Ho STM IS
lGGs1M | ¢ MARKETING RESEARCH REPORT NO. 993
AKEDC s ; :
4

2 "ys
« ial i la
: ion +ichtthral Hort
EEN REZ Peeyg te GGead ge J ®.
OTIOWAL FSi PEUTaEat & —
PME Be “a
iy J

\ \ Will ye Lead
Current Ser al Recol cik
hy

ao oS

DEMAND FOR FARN- RAISED
CHANNEL CATFISH IN SUPERMARKETS:
Analysis of a Selected Market

U. S. DEPARTMENT OF AGRICULTURE: ECONOMIC RESEARCH SERVICE

Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

ABSTRACT

In March and April 1972, a controlled Latin square market experiment was
performed in six Atlanta, Ga., grocery stores to determine the demand for fresh
channel catfish in supermarkets. Prices ranging from $0.79 to $1.29 per pound
in 10-cent increments were used to elicit quantity responses. These responses
were used to estimate linear, log-linear, and log-log forms of a demand curve.
The resulting demand curves indicate an elastic demand for catfish within the
experimental price range. Only about one of every 150 customers purchased
catfish at the current price of about $1.19 per pound. Results of a consumer
questionnaire indicate the possibility of expanding the market by introducing
more convenient product forms at a reasonable cost to the consumer.

Keywords: Catfish, demand, market potential, grocery stores, Georgia.

Washington, D.C. 20250 May 1973

17

PREFACE

This research represents a cooperative effort by the Marketing Economics
Division (MED), Economic Research Service, U.S. Department of Agriculture;
Winn-Dixie Atlanta, Inc.; Goldkist, Inc.; and the Institute of Food Marketing
and Distribution at Georgia State University.

The research was conducted by MED staff at the Richard B. Russell Agri-
cultural Research Center, Athens, Ga., and was made possible by the special
efforts of Lynn Fitzgerald of Winn-Dixie, who supervised the selection of
test stores and provided the necessary liaison between MED researchers and
Winn-Dixie market managers; John Tallent and James Marion of Goldkist, who
managed the difficult transportation and temporary storage arrangements; and
John Wright and Joyce Speck of Georgia State University, who helped in the
selection of enumerators and kept study records. Victor Chew of the Biometrics
Services staff, Agricultural Research Service, USDA, provided guidance and
counsel on statistical considerations associated with the study. The authors
wish to express their appreciation for the excellent help of all participants
in the study.

Mention of firm names in this report is solely for purposes of acknowledg-

ing contributions to the study; this does not constitute endorsement of these
firms by USDA.

iii

CONTENTS

Page
SUMMARY <a cise ols. eae 6a) ore a forte: oilerate ee iat hewatenM ene Temata ets (aye Gee iano vi
EWE PO CUCET MVS 5.6 5a oiss 5 5 os aj ib el cercs)ig cise: acres errs shoe cera atte Sean ree nee ae ee ]
PROCCOUIE i 6: 3 5 a:0:0;6, 5-80 eraifote ayedes sbadegsyetens tele lavopolenerenenteretees tone a-2- rare am 2
MO GGT a cic, ocssnteie (0-0, eyes) a0: 0)'eieenets, eeneiiots. ¢/-ojlecdteteusie.eReMenegeee eepeaeae ere Ras Tea sae arty 5
ES Ti mates “OF Geman. \sicie ci c0iciisistaqs cesfenscclobatel oy aiehereuerc obesetarecetemene te een | Mele 6
Potential. for catfish: farming’. .c «dsiccieve ire ect ae een 10
QUEST TONNAT LE PESUT TS 0575 < cistssel creas’ ciate to ststetaren cate cere arene ove tine 7a. eu ae eee 14
Recommendations. for further yesearchin. sec secmie sects eee ae 14
RE FELCNCOS 6 ais isis ae ta. cne arleeneiece Silakeye: ace ahetete Shaneuar ere enede Moore eon aN SCRA ew naar) anata 16
APPON IX «0: arercie wide. sisailese peso are eser'ehe lane rouoyereseiiaso CARS SyeNsT ONG teticcs Geant Betsey nee 7,
TABLES

Table Page
1--Analysis of variance with a consideration of regression for a

6=x-6 LatinySquarescccsves os 00 cui c ace One ts rae ee ere 6
2--Analysis of variance of pounds of catfish sold per 1,000 customers,

G SitOVeS.: Abul anitas Gd a241.9 7.2.c.5.5: 0 ec ets.a eudeccarevetebavene evenshene aici Romeneyore ena tore
3--Quantity demanded, price elasticity, total revenue, and net revenue

at different points on a linear demand curve for catfish, 6 stores,

Atanas, Gates: 197 Zia ses tee lean Srorgie he tencner ee .ev ov nleai e Qavasie ereeeae eee Meee iene
4--Quantity demanded, price elasticity, total revenue, and net revenue

at different points on a log-linear demand curve for catfish,

GuSLOves . Atanas. Goeeg: 1 DLC vere k core cia se sraraernss wcerstue a ereee eee se eee ote 9

5--Quantity demanded, price elasticity, total revenue, and net revenue
at different points on a log-log demand curve for catfish, 6 stores,
Ateliantars. Ga ss? OW2 Sect. cratslatat ext ete Maciee se ter git areas aac tara heii eaenate eee are ues 10

6--Upper bound estimates of annual U.S. sales of processed catfish in
supermarkets at various prices. 1972... Aon. .iiwswtccemus sebiemaneesc 11

7--Demand elasticity and net revenue based on 3 forms of a retail
demand curve for farm-raised channel catfish, 6 stores, Atlanta,
Geass: PES ci S enc ealasna es elias Sissi Skate ears sage Stee weeks eco GIS tree een een 12

8--Price elasticity and net revenue per pound at different points on
a derived log-linear farm-level demand curve for catfish, 6 stores,
Atlantas Gaws 1972 oc.6 sa aco ae ae odie wee wlacere were recn eee a tena eet oe eres 13

TABLES - Continued

Appendix
table
]1--Raw data for catfish marketing study, 6 stores, Atlanta, Ga.,
TIS 2i Re aU P eerie SiMe Ree te hl cee: bueisots atatewe aN wials ¢: e146 Obes Beales 6 ele es SOS AURE
2--Estimated value added approach to price determination per pound
for farm-raised channel catfish, 19722 .....0 cc ccccdcelecccccccesce
3--Tabular results of channel catfish questionnaire, 1972.............
FIGURES
Figure

1--Tray pack of deep chilled farm-raised channel catfish used in

ESIGN Gaze Morse) sais tated nds sss eds tas WA ere marr aloes ee educate se meas

2--Typical display of farm-raised channel catfish used in experiment..

