AGRICULTURAL MARKET RESEARCH CENTER
FOOD AND RESOURCE ECONOMICS DEPARTMENT
Institute of Food and Agricultural Sciences
University of Florida
Qainesville, Florida 32611
.2 J 2i1 150"1
?1, r r --
A WEEKLY PRICE DETERMINATION MODEL FOR
J. Scott Shonkwiler
Staff Report 7
Staff papers are circulated without formal review
by the Food and Resource Economics Department.
Content is the sole responsibility of the author.
Food and Resource Economics Department
Institute of Food and Agricultural Sciences
University of Florida, Gainesville, Florida 32611
The Florida Agricultural Market Research Center is
a service of
the Food and Resource Economics Department
Institute of Food and Agricultural Sciences
The purpose of this Center is to provide timely, applied research
on current and emerging marketing problems affecting Florida's agri-
cultural and marine industries. The Center seeks to provide research
and information to production, marketing, and processing firms, groups
and organizations concerned with improving and expanding markets for
Florida agricultural and marine products.
The Center is staffed by a basic group of economists trained in
agriculture and marketing. In addition, cooperating personnel from
other IFAS units provide a wide range of expertise which can be applied
as determined by the requirements of individual projects.
TABLE OF CONTENTS
LIST OF TABLES.......................................... ..****. iv
LIST OF FIGURES........................... ................... iv
INTRODUCTION.............................. ..................... 1
Florida Celery Production and Marketing..................... 2
Price Determination Description........................... 4
Market Equilibrium........................................ 6
Model Specification.................... .................. 10
The Estimated Model ......... ................................ 15
Market Implications...................................... 20
REFERENCES................................ ...................* 24
LIST OF TABLES
1 Florida celery: days to maturity by month planted..........
2 Structural results.......................... ............
3 Structural impact and interim elasticities................
4 Reduced form results..................................
LIST OF FIGURES
1 Quantity determination in the disequilibrium market........
A WEEKLY PRICE DETERMINATION MODEL FOR FLORIDA CELERY
J. Scott Shonkwiler
Florida celery production is concentrated both geographically and
organizationally. The two major producing areas in three Florida counties
and a dozen growers account for over 90 percent of the state's celery
marketing during a crop year. Over the last ten years, Florida celery
shipments have comprised about 42 percent of total U.S. celery supply
during the Florida season (November through June) with the remaining
quantities being supplied by California.
Unlike many fresh vegetables, the timing of celery harvesting is
typically flexible within weekly bounds, and once harvested the crop may
be stored for short periods with little deterioration. The marketing of
most Florida celery production is performed by a single sales cooperative
with the capability to set prices. The characteristics of celery supply
and marketing make analysis of the short-run price determination mechanism
for Florida celery rather unique. This report analyzes the physical,
institutional, and economic forces which shape weekly celery harvesting
and pricing decisions. A structural econometric model is formulated to
represent the dynamic operation of the market.
J. SCOTT SHONKWILER is assistant professor of food and resource
economics, University of Florida.
The present study proceeds with an overview of the operation of the
Florida celery market followed by discussions of the price determination
mechanism and its approximation to competitive equilibrium. The final
sections of the paper presents a weekly supply-demand model for represent-
ing the short-run operation of this market and consider its appropriateness.
Florida Celery Production and Marketing
Florida celery matures from field set plants between 70 and 110 days
after planting.1 As shown in Table 1, maturities are longer in the late
fall, but shorten during the spring months. Growers typically have several
weeks in which to harvest their crop depending on temperature and moisture
conditions. Once harvested, the celery is washed, packed and cooled in
preparation for shipment. Total amounts harvested within a given week
are not necessarily marketed since short-term storage is possible. A
recent study has found that most celery shipments go directly to retail
outlets although wholesale celery markets do exist in larger metropolitan
areas (Mathis and Degner). This phenomenon has been observed for other
Florida winter vegetables and has led Bohall to speculate that the price
determination process has shifted from the wholesale level terminal to the
Approximately 95 percent of celery marketing are overseen directly
by the Florida Celery Exchange, a voluntary marketing cooperative which
represents all major Florida celery producers, Members pass complete
market control over their celery to the Exchange by means of contracts.
