An econometric model of structural changes in the U.S. potato industry


Material Information

An econometric model of structural changes in the U.S. potato industry
Physical Description:
xiv, 315 leaves : ill. ; 28 cm.
Jones, Eugene, 1951-
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1984.
Includes bibliographical references (leaves 309-314).
Statement of Responsibility:
by Eugene Jones.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000502564
notis - ACS2280
oclc - 30651011
System ID:

Full Text


"2G-NE JO:;7S

P2. :AL 1 OF TH2 2EQuI52E 3NTS


!98 (

To the memory of my father

who committed his life to providing

educational opportunities for

his children


I wish to express a sincere and profound gratitude to

my committal chairman, Dr. Ronald W. Ward, for his guidance,

counsel and encouragement throughout my research studies,

Dr. Ward always seemed to give me a larger proportion of his

time than that which could be justified, given his other

research and teaching responsibilities. His enthusiasts for

the subject and his dedication to his profession were

kindles of spark that inspired me to keep trying even in the

face of what seemed like insurmountable hurdles,

I am also grateful to Dr. Glenn Zepp for his assistance

in providing data and information through thle economic

Research Service, U.S.D.A., and his comafldRits and suggestions

for improving th3 manuscript. To the other members of my

committee, Drs. 2ichard Kilme:, Frank martin and W. W.

McPhdrson, I as appreciative for their reviews and

suggestions for strengthening the research. I aa also

thankful to Dr, acPherson for his guidance and assistance

during much of my course work, especially assistance in

obtaining fuads, The suggestions and comments of Dr. LSenry

Fishkind during the initial stages of this research are also

apprec:iatsd. The Economic Research Service is due a ispcial

note -.f thanks for funding the research.


I am also grateful to my typists: Kathy Carroll, Linda

Kilmer and Kathleen Laird. The patience and skills

exemplified ay these persons are much appreciated. I am

also thankful for my friends and associates who helped to

make my studies more pleasant and rewarding.

To the most important people of my life, my wife,

Jackie, and daughter, Lauren, I owe a special note of

gratitude for their encouragement, support and sacrifice.

Jackie's uncanny ability to offer words of encouragement

during stressful times proved more effective than any

medicine she ever prescribed for me. Lauren's little gentle

smile was a great source of comfort after a long and

difficult day of work. The ultimate note of gratitude is

owed to my mother, Pauline, who taught me the important

things of life.


ACKNOWLEDGMENTS ...........................1..... .... iii

LIST OF TABLES ...... ......... .... .. ..9.. .3. viii

LIST OF FIG ES .. ..... ..... ...... .... .. ... x

A s .. .ACT ,,,,9. ,9 ,, .., ........ ..... ....,.... xiii


I INTRODO'CTfIO .............. ........ ..... 1
Product Characteristics .... .............. 1
Demand and Supply .... .. ..... .. .... ,, ... .. 2
Structural Attributes within the Potato
market .... ..... .... ... .... ............. 8
Structural Attributes Outside the Potato
j:1ike t .. .*.* ... 3 .. *~ 15
Concept of Performance ....................... 21
Criteria for Analyzing Performance b....,,,sa, 21
General Performance of Potato Industry ....... 23
Problem Statement ,,,..9..9.. ...... ......... 29
Hypotheses ............................... 34
Objectives ......... ... .. ,,.. ....... 34
Methodology and Scope ...........,........ 35
Su sequeat Chapters .......................... 36

II LIT' .AT 2 .. ...................... 38

Introduction ............ ....... ............. 38
Previous Studies .. ......... .. ...... ........ 38
.east Coaprehensi7e Studies ................. 39
."de:rteiy Coaprehensive Studies .,*.......... 42
Comprehensive Stnieas ........................ 46
Present Study .....,,,,....,..,. .,,.9..9. 56

III OVERVIEW OF P?'TATO I:.USTRY .................. 58

latr3ductiozn ,,,,, ... ..).. *. .. ...... ..... 5
Potato Production ......... 9 .......... 58
Potato Consumptio r. .... .... ......... .... 62
Fast Food Esta blis ....,.
t Zstzli,,aea ...,, QI, .... ...


Women in The Work Force ............ ...,.... 67
Population ..................... ........... 68
NHumbr and Size of Households ....... ........ 68
Consumers Taste .......... ............ ..... 69
Implications ,............................... 69
Summary and Conclusion .................., 73


Introduction *..................... .....*.... 76
The Structural Sector ...................... 77
Number of Potato Processing Plants ....... 82
Concentration in Potato Processing ....... 86
Price Cost Hargins ...................... 89
advertising Sales Ratio ................ 90
Change in Processing Capacity ........... 93
Utilization of Processing Capacity ..,.... 93
The Market Equilibrium Sector ................ 95
Production of Baw Potatoes .,......,..... 95
Utilization of Potatoes for Fresh
Consumption .*...*9..,.. ...,..* ....... 97
Utilization of Potatoes for Processing ... 100
Retail Price of Fresh Potatoes .......... 104
Retail Price of Processed Potatoes ....... 105
Wholesale Price of Processed Potatoes .... 105
Farm Price of Raw Potatoes ................. 106
Research and Promotion Tax .,,...,-....... 107
Spread Between Retail and Farm Price ..... 107
Utilization of Total Production ......... 108


Multicollinearity ........................... 110
Principal Components ......,............... 113
Principal Components in Siaultaneous
Equations ...,.... ..,..,,.....,,......*., 117
Principal Components in Present Study ........ 121


Introduction ....................... ......... 127
Pariaeter Estimates and Their Properties .a... 128
Structural Equations Results ................. 135
Number of Processing Plants ............. 135
Concentration in Processed Potato
Production .......*.......... a.... ...... 141
Price Cost Margins ...................... 145
Advertising Sales Ratio ,.. ....,.. 146
Change in Capacity ....................... 149
Utilization of Capacity ..,............ 150
Significance of Structural Results ........... 152
market E-ailibriua Results ............ .... 154



Production of Fresh Potatoes ............. 155
Utilization of Fresh Potatoes ............ 156
Utilization of Processed Potatoes ........ 163
Retail Price of Fresh Potatoes ....,,... 166
Retail Price of Processed Potatoes ....... 167
Farm Price of Fresh Potatoes .......... 168
Significance of Market Equilibrium Results ... 168
Suamary of Structural and Market
Equilibrium Results ......................... 171

VII USING THE MODEL .... ........ .............. .. 175

Introduction ....... ..... ...... ... ... ...... 175
Performance of The Model ..................... 175
Reduced and Final Forms ................ 175
Reduced and Final Forms for Structural
Equations ............ ...... ...... 178
Reduced and Final Forms for Market
Equilibrium Equations ................... 190
Synopsis of Structural and Market
Equilibrium Results ...2.....2............... 212
Model Forecasts .... ... .......... .... ....... 214
Theory of Forecasting .................... 214
Structural Equations Forecasts .......... 216
Market Equilibrium Equations Forecasts ... 222

VIII IMPLICATIONS OF MGD3L .................... 229

Introduction .............. ........ ........... 229
Policy Evaluation ................ ........... 229
Policy Simulation Results .................. 233
Conclusion ..,......... ... .... .... .... 265


A DISCUSSION OF DATA 2.....o....,.............. 266

B DAT. SOURCZS ............................ ... .. 20


EZFEE iCSS .. .. ..... .. .. ... ...,.. .... .. ., 9

BIOG EAPEICAL S KETCH .........., .. ...... .,.,..,,....... 315



Tables Pag<

1.1 U.S. Potato Production and Yield by
Region, 1950-80 *,.,...*.......,,,3.233, 3,... 4

1.2 Changes in U.S. Potato Production by
Region, 1950-80 ..,...... ..... o......,... 5

1.3 Potatoes: Total U.S. Production and Prices,
1950-82 .. ,........................... ....... 6

1.4 U.S. Away-Froa-Home Food Expenditures ......... 20

3.1 U.S. Potato Production by Region, 1950-80 ..... 60

3.2 U.S. Per Capita Consumption of Fresh and
Processed Potatoes, 1950-80 ................. 63

4.1 System of Equations for Analyzing the U.S.
Potato Industry ... .......,...*.. ..... .... 79

5.1 Correlation Coefficients .................. 123

5.2 Eigenvectors ....,.. ................* ........ 124

6.1 Parameters and Statistics for Structural
Equations .. ........................... .... 137

6.2 Parameters and Statistics for Market
Equilibrium Equations ........................ 157

6.3 Estimated Parameters for System of Equations
in U.S Potato Subsector .................. 173

8.1 Simulation Results: Impact of Policy
Variables on Selected Endogenous Variables .. 234

8.2 Simulation Results: Impact of 3 Percent
Changes in Policy Variables on Selected
Endogenous Variables .......................... 235

8.3 Siaulatiou Results: Impact of 6 Percent
Changes in Policy Variables on Selected
Endogenous Varables ,,,........,........... 236


Tables Pag-

A.1 Variables For Analysis of Potato Industry ..... 268

C.1 States Comprising Potato Production 3Eeions ... 286

C.2 Principal Components Analysis ............... 287

C.3 Principal Components Equations ............. 290

C.4 Second and Third Stage Estimates .............. 292

C.5 Demand Elasticities ....,............... .... 297

C,6 Long and Short Terms Multipliers .............. 298

C.7 Boot Bean Square Errors .................... 301

C.8 Raw Data for Analysis of Potato
Subsector ........ .. ..... .. ,,........... 302

C.9 Simulated Effects of Specific Policy
Variables ......... ................. 306







4. 1






Perzcnt Changes in Consumption of Processed
Vegetables with Comparison to Potatoes .......

Changes in Consumption of Seven major
Processed Vegetables with Comparison to
Potatoes ................ ........... ...

Consumption of Frozen Vegetables with
Comparison to Potatoes ,,........*.,,......,

Fast Food Establishments by Number and
Sales ..... **.. *...* o... ..,*>**,

Schematic Diagram of Potato Industry ..........

U.S. Potato Production by States, 1950 .......

U.S. Potato Production by States, 1980 ......

Actual, Predicted and Simulated Values for
Potato Processing Plants ..............,,,

Actual, Predicted and Simulated Values for
Conc tration ,..o..........,,.,,,*........

Actual, Predicted and Simulated Values for
Prices Cost margins ,,....................,

Actual, Predicted and Simulated Values for
Advertising Sales Ratio ......................

Actual, Predicted and Simulated Values for
Changes in Processing Capacity .....,......,

Actual, Predicted and Simulated Values for
Utilization of Processing Capacity ..........

Actual, Predicted and Simulated Values for
Production of Fresh Potatoes ..............

Actual, Predicted and Simulated Values for
Utilization of Fresh Potatoes ....o,......





















Figures Page

7.9 Actual, Predicted and Simulated Values for
Utilization of Processed Potatoes ........... 195

7.10 Actual, Predicted and Simulated Values for
Farm Price of Fresh Potatoes ..,......... 197

7.11 Actual, Predicted and Simulated Values for
Wholesale Price of Processed Potatoes ....... 199

7.12 Actual, Predicted and Simulated Values for
Retail Price of Fresh Potatoes ,........... 201

7.13 Actual, Predicted and Simulated Values for
Retail Price of Processed Potatoes ........ 203

7.14 Actual, Predicted and Simulated Values for
Spread of Retail and Farm Price ............ 206

7.15 Actual, Predicted and Simulated Values for
Research and Promotion Tax .................. 207

7.16 Actual, Predicted and Simulated Values for
Utilization of Other Potatoes ............... 209

7.17 Percent Changes in Farm Price and Farm
Production of Potatoes ...................... 210

7.18 Percent Changes in Retail Fresh Price and
Farm Production of Potatoes ................ 211

8.1 Simulated Effect of Transportation Cost
Potato Processing Plants .................. 239

8.2 Simulated Effect of Hinimum Efficient Plant
Size on Potato Processing Plants ............ 240

8.3 Simulated Effect of Advertising Sales Ratio
on Consumption of Processed Potatoes ........ 241

8.4 Simulated Effect of Advertising Sales Ratio
on Price Cost Margins ....................... 243

8.5 Simulated Efiect of Fast Food Consumption
on Consumption of Fresh Potatoes ,.......... 244

8.6 Simulated Effect of Fast Food Consumption
on Consumption of Processed Potatoes ....... 245

8.7 Simulated Effect of Noa-Fast ?:od Consumption
on Consumption of Fresh ?otatoes ............ 247


8.8 Simulated Effect of Non-Fast Food Consumption
on Consumption of Processed Potatoes ........

8.9 Simulated Effect of Women in the Labor Force
on Consumption of Fresh Potatoes ............

8.10 Simulated Effect of Women in the Labor Force
on Consumption of Processed Potatoes ........

8.11 Simulated Effect of Population on Consumption
of Fresh Potatoes ..........................

8.12 Simulated Effect of Population on Consumption
of Processed Potatoes .......................

8.13 Simulated Effect of Income on Consumption
of Fresh Potatoes ...........................

8.14 Simulated Effect of Income on Consumption
of Processed Potatoes ....................

8.15 Simulated Effect of Hinimum Efficient Plant
Size on Concentration .....................

8.16 Simulated Effect of Minimum Efficient Plant
Size on Price Cost Margins ..................

8.17 Simulated Effect of Minimum Efficient Plant
Size on Advertising Sales Ratio .............

8.18 Simulated Effect of Minimum Efficient Plant
Size on Consumption of Fresh Potatoes .......

8.19 Simulated Effect of Minimum Efficient Plant
Size on Consumption of Processed Potatoes ...

8.20 Simulated Effect of Minimum Efficient Plant
Size on Wholesale Price of Processed
Potatoes ......,,, ..... ......, .............