3--Observed sales at 6 price levels and estimated demand curves

USINGe Se fUNCELONG lt TOVINS s:5 so 40.00 soe sece esd bo 0d ee bGseSeceee deaease

SUMMARY

Better efficiency in production, processing, and marketing, which would
permit lower retail prices, and market promotion should help achieve wider
consumer acceptance of catfish in supermarket outlets. This was indicated in
a 1972 test in six Atlanta, Ga., area supermarkets. In terms of net returns
above the prevailing wholesale price, the optimum retail price was shown to
be $1.19 a pound. However, only one of every 150 or more Atlanta supermarket
customers was estimated to have purchased catfish at this price, which indi-
cates rather low consumer acceptance even in what might be considered a "good"
catfish market. But, if cost-saving technologies were introduced, permitting
a 17-percent price decrease to $0.99 a pound at retail, sales would increase
70 percent.

Six different prices were tested, ranging from $0.79 a pound to $1.29 a
pound, in 10-cent increments. Average sales per 1,000 customers ranged from
10 pounds per week at $1.29 a pound to 40 pounds per week at $0.79 a pound.
These levels occurred without any advertising or promotion other than a sign
at the point of sale identifying the product and its price.

The product used in the test was an overwrapped traypack of four skinned
and dressed catfish weighing about 2 pounds. The catfish were deep-chilled
and of excellent quality and appearance. They were displayed in the regular
fresh meat refrigerated case.

Given the current situation in production and marketing costs, and assum-
ing per capita consumption of catfish greater for Atlanta than for the whole
United States, an extrapolation of potential sales at $1.19 a pound yields an
upper bound of about 109 million pounds, which would require farm production
of about 188 million pounds. The present level of farm-raised catfish sales
through supermarkets is not known precisely, but it is probably less than
10 million pounds (farm weight).

A farm-level demand curve was derived from the log-linear retail curve,
using an estimated constant farm-retail marketing margin. The farm-level
curve indicates that gross farm revenue is maximized at a farm price of $0.21
a pound and a retail price of $0.93. However, considering production costs,
the farmer's net position does not become positive until a retail price of
$1.14 is achieved. This result again was consistent with the prevailing
retail price of about $1.19 a pound and indicated the estimated marketing mar-
gin and farm production costs used were fairly realistic.

A questionnaire distributed at the end of the test was used to derive a
profile of actual catfish consumers and their opinions of the product. The
replies indicated that the common cleaned, fresh catfish product was consumed
regularly by about 77 percent of those who did buy catfish, while persons who
may prefer a product with more built-in conveniences were not willing to buy
the common form.

Vi

DEMAND FOR FARM-RAISED CHANNEL CATFISH IN SUPERMARKETS:
ANALYSIS OF A SELECTED MARKET

By Richard C. Raulerson and Warren K. Trotter 1/

INTRODUCTION

Production of farm-raised channel catfish (Ictalurus punctatus) is an
important farm enterprise in local areas in several Southern States. Ona
regional basis, these local enterprises are relatively important and offer
farmpeople opportunities to more fully utilize their farm resources. In terms
of crop size and value of production, precise data on the magnitude of this
industry are not available, but yearly output ranges from 50 to 60 million
pounds worth $20 to $24 million (farm value).

The major thrust of research for this industry has centered on solving
production problems. As a result, from a purely technical standpoint, the
possibility for expansion of production is virtually unlimited. However, very
little research has been done to delineate the market potential for farm-raised
catfish. This report is addressed to the demand for fresh, processed catfish.
The processed catfish market, while not dominant, is growing in importance and
must eventually provide the main outlet for the catfish farmer's product if
the industry is to expand significantly in the future. There are several rea-
sons for this. Historically, the processing market outlet has produced lower
returns for the catfish farmer than recreational or local markets have. But
the processing market, which did not exist before 1967 has been growing stead-
ily. In 1970, about 6 million pounds of catfish--live weight--were utilized
by processing firms (3). 2/ This increased to 12 million pounds in 1971 (3),
and probably exceeded 17 million pounds in 1972. 3/ The reasons for this
growth are not clear, but they may include lack of other alternatives as the
industry expands, or growing confidence in the processed market. As farm pro-
duction increases, recreational and local markets will tend to become satura-
ted, and increasing market shares will go to consumers who eat at restaurants
and buy through supermarkets. Both of these outlets prefer a constant supply
of a uniform product, which only the commercial processor can furnis’.

The objectives of the research in this study, conducted in March and April
1972, were to:

1. Estimate a demand curve for processed farm-raised channel catfish in
supermarkets ;

2. Obtain the resulting price elasticities;

3. Use the above results in a brief analysis of the potential for cat-
fish farming; and

4. Develop information for use as a planning aid in fucure market re-
search.

1/ Agricultural Economists, Marketing Economics Division, Economic Research

Service. -
2/ Underscored numbers in parentheses refer to items in References on p. 16.
3/ Based on production through June 1972 as reported in 3);

PROCEDURE

The usual time-series or cross-sectional approaches to estimating demand
(or a combination of these) were not available for this study because relevant
retail price and quantity data for farm-raised channel catfish did not exist.
Instead of using secondary data, we selected a controlled experiment in six
stores in the Atlanta, Ga., area as a technique for generating the necessary
primary data. Specifically, the Latin square experimental design was utilized.