The Exchange constantly monitors market developments, growing conditions,
Table l.--Florida celery: days to maturity by month planted.
Season Aug. Sept. Oct. Nov. Dec. Jan. Feb. March April
1972-73 94 102 104 110 105 96 93 92 66
1973-74 91 106 104 104 107 93 98 90 76
1974-75 94 95 98 95 97 93 85 82 70
1975-76 100 94 97 103 105 98 92 81 66
1976-77 96 93 100 106 110 103 88 82 77
1977-78 101 101 109 115 112 108 93 80 70
Mean 96 98 102 106 106 98 92 84 71
and daily production of both Florida and California celery. Then, in an
effort to maintain stable weekly markets, the Exchange usually sets a
Florida FOB celery price on each Monday and Wednesday morning. Infre-
quently, prices are changed on other days during the week in order to
accommodate rapidly occurring market developments. If prices are lower-
ed on such occasions, buyers are given price protection for purchases
made subsequent to the last price quotation. Thus, on a week to week
basis, Florida celery price levels and shipments reflect the operation
of the Celery Exchange in a market with a structure that has been re-
ferred to as a cartel (Brooke and Jung).
Although celery is a farily perishable commodity, the flexibility
in time of harvest and its short-term storability imply that weekly
supplies are not totally predetermined by past plantings. In fact, the
price settingcapability of the Celery Exchange suggests that quantities
supplied and demanded, rather than prices, adjust to produce a market
equilibrium. Of course, prices set by the Exchange each Monday do not
necessarily correspond to actual average prices received by Florida
celery producers the entire week; the Exchange may adjust price on
Wednesday (or more frequently) to accommodate unforeseen changes in
Florida and California supplies or national forces such as variations in
demand conditions. Two important prices then are generated by the
operation of the Florida Celery Exchange. The Monday price may serve as
a signal to Florida growers. The subsequent adjustments to the Monday
price may prompt different grower responses depending on whether prices
have changed significantly and whether physical factors allow harvest
These features and their interpretation imply.that weekly Florida
celery price determination must account for producer response, the short
term storable nature of celery, the operation of the Exchange, components
of demand and California competition. The formulation of a model to
appropriately replicate the operation of this market is strongly conditioned
by assumptions concerning the degree of competition existing. Specific
concerns are directed to deciding whether prices adjust sufficiently
during the week to maintain balance between quantities supplied and
demanded. Therefore, the exact process by which prices are determined in
the Florida celery market must be further analyzed before modeling
decisions can be made.
Price Determination Description
The fixing of a non-negotiable market price by the Florida Celery
Exchange represents a substantial departure from the theoretical price
determination process of a competitive market. The unique characteristics
of the Exchange's pricing system, however, imply that the weekly average
prices generated during the period under study closelyapproximate the
price patterns which would likely exist given competitive equilibria.
To understand the pricing mechanism involved and its relation to theoretical
price determination conditions, the pricing activities of the Exchange
will be developed more fully.
The Manager of the Celery Exchange meets with the three-member
pricing committee every Monday and Wednesday morning and presents detailed
market information collected during previous days. Specifically, he has
compiled the previous days' harvesting, sales, and inventory levels of
the member celery producers. These data are collected daily via the
private communication system linking major producers and shippers with
the Exchange. Further, he has current estimates of that day's production
as well as production estimates for the entire week. This information
fairly well summarizes the supply side of the Florida market when merged
with observations on weather and labor availability.
To collect information on demand conditions, the manager usually
contacts Florida celery salesmen a day or two before the pricing committee
meets. From conversations with them, he elicits the number of new
buyers in the market, changes in size and frequency of orders, and other
information about distribution and demand developments. A similar
method is also employed to infer the general tone of the California
celery market. Calls are made to a select group of California shippers
and a subjective appraisal is made concerning their current market
conditions. To augment this information, U.S.D.A. Market News reports
are constantly checked in order to monitor shipments and representative
prices for the California crop.