8.21 Simulated Effect of Minimum Efficient Plant
Size on Retail Price of Processed Potatoes

















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



Eugene Jones

August 1984

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

The U.S. potato industry during the past two decades

experienced dramatic economic changes in both its structural

and market equilibrium characteristics. Exogenous and

endogenous forces interacted to shape these transformations

within the industry.

The primary objectives of this study were (a)

development of an econometric model that captured and

explained the adjustments noted above, (b) determination of

the future of the industry, and (c) evaluation of the

impacts of policy variables oa the industry. A sixteen

equation simultaneous system was developed and estimated

using three stage least squares (3SLS). Because high

correlation was noted among many variables, principal

components were derived from the original variables and used

in the first and second stages of estimation.

Identification of the endogenous variables estimated with

principal components was maintained by carrying the

endogenous parts of the principal components as additional

equations in the system.

The empirical results indicate a relationship between

the structural and market equilibrium characteristics, but

not always a direct relationship. Increased demand for

processed potatoes, for example, served not to increase but

to decrease the number of plants or firms processing

potatoes. As firms left the industry, concentration among

the top four potato processors increased. The results

further indicate increased advertising intensity with

increased concentration. Empirical results for the demands

for fresh and processed potatoes indicate a strong

preference by consumers for processed over fresh potatoes.

The retail price of fresh potatoes has little impact on

demands for either fresh or processed potatoes. Demands for

both fresh and processed potatoes, on the other hand, are

influenced by the price of processed potatoes.

The quantified model was used to both forecast supply,

demand, and structural changes through 1990 and to study the

sensitivity of the industry to specific exogenous changes.

Explicit policy issues were identified and simulated.



Pro.-ict C characteristics

Agricultural commodities are bought and sold in

markets. These markets are in essence a system of

communication for transmitting consumers' 1"minds to

producers and producers' supplies to consumers. During this

communication, market prices are established. which, in turn,

become the primary m.iuium of communication between producers

and consumers. A siijle price may be establish-d for those

commodities produced and sold in closely defined j-eor-il1ic

areas where buyers and sellers are in ia mdiate

communication with each other. Multiple prices, however,

most after. characterize those commoditie. which are bought

and sold in widely dispelrsed rj-o.caphic areas where intimate

communication between buyers and sellers is absent.

Potatoes, the commodity of interest in this study, are

characterized by multiple prices with respect to time, place

and form. Given a perfect market for potatoes, that is,

"all buy-rs and sellers in it have perfect knowledge of

demand, supply, aul prices and act rationally upon that

knowlAd:r: the price differential in time, place, and

form would be equivalent to the corresponding differences in

costs" (Shtephard and rutrell, p. 24). ;iven that perfect

markets for potatoes do not exist, one would expect some

disparity between price and production cost. Price-cost

disparity may be due to incomplete or inaccurate

information, the existence of market power by buyers and

sellers, preferences on the part of buyers for particular

commodities, risk and uncertainty regarding the future, and

so forth. To the extent that price-cost differentials

exist, economic resources are allocated in a nonoptimal or

inefficient manner.

Price discovery with respect to product form will be of

considerable interest in this study as fresh and processed

potatoes are analyzed. Ideally, one would expect

competition in the potato market to equate the net value of

processed potatoes, less processing cost, to the price of

fresh potatoes. This ideal situation may be inhibited by

several market imperfections. Among these are processors'

market power, potato growers' incomplete knowledge of

processing cost, negotiated contract prices for frcsh

potatoes, advertising persuasion, and consumers'

susceptibility to fancied product differentiation.

D-gand and Supj2Iy

The supply of potatoes in the U.S. is limited primarily

by demand constraints. Potato producers have continually

had to restrict their acraaje as technological advances

have greatly augmented yields. Average yields per acre have

increased from 153 hundredweights (cwt) in 1950 to 262 cwt

in 1980, an annual increase of 2.3 percent; acreage in

production has declin-d 1 percent annually since 19=ln.

Total consumption cf potatoes also has increased by an

annual rate of 2.3 percent since 1950. And while total

increases in demand an1 supply have been moderate, there

have been dramatic shifts in supply areas and supply


Historically, potato production in the Northeast has

been amonj the highest in the nation, with Maine being the

primary producing state. In 1950, for example, 29 percent

of U.S. potato production occurr-d in the Northeast; aver-gje

yields were 212 cwt as opposed to 197 cwt in the Pacific

northwest the major compt-titor of the Northeast (Tables 1.1

and 1.2). 3y 1980, average yielic in the Northeast had

increas-ed to 23b cut while its production amount-! to only

14 percent of U.S. total. By contrast, avera-je yields in

the Pacific Northwest nad risen from 197 cwt in 1950 to 331

cwt in 19.0; production as a rercentaqg of U.S. total had

risen from 17.6 :,-rcent to 47 percent. These relative

changes to'jether with those occurring in other regions

increased U.S. total production from 259. 1 million cwt in

1950 to 302.8 million cwt in 1980. Tnis 1980 production was

coisidierably below 1979 production of 342.5 million cut, and

even turthlir below 1978 production of 365.2 million cwt

(Taole 1.3). The low potato production of 1980 couple, with

an inelastic demaad for potatoes generated net farm income

to potato producers far in excess of the income generated

the previous two E-ars. This higher income result2,i from



< w
c o

eO -


Ll >

D e
c 0

0 o

c< e

. 0


C, ei

N a, Ir'. 0 f N 1

M. Ar ar 0 0n A % 0 tD 0 f
-a 0 '.N I
i 0 r4 N rq fs M0
SM % CP 0 at N

' c N N c N N C I

N mN N e Nq N I

D- a D I
S p N N


- BN I iN D 0 N D t

N N fs I
N N N N UIt t
p* m n N r

f I

B t N U & \

N r~-- h~ hlI

~CI~Dlc c~CDI
PrO N CbC ~ I
rC~VIO~lc P I I

NrhrO P~rI
r NID ~ ~imNI

~P -~~' NcI

I 0 ao 4 I C
I 41 4 j w 4 0 1 U

0 I AI ) I ha 0 00
0 .00 4 h c

' u 'v 0 Ad 4:1 u .4 r iU
S t 0 .4 0 m B & 0
= I E E = C 0 ) I



Oa I

0 I
00 I
00 I

I- -

U 4

0 .


O u

Srr 9


c Io
I ) I I


Io I I

I I Ul IC 1 I 1

m I CI I I
I LI I I 0
S U. IO I I *

1 1 I
aol o L

H1 4 1
I 1 I I
I l l -I I

EI 1 I I a
I U I C 1 L

0I 1 II CL
S I OI 1f Io C

0 I 0I I

1 1 1 0**, 1

W I t 0 1 0 0 1 C N D C) l
di %i D 1 4 4 I 6
S I %D -O I 0

0 I I' I IN
SI I I I a
o l I I 4

010)C rz 1
c* 'N -) .4I I I-
NI I =I I I 0

o~ 0 I I O eo tN ar 0n y 0 I 0

0 I o I 0 I
. I I & O I 1-1
o I I I U

OCD 01 C0D0 C C .0 0
't I b l I I WI
0 c I I 01
SI 2 0 0 0 C N ury 0
B te *<4o

aN 4 4I o1
I | Ca cI a c a q ) I VI

a I 01 N C0 t 0l
S I il I oI
C I I Gl

0 I 0I U1
.. I oc 0) 1 c Ln or T- sr O

EU I z0 *M 0 0 I U0 -4
^ 01 a01

-B *) ,l. ) I l

U I I o 4 l
* I ) 0 I al

) I 4- > ) \ I l a
a- I 8 4) O U L I 1
3 I U0 0 ) ) -I I) 0

0) I 0 -+ .C z 4-a '- a -4
- I 4 ? 0 *-M *\ 4 I

r* I0 3 (0 a D C) ^ a 0
8 I 2 a 3 r 3 U U E-C

Table 1.3--Potatoes: Total U.S. Production and Prices, 1950-82

Percentage Change from Previous Year





average Farm


Source: Vegetaole Outlook and Situation, USDA, July, 1983.



-8. 1
2. 5
-11. 1
-3. 1


-59. 2
22. 8
-1. 1


average potato prices of th.55 Fer cwt in 1980 as opposed to

13.40 for 1978-79 (Table 1.3).

Just as .rliWmatic as the supply shifts from the lTortheast

to the Pacific Northwest are changes in utilization of this

supply. Processed potatoes (chits, dehydrated, frozen,

etc.), while comprising 6 percent of total potato

consumption in 1'150, rose to 61 percent in 1980. This

represents an increase in per capital consumption of

processed potatoes from 6.3 pounds to 64.4 cunds, fresh-

weight -, uivalent. By way of comparison, fresh potato

consumption declined dirinrj this period from 100 pounds to

51.4 iouils per capital.

From 1950 to 1965, increases in per capital consumption

of processed potatoes just offset declines in per capital

consumption of fresh. Since 1965, per capital consumption

increases of procLssel potatoes have more than compensated

for the decline of fresh consumption, yielding a steady

increase in total per capital consumption. Much of this

increase can be attributed to tne growing consumption of

frozen potato products, primarily frozen french fries.

Whereas -cr zanita consumption of frozen potatoes amounted

to 14.3 pounds in 1 65, this total had increased to

approximately 40 pounds in 1980. Frozen french fries

comprised 83 percent, roughly 33 pounds, of this 1980 frozen

potato consumption.

It should be emphasized that some growth of the potato

industry occurred even during the 1q50-65 period when


increases in per capital consumption of processed potatoes

were virtually offset by declines in per capital consumption

of fresh. Growth occurred as total quantities consumed

increased with population growth. Continued population

growth coupled with more than offsetting quantities going

into processed potatoes since 1965 has accelerated the rate

of growth of the industry.

Structural Attributes within the Potato Market

Structural attributes are those forces or conditions

either inside or outside a market which influence the

behavior of firms anL. ultimately market p-rformance. Since

demand and supply are the crucial elements of a market,

structural attributes of significance must influence these

elements (Cochrane, 1957). Tais section will focus on those

attributes within the processed potato market that are

influencing or can influence the performance of the potato


A comprehensive list of structural attributes relevant

to agricultural commodities has been provided by Cochrane

(1957). Other authors, Bain (1942), Lanzillotti (1960), and

Clodius and Mueller (1961), have also provided fairly

comprehensive discussions of structural attributes. This

discussion, however, follows Cochrane most closely.

According to Cochrane, seven structural attributes are

pertinent to agricultural markets: (1) number of

operators, (2) product homogeneity or product

differentiation, (3) product durability, (4) bulkiness of

product in relation to its value, (5) ratio of overhead

costs to variable costs, (6) continuity and length of the

production process, a;. (7) adeuiiacy of grade description

for buyiu. and selliarj. Each of these attributes will r

counidered tar its relevance to the potato industry.

Cochrane (1957) argues that the fever the number of

operators, the greater tl.b likelihood ci market collusion.

Firms recjnize that one firm could r-du-e its price and

affect the sales of others. Consequently, rccoquizir.c their

interdlp- nlience, they are likely to collude (either formally

or tacitly) on price and the probability of this collusion

increases as the number of operators decreases. Fdii (1942)

and Lanzillotti (1960) argue that other factors such as the

number and d-jre? of concentration among buvyrs and the size

distribution of firas in relation to numLbr are also

sigiiiiicant in determining the likelihood of market

collusion, Clodius and MuWiller (1961) emphasize that the

interdependence of structural attributes cannot be

cona.ser'i iiiependently. Tr.ey aLJue that firms, when few in

numbers, are likely to differentiate their product and use

non-price competition instead of price collusion.

As suggestive of market influence, the number of

operators in the potato industry has declined continually

since the 19JO's. This loss of operators is reflected in

the loss of opratihj plants, although there may not be a

one-to-one correspondence. The number of potato chip plants

has declined from 430 in It0 to roughly 177 in 1980.


Fiozen plants declined from 97 in 1964 to only 19 in 1980, a

decrease of 80 percent. Reversing tl'is trend is the number

of dehydrated plants; available statistics show an increase

from 28 in 1964 to 33 in 1980. This increase probably has

resulted from the entry of potato chip and frozen processors

into potato dehydration as they recognized the feasibility

of processing dehydrated products from potatoes otherwise

unsuitable for chips or frozen products.

Reductions in the number of plants and operators are

believed to have influenced market power within the

industry. Frito-Lay, the largest processor of potato chips,

reportedly increased its market share from 20.9 percent in

1970 to 31 percent in 1982 (Warwick, 1971; Snack Food,

1982). Wise, the second largest processor of potato chips,

purchased several smaller chip processors in 1980 and these

acquisitions are expected to increase its market share.