In the usual controlled market study, relevant market data are collected
from a specified number of market participants for a specified time period.
In the Latin square experimental design, the number of columns (stores) must
equal the number of rows (weeks) and the treatments (prices) should be so
assigned that each treatment appears once in each column and in each row. In
this study, the technique allowed the separation of the price effect from the
effects of different stores and weeks. 4/

The size of square chosen was dictated by cost and statistical considera-
tions; a 6-by-6 square allowing 20 error degrees of freedom was chosen. Test
prices ranged from $0.79 to $1.29 per pound in $0.10 increments. To eliminate
volume of customer traffic as a variable, the quantity of catfish purchased
was stated in terms of pounds per 1,000 customers. To minimize the effects of
variables other than price, the location and size of catfish displays within
each store were held constant throughout the study. In addition, store man-
agement agreed not to advertise catfish during the study.

The product used for the test was deep-chilled, cleaned, headed, and
skinned farm-raised channel catfish. 5/ Figure 1 depicts the product and fig-
ure 2 a typical display used in this experiment. The product was identified
at point-of-sale as "farm-raised channel catfish."

4/ One underlying assumption was that variations in week-to-week sales
caused by changes in the level of general business activity, weather, and the
like were similar for all six stores. Another assumption was that store-to-
store variations caused by differences in the effectiveness of merchandising
by the individual stores were reasonably constant from week to week.

5/ Exploratory work on quality aspects of the product was conducted by Ruth
Durning of Goldkist, Inc. For this work, enumerators periodically sampled
catfish that were actually on display during the study. James Marion, Director
of Research, Goldkist, Inc., indicated that deep chilling to 28° F. resulted
in a sizable increase in shelf-life of the product compared with that of con-
ventional ice-packed fish. The product maintained a fresh color and aroma for
10 to 12 days, after which deterioration was fairly rapid. Quality deteriora-
tion was caused primarily by the development of bacteria, since no molds were
isolated from the fish. Bacteria were characterized mainly as Pseudomonas,
Proteus, Citrobacter, Myxobacter, and Peptostreptococcus species, aS well as
Escherichia coli, Serratia marcescens, and Chromobacterium violaceum. When
deterioration occurred in storage, the growth of Pseudomonas and Proteus
Species increased while viable organisms of other species decreased. Taste
panel evaluation of fish indicated that taste, aroma, texture, and appearance
of the fried product remained very good up to 14 days.

Figure 1

Tray pack of deep-chilled farm-raised channel catfish
used in experiment

The product used in the study was purchased from Goldkist Fish, Quitman,
Ga.; transported to Atlanta weekly by truck, and stored at a central point
until transferred to store displays. Three marketing students from Georgia
State University transferred the product to stores as needed, maintained store
displays, and kept store sales, inventory, and customer records.

The choice of market area was based on the criterion of using a market
where consumers had prior knowledge of the product and yet where costs of con-
ducting the experiment could be minimized. The Atlanta area met this criter-
ion. The six particular stores selected sold to a cross-section of urban,
suburban, and rural consumers.

During the last 2 days of the pricing experiment, and on the weekend fol-
lowing its termination, questionnaires were given to catfish purchasers. The
object of the questionnaire was to learn something about the consumers' atti-
tudes and opinions about the product and to gather information concerning
repeat purchases, substitute products, catfish cooking methods, and consumer
socio-economic characteristics. 6/

6/ The questionnaire is reprinted in the appendix.

3

‘SSN URNA aR ES LOS lls

hh
Lb |

Peewee: —

em meee aes a ee ce on on eal Gh Ga Us os es Ge Mat et not do TE ET Uk nd nk
ee HE EE a TT ee Se ee Se
ee ee ee ee Pe
| 22 2 9 ee ee oe oe RERSES ASRS ae

A EE Ne Oe YE Eee ee en ‘So oom

we

Pe

Bee”

fake Sarr RR Sri aes

Figure 2

Typical display of farm-raised channel catfish
used in experiment

MODEL

The Latin square experimental model can be expressed as:

V5 5k =ut p. + Ys + Ty + £45k? 1s Js. k= 1.4.6

where

u = grand or overall mean,

p; = row (week) effect,

Meike column (store) effect,

t = treatment (price) effect, and

Ci ik - random error term associated with Y4 jxEh observation.

It is assumed the C4 jk's are normally and independently distributed with com-

mon variance and zero mean.

Usually, the Latin square is investigated through analysis of variance
(ANOVA), and the 6-by-6 square used in this experiment would be analyzed as in
table 1.

In this study, the concern was with prices (treatments), while the stores
and weeks were of only passing interest. Specifically, there was interest in
whether a price-quantity relationship existed. Regression can be tested by
dividing the degrees of freedom for prices (five) into a single degree of free-
dom for regression and four degrees of freedom for deviations from regression.
For the purposes of this study, the F test for regression had to yield an F
value sufficiently large to assure that regression was an important (signifi-
cant) factor in determining the size of Yak?

Within this framework, three price-quantity functional relationships--
linear (Q = a + bP), log-linear (log Q = a + bP), and log-log (log Q = log a
+ b log P)--were investigated 7/ to provide comparisons of the different forms.
There were some a priori notions as to which form would be most desirable, pro-
vided it met the usual statistical criteria. [It was expected that the two
relationships involving log forms would provide a better fit than the natural
numbers relationship because most demand functions are curvilinear. It was
also thought that the two log forms would better satisfy the specification of
homogeneous variance. A consideration of direct price elasticities led to the

7/ Due to the nature of the experiment, the relationships correspond to esti-
mates of a demand curve. At each given price, supply is perfectly elastic, so
consumers are allowed to purchase the maximum quantity they desire at any given
price. OQbservance of the maximum quantities desired (purchased) at various
prices results in a demand schedule by definition. Regression is then used to
estimate a demand curve.

Table 1--Analysis of variance with a consideration of regression for a 6-x-6
Latin square

Source of Degrees of ; Sum of

variation > freedom isquanes,. 4 pcatesddane VG
Tovah ern cel: 35 TSS af -
Weeks (W)....ceceeeee! 5 WSS WSS = WMS WMS
; 5 EMS

Stores (S)....eeeeeee : 5 SSS SSS = SMS SMS
; 5 EMS

Prices (P)....ssseeee! 5 PSS PSS = PMS PMS
; 5 EMS

Regression (R)..... 1 RSS RSS = RMS RMS

; 1 EMS

Deviations (D).....: 4 DSS DSS = DMS DMS

: 4 EMS

Error (E)ss.sssseeees : 20 ESS ESS = EMS aS

Supposition that the linear form would probably yield unrealistic elasticities
toward the extremes of the data. 8/ It was also felt that the log-log form
would be somewhat unrealistic because of the wide price range over which its
constant elasticity would have to be operative. The log-linear form was known
to yield an elasticity estimate which would change with price, and it was ex-
pected that the elasticity estimates would not be completely unrealistic at the
data extremes.