With this information base, the Exchange sets non-negotiable prices
for the major size categories of celery. Because much of the information
used is non-quantitative, the Exchange employs no empirical models in
its price setting activities. Rather, the collective experience and
knowledge of the manager and the pricing committee translate current and
projected market conditions into a price that is expected to clear the
market. If market conditions are so uncertain that rapid developments
are expected, then the pricing committee may decide to meet more often
than Monday and Wednesday. In this manner the Exchange closely monitors
Florida cold storage inventories, the California market, and exogenous
developments such as weather in order to set a price which presumably
clears the market.
The concept of a market clearing price is, of course, fundamental
to the competitive equilibrium presumed to occur on a week to week
basis. The market clears in the sense that the Exchange does not control
supply in the short-run.2 During the period under study, there were no
harvesting holidays or other supply restrictions imposed. However, the
Exchange is constrained in its ability to manipulate prices for long
periods because of the perishability of the crop. This fact is evidenc-
ed by the considerable attention given by the pricing committee to
unsold celery inventories held at the beginning of each price setting
period. Breimyer (p. 116) terms the price determination mechanism
employed by the Exchange as "supply-demand estimation pricing." This
terminology implies first that supply curves are relevant concepts for
identifying output and secondly that the set price actually acts to
clear the market. Because no supply restrictions are imposed and for
other reasons discussed in the following section, it will be assumed
that a supply curve exists for the weekly observational period. The
second implication concerning the ability of the administered price to
equilibrate supply and demand suggests the need for a more thorough
If the Exchange price does not permit the market to clear, then the
quantity actually transacted either lies on the demand or the supply
curve, but not at their intersection. Fair and Jaffee3 represent this
market with a demand and supply equation of the form
Dt = caXDt + alPt + t (1)
St = oXSt + B1Pt + mt (2)
where XD and XS represent demand and supply shifters respectively and P
represents the exogenous price. The quantity transacted in any period,
Qt equals either Dt or St because price does not necessarily adjust to
permit Dt = St = Qt. A third equation to complete the disequilibrium
model then is
Qt = Min (Dt, St), (3)
A graphical representation is given in Figure 1.
The heavily shaded lines in Figure 1 represent price-quantity com-
binations which may occur in a disequilibrium market. When demand
Figure 1.--Quantity determination in the disequilibrium market.
(supply) at Pt exceeds supply (demand) at Pt then shortages (gluts)
exist and transactions occur along the darkened supply (demand) curve.
Therefore, in a given period either supply or demand is unobservable,
and concern is focused upon whether the quantity transacted follows from
equation 1 or equation 2. An indicator of which relation is responsible
for the observed transacted quantities is the direction of the price
changes from period to period. Fair and Jaffee point out (p. 501)
Most dynamic theories of price-setting behavior
formulate the change in price as some function of
the excess demand existing in the market. If the
change in price and excess demand are related in this
manner, then the change in price may be used as an
indicator of the amount of excess demand (or supply)
in the market.
If the.price change is instantaneous and complete then supply and
demand will be equilibrated. But, the magnitude of the response rate
must be referenced to the length of the observational period. In the
present context, the celery market may be in disequilibrium from Monday
to Wednesday until prices can be adjusted to mitigate the consequences
of excess demand or supply. Viewed over a weekly period then the market
may nearly replicate the competitive solution.
To illustrate this feature of lengthening the observation period so
as to allow more adjustments, consider the price adjustment mechanism
suggested by the present disequilibrium analysis
Pt = (Dt St). (4)
Clearly, market adjustment depends not only on the magnitude of y and
the length of the observation period, but on the number of adjustments
made (or allowed) per observation period. Because the weekly Florida
celery market permits at least two adjustments to equilibrate supply and
demand, the assumption that the Exchange's price represents a true
equilibrium value over the weekly period appears justified.