While the national market share of Wise is not known, it

supposedly controls more than 50 percent of the market in

the New York metropolitan area, and a larger market share

than Frito-Lay all along the east coast (Warwick, 1171;

Snack Food, 1982). Thle frozen potato market is equally if

not more concentrated than the chipping market. One firm

alone, Ore-Ida, currently controls 54 percent of the retail

market for frozen potatoes. ()uick Frozen Foods, May 1082).

And although the retail market represents only 24 percent of

the total market for frozen potatoes, this same firm is a

major supplier to the institutional market. The number one


and three frozen potato processors, J.R. Simplot and Lamb-

Weston respectively, control a major share of the

institutional market.

As suljested in the theoretical framework d&velopLId rv

Clodius aLu Mueller, potato processors, recognizing their

numbers acird market shares, have ostensibly resorted to

product difirentiaticn as a means of increasing their sales

aiJ profits. This product differentiation undoubtedly has,

in eiiect, become a form of conduct to further reduce the

number of operators. -,cause there is limited buyer

concentration in the retail market, consumers apparently

have little couatervaili ng jower. While the number of

potato processors may be relatively large, it seems that

their size distribution may be the most imluortant

determinant of their conduct and performance

(Lanzillotti, 1'6) .

troiuct ii terentiation in the d r ot ato

industry could have chang'j-i the type of competition from

price to non-price and greatly augmented the intensity or

this non-pri:e competition. This would be consistent with

the theoretical argument advanced by Cochrane. The

ex[.nsion of Ore-Tda in the retail frozen potato market is

an excellent example of this non-Trice competition. Ore-Ida

accounted for 28 percent of retail potato sales in 1467-r6;

by 1981 this share had increased to 54 percent as the

company introduced many new products and greatly expanc..

its advertising expenditures on existing products.

Althuuqh product differentiation undoubtedly has

enhanced the de'jree of non-price competition in the

processed potato industry, its traditional role has been one

of maintaining excess profits in industries through its

ability to act as an effective entry barrier. Consumers

develop preferences for particular brands and continue their

consumption of these brands even when other lower priced

brands of similar quality are available (Comanor and Wilson,

1967). Monprice competition, of course, is a method firms

frequently use to maintain this consumer loyalty.

Cochrane (1957) argues that the elasticity of supply of

an agricultural commodity is influenced by its durability .

of all vegetables, potatoes are probably the most durable.

Fall potatoes, for example, can be stored from eight to ten

months. This quality of potatoes should have enhanced

revenues for raw pctato producers and contributed to more

efficient production scheduling for processors. Frozen

potato processors, because of their large size and loni

operating periods, probably benefit the most from the

storability of raw potatoes. Processors of dehydrated

potatoes should also benefit from the durability of fresh

potatoes but to a lesser extent than frozen processors

because of shorter production periods. Potato chip

processors, being generally small with limited production

periods, probably r-ceive the least benefit from the

storability of potatoes.

The durability of the final products produced by

processors influences the elasticity of supply for these

products and ultimately processors' performance. Ochydrate-

products have an almost unlimitc3 shelf life. Frozen

potatoes also have an exteii--d life but storage cost is a

more significant factor than for dehydrated products.

Potato chips have a limited shelf life and receive excessive

breakage from handling and transporting. These product

characteristics should yield a rather elastic supply

rcspcnse for dehydrated potatoes .nd somewhat less elastic

restusipes for frozen potatoes and c iips.

The weight of fresh potatoes in relation to their value

has had a major influence on the current structure of the

processed potato industry. Because of the weight-losing

properties of potatoes, processors of frozen and dehydrat-d

potatoes have located their plants proximate to major

production areas. This minimizes transportation cost as

processed products have higher value/we ratios.

Increasing these value/weight ratios expands the geographic

area over which these pro'i'cts will be utilized, influences

the number or processors and ultimately influences the

elasticity of demand facing processors (Cochrane, 1957).

Potato chip processors have received limited if any

benefits from the weight-losing properties of potatoes.

This results because minimum pruiouct distribution costs for

potato chips occur from plant location near larjg population

centers as opposed to major potato producing areas. Plant

location near producing areas would result in excessive

product breakage as chips are transported to consuming

areas. And even if chips did not experience breakage in

transport, their bulkiness relative to value would still

make plant location near production areas nonoptimum unless

these areas corresponded to large population centers.

According to Cochrane, the ratio of fixed to variable

costs is a net determinant of supply response as well as the

slope and position of an industry's supply curve. This

ratio, although currently unknown, is hypothesized to be

increasing for the processed potato industry. This

hypothesis arises from the rate of exit of smaller firms

from the industry. Larger firms, having accelerated their

rate of adoption of new processing technology, may have

raised the minimum -fficient plant size beyond that

attainable by smaller firms.

The length ot the production period in the processe-

potato industry is directly proportional to firm size.

Larger firms operate more hours per day and more days per

year than other firms in the industry. The supply curve for

processed potato products is therefore more elastic as a

result of larger firms extended production periods

(CochiLane, 1957). It is of interest to note that larger

potato processors generally face a constant long run supply

price for raw potatoes. Per unit variable cost remains

fairly constant as production increases because larger

processors purchase a major proportion of their raw potatoes


through contracts at fixad unit cost. Lon'-run supply price

increases significantly only when processors urnerestimate

requisite -uantities and have to make sizable purchases on

the open market at higher than contract prices.

If the seven major structural attributes iliscus~Pd by

Cochrane, perhaps none is more applicable for the process--

potato industry than the adequacy of gralinj. Each

processed potato product, with slight exceptions for some

dehydrated products, is best produced from a specific jgri]e

and quality of potato. Contract purchases have arisen

partly as a result of processors' need for sp-citic quality

characteristics in their potatoes, nnc- the proc?-sed

products are pcoduccd, .-jradin7 ceases to be an important

structural attribute as grading becomes standardized and

consumers make purchases Ly package size.

Strc)CturJl Attribut-3 Outside t.L Potato Mar?.et

Several forces external to the potato industry have had

and continue to have significant effects on the industry.

Uncertainty exists as to whether these forces are pulling

the industry, internal forces are pushing the industry, or

both sets of forces are interacting to pcopel the industry

forward. This section will provide a brief discussion of

the following external forces which appear to be significant

in altering the structure of the potato industry: (1)

changing consumers' tastes, (2) growingg percenta-g: of women

in the labor force, (3) increasing population, (4) rising

income, and (5) increasing avay-from-howme fcou consumption.


Changing consumers' taste is believed to have played a

role in shifting potato consumption from its fresh to

processed form. Quantitative measurements of taste being

next to impossible, only qualitative discussions are most

often provided. Observations on potato consumption seem

suIggestive of changing consumers' taste. Per capital

consumption 2f fresh potatoes in 1910 amounted to 198

pounds; it declined to 100 pounds by 1950, and to 56 pounds

by 1980. Consumption of processed potato products, however,

increased from an unmeasurable amount in 1910 to 6 pounds

per capital in 1950, and then to 62 pounds in 1980.

These changing consumption patterns are not in

themselves indications of changing demand as supply also

could have changed. Previous studies though suggest that

there have indeed been changes in demand. Scott and Mumford

(1949) attribute these changes to (1) a rising standard of

living associated with increasing incomes, (2) increase-

availability of other vegetables and a desire for variety in

the diet, (3) a lesser need for high-calorie foods in

present day occupations, and (4) progressive urbanization of

the U.S. population. Gray, Sorenson, and Cochrane (1954)

ascribe these consumption changes to (1) declining

immigration of people accustomed to heavy potato

consumption, (2) the flight from calories explanation, and

(3) variations of a civilization argument which suggests


that modern housekeppers prefer convenience foods re-uirii;g

little pre-aritions, and modern housewives with outside jobs

have little time for meal preparation.

-Secause much tie has elapsed since the above studies,

fi'e (1967) suggests that the enumerated factors are less

applicable for the period since the early 1960's. Two

additional factors, in Lis opinion, have had a more

pronounced effect: (1) the discovery and dissemination of

factual information on the nutritional value of potatoes,

and (2) the rapid introduction of processed potatoes.

These two factors along with those enumerated by the other

authors would probably prove significant in explaining

changing ptato consumption especially before the decade of

the 1970's. Since 1970, four factors enumerated in the

first paragraph of this section--growth of women in work

force, population, income, arni away-from-home food

consumption--probably would provide] more insight into

changing potato consumption.

The growing ijercentage of women in the work force has

serve .i to alter considerably the market basket of goods

cons'iiu by households. Market baskets contain more

convenience foods as households have less time to devote to

meal preparation. Since potato processors presently perform

many of the services traditionally performp.i by women, one

would hypothesize a relationship between consumption of

proczsesI potatoes and women in the labor force. Ava ild.le

data support a direct relationship. Fr-m 1950 to 1980,


women in the work force increased from 29.6 percent to 42.4

percent; consumption of processed potatoes increased during

this period from 6 percent of all potatoes consumed to 61

percent. While not all this increase has been due to women

in the labor force, recent studies suggest that they have

accounted for a significant proportion of it (Beilock, 1981;

Cardwell and Davis, 1980; Estes, 1979).

Population growth could have an impact on the potato

industry simply by providing an expanding market. During

the 1950-65 period, this factor seems primarily responsible

for increased potato consumption as par capital consumption

remained relatively constant. Many potato processors are

expecting population growth to provide a continued source of

industry expansionn as they focus their advertising messages

on attracting new users ot their products. Available 1981

statistics suggest that 40 percent of American househclls

are nonusers of the major processed potato product, frozen

potatoes. If compardblE_ rates exist for other products,

current nonusers together with prospective users from

population growth could represent significant growth

potential for the processed potato industry.

The effect of rising income on potato consumption has

not been clearly established. A study conducted by the U.S.

Department of Agriculture (1955) found a positive

relationship between potato consumption and rising income

for low-income households, but a n ejative relationship for

high-income households. Shuffett (1954) found an income

elasticity of demand for potatoes of .42. "lost studies,

however, have round a very weak cr no income effect. These

relationships have been established primarily for fresh

potatoes as the studies were conducted before the

introduction of the many Irocessed forms. In a more recent

stuiy, Estes (1979) found an income elasticity of demand for

processeJ and fresh potatoes of .67 and .11 respectively.

With processed potatoes becoming an increasing proportion of

total consumption, risiuj incomes could have greater

significance for the potato industry.

A close parallel to the income elasticity of demand for

potatoes is the income elasticity of demand for away-from-

home food consumption. Of particular significance to the

potato industry is the rate of increase of food consumption

at fast fool establishments. Consumption of processed

potato products, particularly frozen french fries, is

believed to be fir more prevalent at fast food than non-

fast food establishments. Available statistics indicate

that cha!i s in food consumption at fast foo.] establishments

have greatly outipJcel that of non-fast food establishments.

Food consumption at fast food establishments incr-ased

between 1972 and 1977 at an annual rate of 10.3 percent as

compared to 1.9 percent for non-fast food establishments

(Table 1.4). Fast ::-od t:xp.=nditures in constant 1967

dollars have increased from $1.1 billion in 1958 to $10.4

billion in 1977, an increase of over 800 percent. By way of

comparison, non-fast food expenditures increased during this



CO (


( U



0 rf




period' from $12.1 billion to $17.3 billion, an increase of

43 percent.

Concept of Performince

Market structures within the U.S. are quite diverse,

ranging from near monopoly to almost pure competition.

Within taese structures, there are recognized levels of

performance. However, little knowluJdge exists as to how

various structures impact performance and, in turn, are

impacted i.y performance. Similarly, disagreement exists on

the recognition and measurement .Jf p-rformance. As usei in

this study, pectormance relates to (1) stability of

margins; (2) profitability :z production, (3)

progressiveness of the industry, and (4) growth of

rL-oduction. Applying these concepts to the potato industry

requires the articulation of Ijiiantitative measurement


Crit~trir for Analyzin P-rformance

In addition to the aforementioned performance measures,

other measures for evaluating p-rformance also have been

propos-.2. Some of these incluJe (1) differences of capacity

utilization from available cap.icity, (2) percentage of total

market held by top 4, 8, or 10 firms, an,. (3) rate of change

of expenditures on advertising and research and development.

While each of these measures will be an integral part of the

specified econometric umoiel, they will be used only in

conjunction with the previously specified performance


concepts to evaluate the overall performance of the potato


Stability of marjins as a performance measure may appear

counter to the ostensible profit maximization objective of

firms, but its use here is intended to measure the extent in

which larger rirms pass on cost advantages in the form of

lower prices and subsequent increased output (Collins and

Preston, 1959). High variability in price-cost margins

would suggest abnormally high prices during periods of

industry adjustments. Stability of margins, on the other

hand, would suggest profits, but the absence of supernormal

profits. Thus, the variability of margins is one measure by

which performance of the potato industry will be evaluated.

Althniyiqh a positive price-cost margin would indicate

profitability, an issue of concern here is the extent to

which profits have bien enhanced by the structure of the

industry. With a taw large processors controlling dlrg*-

perczntages of the market for all processed forms of

potatoes, it may be hypothesized that these processors'

market power would lead to some discrepancy between price

and cost. It may also be hypothesized that concentration

among institutional oo] establishments Frovides some

countervailingr power for those products which are market.rd

primarily through institutional establishments. Price,

according to general economic theory, would be determined in

accordance with the relative bargaining strength of

rsp-ctive parties in the absence of collusion. If, as a

minimum, tacit collusion exists between processors and

institutional buyers, consumers most likely would be

subjected to hiqir-r prices than those which wouill prevail

under countervailing barjainirg. Lackin., information on the

bargaining power or re-spfctive parties, the influence of

concentration on mar.jins will be us-d as a measure of the

extent to which market power has enhanced profits.