A comparison of the predicted versus actual sales, using the three func-
tional forms, is shown in figure 3.
ESTIMATES OF DEMAND

An analysis of the experimental results in natural numbers resulted in the
ANOVA presented in table 2. 9/

8/ The interest here was to look at elasticity estimates throughout the data

range, and not only at some arbitrary point such as the midpoint of the data.
9/ The raw data obtained from this experiment are shown in app. table 1.

6

OBSERVED CATFISH SALES AT 6 PRICE LEVELS

Estimated Demand Curves Using 3 Functional Forms, 6 Stores, Atlanta, Ga., 1972
PRICE $/LB.

1.30
1.18
1.06
0.94
0.82 LINEAR LOG-LINEAR LOG-LOG
: Q = 80.15 - 55.87P LOG.Q = 5.72 - 2.62P LOG.Q = 3.06 - 2.67 logP

O 12 18 24 30 36 0 12 18 24 30 36 0 12 18 24 30 36 42
POUNDS SOLD PER THOUSAND CUSTOMERS

U.S. DEPARTMENT OF AGRICULTUF = NEG. ERS 125-73 (3) ECONOMIC RESEARCH SERVICE

Figure 3

Table 2--Analysis of variance of pounds of catfish sold per 1,000 customers,
6 stores, Atlanta, Ga., 1972

Source of Degrees of ; Sum of

variation : freedom "squares Mean square i F

TORU Gh ee eee es 35 7,830.07 -- --
WeekSK( Wiese iscascest 5 608.62 1213.73 4.94
SORES CO) iecieca.creise et 5 35208. 30 641.66 26.03
PYeGCSMUR) a3 64 oo eas 0. 5 3,520.11 704.02 28.56
Regression (R).....: ] S527 0503 35278203 132.98
Deviations (D).....: 4 242.08 60.552 2.46

0 493.02 24.65 --

EVROVE Vices to «eeecas 2

-- = not applicable.

All of the calculated values of F were significant at the 0.01 level.
Only the effect of prices will be discussed here. 10/ The regression equation
corresponding to the ANOVA in natural numbers is:

Q = 80.1526 - 55.8743 P po eaitos
(t = -7.38)

where

Q = pounds/1,000 customers

P = retail price

An analysis of this equation yielded the information contained in table 3.

Table 3--Quantity demanded, price elasticity, total revenue, and net revenue at
different points on a linear demand curve for catfish, 6 stores, Atlanta, Ga.,

1972
Price per ; Quantity demanded re Total Net revenue 2/
pound . per 1,000 customers : Elasticity I/ " revenue | Pw = $.89/1b. 3/
Pounds Dollars/1,000 customers
SO 4/9 Gsaareaee 36.0] -1.23 28.45 -3.60
BOO ie erm fer tucves : 30.42 -1.63 27207 0
BOs alain 24.84 -2.23 24.59 2.48
| 20 2 eee 19.325 -3.16 20.98 SInet)
eee gee gees 13.66 -4,87 16526 A. 10

liZ Olereaterere aye : 8.08 =8..93 10.42 Bu23

2/ Returns to all retailing inputs.

1/ Elasticity for linear demand curves is N = b (5)
Q
3/ Wholesale price.

Note that demand was elastic throughout the actual range of prices used in the
study. This was a reasonable outcome, because at these prices, there were sev-
eral seafood and meat substitutes for channel catfish. Total revenue decreased
as price increased because demand was always elastic. The net revenue schedule

10/ The weeks were significant because of an apparent overall buildup of
store sales during the course of the study. The store effects were significant
because of an apparent difference in sales between urban and nonurban stores.
However, these observations came to light after the data were collected; there-
fore, any further statistical analysis of weeks and stores to test these obser-
vations would not be valid. Still, these observations should help to guide
future studies.

was based on a wholesale price of $0.89 a pound which was the approximate cur-
rent wholesale price for the deep-chilled product used in the study. Assuming
constant costs of retailing, this schedule shows that the profit-maximizing
price from the retailers' point of view would be about $1.19 a pound. This was
a common retail price in the Atlanta area during the period immediately preced-
ing tnis study.

The regression equation for the log-linear functional relationship was:11/

log, Q = 5.7177 - 2.6211 P Ro = .97
(t = -11.33)

Note that this relationship fitted the data better than the linear relationship
(as expected). An analysis of this equation yielded the information in table
4. The demand was elastic but not constant throughout the data range. The
demand elasticity increased as price increased because of the functional form
(as was the case with the linear relationship). However, the change in elas-
ticity as price increased was less abrupt, so that a reasonable elasticity
estimate of -3.38 was obtained at the highest study price of $1.29. An exami-
nation of the revenue position of the retailer indicated that a retail price
Slightly greater than $1.29 would maximize his net revenue.

Table 4--Quantity demanded, price elasticity, total revenue, and net revenue at
different points on a log-linear demand curve for catfish, 6 stores,
Atlanta, Ga., 1972

Price per Quantity demanded Bete Total Net revenue 2/
pound . per 1,000 customers ; Elasticity U/ * revenue : Pw = $.89/1b. 3/
Pounds Dollars/1,000 customers
$0.79 26. S 38. 36 22507 30. 30 -3.84
1619 eee ema 29.52 -2.33 26.27 0
OQ Stes e ead 22:7 | -2.59 22.48 Zoey
(PAO Ss ane ee a le 47 -2.86 19.04 3.49
ida UC ea eer rete 13.45 -3.12 16.01 4.03

(fi AC eS ie See 10.35 #3500 13505 4.14

1/ Elasticity for the log-linear relationship is N = bP.

Proof: log Q = atbP > Q = e@*>P, ag p = (bettPyeatbP) p = bp, ged.
dP Q
2/ Returns to all retailing inputs.
3/ Wholesale price.

11/ ANOVA tables for the log-linear and log-log forms will not be presented.
The logarithm forms use natural (base e) logarithms.