The few studies of short-run price determination in the celery
market have not taken a structural approach which allows simultaneous
determination of supply and demand relationships. Of course, the explicit
assumption that supply is actually determined, in part, by current
prices typically is suppressed in modeling many agricultural markets.
The advantages of a full structural model stem from the requirement for
a formal description of the price determination process and the correspond-
ing restrictions this structure imposes on the reduced form parameters.
The following discussion of the structural specification adopted, there-
fore, takes into account information regarding physical constraints,
perceived market operation, and relevant economic theory. The model is
dynamic in the sense that current Florida conditions affect current and
future levels of key variables; but because lagged California shipments
and prices enter the model, these variables must be set at pre-specified
levels to analyze periods longer than a week. Thus, interpretation of
other than impact (i.e., single period) multipliers or elasticities must
be conditioned on these features.
As mentioned previously, celery production is largely concentrated
in the hands of a few Florida producers, which suggests that an oligopolistic
treatment of the market may be warranted. In imperfectly competitive
markets, supply functions are not theoretically valid concepts because
the output levels of individual producers are assumed to influence the
market price. In the present context, however, the price to which
Florida producers respond is, to some extent, given by the Exchange.
The Monday celery price declared by the Exchange represents a parameter
(rather than a variable) on which growers may base harvesting decisions.
In this case the supply price perceived by producers is assumed to be
the average weekly price which is, of course, highly correlated with the
Monday price. A relatively high current supply price may induce accelerat-
ed or more thorough harvesting of the celery crop, reductions of stocks,
and thus, larger marketing. Yet, the crop base is essentially fixed
for periods of less than three months, so a substantial limit to supply
response must be imposed over the short-run.
The discrepancy between quantities harvested and quantities shipped
is accounted for by adjusting levels of celery storage. Data on the
response of stock holdings to price changes were not readily available
for the analysis so this effect must be implicitly incorporated into the
postulated supply function.4 Stock reductions, hence a corresponding
supply increase, are expected when current prices rise. Because stocks
may not be immediately replenished, the effect of higher prices over
several successive periods would preclude supply expansion from stock
Besides including average weekly prices in the supply equation,
additional variables are required in order to reflect short-run rigidities
in supply. Although data are available on weekly plantings of celery
for every season analyzed, experimentation with alternative lag forms
produced generally poor results. Probable difficulties with this approach
include the fact that the correlations between plantings from week to
week are typically large and the effects of weather in terms of varying
maturities in yields obscure the relationship between lagged plantings
and supply. Therefore, to approximate the traditional seasonal production
pattern, a weekly low order polynominal in time was specified. It gave
generally superior results relative to combinations of lagged weekly
In order to account for the fact that the crop's perishability
probably limits short-run supply response to a one- or two-week period,
lagged prices were included in the supply equation. As expected, it was
found that high prices the previous week tend to reduce the current
week's supply due to the fact that stocks were likely reduced the previous
week and celery harvesting was accelerated. To complete the specification,
lagged celery production was included to capture production inertia and
two seasonal dummy variables were introduced to represent poor growing
conditions during 1973 and 1976.
The quantity of celery demanded by wholesale and retail outlets is
postulated to depend on the current Florida price, the amount of previous
California celery marketing, prices received the previous Friday in
California, and seasonal or annual variations in demand.5 The responsive-
ness of demand to short-run price changes is generally constrained by
the perishable nature of the commodity. That is, the relatively short
interval between purchases and ultimate consumption limits adjustments
to prices. Also, demand for Florida celery is discontinuous because
commercial marketing do not begin until the late fall; which requires a
decided change-over by wholesalers and retailers to the Florida product
and reduced purchases of the California product (Mathis and Degner).