A pri.)jessive industry may be defined as one which not

only meets the demands of consumers but also influences or

shapes these i-maads. An industry may use such methods as

product development, pricing techniques, advertising, and

product promotion. As quantitative data on many of these

efforts are not available, advertising e(pt-nitures will be

used here to evaluate the progressiveness of the potato


Analyzing growth as a performance measure pLobirly will

require some relative comparisons of the potato industry

with other industries. An alternative method would simply

measure the growth of the .rinustry in terms of its provision

of the products demanded by consumers, that is, product

availability. If the industry has made available many

products with efficient allocation with res-ect to time,

place and form, then the growth rate achieved should closely

parallel optimum .jrcwth.

Gei:eral Pcrforuianct of Potato Industry

Potato production in the U.S. has increased

considerably during the past three decades--rising from


259.1 mil. cut in 1950 to 325.7 mil. cut in 1970, and then

falling to 302.8 mil. cwt in 1980. This increased

production has come primarily from tremendous increases in

yields per acre; total acreage in production has actually

declined from 1,679.9 thousand acres in 1950 to 1,154.5

thousand acres in 1980, a decrease of 32 percent.

Utilization of this potato production changed considerably

during the 1950-90 period. Whereas nearly all potatoes were

consumed in their fresh form in 1950, less than four-tenths

were so consumed in 1980. This shifting of consumption has

greatly altered the structure of the potato industry.

Per capital consumption of fresh potatoes has declined

continually since 1950. Despite this continual decline, the

rise in consumption of processed potatoes since 1965 has led

to an overall increase in per capital consumption of total

potatoes. Besides the per capital consumption decline of

fresh potatoes, there also has been a decline in absolute

consumption, a decline from 242.8 mil. cwt in 1950 to a

1980 level of 97.2 mil. cut. The consumption of processed

potatoes increased during this period from 16.3 mil. cut to

152.9 mil. cwt.

Potato chips were the dominant processed potato product

conslimed until 1965. Since 1965, the volume of potatoes

processed into frozen potato products as well as the per

capital consumption of frozen potatoes has greatly exceeded

that of chips. RB-flicting back it may bIe recalled that 1965

was approximately the time that total per capital consumption

of potatoes bijaa to reverse its declinir.n trend. Thus,

despite the high utilization of potatoes in chips for

earlier ..riods, it should be clear that rising consumption

of frozen potatoes has played a significant role in

generate.,, the growth needed to reverse diminishing per

capital consumption ci potatoes.

Next to frozen french fries and potato chips,

dehydrated potatoes (largely flakes and granules) constitute

the third largest component of processed potatoes. In fact,

these three components have consistently comprised from 75

to 87 percent 'f all process1-i potatoes. While the quantity

of ircsh potatoes channelVI into dehydrated potato products

has risen considerably during the past three dec aes,

d-nydrated products as a percentage of all processed

potatoes have shown no consistent pattern ot change.

DehydLated prLolucts rose from an unmeasurable amount of all

processed potatoes in 1950 to 17.1 percent in 1960, to a

high of 23.1 percent in 1976, ind then declined to 18.4

percent in 1980. Since 1965, per capital consumption of

dehydrated products has increased from 7.0 pounds in 1965 to

a high of 14.5 pounds in 1974, decline=1 to 10.1 pounds in

1976, and then rose to 11.5 pounds in 1980.

Canned potatoes have always represented a small

percentage of processed potatoes. Total and pvr capital

consumption rose continually, although minimally, from 1950

to 1974. Since 1974, per capital and total consumption have

remained relatively constant. Starch and flour, the other

processed potato products, are no longer significant

processed potato products. The amount of potatoes used in

the production of starch and flour has declined from a high

of 20 million cwt in 1961 to a 1980 level of 2 million cut,

a decline of 90 percent. These changes suggest that starch

and flour are residual products with production highly

dependent on prices. This relationship has been partly

confirmed by Hee (1967) as he found the elasticity of demand

for potatoes used in starch to be 1.01. This elasticity

estimate suggests that starch producing plants are operated

below capacity during periods of low production and

relatively high potato prices, but these plants approach

normal capacity as production increases and prices fall.

Despite fluctuations in the patterns of consumption of

some processed forms of Ictatoes, continual increases in the

consumption of frozen potatoes have served to greatly expand

the growth rate of processes potatoes relative to that of

seven major processed vegetables (Figures 1.1 and 1.2).

These vegetables are asparagus, cabbage, cucumbers, snap

beans, spinach, sweet corn, and tomatoes. Although the

percentage of these vegetables going into processing has

always exceeded that of potatoes, processed potatoes as a

percentage of all potatoes are continually increasing. The

processed percentage of the seven major processed

vegetables, on the other hanr, is decreasing. Stated

differently, processed potatoes seem to be more readily

2.75 1




2I. i-
2 -1


I 75.


I 1.25-

1 ,-Ot

19 -;0



Figure 1.1.

Percent Changes in Consunptian of Processed Vegetables
with Comparison to Potatoes




x ----






Q .7

li. 4I

0. 3






Figure 1..2.

Changes in Cbnsumption of Seven Major Processed
Vegetables with Comparison to Potatoes



I I T r 1 I I I I I1 I I I I I I



meeting the dena.ds for convenience foods than other

procrsss-d vegetables.

Relative gains of pro:-Ssad potatoes on other

v-getables have resulted largely from the gains frozen

potatoes have made on other frozen vegetables (Figure 1.3).

Leading the growth of frozen potatoes are frozen french

fries. Presently, frozen fries account for 82 percent of

all frozen potato products; they account for 53 percent of

all frozen vegetables; and they are consumed at more than

twice the rate of tUe seven leading frozen vcgj-tahles.

These vesi-etables are broccoli, carrots, corn-on-cob, cut

corn, green beans, green peas, and spinach. Frozen rench

fries in 1980 constitute 30 per-cent of all frozen retail

vegetables and 65 percent of all institutional vegetables.

Commercial development of frozen french fries bejan in 1q47,

and as the precaiing discussion su-jests, they have since

shown remarkable growth.

?troLlem Sfatement

Structural changes in the U.S. economy since the early

19503s together with some specific to the potato industry

have had and continue to have significant impacts on the

potato ii-dustry. Chan'js include (a) an increasing dEsj~ij

for away-frooi-home foD0 consumption, (b) a changing labor

force composition, (c) a rapidly expanding fast food

industry, (d) a rising pcr capital income, (e) a reduction

in household size, but an increase in number, (f) a growth

o: population, and (g) a change in consumers' taste.



6000 -


L i
L -r


f j
F 1

D 2000-


1 -0

~i ~ / .7 -. -3


Y -






Figure 1.3. Cbnsumptian of Frozen Vegetables with C~mparison
to Potatoes






Changes sp-cific to the potato irl-ustry are (a) a rallin

demand for fresh potatoes, (b) a rising demand for

proc-ssel potatoes, aii (c) a spatial redistribution of

production, both raw and processed forms. As a result of

these changes, there have been (a) a si-nificant shift in

the utilization of potatoes from their fresh to process.I

forms, (C) a rapid development of process] potato

products, (c) a decrease in the number of potato

processors, but an increase in their size, (d) a

significant increase in advertising expenditures for

proc.ssc~d potato products, and (e) an expanding expcn.liture

of funds for potato research and J-velopment.

Alth-'u.jh the above are well recognized, little

knowld,: exists as to how these external and internal

forces have interact-.1 to affect the structure and

performance of the potato industry. Changes external to the

industry could have precipitated as well as continued the

changing structure cf the industry. Unisr this scenario,

the inJu;try shouia stabilize when these forces stabilize.

Alternatively, it forces internal to the industry !,ive been

the primary determinants of chdrnje, then the long-run

stability of the industry becomes dependent upon the extent

to wv'chl these forces continue their pattern of chan-?.

Structural changes, although important, are secondary

to identifying tand the performance of an

industry. Performance je-nerally will have some influence on

the ultimate structure of an industry. For the potato


industry, performance is important as it relates to three

separate but not necessarily distinct groups. These groups

consist of (1) raw potato producers, (2) consumers and,

(3) potato processors.

Raw potato producers, having an increasing percentage

of their potatoes purchased through contracts, are less

dependent on the direct operation of the forces of demand

and supply in determining. prices and revenue.

Theoretically, contract purchases should have reduced price

risk to growers and, under certain conditions, increased

their cr-Aitworthiness as these contracts may be used as a

form of collateral in securing business loans. This reduced

risk ani increased creditworthiness could come at the

expense of total revenue. 3uch trade-offs must be analy??d

from the perspective of the overall performance of the


Changes in the p[ tato industry have provided consumers

with more and a greater variety of products from which to

choose. Some economists would arque that this has to

represent an increase in welfare because more is better.

However, if we equate consumer welfare with consumer

surplus, it may be shown that more and a greater variety of

products are not cyuivalent to a gain of welfare for

consumers when these products are supplied at prices in

excess of their costs. Consequently, analysis of price-cost

margins is one standard by which performances of the potato

industry will be measured.


Potato processors, like raw potato producers, should

have experienced some reduction in risk from contract

purchases as this metnocd of acquisition assures available

supplies for more efficient production scheduling. Such

risk, and therefore cost, rcducticn could be reflected in

prices paid for raw potatoes, prices laid by consumers for

processed products, processors' profits, or some combination

of these elements. Evaluating the impacts of contract

purchases on various industry participants may be best

performed by determininrj the distribution of income

resultin- from the changing structure and performance of the

potato industry.

In summary, the problems involve the determination of

those factors which have caused the transformation of the

potato in oijtry, and then measuring the imlpacts of these

factors with r--s:ect to time. For example, it may be

hypothesized that variables relating directly to demand and

supply will have a more immediate effect on the industry

than those re-iting to structure and performance. Thus, an

intriguing aspect of this study will involve separating out

the time dimension of the imFicts of both endogenous and

exogenous factors. A related problem involves the

determination of the future of the industry. This problem

will be addressed by deriving the r2duc,-d form of the model

and taen shockiiu the system to determine the impacts of

these shocks.

H yotheses

1. Because or structural changes external to the

potato industry, the industry has evolved from an elementary

structure with nearly all potatoes being consumed in their

raw form to a very complex structure with processed products

dominating consumption.

2. Expenditures on research and development by large

potato processors have led to the discovery and adoption of

new processing technology. hIis technology has raised the

minimum efficient plant size beyond that attainable by most

small firms, thereby causing their exit from the industry.

3. The potato industry, recognizing and responding to

demands emanatirj from external changes, has undergone

significant internal changes in accommodating these demands.

Trh&es internal changes have altered the structure and

performance of the industry; continued interaction of

structure and performance should serve to further shape the


4. The changing structure and performance of the

potato industry have altered the distribution of income

accruii.j to various participants within the industry.

5. Determining the direction and extent of future

external change should allow for the simultaneous

determination of the future structure and performance of the

potato industry.


AIdny factors have contributed to the changing structure

and performance of the potato industry during the past three


decades. The primary objective of this study is to identify

these factors and evaluate their relative impacts. A

corollary of this objective, of course, is to evaluate the

impacts of industry charnjes on consumers and various

participants within the industry.

Specific objectives necessary to the accomplishment of

the primary objectives are

1. Development of a theoretical model which

illustrates how the potato iiiustry could be affected by

various exoqgenous andl endogenous factors.

2. Development of the empirical counterpart to the

model in (1) which luantifies the effects of both exogenous

aiid endojenous factors.

3. Interpretation of the empirical results in terms of

the impacts industry char.;es have had on consumers and

industry participants.

4. Zviluation of the ultimate performance of the

industry in lijht of the developed theoretical model.

!Ot hodoloIy and Scoj.-

Sixteen endojenous and twenty exogenous variable are

identified' to explain the changing structure and performance

of the uptato industry. These variables are incorporated

into an econometric model which allows for a simultaneous,

two-way relationship between the structure and performance

variables. For example, the structure variable

concentration may influence the performance variable

profitability while profitability simultaneously influences

concentration as well as other structural variables.

This study encompasses all raw potatoes and all

processed forms. Processed forms have been grouped together

for estimation purpuoss. For example, no attempt will be

made to estimate the degree of concentration in the chipping

industry separate from that in the frozen potato industry.

Aggregation of dita is not expected to distort the results

as all product forms generally have exemplified the same

trends in terms of such factors as concentration, firm size,

number of firms, and advertising expenditures.

Once the econometric model has been estimated, various

simulations will be performed to illustrate the industry

response to certain economic shocks, whether endogenous or

exogenous. Also, the model will be used to provide

projections of changes over time. one area of particular

interest is to simulate the industry response to continued

food consumption away-from-home.

Subsequent Chapters

A review of literature applicable to the potato

industry is provided in Chapter II. This review is followed

by a brief overview of the potato industry in Chapter III.