The regression equation for log-log functional relationship is:

log, Q = 3.0593 - 2.6708 log, P po = .97
“" (t= -11.09)

This relationship fitted the data as well as the log-linear relationship. An
analysis of this equation yielded the information in table 5. The outstanding
difference between this relationship and the log-linear was the constant elas-
ticity associated with the former. From the retailers' point of view, net
revenue was maximized at about $1.42 a pound (assuming a correct prediction
beyond the range of the data).

Table 5--Quantity demanded, price elasticity, total revenue, and net revenue
at different points on a log-log demand curve for catfish, 6 stores,
Atlanta, Ga., 1972

Price per : Quantity demanded las ; Total : Net revenue 2/
pound * per 1,000 customers ° Elasticity I/ " revenue : Pw = $.89/1b. 3/
Pounds Dollars/1,000 customers
On / oN ee 40.00 267 31.60 -4.00
“ot ee Serene: 29.09 -2.67 25.89 0
POU ES & tei, : 21.89 -2.67 216607 2.19
OO eRe... ee : 16.93 =2207 18.45 3:39
ee ne ee 13.39 -2.67 15.94 4.02
VAZ Os tenes : 10.80 -2.67 13293 4.32

2/ Returns to all retailing inputs.
3/ Wholesale price.

1/ Elasticity for the log-log demand relationship is N = b.

A striking similarity among the elasticity estimates derived from the
three functional relationships occurred approximately at the midpoint of the
data. The ranges in elasticity based on the prices of $.99 to $1.09 a pound
were -2.23 to -3.16, -2.59 to -2.86, and -2.67 (constant). Actually, this re-
sult would be expected since all relationships fitted the data well and there-
fore tended to have the same slopes at the approximate midpoint of the data.

POTENTIAL FOR CATFISH FARMING

Although results of a marketing study for one geographic region are not
statistically valid for predicting sales in other localities having different
consumer characteristics, some useful conclusions concerning the overall po-
tential of processed catfish can be drawn. A recent USDA study showed that
consumer acceptance of catfish was higher in the South and Midwest than in the
Northeast and West Coast Region of the United States (4). Sales of catfish in
the Atlanta area should be reasonably representative of the South and Midwest,
but would probably be somewhat higher than the average per capita sales for
the United States. Thus, an extrapolation of sales for Atlanta should yield
an upper bound on potential sales for the entire United States under the
present level of consumer acceptance (table 6).

10

Table 6--Upper bound estimates of annual U.S. sales of processed catfish in
supermarkets at various prices, 1972

: Weekly sales :

Price WieS? ; UeS2 Annual :
per. 1,000... : : > Annual farm
Der Atlanta : cus tomers : sales : U.S. > equivalent 3/
pound us toniens . per week 2/ . per week ° sales ° =
Pounds Thousands ---------- Million pounds----------
SON 7 : 9381.36 156 ,000 5.98 311.0 536.2
rote i Re meer 29.52 156,000 4.61 239.7 Ale
SOON cee. aus 2e3f\ 156,000 354 184.1 B17 4
TOSS es a 17247 156,000 243 142.0 244.8
LO ase 13345 156,000 22110 109.2 188.3

ASS as Per are 10:35 156,000 1.61 Sid 144.3

1/ Based on log-linear demand curve.
2/ Based on shopping habits in the Atlanta area.
3/ Based on 58-percent dressout.

At the current Atlanta price of around $1.19 a pound, the maximum esti-
mate of sales is a little over 109 million pounds, and would require farm pro-
duction of about 188 million pounds. Relative to other meat production, this
is quite small. For example, in 1970, Arkansas alone produced over eight
times this many pounds of broilers (2). In the Atlanta study, where no pro-
motional effort was made, less than one of every 150 customers purchased cat-
fish at $1.19 a pound. 12/ Even at a price of $0.79 a pound, only one of
every 50 customers purchased catfish, which indicates that, at the present
time, catfish does not have wide consumer acceptance in what might be consid-
ered a "good" catfish market area.

A more promising result of this study for the catfish farmer is the rela-
tively high demand elasticities over the range of prices tested (table 7).
For example, if the industry becomes more efficient so that price can be re-
duced from $1.19 to, say, $0.99 a pound, this study indicates that catfish
sales would increase by approximately 70 percent. Also, results from other
studies show that products with elastic demand are responsive to promotional
activities. Thus, it can be hypothesized that catfish sales would be respon-
sive to promotion. A combination of promotion and lower prices (through
better efficiency in production, processing, and marketing) may allow for a
fairly substantial increase in output. 13/

12/ Individual packages contained about 2 pounds of catfish; so, assuming
each catfish customer purchased only one package at a time, a maximum of 6.7
catfish purchasers per 1,000 store customers is indicated, or only one of
every 149 store customers.

13/ It is recognized that information-promotion activities may change the
elasticities as estimated by this study. Additional research is needed to
furnish further evidence regarding this matter.

1]

Table 7--Demand elasticity and net revenue based on 3 forms of a retail demand
curve for farm-raised channel catfish, 6 stores, Atlanta, Ga., 1972

peice Linear Log-1inear Log-log
per : Elas=4 = Net EL Blasa ai Net NpeliaS= 40. Net
pound , ticity } revenue 1/< ticitys, revenue l/s .ticity revenues |/
$/1,000 $/1,000 $/1,000
cus tomers customers cus tomers
SOu/Onceaeat) Sees -3.60 -2.07 -3.84 -2.67 -4.00
73) ane ea 2 =1 763 0 -2.33 0 -2.67 0
ROO sce nan ene eee 2.48 -2.59 ee | -2.67 2.19
1(10)° eee ere eS ea {L6) 3.05 -2.86 3.49 -2.67 3.39
VAG ea ct Hae Oy 4.10 -3.12 4.03 -2.67 4.02
(29S ata HO. 9S 3.23 -3.38 4.14 -2.67 A.32

1/ Assuming a constant wholesale price of $0.89 per pound, net revenue refers
to returns to all retailing inputs.