Thus, lagged levels of demand may be important indicators of how rapidly
wholesalers and retailers begin handling Florida celery. The pace of
the change-over is probably conditioned by custom and habit as well as
Previous demand levels may also indicate the size of wholesale and
retail inventories or the extent of market saturation, and thus are
hypothesized to depress current weekly prices. As suggested by consumer
theory, the elasticity of demand for Florida celery should increase over
time due to adjustment to California marketing and demand substitution.
Prices of possible vegetable substitutes and consumer income are not
included in the present analysis because price determination is presumed
to occur at the producer-shipper level rather than at the retail level
during the weekly observation period.
Some seasonality in demand has been suggested by Brooke and Jung
who found that demand tended to erode toward the end of the marketing
year. Monthly dummy variables are used to capture this presumed season-
ality. Year to year growth in consumption due to increasing population
and incomes as well as changing preferences are included in the demand
specifications via a yearly time trend. An admittedly ad hoc specifi-
cation of the demand relation is the result of limiting observations to
only a week's duration. Finally, a variable representing the Christmas
holiday week is incorporated to account for interruptions in product
distribution and demand. With these considerations in mind the following
equations represent the formal model adopted.
t = f(At' APt-1 Qt- Wt D73, D76)
t = fd( t- PC t-, QCt-,, YEAR, DH, Mj)
APt = Pt Pt-
t = 1, 2, ..., 26 j = 1, 2, 6
Q = Florida celery shipments in carloads
P = Average Florida celery price in cents per 2 2 1/2 doz.
PC = Celery price in California on Friday
QC = California celery shipments
W = Weekly trend term for each season, second week in
December is 1
W2 = W squared
D73, D76 = Dummy variables having values of one during 73 or 76
season and zero otherwise
YEAR = Annual trend term,1972-73 season is year 1
DH = Dummy variable for Christmas holiday week
Mj = Dummy variables for each month except March
The Estimated Model
The weekly Florida celery price determination mechanism was analyzed
using 156 observations for six December-June marketing seasons (1972-73
through 1977-78). Table 2 presents the parameter estimates for the be-
havioral equations using both two and three stage least squares estimation
techniques. Both estimation methods yield consistent structural parameter
estimates, whereas three stage least squares also produces efficient
parameter estimates. Calculated 't-values' are conditional in the sense
that the desirable statistical properties of both methods depend on large-
sample behavior. For the two-stage least squares models the calculated
value of first order serially correlated residuals, p, is presented
Because reduced form predictions derived from the three-stage least
squares parameter estimates were substantially better for the price
equation and only sightly poorer for the quantity equation, the follow-
ing discussion of the structural estimates will refer to the three-stage
The estimated supply equation includes current and lagged price
changes in an effort to capture short-run supply responsiveness and its
regidities. A 10 percent increase in current price causes about 5 percent
increase in supply. But, a 10 percent increase in prices the previous
period calls forth a 2.5 percent decrease in current supply, reflecting
the fact that stocks may have been reduced and harvesting accelerated
the previous period. Structural impact and interim mean elasticities of
supply are presented in Table 3. The interim elasticities show the
supply effect of a price increase diminishes rapidly.
Coefficients on the weekly polynominals in time suggest that celery
production peaks about the 12th wek of the season each year. This
Table 2.--Structural results.
Supply Demand Supply Demand
Variable "t" B "t" ( "t" g "t"
Qt1 --- 1 ---- -- 1
P ---- --- -.242 3.14 --- -- -.271 3.62
APt .284 3.34 ----- ---- .282 3.45 ----- --
INTERCEPT 79.1 4.95 -2.03 .01 71.5 4.59 166 .84
APt-l -.238 3.91 --- --- -.141 2.69 -- --
Q- .635 9.81 .473 7.95 .646 10.30 .529 9.17
PC --- --- .172 2.71 ---- ---- .200 3.29
QC ---- -.222 5.52 ---- -- -.129 3.83
Wt 8.70 3.26 --- --- 8.74 3.38 -- --
W2 -.356 3.89 ----- --- -.358 4.03 -- --
YEAR ---- ---- 4.23 1.30 -- ---- 5.61 2.00
DH ---- ---- -82.8 4.38 ---- --- 36.60 2.32
DEC ---- ---- 20.2 1.15 -- -- -3.20 .20
JAN ---- ---- -14.9 1.18 -- -- -16.8 1.52
FEB ---- ---- -22.0 1.87 ---- -- -8.47 .86
APR ---- 5.87 .50 -- ---- 4.86 .48
MAY ---- -49.8 4.12 -- -- -41.1 3.88
JUN ---- --- -69.0 4.44 -- -- -58.4 4.16
073 -13.4 1.27 --- --- .37 .04 -- --
D76 -22.1 2.07 ---- -14.4 1.60 --- --
p .03 ---- ----- ---- .01 -- -- --
Table 3.--Structural impact and interim elasticities.