A theoretical model, delineating the exogenous and

endogenous variables which most likely impact the industry,

is developed and presented in Chapter IV. An econometric

model is developed and estimated in Chapter V. A discussion

of the econometric techniques and problems associated with

the model is also provided in this chapter. The econometric

results are discussed in Chapter VI in light of the


theoretical model developed in Chapter IV. A discussion of

the uses for this type of simultaneous equation system is

provided in Chapter VII. The study concludes with a

discussion of policy implications in Chapter VIII.



This chapter will provide a discussion of previous

studies of the potato industry for the purposes of (1)

shedding insights on the approaches and results of other

authors and (2) providing a foundation for evaluating the

contribution of the present study. Because of its high

value as a farm commodity, many theoretical and empirical

analyses have been completed on potatoes. This review,

however, will be limited to econometric studies that contain

at least an empirical content on potatoes.

Previous Studies

The econometric studies discussed here may be group]

into three categories: (1) comprehensive, (2) moderately

comprehensive and (3) least comprehensive. Studies in the

last category are so classified as they pertain to potatoes;

these studies typically addressed several agricultural

commodities but with no indepth analyses of any.

Comprehensive studies include the work of Zusman (1962), Hee

(1967), Estes (1979), Cardwell and Davis (1980) and Beilock

(1981). Studies by Gray, Sorenson and Cochrane (1954) and

Simmons (1962), while classified as moderately

comprehensive, are actually comprehensive in their

tnitoretical analysts, but less comprehensive in their

empirical or econometric analyses. Ihe studies of Fox

(1953), Shuffett (1954) and Waugh (1964) estimated the

demand and price structure for several commodities but with

no specific focus on potatoes. A discussion of these ten

studies in the reverse crder of their classification


Least Comprehensive Studies

Fox (1953) estimated demand equations for several

agricultural commodities using single-e.juation least-squares

estimation. he sugjtsttd that this procedures was most

appropriate for potatoes, since the price for any given year

was not materially affectcu by production, consumer income,

or supplies of competing products. Estimates for the farm

demand equation with price as the independent variable

showed a hi-hly inelastic demand with respect to price tut a

less inelastic demand wth respect to income. Specifically,

price flexibility coefficients of -3.51 and 1.20 were found

with respect to production and income, respectively. In

addition, the author concluded that elasticities of demand

for potatoes at the farm and retail levels were proDably

equal. He also founa that 98 percent of the variation in

farm prices of potatoes was explained by changes in retail


Shuffett (1954) was the first to estimate seasonal

demanau for potatoes. The total U.S. potato crop was

divided into early potatoes, production from December

through August, and late potatoes, production during the

remaining months. Early potatoes comprised an average of

only 18 percent of production for the 1920-41 study period;

late potatoes represented the remaining 82 percent. The

motivation for separating production stemmed from the nature

of the product. Early potatoes are more perishable and must

be moved through the market uFon harvest. late potatoes, on

the other hand, can be stored and therefore marketed more


Using the single-equation approach as justified by his

predecessor, SiLuffett found the demabLd for early potatoes

to be inelastic but less inelastic than that for late

potatoes. The cross elasticity for early potatoes was found

to be less than the direct-price elasticity assuming, of

course, that elasticities are derivable as reciprocals of

price flexibility coefficients. This suggested that while

late potatoes do compete with early potatoes, they are less

than perfect substitutes. The income flexibility

coefficients suggested a higher income elasticity for early

than late potatoes. Also, like his predecessor, Shuffett

concluded from his attempts to calculate price elasticity of

demand at the retail level that it was probably equal to its

farm level value. A higher retail than farm level

elasticity was actually derived, -.40 versus -.25, but

Shuffett concluded that these were essentially equal because

not all quantity produced moved through retail channels.

Unlike nis predecessors, ia'unh (1964) found what he

considered to be considerable differences between price

elasticities at the farm and retail level. In fact,

determining differences in elasticities for farm commodities

at the farm and retail levels were the primary focus of his

study. UIing a 1948-62 data period, he found that a 1

percent increase in consumption resulted from a 5.3 iErcent

decrease in price at the farm level but only a 2.6 percent

decrease at the retail level. It should be emphasized

thlouqh that his farm level demand was a derived demand.

Waujh estimated demand at the retail and wholesale levels

and then estimated iarm demand as the difference of these

demands. Income was found to have a positive effect on

retail price but an unexpected negative effect on farm


To recapitulate, all authors used the single equation

approach to estimate the price and demand structure for

potatoes. Waugh (1964) concluded from his findings that

demand was far more elastic at the retail than farm level.

Fox (1953) and Shuffett (1954), on the other hand, concluded

that elasticities at the twc levels were probably equal. In

their opinions, an appearance of a higher elasticity at the

retail level resulted because the quantity used to derive

this measure understated actual quantity. In estimating

seasonal demands for potatoes, Shuffett (1954) found some

substitutability between early and late potatoes, but a

Higher preference for early than for late potatoes.


Positive income elasticities were found for all estimated

demand equations, though for a derived farm level demand

equation, Waugh found a negative income elasticity.

ModeratelyComprehensive Studies

Gray, Sorenson, and Coclrane (1954) developed primarily

a theoretical analysis which addressed the problems of price

and income instability for potato farmers. These problems,

in the authors' opinions, resulted from demand inelasticity

for potatoes and price expectations of potato farmers (p.

172). That is, an expected high price led to a large

production which, in turn, led to low prices and income

because of the inelastic demand for potatoes. Though the

authors recognized that many exogenous factors had impacted

the potato industry, they suggested that these had not been

the major cause of price and income instability. Even if

these exogenous forces continued, the authors suggested that

income and price stability could be obtained through either

controlling production or influencing the elasticity of

demand. Production or supplies could be controlled, for

example, through acreoge allotment or marketing orders. A

more elastic demand could be obtained through various price

support programs. The authors discussed the merits of these

stability projraas atter empirically deriving the demand and

price structure for potatoes.

Unlike their predecessors, the authors did not include

income as a relevant explanatory variable. The farm price

oi potatoes was estimated as a function of consumption and a

trend variable. The estimates revealed that farm price

decreasAd by 2.7 percent for each 1 percent increase in

consumption. The trend variables revealed that, for the

1910-42 period, the declining per capital consumption of

potatoes had not been completely offset by the growth of

population. Estimates for a previous peric3, 1E70-1909,

showed that population growth had increased total potato

consumption by three times the rate of increase in per

capital consumption. With the population changing at similar

rates over both data periods, the authors concluded that

forces which coull iprcipitate greater instability had been

introduced during the latter period. lo contain this

instability, the authors developed several theoretical

arguments which encompassed the aforementionied factors.

Simmons (1962) also developed a fairly comprehensive

theoretical analysis of the potato industry. ue stressed

methods or amelioration {rice and income instability for

potato producers. For the 1950-61 period, he had observed a

continuation of those instability measures noted by Gray,

Sorenson and Cochrane (1954). Simmons emphasized a nTrEd for

the government and the potato industry to superimpose

additional demands upon the free market demand. One

sujgested method of iiiposin; an industry demand was throuljh

product promotion, a measure eventually adopted in 1973.

Controlling or influencing the elasticity of demand, in his

opinion, would be more effective than supply managqyecrit


because much variation in supply is due to uncontrollable

factors such as weather and pests.

While both Gray, Sorenson and Cochrane (1954) and

Simmons (1962) reccynized possible differentials in supply-

pcice relationships by seasons and geographic areas, only

Simmons attempted to capture them in his econometric

estimates. He estimated four seasonal and three regional

deiadnd equations with price as the dependent variable.

Supply equations corresponding to each demand equation were

also estimated with acreage as the dependent variable. In

addition, the author estimated intraregional demand and

supply equations as well as total U.S. demand and supply.

Supply was found to be inelastic at the national and

regional levels.

Unlike Snuffett who estimated early and late seasonal

demands, Simmons divided early potatoes into three groups:

winter and early spring, late spring, and early summer. For

the 1950-~1 data period of this study, these three groups,

in the order in which they are listed, represented

respectively 3, 10 and 6 percent of total production. The

remaining 81 percent for the late crop was essentially equal

to its level for the 1920-41 period of the Shuffett study.

The Jemand equation for winter and early spring potatoes

revealed that late potatoes were very competitive with them;

in fact, a 1 percent change in production cf winter and

early spring potatoes had a smaller effect on its own price

than that of a 1 percent change in production of fall

potatoes. TLhse effects were reversed in the demand

equation for late spring potatoes. Because ci product

perishability, late fall potatoes were not considerzj

competitive with early summer potatoes. Late spring

potatoes though were considered competitive with early

summer potatoes and they were found to have an unexpected

positive impact on them. However, because two of the three

parameter estimates in this equation were nonsignificant,

the authors concluded that the regression results were


Estimates For the late summer and fall crop revealed

that a 1 percent change in its production resultEl in a 5

percent decline of price in the opposite direction. A

nearly identical price effect was found for total U.S.

production, a result which may have been expected as the

late summer and fall crop reiresenr-e over 80 percent of

production. Income was found to have negative but

nonsijnificant impacts on total as well as late summer and

fall demands, a findiiij opposite that noted in previous

studies. The -.21 price elasticity for total potatoes,

however, was fairly consistent with that found by Shuffett

(1954), Fox (1958) and later observed by iiee (1967).

Separate regional demand equations for the eastern,

central and western regions revealed that only prices in the

western region were materially influenced by production

changes in .:owi1eting regiions. A 1 percent change in

production in the eastern and central regions resulted in an

opposite change in prices in the western region of 5

percent. The author concluded that prices in the western

region were influenced by production in competing regions

because a large proportion of its production was shipped out

whereas the competing regions consumed most of their

production. As in the seasonal demand equations, the author

found nonsignificant income effects, though both positive

and negative income parameters were observed.

Comprehensive Studies

Zusman (1962) developed and estimated an econometric

model of the entire U.S. potato industry, even though he was

primarily interested in California potato production as an

element of total U.S. production. Specifically, he wanted

to ascertain that level of production of California potatoes

that would result in profit maximization for its growers.

Because of the competitive nature of the industry, he noted

that this could be obtdiied only by observing the

interrelations of California potatoes with total U.S.

production. To this end, he empirically estimated his

econometric molel with a view toward framing a policy for

California production.

Zusman, like his predecessor Shutfett, divided the

potato industry into two seasonal markets, though the

markets encompassed dissimilar time periods. A spring

market which encompassed all of Calitornia production and a

winter market which encompassed the bulk of U.S. potato

production were defined by Zusman. Three demand equations,

three supply equations, three margins equation, three

supply identities and one production carry-in s.uation were

estimated primarily thrcujh a simultaneous equation

approach. As only the acree'g. of potatoes was estimated and

then multiplied times average yield to get total supplies,

these results are not emphasizeJ here. The supply c:

potatoes was found to be inelastic.

As with previous authors, the estimated results cf

Zusman s~zLwed a more elastic demand for the smaller spring

crop than the larger winter crop. Moreover, separating the

winter crop into food and nonfood uses, Zusman found a

higher demand elasticity for the latter than for the former,

a fidiinj later confirmed by Hee (1967). A much lower

elasticity of demand for potatoes was found by Zusman than

that found' by other authors (l?e 1967; Shuffett 1954; and

Gray, Sorenson and Cochrane 1954). Zusmia attributed these

differences to the fact that his estimates were for potatoes

use-d only as fool, whereas other authors included all

potatoes. The d~emald. equation for spring potatoes revealed

that significant differences existed among California spring

potatoes, other spring potatoes, and carry-in potatoes. Ihe

associated short and lor, run multipliers derived from these

differences together with other multipliers provided the

framework for Zusman's policy analyses on profit

maximization for Califcrnia producers. By varying

California production through assumed collusive action of

producers, the author determined the resulting profit for

alternative production levels. In general, the results

revealed that short-run gains in profit could be obtained,

but only through increased market instability. Long run

gains in profit were not obtainable.

By a similar seasonal division of the Fctato market as

utilized by Simmons, Bee (1967) estimated the

interrelationships among seasonal production. Potato

production for one of the four seasons, summer and late

fall, was not considered competitive with other seasonal

production. Utilization of this fall crop, however, was

considered competitive as alternative uses such as food,

seed and feed competed for it. Nonfall potatoes, on the

other hand, were hypothesized to be noncompetitive in

utilization because they were used almost exclusively for

food. Two-stage least squares as well as a limited

information method were employed to estimate these


Separate demand equations for utilization of the late

summer and fall crop revealed that distinct demands did

exist for potatoes used for food, livestock feed and starch.

As expected, the demand for food was more inelastic than for

other uses. Disposable income was found to have virtually

no effect on the demand for potatoes as food. Moreover,

processed vegetables, believed to be a primary substitute

for potatoes, were also found to have a nonsignificant

effect on demand. No estimate of supply was obtained as it

was assumed fixed for the late summer and fall crop.

Demand estimates for the remaining three equations,

intended to reveal the degree of comstition and

substitutability between dirterent seasonal potatoes,

invariably showed higher direct than cross elasticities in

absolute value. Even though positive and relative high

cross elasticities were tound, the estimates clearly

revealed higher preferences for the dominant than competing

crops. Stora4E potatoes from the late summer and fall crop

served as the competing crop in the demand equations for

winter abiu early spLiiig as well as late spring potatoes.

