Between 1952 and 1973, the new industry growth rate

has averaged 16.5 percent but the total employment in 1952 was only

514 persons (Figure 33).


Leather products industries (SIC 31)

The leather product industries are few in number (15), most

occurring in Knoxville, Dandridge (near Jefferson City), and Morris-

town (Figure 34). The average expansion rate has been significant,

24.6 percent even though the base year employment (1952) was only 546




























01


DI -
a)



.0


(0

Ur)

U')
a),



Lr)


CY)
t p









\es 0
II 001


1111 1 1 1 1 1


lilt I IO I I0 I IS


C.,
)"





1-
Cr


I\


00 4

z




0
E-4
zO
cn


:
-


.


,o


I I I I I I l l I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I


sOI s h01 s


cOI S 20I s
































































_________-T,________ /______\_ _____________ _______________ ___________________________

I.

L i
i^ k,-^r -~------ --- s
e- L L_________


























































os si Oh SE oE SZ OZ SI 0 S


I 1 l i l l i l i 1 I l i I 1


0
z



Z1-



















I


CN
Pri
1-4










(n
Oi


-""''' 'i


i
i
i
I r


I i I1 i


I I !


11i i i


sOl s 01 s


OI s 1O







78


















C-










000
0 Po






to













I IV










I i iI I I Ih SP

























































OS Sih A SE O S 0 S O[ S


11111l(1 I IiiilIi


liii1 I I I


. . . .. I I I I3
SUL #"I I LL .


liii i


C.3
ZZ


SH
m g











.-I E-I
0
O


On

O^


0
W1





H
3




04















1-4
IM
fc






Ub
COi


U L


' ' "


SUl


hUL






































liz
Hr


k~









persons. Employment per plant averages 100 persons as does the

median, indicating a fairly symmetrical distribution. Expansion in

the future may be expected to continue but probably not at such a

high rate (Figure 35).


Stone, clay, and glass products industries (SIC 32)

These represent strong local raw material, local market indus-

tries, and thus their location is determined by the chance occurrence

of resources and the distribution of urban places. The spatial

pattern supports this observation (Figure 36). The greatest concen-

tration, as expected, is in the Knoxville vicinity with 2/3 of the

total number of plants (68). The average employment per plant is 32

persons. The median is 14 employees per plant. Expansion has kept

pace with population growth in recent years and probably will con-

tinue to expand in response to total regional growth (Figure 37).


Primary metal industries (SIC 33)

This industry type is distributed throughout the region simply

because of the large number of foundries which historically developed

in East Tennessee (Figure 38). The largest employment industry is

Alcoa Aluminum in Maryville-Alcoa with 5000 employees (70 percent of

the total employment). Median employment, however, is only 50 persons

per plant. Employment has dropped steadily since 1963 but is expected

to level out in the future (Figure 39).


Fabricated metal products industries (SIC 34)

This classification includes welding and machine tool industries

which usually develop locally and are local-market oriented. This


























o







o H












01
0 m


















In


r)
CVl













=r








H H
*o M
















0r)



o a













In
N E- 0

































Sol hol S co Z I







83



















'000




0 0e









1 CN
I ---~F -4





pC- __ __






F I,


~-: I I I+T




I ~ i I -I
























































IO S. Oh SI I SZ I OZ SI ... S I
OS Sh Ob SIE OE Sa 02 ST l s F!S


I I I Ii I I I 1 lI II I I I I I


Iii 1) ii i i1 i I


DI s s0i s Oi s s


cn


u
Z;














.z0
E-4

a^


u
I
v,







85


















---- -K
- -------------------










E-4~








CU,
A 1-41
11~- --1con

























































OS Sh Ot SE OE SZ o. ST 0o s


I


/11111 I I lii ii r H


1 c 1"V
s1L LU


ZZ













E -4
az
1.4
Q 5



o5
en F


. -. I=


*r-4
HM
I






2 -,
I


cj,



en


1ll1111 I


~
u


I . .


hul


EU









explains the wide distribution of the industry Figure 40). Most of

the 73 plants are found in Knoxville (upproxi ,:itely 75 percent) indi-

cating some response to raw material availability which is greater in

the Knoxville area. The average number of persons employed is 45 but

the median is only 14 employees per plant. The industry has grown

significantly between 1952 and 1973 probably in response to national

growth in the transportation and recreational vehicle markets (Figure

41).


Machinery (except electrical) industries (SIC 35)

The machinery industry is heavily concentrated in the Knoxville

region and to a lesser extent in Oak Ridge and Morristown (Figure

42). Many of these industries perform supporting functions such as

the refurbishing of industrial tools (for example, the woodworking

tool industries in Morristown). The largest employer has 639 employees

but the average for the region is 27 persons per plant and the median

is only 7 persons per plant. Total employment has remained fairly

constant in the last decade (Figure 43).


Electrical machinery, equipment, and supplies industries (SIC 36)

The electrical machinery industry has grown from 578 employees

in 1952 to 2726 employees in 1973, averaging a growth rate of 60.9

percent. The largest employer is Magnavox in Morristown with 716

employees in 1970. Average employment is 120 persons; the median is

40 persons per plant. Most of the industry is concentrated in Knox-

ville and dispersed lightly elsewhere (Figure 44). In recent years,

employment has expanded significantly (Figure 45).




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 - Version 2.9.9 - mvs