Period Impact elasticities Total elasticities
Supply equation (3SLS)
t .501 .501
t+l -.428 .073
t+2 -.026 .047
t+3 -.017 .030
t+4 -.011 .019
Demand equation (3SLS)
t -.481 -.481
t+l -.254 -.735
t+2 -.135 -.870
t+3 -.071 -.941
t+4 -.038 -.979
would correspond to the period between the.end of February and the first
part of March. The dummy variables D73 and D76 were included to reflect
the poor growing weather experienced during these seasons, and are both
expected to have negative signs. Apart from the calculated significance
of these latter variables all other variables enter the supply equation
at highly significant levels and with the expected signs.
The estimated demand equation shows that demand is inelastic at the
producer-shipper level with a 10 percent price increase causing a 4.8
percent reduction in demand. Demand does become more elastic in the
long run as illustrated in Table 3. Of course, these structural elastic-
ities represent only partial effects because for periods longer than a
week there appears to be a strong relationship between the California
and Florida markets. In fact, a 10 percent increase in either the
previous week's shipments or the Friday price in California will re-
spectively reduce demand by 2.1 percent or increase demand by 3.4 percent.
Some seasonal factors enter the demand equation with high levels of
significance. As expected, demand tapers off toward the end of the
season (Brooke and Jung) but otherwise there were no other a priori
notions about the magnitudes of the monthly dummy variables; and their
significance levels suggest little contribution to the equation. The
dummy variable representing the Christmas holiday shows how demand drops
from previous levels during this period. Finally, the coefficient on
the annual trend variable has the interpretation that, all other factors
held constant, demand for Florida celery grows about 4 percent per year.
The usefulness of the structural approach arises from the comparison
of the least squares estimates of the unrestricted reduced form to the
restricted reduced form paramter estimates implied by the structural
model. Table 4 presents the reduced form parameter estimates and calculated
standard errors for both estimation methods. Typically, these parameter
estimates would be used for forecasting and control because the direct
effects of changes in predetermined variables on price and quantity can
The results in Table 4 point out the inverse relationship between
lagged prices and current shipments and between lagged shipments and
current prices. Reasons for these effects were mentioned in the previous
section. By and large the unrestricted and restricted reduced forms
correspond closely with respect to coefficient signs and relative magnitudes.
Notable exceptions are the coefficients on QCt-1 in both the quantity
and price equations. Essentially, the 3SLS restricted reduced form
discounts the importance of lagged California shipments on Florida
shipments and shifts this effect to the Florida price variable. The
Table 4.--Reduced form results.