Early summer potatoes were found to be competitive with late

spring and late summer potatoes.

The otfect of income on J-maiaj was nonsignificant in all

four equations, thou.jh a positive and high income elasticity

was revealed for winter potatoes. Specifically, a 1 percent

change ii income a .3 percent change in demand for

winter potatoes. It should be :mpihasiz*d though that even

the sign of this income parameter varied with the estimating

technique. Logic then would suggest that income had a

nonsignificant, if any, etfect on demand. This pattern of

relatively iioisijrificant income i=ftcts has persisted for

all the studies which to date have examinc3 data periods

from 1920 to 1'ib6.

It may be noted that the three estimation techniques

used by dee generally yielded entirely different results.

Though the author notEd the appropriate uses for two-stage

least squares, limited information maximum likelihood, and

ordinary least squares, he emphasized the best results

irrespective of which technique yielded them. Such varying

emphases naturally compromised consistency.

Estes (1979) conducted a more comprehensive study of the

potato industry than those discussed heretofore. while the

primary goal of the study was to examine the impacts of

changing potato production in the U.S. Pacific Northwestern

region, the model developed to evaluate these impacts

encompassed the entire industry. Unlike Gray, Sorenson and

Cochrane (1954) who estimated potato acreage, Estes

estimated acreage and yield and then derived production as

the product of the two. These equations, estimated by

ordinary least squares and a maximum likelihood technique,

revealed that supply was generally price inelastic. Supply

elasticities varied among six identified production regions.

A variable included to capture risk, variance of potato

prices, was found to have only a modest effect on planted


Demand e'iuations were estimated for potatoes used as

food, but for the first time these potatoes were separated

into fresh and processed uses. An additional equation for

potatoes used as livestock feed was also estimated. All of

these equations were estimated at the national level.

Retail level demand equations for potatoes as food were

derived by using farm price together with price and cost

indexes as proxies for margins. Demands were found to be

less price inelastic for processed than for fresh potatoes


(-.235 versus -.124). Positive but nonsignificant income

effects were found for both fresh and procEssue demand

tquations. These findings were consistent with those of

previous studies.

Jo capture the effects of changes in demand for

convenience foods, Estes included the percentage of women in

the labor force in his fresh and processed demand equations.

The estimated results showed a 1.1 percent decline in fresh

potato demand but a 4.7 percent increase in processed potato

demand for each 1 percent increase of women in the labor

force. The demand for iresu potatoes was not influenced by

substitute commodities, a finding consistent with those of

He (1967), Zusman (1h62) and Simmons (1962). These authors

though did find substitutability amorni seasonal cros.

rnouj- Estes recognized the qains in efficiency from

using three-stage least squares, he refrained from using

this method to avoid transmittirAj possible misspecification

errors throughout the system. As parts of the modtl were

estimated lising ordinary least squares, maximum likelihood

estimates and two-stage least squares, a truly simultaneous

analysis was not obtained. Thcugh given a correct mapping

of the various equations, no distortions should have

resulted in the system. From this system, the author found

that an increase in acreage planted in the Pacific Northwest

would generally lead to decreased acreage in competing

regions, decreased total U.S. acreage, but increased total

U.S. production because of higher yields per acre. Pctato


prices were adversely atrected by these market share gains

for the Pacific Northwest.

Cardwell and Davis (1980) examined product forms and

interseasonal competition in the U.S. potato industry. The

authors estimated demand and supply equations for four

production seasons: fall, winter, spring and summer. This

study is significant in that it is the first to estimate

separate demand equations for the four main product forms of

potatoes: chips, frozen, dehydrated ard fresh. To complete

the model, potato storage and acreage were also estimated.

The estimated system of equations consisted of a

general demand and supply model, a product form suLmodel and

an acreage submodel. Estimates from the sulmcdels were used

as iinut into the ye.nral model. For example, the estimated

demand quantities of the various product forms of potatoes

were regressors in the seasonal demand equations. Most

parameters in the suLmodels had sign as hypothesized.

Several parameters in the general model had signs opposite

of those hypothesized.

The own price parameter for each seasonal supply

equation showed a downward sloping supply curve. The fall

supply equation also showed a direct relation between input

prices and production. Parameter values for the seasonal

demand equations, save summer, had signs as hypothesized.

Parameters for consumption of fresh potatoes and potato

chips in the summer demand equation showed a decrease in the

price of summer potatoes for increases in consumption of

chips and fresh potatoes, These results were unexpected

because most summer potatoes are consumed as chip: and


Furtahr findini;s of the study showed that fresh

potatoes were more readily substituted for dehyrated

products than for frozen products. Frozen products were

more readily substituted for fresh potatoes than fresh for

frozen. Income was found to have a positive effect on all

processed foras, but a negative effect on fresh potatoes.

Income elasticities for chiej, dehydrated and frozen

potatoes were .10u, .280 and .451 respectively. An income

elasticity o; -.622 was found for fresh potatoes. Seasonal

price flexibility coetticients were reasonably consistent

with those of Hee (1967) and Zusman (1942) 'Te price

flexibility coefficient for fall potatoes showed' a 3.44

percent decrease in price for each 1 Fercent increase in

production. The authors noted that this fall coefficient

was much larger than that found ty Hee and Zusman. This

difference was attributed to the larger share of fall

production to total production in their 1958-77 data period

as compar-'1 to earlier data periods used by Hee and Zusmaij.

Beilock (1981) conducted an econometric study of the

potato industry which incorporated various features of the

aforementioned studies, but also added new dimensions. Like

Zusman, Beilock divided potato production into two six

months seasonal markets; he then estimated national demands

for them. Additionally, he attempted to incorporate risk in


his supply equations and convenience preferences in his

demand equations as his predecessor Estes had done.

Further, potatoes for food were divided into uses similar to

those used by Cardwell and Davis (1980). New dimensions

included estimates cL supply not only through the product of

acreage and yield but also through direct estimates;

estimates of national supply and demand equations for

several product forms; estimates of regional supplies for

these same product forms; end estimates of the impacts of

energy cost on regional production.

The estimates, from both ordinary and two-stage least

squares, generally revealed an inelastic supply at both the

regional and national levels. Risk, defined as the variance

of farm level potato prices was found to have a ne-gligihle

impact on production. Both of these findings are consistent

with those of Estes. Supply equations for the various

product forms generally showed a shiftirn of supplies away

from tresh toward process especially frozen potatoes.

Additionally, as production in a region increased, supplies

were allocated to all product forms as opposed to a specific


Unlike estimates from previous studies, Beilock's demand

equation generally revtaid larger demand elasticities for

the dominant fall crop than for the smaller early crop. For

example, a -.82 price elasticity was found for fall potatoes

used as fresh and chips, but a -.34 elasticity for early

potatoes tor the same use. Yet, for the 1961-78 data period

of this stuly, fall potatoes averaged 85 percent of

production; early potatoes represented the remaining 15

percent. In estimating price elasticities for all potatoes

used as food, Zusmdan (1962) found elasticities for the early

crop to average two to tour times higher than those for the

fall crop. Though production as defined by Beilock (1981)

and Zusman (1962) differcJ, the results seem to suggest that

the demand for chips is much more elastic with respect to

fall potatoes than with respect tc early potatoes. As did

Estes, the author founJ that no substitute product or

commodity influenced the consumption of potatoes.

Unlike Estes, the author found very elastic as opposed

to inelastic income elasticities. Additionally, though the

products were not homogeneous, significant differences

exist-a among the convenience elasticities of the two

authors. Usini the percentaj- of women in the labor force

(4LF) as a proxy lor convenience, Beilock's results reveal

that the demand for frozen potatoes with respect to the

dominant fall crop was essentially unaffected by changes in

WLF; estimates by Estes, on the other hand, showed a 4.7

percent change in demand tor processed potatoes for each 1

percent change in kLF. Though the estimates were for

different commodities, it may be noted that frozen potatoes

were the dominant processed form for both studies.

As the primary goal of Bcilock's study was to examine

the imldct of rising energy cost on r--ional production, he

simulated alternative energy scenarios. The results showed

increased concentration of potato processing in the

Northwestern c-'Jioi of the U.S. The author concluded that

this result may have been expected as production of

processed potatoes requires high energy costs, but

transportation of these products requires relatively lower

energy costs.

Present Study

Tne present study incorporates features from several of

the aforementioned studies, but it is not a logical

extension of any. For example, it attempts to account for

risk in potato production and convenience preferences in

demand, factors previously incorporated by Beilock and

Etes. Also, as with these authors, separate demand

equations are estimated for fresh and processed potatoes.

Beyond these measures, similarities end. While the primary

focus of all the reviewed studies has been some aspect of

supply and demand, the present study focuses on the

interaction of structural forces with supply and demand


On the structural side, six endogenous variables are

estimated: number of Fotato processing plants (NPP),

concentration (CRP), price cost margins (PCM), advertising

sales ratio (ASR), change in Crocessing capacity (CCA), and

utilization of processing capacity (UCA). These variables

are linked to the supply and demand variables and estimated

using three-stage least squares. The data period is from

1960 to 1981 witu one year being the unit of time.


The estimated model is used to simulate the potato

iLdustry for alternative rates of growth of the exogenous

variables. Additionally, policy variables are simulated and

their impacts on the potato industry evaluated.



Changes in the U.S. potato industry have been rapid and

widespread. All facets ot the industry from production at

the farm level to consumption at the household level have

changed dramatically over the past three decades.

Production of potatoes at the farm level not only has

increased significantly but has shifted considerably among

regions. Consumption of potatoes has shifted precipitiously

from a predominance of fresh to a predcminance of processed

products. All ot these described changes as well as others

were precipitated primarily by the growing demand for

processed potatoes. Tnis chapter will therefore provide an

overview of the potato industry with an emphasis on changes

resulting from or associated with the growing demand for

processed products.

Potato Production

During the past three decades, yearly potato production

in the U.S. has increased by approximately 100 million

hundredweights (IaLie 1.3). These increases occurred

despite a more than 30 percent decline in acreage planted, a

more than 45 percent decline in consumption of fresh

potatoes, and a more than 95 percent decline in potato

farms. Lisinj potato production obviously has been in

response to factors which have at least offset the negative

factors. Potato yields per acre, for example, have

increased by more than 70 percent; consumption of processed

potatoes has increased by over tenfold; and average farm

size has increased by nearly thirtyfold. Prices received by

potato producers have incre+ sed considerably during the most

recent d-cade, but before this time they fluctuated

inversely with production with no upward or downward trend

discernible (Table 1.3).

Much of the increase noted in production has been due to

shifts among production areas. Potato production in the

Northeast, for example, has fallen from 29 percent of total

U.S. production in 1950 to only 14 percent in 1980.

Production in the Middle Atlantic region has fallen during

this same period from 8 to 2 percent of total production.

By comparison, production in the Pacific Northwest, as a

share of total U.S. production, has risen during this period

from 18 recent t to 47 percent (Table 3.1). States

comprising each region are declined in Table C.1. These

shifts naturally have been from relatively low-yieldina

areas to high-yieldinj ones.

The economic stimuli for the above production shifts

have been (1) higher relative productivity of land in the

Pacific Northwest, a.d (2) lower transportation cost an]

higher value per pound of processed versus fresh potatoes.

Both factors are significant because despite the greater

-4 60

I 01 mI N N % 0 r- o0 I
I El -l 0 I

I m
I o I In o
I 'I I
01 I I
I I 0 -- 0 0 I
I 1 N 0 I
0 OI I I

= _-I I Ia r o

01 I r 0
I i I

Il I 00 00 I' 4~ (N n 0 I
I n I * o.

I Ci I o '.0 s I 0 _- 0o -4

S 1 1- N N .I
I *dI I I 6

I 01 N I 0C

I ,I 0 1 LM 0 0 I ,-
I l 4 I


H I 1 01
I I I mc

I I U)

.Q =1(- 1 M C On N N 0 (01
0 I i o fI r0

I I I 4 w
II T w U
-I I 0 O N 0 N eN I- ol
I o 0 o o m ., L.A a a ,.

0 I 1 c r i I
0 I -I I rN
I I 01 I (|I
O I I I 0 N I 0
A I 1 I N I

a- I 1 I f a
0 0 1 co N C p N I Ln I
0 I 0 rII I

)4 1 I C)I

3 I 01 I 01
1 I I 00 0 L I

* 01 1 0 N O I 4)1

o I I *I N I oI

1-4 u| I 40 r- '- '.0 L (N4 N a- W '|
I O WI(a I 01
3 0| r4 1
*3 al r (N (N (N N L MI

S I u I (N I I 3

rrj 01 u
r I I I WI
4. Ig I I 01
0 I I -I
4 I I 4) (0 I (0l
t a 1 0 ) i I 01

SI I M 0u Q1
I I ) L I + .
I I I 00 f H I )
I- I 4 t 4 4)> I .
* I I U '3 0) 3 ) I I-1
SI (0 0 ) > U a H 0 1
4 I I a -U > I LI
I I I 4) 43 ) + ) ...4 *i4 4 a I
0 I 3'I 1-J I'L
n 1 H 3 3* 1 -* 3 U 0 3 [n
(0 I < 0 --4 0 ( 2 a0 I


productivity of land in the Pacific Northwest, the northeast

undoubtedly would have remained the largest producing rryion

in the absence of differentials in transportation costs.