This study was designed to estimate a demand curve for processed catfish
at retail. However, catfish farmers would be rightly concerned with the farm-
level demand. As pointed out earlier, price and quantity data appropriate to
estimating demand are simply not available. However, by considering estimated
processing and marketing charges and taking into account processing and market-
ing losses, the farm-level demand can be derived from the retail demand. The
farm price may be estimated from the retail price by use of the following con-
version: 14/

Pe = i - $0.532
1.91
where

Pe = farm price and o = retail price.

14/ This conversion is based on app. table 2. A careful examination of this
table reveals that the farm price can be stated in terms of the retail price
according to the equation

2 (C2 cer)
i aoe fue eee
(i/a)/(1 - b)

a = retail price and Pe = farm price,
C5 C, and C. = costs of processing, wholesaling, and retailing,

a = dressout fraction, and
b = fraction of product loss at retail

12

Using the schedule of estimated farm prices and quantities, the resulting
farm-level demand curve is:

Tog, Qe = 4,92 - 4.85 Pe

This equation can be analyzed in terms of demand elasticity and revenue
position of the farmer. These results are presented in table 8. Since the
derived demand curve is presented in terms of pounds per 1,000 customers, only
the changes in the revenue position of the farmer are of interest. Note that
the demand is inelastic in the lower price ranges, whereas the retail demand
was always elastic over the price range used in the study. Total revenue is
maximized at unitary elasticity 15/ or a farm price of $0.21 and a retail price
of $0.93 a pound. However, the farmer's net position does not become positive
until a retail price of $1.14 a pound is achieved. The farm-level demand curve
derived above is based on sales of six supermarkets in Atlanta; however, the
results can be thought of as ball-park estimates. These results are consistent
with the retail price of about $1.19 per pound that was current in Atlanta at
the time of the study.

Table 8--Price elasticity and net revenue per pound at different points ona
derived log-linear farm-level demand curve for catfish, 6 stores,
Atlanta, Ga., 1972

Retail price : Farm price ae Gross farm Net farm

per pound > per pound I/ : Elasticity > revenue : revenue 2/

Dollars Dollars/1,000 customers
GOI9 OG bs. sated dees > 0.135 -0.65 9.38 -12.84
SOO Es oes eases eee! 0.187 -0.91 10.54 -7.50
O}S) ics a er 0.240 -1.16 10.41 -3.47
1a OS) Ae ee : 0.292 -1.42 9.74 -0.94
PEs caters arcsec 5 'd 0.345 -1.67 8. 86 +0.64

HO eee naeee are te! 4.0 \sle wr eres 0.397 -1.93 7-89 +1,52

1/ Price paid to farmers at processing house door.
2/ Returns to farmer labor and management assuming constant costs of produc-
tion, harvesting, and transporting at $0.32 a pound.

In the analysis of retail sales presented earlier, it was pointed out that
there is probably some room for expansion via market development and cost-
saving efficiency. The analysis of farmer returns assumes that no market de-
velopment occurs and that technology is constant. Under these assumptions,
further allocation of farm production to supermarkets would not be desirable.

15/ Unitary elasticity indicates maximum revenue for all demand curves which
are less convex than a rectangular hyperbola.

13

QUESTIONNAIRE RESULTS

Although the questionnaire used at the end of the study did not yield
statistically valid results because of problems in sampling the nonrespond-
ents, there was a 45-percent return (48 of 106) and some of the results are
presented. 16/

Of those who returned questionnaires, 77 percent indicated that they had
purchased catfish from a grocery store before. Since the demand results indi-
cated that less than one of every 150 customers purchased catfish at the nor-
mal price of $1.19 a pound, then only one of about every 650 customers was a
"new" customer. At the relatively low price of $0.79 a pound, only one of
about 200 customers was willing to purchase catfish for the first time. This
strengthens the observation made earlier that catfish has a marketability
problem, even in what is considered a good catfish market area.

Answers to other questions indicated that consumers purchasing catfish
tended to be traditional catfish eaters. For example, only 6 percent indi-
cated that catfish need to be easier to cook, 94 percent indicated that they
preferred fresh catfish (while no respondents preferred precooked or prebread-
ed forms), and 98 percent indicated that they fried their catfish--the tradi-
tional Southern cooking method for this fish.

The product used in the study was apparently acceptable to consumers. Of
consumers who answered the questionnaire, 58 percent indicated that they pre-
ferred farm-raised catfish, 8 percent favored wild catfish, and 33 percent
were indifferent.

RECOMMENDATIONS FOR FURTHER RESEARCH

The controlled market experiment using a Latin square design appeared to
work quite well in an investigation of the demand for catfish in the Atlanta
area. Similar research in other geographical areas is needed to better esti-
mate the demand for farm-raised channel catfish in supermarkets. 17/ Future
studies could incorporate more objectives with little extra investment. For
example, the possible difference in sales between urban and nonurban stores
could be readily investigated.

The questionnaire analysis indicated that most of the customers who pur-
chased catfish had previously purchased catfish in supermarkets, and that vir-
tually all of those customers preferred fresh, fried catfish. Therefore, a

16/ Telephone interview was attempted as a means of sampling the nonrespond-
ents. Unfortunately, the procedure did not work well. Of 16 nonrespondents
contacted by telephone, five said that the party who could answer the ques-
tions was not at home, four were wrong numbers, three directly refused, one
was not a working number, and three answered the questionnaire. App. table 3
Summarizes questionnaire replies.

17/ The statistical results obtained in this study indicate that a 5-by-5
Square allowing 12 error degrees of freedom would probably provide statisti-
cally significant results in future studies.

14

study of different product forms (precooked, breaded, and so forth) may reveal
a segment of consumers who were not reached by this study and who prefer a
convenience product.

Since so few customers purchased catfish during this study, and since
most consumers who answered the questionnaire indicated that they liked the
product, the possibility exists that the market for farm-raised channel cat-
fish could be expanded if additional consumers tried the product. Therefore,
a study of the response to catfish promotion and the nature of the repeat
business (the advertising decay curve) is recommended.

REFERENCES

Greenfield, J. E.
1970. Economic and Business Dimensions of the Catfish Farming
Industry. U.S. Dept. Interior, Ann Arbor, Michigan.