DEiogyrdphers Vill quickly discern that the loyic of this

argument stems Lrom the proximity of the Northeast to large

population centers.

The decline of farms 1.roducinj potatoes far exceeds the

decline of farms producing aggregate U.S. production.

Whereas 3.7 million farms were producing agricultural

products in 1959, this figure had declined to 2.5 million in

1978, a decline of 32 percent (Census of Agriculture). In

contrast, Eutato farms declined during this period from 686

thousand to 28 thousand, a decline of 96 percent. These

drastic losses of potato farms were primarily due to

shifting de:manids from fresh to processed potatoes. These

shifts have I~'c not only to an abandonment of many farms

which prcuced primarily for in-home consumption, but to an

abandonment of farms which lost marketing outlets.

In summary, shifting potato production among regions has

not just accompanied risiun production, but has actually

accounted for it. Sniftiiig production also has led to fewer

marketing outlets, especially outlets for fresh potatoes in

the decliuinj Northeast and Middle Atlantic regions. With

fewer marketing outlets, marketing costs naturally increa-ed

because of increased distance to markets. Potato farms

realizing prohibitive increases in marketing costs naturally

ceased production.

Potato Consumption

Potatoes consumed in processed forms have increased

tenfold during tie past three decades. Consumption of fresh

potatoes, by comparison, has plummeted to only one half its

level three decades ago. An increase in consumption of

processed potatoes may have been predicted as converting any

product from its natural form to several processed ones

usually increases total usage because of greater variety and

increased convenience. Potatoes are not an anomaly. What

is unusual about potatoes is the magnitude of decline in

consumption of fresh potatoes coupled with a more than

offsetting rise in consumption of processed potatoes.

Per capital consumption of fresh potatoes declined from

100 pounds in 1950 to 50.5 pounds in 1978 and then rose to

51.4 pouuds in 1980; per capital consumption of processed

potatoes rose froi (.3 pounds in 1950 to 70.6 pounds in 1978

and then d-clined to 64.4 pounds by 1980 (Table 3.2). While

several processed potato products have generated this

growth, frozen potatoes have played the most pronounced

role. Much of this growth of frozen potatoes has resulted

from the increased usage of these products at institutional

establisLments (restaurants, hotels, lunchrooms, etc.).

Frozen french fries, dehydrated mashed, and potato chips

are the major processed potato products. Small quantities

of potatoes are canned and processed into starch and flour.

Potato chips, the first processed potato product, led the

growth or potato processing~ until the mid 1960's. By 1966,


,n 6n %D m %D %0 %r0 -%D Lrl- ~o o% I,- U 4a 0 mV'- 9-

M~6~~066~~~NmN N CN m M M en

I I I I0~ll l lUI ,
IIIl *I *iii a ** *a*** *****rc

111111I I1

I I I I % % r-- r- r- a- p- I-- %Z a- kz a %C r- p-




V, -

-4 -




I 4
u: A

OO *

in *
* &
h- I



c =






U 5




M '




0 01

0W 31


M (Li


C >|

5 %
*o -

0) f
U S30
=r CO

m. S C o 0 (N r4iS M i 4 n rn U, n a 'Q- (N % U r- u i -D%0

* ** 4 Z' O0a; a* a a -: a*
o= i 0 0 0 0 0 a 0 -a-

m- 0 0 0C- -- --

Un L, Ln L, L, LUn LM LU,) \L3 U, U, %C U, %. 10 %0 %.0 o r- r.- r- r- flr r- f s
o a a- a- *\ -o-ca- a- -) C a- L Yo a-I
r CC. ** e C ~ CC~

r- 0D r- p 0co C% N. %D r 00 0% f u -M IN %D .0CY 0 N 0 .o 1LJ .
0 ^- N O n00 00 :v j N


per capital consumption of frozen potatoes had surpassed that

of chips, had more thdn doubled chips by 1975, and has

maintained this 2 to 1 ratio into the 180's (Table 3.2).

Although a variety of frozen potato products are produced

(hash browns, diced, scalloped, etc.), frozen french fries

constitute over 80 percent of all frozen potatoes.

Moreover, frencn fries account for more than half of all

frozen vegetables.

While the preceding description of changes in potato

consumption may reveal much about the potato industry,

identifying and then discussing those factors which caused

these changes is obviously of importance to researchers.

While many factors have precipitated these changes, some of

the more significant ones are believed to be changing

consumers' tastes, rapid growth and expansion of fast food

establishments, an increasing number of women in the work

force, a growing population, and an increasing number but

decliniing size of households.

tast Food Establishments

Frozen french fries, the processed potato product which

has shown the most significant growth, are channeled

primarily through fast food establishments. According to

potato marketing specialists, roughly 60 percent of all

frozen french fries is clannelEd through fast food

establishments. With the 1981 production of frozen french

fries totaling 3.6 million pounds, this means that roughly

2.2 billion pounds of thesa fries were sold at fast food


establishments. The Leaiing fast food chain, 1cDcnald's, is

estimated to account for 25 percent of all frozen fries

(Cox, 1982). These figures then translate into sales of 900

million pounds of trench fries in 1981 at McDonald's


Frozen french fries are not only marketed primarily

through fast-food estaolishmerts but also account for

roughly 20 percent cf fast-food sales (Purcell 1976; Cox

19d2). Cox (1982) reports that seven of every ten customers

arriving at fast food establishments after breakfast are

estimated to order fries as part of their entree. Given

this importance of frozen fries at tfat food establishments,

some correlation between the growth of frozen potatoes and

the fast food industry may be expected. Indeed both have

shown tremendous growth.

The number of fast food establishments increased

fourfold during the 1958-77 period. In actual numbers,

these establishments rose from 25.1 to 100.5 thousand.

Sales at these estatlisaments increased ninefold, rising

from $1.1 billion in 1958 to $11.8 billion in 1982 (Fiqure

3.1). with 20 percent of fast food sales being comprised of

frozen irC'ich fries, dollar sales cf fries at these

establishments then totaled $2.3 billion.

Tre descriLbe growth of the fast food industry has given

a tremendous boost to consumption of processed potatoes, via

frozen french fries. Otrhr processed potato products have

yet to Jaili marketability at fast food establishments,

Number of Fast Food Establishments, 1958-77






15 30 45 60 75 90 105


Sales at Fast Food Establishments, 1958-1982







1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0

Figure 3.1.

(Billion of 1967 Cbnstant Dollars)

Fast Food Establishments by Number and Sales

though it is not inconceivable that chips and dehydrated

potatoes coui. become competitive products at these

establishments. Fresh potatoes already appEar to be gainingg

a reasonable share of fast food sales. Ib1is marketing

outlet for fresh potatoes may prove to be the requisite

stimuli to arrest the decline in consumption of fresh


WoJmllrn i; tlthu Work Force

An increasing pec:-centage of women in the labor force

supposedly shifts demand toward food products which require

little prc l artion time at home. In essence, women in the

labor force increase the demand for convenience foods.

Because potatoes represent or contain convenience,

growth of women in the labor force is expected to stimulate

growth of procEssed potatoes. Labor force data reveal that

women as a percentage of th labcr force have grown from 29

percent in 1950 to a 1982 level of 43 percent. growth

then represents an increasing demand for the convenience

reflected in processed potatoes. 3-cause women have less

time aLd energy to devote to food preparation, market

baskets of today clearly contain more products reflecting

increased convenience.

This described growth in d&maul for convenience has not

only influenced the characteristics of market baskets, hut

has accelerated the growth of fast food establishments.

Diners o0 fast food restaurants list convenience as the

primary reason for their patronage. These diners ILtrine


convenience as not having to (a) cook, (b) do dishes, (c)

wait for service, or (d) travel far (Restaurant Business,

May 19S2). Thus a growing demand for convenience alters

both in-home and away-trom-home food consumption.


Growth of total potato usage during the early 1960's

resulted primarily from an expanding population as per

capital consumption remained fairly constant. Since the aid

1960's, per capital consumption of potatoes has increased but

population growth is hypothesized to have continued its

positive effects. These effects should be especially

noticeable for frozen potatoes since considerable growth has

occurred in that segment of the population which Ereguently

consumes these products at fast food establishments. That

is, surveys by Restaurant Business (May 1982) have fund

that the 25 to 44 age group, which has a greater tendency to

eat out, has shown and c ntinues to show considerable


Number and Size of Households

Consumption of processed potatoes during the past two

decades has been influenced by a growing number but

declining. size of households. Declining household sizes

supposedly have increased the potential for away-from-home

food consumption by decreasing the cost of eating out while

simultaneously increasing the inconvenience of eating-in.

These changing consumption patterns increase the expected

growth of frozen potato products because of their heavy

institutional uses, but decrease the expected *-rowth of

other processed potato products because of their primarily

in-homc uses.

Consumers Taste

Chanjiny consumers' taste is believed to be a factor in

influencing the shift in consumption from fresh to processed

potatoes. While it is impossible to quantify taste or its

impact on demand, it may be noted that potato processors

have stressed improve taste and nutrition of its products.

So there is at least a perception that consumers' decisions

are a function of taste. However, it should be recognized

that differentials exist in the time periods for which fresh

and procrc' sed potatoes have been available. That is, it may

be appropriate to distinguish between a Lsifting taste and a

developing taste for processed potatoes. Since neither can

ba -uaintified, suffice it to say here that taste is believed

to have influence the demand for potatoes.


The changing composition of potato consumption from

fresh to processed forms has implications for consumers,

producers and processors. From one perspective, higher

proportions of potatoes consumed as processed mean more

consistent quality, longer storability, and increased

convenience for consumers (Quick Frozen Foods, June 1977).

From another perspective, consumers are confronted with

higher costs and] a loss of nutrients from increased

consumption of processed potatoes (Tilburt and Smith, 1975).


Producers, because of increased sales to processors through

contracts, are believed to have realized a reduction of

price risk and an enhancement of creditworthiness with

bankers. Processors are believed to have realized a

reduction in production costs, a higher quality of raw

products, and a more timely delivery of raw products.

While the positive attributes of processed potatoes have

undoubtedly increasEl utility for consumers, the supposedly

negative attributes--increased cost and fewer

nutrients--must be evaluated with caution when making

comparisons with fresh potatoes. Most comparisons of cost

per servings of processed with fresh potatoes show a higher

cost for processed. These comparisons though, almost

invariably fail to account for the differentials in labor

requirements for preparing each form. Talburt and Smith

(1975) have found that imputing a mere $.50 per hour to the

cost of time for food preparation results in a favorable

comparison of cost per serving of processed with fresh

potatoes. Additionally, nutritional comparisons must also

be made with caution. Most nutritional comparisons involve

processed potatoes with fresh-dug potatoes. However,

potatoes stored for a mere three months, not uncommon for

fall potatoes, can lose as much as 50 percent of its primary

nutrient, ascorbic acid (Vitamin C). Processed potato

products, on the other hand, are virtually unaffected by

storage. Thus, on balance, processed potatoes could ccmrare


favorably with fresh potatoes when the cost ot time and the

losses from storage are considered.

Increased potato procesiinj is believed to have

favoraLly impacted the quality cf potatoes marketed as

fresh. As the potato processing industry matured and

processors became larger and more diversified, potatoes

which were most appealing in size, shape and form were

sorted by processors for fresh marketing. This selected

sorting resulted not only in higher average quality of frssh

potatoes, but in reduced time and expenses of consumers'

search for quality. Simply put, it led to more efficient

marketing of potatoes.

Before the tremendous expansion of potato processing,

raw potatoes were sold primarily to potato chip processors

and inrst. buyers. Both raw potato producers and chip

processors were concentrated around large population

centers, particularly the Northeast region. These locations

were Jetermined from the economics of transportation costs.

With demand shitting from fresh to processed potatoes, raw

producers have found less populated but more productive

areas to concentrate their production. Potato chip

processors, on tae other hand, are still optimally served by

larje population centers.

The dramatic growth of frozen and dehydrated potato

products, as noted previously, has been accompanied by a

significant loss ot potato farms. And while there has been

significant increases in yields per acre, tremendous


increases have also occurred in average farm sizes. From

1959 to 1978, the average sized potato farm jumped from 1.75

acres to 48.9 acres. Part of the economic stimulus for

these increases in farm size resulted from raw product

specifications of processors. Small producers, because of

unavailable technical assistance, generally have been unable

to produce potatoes wit h consistent processing

characteristics--high solids, low and uniform sugars, and

good type (Gomena, 1968). Because the probability of

technical assistance from processors increases with farm

size, many producers have either expanded or ceased


Potato processors have suggested that increased potato

processing has reduced their production cost, improved the

efficiency of their operations, and enhanced the quality and

uniformity of processed products (Conrad, 1975). Processors

contend that because they have become increasingly larger,

contract purchases are imperative for efficient operation.

Contract purchases, according to processors, reduce the

price risk associated with open-market purchases; since

price risk is an element of cost, a reduction in price risk

should therefore reduce production cost. Efficiency of

operations, according to processors, has been enhanced

because of the timely deliveries generated through

contracting. lhat is, processors incur less search cost in

obtaining raw potatoes. Processors further contend that

their smaller number coupled with their increased size has


result in improved product yu.lity. Improved quality, it

is argjuc, results from competition amoLy processors for

consumers' dollars. Because consumers will readily switch

brands upon variations in quality, processors have

consistently provided uniform quality. Before the decline

in number and increase in size, processors contend that

uniformity and consistency of quality were sacrificed

because smaller processors maintained variable production


In the movement to fever and larger potato processors,

it may be notel that larger processors often acquired

smaller firms, marketed the acquired product brands, but

closed the facilities produciu-j these brands. An indication

of the impacts resulting directly and indirectly from

actions of larjer processors is reflected in the changes in

potato processing plants. Whereas 400 plants were producing

potato chips in 19o0, by 1980 this number had declined to

177. Plants processing frozen potatoes increased from

approximately 30 in 1960 to 91 in 1964, but declined to only

33 in 1960. Dehydrated potato processing plants increased

from an estimated 21 in 196) to 28 in 1964, but declined to

19 by 19'0. These ilant changes are believed to reflect the

impacts of differentiated product advertising and the

increased cost of adopting new technology.

Surmary and Conclusion

Dramatic economic chances have occurred in the U.S.

potato iindlustry durij the past three decades.


Technological advances in agriculture coupled with regional

shifts in potato production have greatly augmented fara-

level potato production. Additionally, technology has

increasingly added convenience to potatoes via the

development of new varieties of potato chips, frozen french

fries, dehydrated mashed, and other processed products.

Changing socio-econcmic conditions have influenced taste

preferences of American consumers to an extent that they are

now expressing a stronger preference for processed than for

fresh potatoes. A question of relevance here is whether

these described changes also characterize the future of the

potato industry?

Potato production at the farm level is constrained more

by demaii conditions than by technological know-how.

Advances in technology therefore, are not likely to greatly

increase total production. Potato farms, however, are

expected to continue increasing in size because farm

technologies are generally land-using. Hence, there are

likely to be fewer farms with potato production further

concentrated in the Pacific Northwestern states of Idaho,

Oregon and Washington. Further concentration of production

in these areas also is likely to lead to further

concentration of processors

Technological advances in the processing of potatoes are

expected to add more convenience to processed products.

ThesE convenience aspects most likely will be reflected as

less food preparation time, more fortified and retained

nutrients, more palatable taste, and longer storability.

These aspects probably will increase in significance as

women Lb.uome a larger share of the labor force. Fast food

establishments also are likely to find these attributes of

growing sigjuificance because of the importance of

convenience to their patronage.

If processors can add more convenience to processed

products, consumption of these products undoubtedly will

increase. Increased advertising and promotion of both fresh

and processed potatoes may also further enhance potato

consumption. Indeed advertisers and promoters of fresh

potatoes arjue that the recently enacted Research and

Promotion tax has not only arrested the decline but actually

increased consumption of iresn potatoes. Accelerating the

per capital consumption of fresh and processed potatoes

should load to tremendous growth cf the potato industry.



In this chapter a simultaneous equation model is

specified which captures and explains the dramatic economic

changes that have occurred in the potato industry. Figure

4.1 provides an illustration of the vertical market

structure of the potato industry. Excgenous and endogenous

factors which have both structural and market equilibrium

dimensions impact the entire vertical system and the overall

performance ot the industry. Performance, in turn, impacts

the endogenous factors.

Ihe ,model ;peciiied in Iable 4.1 captures the

schematics which Figure 4.1 illustrates. The model consists

of sixteen endogenous and twenty exogenous variables; all

variable symbols and definiitions are provided in the notes

to Table 4.1. Following the equations and variables

specification is a theoretical discussion ot each equation

and variable. For expository purposes, the model as

specified in Table 4.1 is subdivided into two sectors: (1)

structural, and (2) market equilibrium (price-demand-

supply). Equations 4.1 through 4.6 are structural while the

remaining 4.7 through 4.16 express dimensions to market

equilibrium. Tnese two sectors are hypothesized to have

impacts on the model that litter with respect to time.

hiile only a suptrlicial discussion of t:is time-Nimension

problem is presented here, a thorough and more profound

discussion follows in Chapter V.

The Structural Stctcr

The field of industry 1 organization is rich with

studies on the structure and performance of industries.

These studies pLr-.ominantly estimate single-equation models

in which causality runs from structure to performance. The

moiel specified in Table 4.1 is both dynamic and

simultaneous with respect to the structural variables.

Equation 4.1 attempts to capture those factors influencing

and .etermiinj the number of processing potato plants

(;PP)). The number or plants not only is influenced by

several variables, but N'P, in turn, influences several

variables. Concentration (CPiP), as measured by equation

4.2, is also impacted by Wud impacts several variables

within the model. Tiiation 4.3 captures those factors

determinini the price-cost-marjiLs (FCM). These margins

influence other variables, particularly those within the

market equilibrium sector. An advertising-sales ratio (ASP)

is determine by Equation 4.4. Equation 4.5 measures

change in processing capacity (CCA) while 4.6 measures

utilization of capacity (UCA). These six equations are

highly interrelated and are intended to capture the

structural dimensions of the potato industry. Among other

relationships, it may be observed that PCM intiuences NPP;

Exogenous Factors




Figure 4.1. Schematic Diagram of the Potato Industry

Potato Producers


Market Equilibrium

Processors Fresh Handlers


Market Equilibrium


Haoe Away-from-home

r0 0 0 0 0 0 0 0
SC' m ( 1i>n w i oo 01%

r-C -
t-, [-


Hl I

r-I H-
0- -a,

o 0

-. (-



H C' *ct

I I-

r-q r-I
[- ) c-

o r-4 CN M L)in L
r-i N (C ) L 'in r- H H H ,-4 H r- -
iq ^ ^ ^ vs ^ cy *^ < *i cy < ^ 'y v


co o o

-4 -

CS -1

A, C4" CN
00 c0

-4 14

CD Ch(
0L ri"


ul ,

in In)
N rn

I- r

m m

-4 CNN

'I m Lw

Kf oa 0
w e= e
13 m0
E:~ a,
= r


= =
-.4 f I

r 0

W r
0< -4

Cu a e e

a r- r
= =

NL r r. LA '

=~ =

0 3-I N m '0 LA '
-I N Im '0 LA '0 r- m c, .-I r- -4 .4 4 -

Table 4. 1 ContinuLd

Enldo.=rn cus Variables

NPP Number oZ processing plants
CCP Concentration in [otato processing
PCM Price cost mar,ins
ASE Advertising sales ratio
CCA Chang]_ in processing capacity
UCA Utilization of processing capacity
QrP Production cf fresh potatoes
UZ1 Utilization of potatoes for fresh consumption
UZP Utilization of potatoes for processing
iPF? Retail price cf fresh potatoes
HtP .Rtail price of pro cssed potatoes
P[' Wholesale price of processed potatoes
FPi Farm price ci fresh potatoes
17T Research and promotion tax
SPR Spread between retail and farm price
UZ- Utilization oi other potatoes

Exoycli.ous Variables

MS Minimum efficient plant size
PCEL Price cost waryins layIgd
CT Tranispurtation cost
GD -ojraiphic dispersionof potato production
CK Cost of capital
FS Price of sudar beets lagged
FirL Farm price of frest potatoess lagged
MC Marketing cost
TR rrond variable
RS Risk
GC ProJuce sales through retail grocery stores
TP lotal avay-from-home restaurants food sales
FF Fast-tool sales as a Iprcent of total food
PN Population
IN Income
14N Women in labor force
Gs Exp cted growth
UZPL Utilization of potatoes for processing lagged
QBPL Production of fresh potatoes layged
ASSL Advertising sales ratio lagged


IPP influences both CRP and CCA; CRP influences ASR; and CCA

influences UCA. Other relationships among these structural

and market equilibrium variables follow in the discussion of

each equation.

Number of Potato Prccessinq Plants

Equation 4.1 estimating the number of potato

processing plants (NPP) represents a proxy equation for the

more appropriate one, the number of firms. while there is

unlikely to be a one-to-one correspondence between changes

in plants and firms, a sufficiently high correlation is

expected to exist so as net to distort the explanatory

proficiency of the model. This expected result stems from

the entry and exit patterns of potato processing firms.

Potato chip processors have always dominated the potato

processing industry as measured by the number of firms and

plants. Entry intc potato chip processing was rapid and

extensive following world War II, consisting mostly of small

producers. As the industry grew and matured, consumers

began demanding fresh, unbroken, and ungreasy chips

(Warwick, 1973). To satisfy as a minimum the first two

demand characteristics, potato chip processors began

implementing a marketing system of delivering their

products from plant to retail outlets. Because the volume

of production of many smaller processors could not sustain

these additional marketing costs, these firms were either

purchased by larger ones cr exited the industry. Similar

changes also occurred in the frozen potato sector, tut exit


of smaller firms was due primarily to shifts in marketing

trom retail to institutional establishments anj increases in

advertisiing expealitures by larger producers. Institutional

establishments demaiided from processors a level of product

uniformity and specificity that smaller firms often found

unprofitable to produce. The limited diversification of the

smaller firms typically could not support production of by-

products which helped to sustain the profitability of larger

firms. Advertising by larger producers also helped to erode

market shares of smaller producers. Unlike chips and frozen

potato production, dehydrated production was never inun'dted

by smaller producers.

The equation as specified in Table 4.1 to measure the

number ao potato processing plants is

(4.1) NP? = f(PC", UCA: MS, PCML, CT, GD, H, UZPL)

Price-cost-margins (PCM) are expected to be significant in

firms' exit decisions, but PCMI are hypothesized to be of

greater si quilicrince in firms' entry decisions. Stated

differently, low PCM will most likely influence the decision

of marginal firms to exit the industry while Icnger time

periods will be iei.d.d for firms to enter the industry

because or time rt4uired to bring new plants into

production. Positive coefficients are expected for both ?cM

and PCiL as firms are hypothesized to respcnd aftirmatively

to improved profit conditions.


Theoretically, rising margins should induce firm entry

or, as a minimum, increase firm size. Similarly, falling

margins should precipitate or accelerate the exit of firms

as marginal firms' costs rise above product prices.

Moreover, an increasing demand which raises capacity

utilization (UCA) and profitability of an industry should

escalate the number and size of firms. That is, lower rates

of capacity utilization are likely associate d with

slackening demand and therefore lower PCM, while higher

rates of UCA are likely associated with increasing demand

and a higher profit rate (Sawhney and Sawhney, 1973). This

hypothesized positive relationship between NPP and PCM is

also expected to characterize that between NPP and UCA.

Several exogenous or predetermined variables are

hypothesized to impact NPP. As the minimum efficient plant

size (MS) increases, small firms with limited shares are

likely to exit the industry as their competitiveness with

larger firms diminishes. minimum efficient plant size may

be affected by several factors. Among these are

transportation cost, capital cost, volume discounts, and

market size. These factors are believed to be reflected in

changes in average plant size. And while the average size

of existing plants changes, size of new plants is considered

the most appropriate measure. MS therefore reflects the

average size of plants coming on stream within a particular



La'jged price cost marjins (PCLY) have been discussed in

conjunction with PCL. A one period lag is expected to be

sufficient for capturing the requisite time horizon for

adding new plants. Lagged utilization of processed potatoes

(UZPL) also represents a one period lag as processors are

expected to adjust their plants according to the previous

year's changes in processed potatoes. In general,

processors should be exFcctci to increase NPP as UZPL

increases. However, NPP could decrease with increased UZL

if significant product differentiation, advertising

expendit,'LLs, cost differences, and other entry barriers

should effectively hamper the competitiveness of smaller

firms with larger ones. These factors, which are accounted

for in the molel, indirectly affect NPP and could lead to a

relatively insignificant coefficient for OZPL.

Since firms jtaLrally expdnd dind contract their

operations based on some expectation of future growth, NFP

should be directly related to expected growth (GR). For the

processed potato industry, GB is likely to be based on a

number of factors such as recent changes in sales, income,

per capital potato consumption, women in the labor force, and

away-from-home food consumption. Changes in demand for

processed potatoes during the previous three years should

capture most or these forces. Hence, G5 is defined as UZP

lagged three perioas with equal weights per period.

ihe impact of transportation cost (CI) on NPF cannot be

determine apriori. With CT being a significant proportion


of marketing costs, firms may be expected to add plants in

an attempt to minimize total marketing cost. Transportation

cost, however, is only one component of marketing cost.

Other components may be changing at a rate which would

offset advantages to bc gained from a more efficient spatial

distribution of plants. Furthermore, economies of scale in

production may favor concentrated production over more

dispersed production.

Although the sign of Cr is indeterminate, geographic

dispersion ot raw potato production (GD) is expected to

have a significant and positive effect on NPP.

Transportation cost relative to the value of raw potatoes

virtually dictates the spatial allocation of processing

plants in some proportion to GD. With raw potato production

becoming increasingly concentrated in the Pacific Northwest,

a growing concentration of plants in this area may be


Concentration in Potato Processing

Althuug L there are many indexes for measuring

concentration, this study utilizes a four-firm concentration

ratio. This measurement is chosen based on data

availability rather than theoretical reasoning which

precludes other measures, while the consideration of only a

point on the concentration curve is a valid criticism of

this measure, the use of time-series data over twenty-two

years is expected to negate this criticism. Concentration