U.S. Department of Agriculture.
1971. 1970 Agricultural Statistics for Arkansas, Fayetteville, Ark.
Statis. Rptg. Serv. in cooperation with Univ. Ark. Agr. Expt.
Sta. Aug.

U.S. Department of Commerce.
IZlETZ:
Catfish Processors’ Report, Little Rock, Ark. National
Marine Fisheries Service (various issues).

White, Elizabeth D.

1972. The Market Potential for Catfish, The Catfish Farmer, 4 (6): 24.
Nov.

16

APPENDIX

Appendix table 1--Raw data for catfish marketing study, 6 stores, Atlanta, Ga.,

1972
Weaie Store Store Store Store Store Store
ae | 2 cd .

Week 1: :

Sales (pounds)........: 79.84 74.90 67.79 59.43 65.43 71.68

Customers (number)....: 6,541 4,404 4,245 2,300 5,278 5,810

Price: (S/ui bis) ss noes cess . 89 79 1.09 1.19 1.29 .99
Week 2: ;

Sales (pounds)....-...2 27.14 60.63 85.41 148.19 124.35 28.88

Customers (number)....: 6,665 3,978 3,863 2,863 5,294 6,015

Pisticet$/1bit) scot occ: 1.29 89 99 .79 1.09 1.19
Week 3: ;

Sales. (pounds)........: 138.45 33.07 30.43 91.46 126.87 41.09

Customers (number)....: 6,123 35550 S5002 25330 4,446 5,170

Picea SPD) ccccc0 ees 79 1.19 1.29 99 . 89 1.09
Week 4: :

Sales (pounds)........: 49.34 Cyd 79.07. 119.86 165.08 156.23

Customers (number)....: 4,672 3,591 3,727 2,191 M23 4,095

PriceUS/1be).cciccssst 1.09 1229 1.19 . 89 99 .79
Week 5: :

Sales; (pounds)........: 39.17 71.13 130.64 70.38 305.16 48.54

Customers (number)....: 6,724 4,237 4,089 2,471 5,344 5,860

Pricer (G/UDs) 6 e660 Se TS .99 . 89 1.09 .79 1.29
Week 6: :

Sales (pounds).........: 75.91 27.38 206.44 A224 139.59 156.67

Customers (number)....: 6,524 4,194 4,011 2,416 5,630 5,808

Price IG GPUDI) «cies vee set 299 1.09 19 1.29 BIBS “09

AZ.

Appendix table 2--Estimated value added approach to price determination per
pound for farm-raised channel catfish, 1972

Item Amount
Producer: :
Production -EXPeNnSes <..4006 st cakes een eee $0.26
Harvest and tranSportatiOnies.. ca «sees oer 06
PYPOF 1:G,,cs ort % ine @ acats Quatre siecle eee eusnecdl oh Ieee ots .02
GrOWET: I CO%.2.10, 005s ernie eas cenee see setae ; . 34
Processor: :
PRACEOT PROCUGE, 1) Ratclowe jucok7srctene srann emia enerrenst 09
PROCESS INGPEXPENSEs ce ieic «cia che choueieye overt rNoterere cases lid,
PRONE aA Spee farce en alasestiets Onle ere anes ge Nera sreeaes eeegnie tae ent anare 0]
F.0D. *PhOCeSSANG) PlanGeads auc oe tate 6 aretats ; wy)
Wholesaler: 2/ ;
F.0D. PROCESS ING) PlaNtcas.cs ewe ee cen ure ce cc : AE,
Selling raid tranSpOrtatkiOliea:. «cues eae acc san0)
FPO Spl Certara k atschiaity Se ssmesisueeer name oe aaba ws See aa epeio eee ieee : 20)
Wholesale DRICELe eee eee .88
Retailer: :
PRiCerOt sprOGUC Emo) ft iccisurms dg eicoe ease ceretes s : 98
EXONS Crardatirccerstiers arava > i aaeis aroe le marnns a cree los cet : 20
POG scrote terareterece ue sttzenelain'e sucbe cits cuales Mencieen siete 01
RECA OV CO cris: 2.5.2 a Sisal ar siaiteerere ereee Stee eee 1.19

1/ Assuming a dressout of 58 percent.
2/ Processor may assume this function.
3/ Assuming a 10-percent product loss.

Source: Adapted from (1).

18

Appendix table 3--Tabular results of channel catfish questionnaire, 1972

Percent
Question : or

average I/

Have you (or someone in your family) ever bought catfish from :
a grocery store before: :
N

Ones revere arantrenatone sols chi ecarsM sane en aiS, 9 ae aoe ees os ee adiadse 23
ME Siem tcvenet cuatie otopeirsiiareierbcsvesensiale wie d.Gha ah Ole o.6-elene mie 6 oo ead aie Vi).
(If yes) How many other times in last 6 weeks?...... ; 1.8
Did you buy catfish anywhere else in the last 6 weeks?

Oa amentne Gupaaline nhs ain Oiaae sdb ane D stare a eae S anes Reese: 79
Me Stamenersyenetotnetus er aisausieoiets 6 auch clave Wile Bow ineca umes ah als wa peta se 2]
(Gives Sedhood marketesis. ce ceacececctsastueecces’ 8
CUipaieS)) s OtheY saci cies etmek oceans shou cowascsees eet 12

In the future, how often do you intend to buy catfish from
a grocery store?

We deast once every 2 WeekSs....iisi06ssbeceses ceca : 60
INE MICAS JONCE ia KINOIVEN. 4 4 o/ecnssssce setup wag ea ates es eR oes 35
PiphedS it ONCE a VEAN. giaciu tp ocd ae aie see eee bake ee BS ans : Z
INEVIG HR apetensecrers tye cisieceieass occ ele ewe w ae uae + Rais wwe ace a ee 2

Concerning the catfish you bought recently in your grocery
store, wnich of the following product improvements are
needed most? (Check .as many as apply)

Ei siieie EO CO OK iatnects 2 arc. e tae ese coals aoe GU de hue wees ae 6
BOWEE NICE Seana Atkin tthe wen eGtl ana oe Ra wateley ae 44
DEIEC Ip SEG Sit Che ee eta ao ce etapa Ebciaee.s: se ted wd 2 BERRI Aloe tow Sac eee 17
PSS Set OMG Sucre stspecayteare ters eis cnn s 210 Gens aucleteceneis Wanaka Slave 5a : val
BCHAE Nees SINIE lh lasnctienstouesst eteiancite-s eeifecaus asec ue ur eto aiaue a Rvee whats : 8
AOC meOntmeenite S Coen cwen pacer eva ore ters uc sece. ct caver cue on oad levapa acheive arate eueanere 29
How would you prefer to buy catfish? (Check as many as apply):
Fave as ene otis aa tor ote cco scarab allasosw: uate rere a bb ewe sate «bo wie wats 19
FATES Pepe arcater ates sate ate ato carter Eee ak aria 5 are ve erat ga ams Saedasesnoen See 94
fea ne nthe ten as crn Guise. chinvatrels Gaseela sea @ Wie ere wteGle ene we eee 15
PLE TAMe ANG SICH WIG iavieraitesa.ce elec cos ciate: a.cle aod nueva cee 4-8 erate oa eres 0
Bre aGederVOzelliaca waGieed ic css Coes ok bee Chie uae cys 0
BYCAGEE TCESIos enews oisaresiae ceaiee se Cade seer eaeee eee ee 0
See footnote at end of table. Continued

Appendix table 3--Tabular results of channel catfish questionnaire--Continued.
Percent:
Question f or
average 1/

ee -Y’- "rer
°

Which kind of catfish would you rather buy?

Farm-1aised. ic. ses esssc-eensscc.o cen mtce ek eee 58
Caught: from a iriversor. laken.. wachaeercss oe eee: 8
NO. Ci FTEVENCCs 92.05 52ers b arsus secant alerdeeialcas Cee eee 33

If you had not made your recent purchase of catfish, which
of the following would you have bought in place of catfish? :

BOOT sepia Sra: a igs di ehialeeutns vecolen aera esetene, stale cael cares uccepeeseteeas eee 33
GRIN CKOM ysis to 59 .d-e a eespces OP atecnicze cals hee eros aerate : 19
Other ‘fish Or S@afOOd.0s o-oo aes tee ee ee oe 48
POVEK niscd-cavons Sin Sve bes’ Snes eke otra es eee 10
How is catfish usually cooked in your home? :
RYGds S102. aos ac cease Gets oP Date Rocce ee eee ere eee ; 98
BAKGG oc sve Sone. i srerhce tues Sacepiss Sse She SS Oued ene ate 2
BY OT VEG soo Si citiene uit a.b keels one castesaue lo trcheasierah ts eet ee 6
BaWDC CU sia ae alo cctccave oo 5 910 eesate «, sheteke fetes stearate Mareen 0
What is the age of the head of your family? :
ESS: CMAN: (25 cic 2a restate serene ee iste Guctel Skeuarsee, Sysaucie tite renee c ; 8
DSRS Mei arataen telatins totoue Or aie ter atta oscks Oe a eee IO IE 12
BH Ac Mca alas ora gs duarenah wang. aces Piaoheraiatal one alain Ga pl anenk Bigiacerace te D hase anS 42
BO! AM OVE Rs iad i cee a eit be.c0h0'% voralaldeioye ce te vale ealtaab Aaa eee ait ey)
How many people in your family are living at home, including :
VOUNS © IGE o axstcre av stores cake asso aectoue tas Wins ate tiaeie ee aacleusie eens Mico ee B25
What is the approximate annual income for your family? :
ASS) Chiat S5:s 000s. 2c cccuentisvonvciaienaroureietereusncce eee ree 13
G5 OOO S95 99s rsd carat 6 toe vorsvasayocs ean-eGoievubsin aun tuoteyenexs seasuetst bes 33
GTO OO Os atid ON Ci yor et d overs co ncvaxsseuceon levstonescuevaveucy nisl neck Oe 54

1/ Percentages may not total correctly for some questions because of round-
ing or multiple answers.

20

Questionnaire

PURCHASES OF CHANNEL CATFISH

Dear Consumer:

By purchasing farm-raised channel catfish, you have participated in a market
study of this fish. Your answers to the questions below will help us to com-
plete this study which is designed to help in providing you with food that is
wholesome and enjoyable. Please answer all the questions even if you or mem-
bers of your family may not eat catfish regularly, and mail this questionnaire
promptly in the envelope which needs no stamp. Thank you.

10.

Atk.

Richard C. Raulerson
Agricultural Economist

Have you (or someone in your family) ever bought catfish from a grocery
store before? No / / Yes / /
(If Yes) How many other times in the last six weeks? Number __

Did you buy catfish anywhere else in the last six weeks?
NowLa= Yes /., _/ (If Yes) Where? Seafood market / , 7 Other

How many pounds of catfish would you prefer to buy at one time? pounds

In the future, how often do you intend to buy catfish from a grocery
store? At least once every two weeks / / At least once a month / /
At least once a year / / Never / /

Concerning the catfish you bought recently in your grocery store, which
of the following product _improvements are needed most? (Check as many as
apply) Easier to cook / (of Lower price / , ey Better taste / /

Less bones / / Better smell / / None of these / /

How would you prefer to buy catfish? (Check as many as apply)
Frozen / / Fresh / / Fillet / / Heat and serve / / Breaded frozen / /
Breaded fresh ee:

Which kind of catfish would you rather buy?
Farm-raised / / Caught froma river or lake / / No difference / /

If you had not made your recent purchase of catfish, which of the follow-
ing would you have bought in place of catfish? =
Bee te, ~7 Chicken iis _/ Other fish or seafood / , “7 Pork Lay

How is catfish usually cooked in your home? —__
Fried / / Baked / / Broiled / / Barbecued / /

What is the age of the head of your family?
Less than 25 / / 25-34 / / 35-49 / / 50 and over / /

How many people in your family are living at home, including yoursel f?
Number

What is ie approximate annual income _for your family?
Less than $5,000 / 7 $5,000-$9,999 / / $10,000 and over / , ws

21

UNITED STATES DEPARTMENT OF AGRICULTURE

WASHINGTON, D. C. 20250
POSTAGE AND FEES PAID
U.S. DEPARTMENT OF
AGRICULTURE
OFFICIAL BUSINESS

PENALTY FOR PRIVATE USE, $300 FIRST CiASS
AGR 101