Qt Pt Qt Pt
A A A A A A
Variable H1 SE Tip SE 1Q SE ip SE
INTERCEPT 303 196 620 275 49.5 93.8 429 303
Pt-i -.098 .041 .461 .057 -.138 .032 .510 .073
APt_1 -.207 .048 .166 .067 -.069 .025 .256 .085
Qt-, .493 .060 -.163 .084 .587 .050 -.211 .108
PCtl .079 .035 .416 .049 .102 .027 .361 .063
QCt-1 -.203 .039 -.042 .054 -.067 .017 -.233 .058
Wt .284 6.18 -7.12 8.69 4.28 1.23 -15.8 4.54
2 -.074 .206 .365 .289 -.175 .043 .647 .163
YEAR -.228 2.64 10.4 3.71 2.86 1.32 10.1 4.19
DH -87.4 17.2 28.2 24.2 -18.6 7.69 -66.1 26.7
DEC 10.8 46.0 -10.3 64.6 -1.63 7.77 -5.79 27.6
JAN -18.1 26.9 -12.0 37.9 -8.55 5.33 -30.3 18.6
FEB -26.4 15.6 17.7 21.9 -4.32 4.88 -15.3 17.2
APR 13.4 16.1 10.7 22.7 2.48 4.98 8.79 17.0
MAY -25.7 30.3 -40.5 42.5 -20.97 5.47 -74.4 21.1
JUN -45.4 44.2 -63.7 62.2 -29.8 7.36 -106 29.2
073 -11.8 9.83 -22.7 13.8 .181 4.61 -.669 16.3
D76 -16.9 8.84 18.5 12.4 -7.06 4.42 26.0 15.9
R2 .805 ---- .951 ---- .713 ---- .936
p -. 04 --- .02 -- -- -- --- --
effect of the weekly polynominal in time is much greater in the 3SLS
reduced form since this seasonality is more systematically introduced.
Also, the dummy variable for the Christmas holiday, DH, reflects a
depressing effect on both shipments and prices in the structural model's
Questions now focus on why the Exchange exists. After all, no
supply restrictions have been imposed in recent years and its price-
setting behavior appears to replicate the competitive equilibrium. The
empirical results do provide some indication. It was found that structural
elasticities of supply and demand are almost identical in the short-run
(one week). If during periods less than a week supply is more elastic
than demand, the market price determination process would be unstable
given some lag in the short-run adjustment of supply to price. Thus,
the Exchange, by pegging a price, may reduce a short-run tendency of
divergence from equilibrium.
An alternative view is that since there are only a few sellers in
the market with different amounts of information concerning market
conditions, an ologopolistic interdependence may lead to varying forms
of non-competitive equilibria based on conjectural variations, a dominant
firm's price-setting power, or the cartel model. Or, market uncertainty
may generate price wars and non-price competition and lead to a misallocation
of resources and price instability (Wu). Certainly in a market with a
homogeneous, perishable commodity of this type, instability seems more
likely to occur in the absence of an institution such as the Exchange.
And the cost of stability may not be the existence of prices much above
the competitive norm. Wu (p. 70) states, "I suggest that while stable
prices are found in conjunction with monopoly power, stable price itself
does not necessarily lead to inefficient allocation of resources."
Recent work in the economics of information illustrates the uncertainty
faced by firms and the implications of market structure on price determination
and dispersion.6 It turns out that price equilibrium conditions and the
informational value of prices depend largely on how information is
transmitted and collected (Garbade, et. al.). The costs associated with
the collection and analysis of information concerning a market or commodity
reflect an expenditure of resources. If prices are widely dispersed in
an imperfectly competitive market, buyers of inputs may lose market
opportunities if they purchase inputs at higher prices than their competitors.
Thus, there may be substantial returns to buyers who incur search costs
in such markets. Finally, even fairly competitive markets may still
yield significant price dispersion despite relatively available and
costless information (Pratt, et. al.).
The present analysis has focused on providing a structural represent-
ation of weekly Florida celery price determination. The results obtained
appear consistent with an equilibrium market in which price adjusts to
equate supply and demand. For intra-weekly periods some evidence of
disequilibria may be found, but this would require the analysis of daily
data, some of which is not available.
Prices observed may closely approximate a competitive market's
operation despite the fact that the price-setting features of the Exchange
seem to suggest a monopoly position. Because the observed prices correspond
to a structural model developed from the assumption of a competitive
market, an apparent conclusion is that "rational" price setting behavior
is manifested by the Exchange. That is, the projections of prices are
rational in the sence that unbiased estimates of the actual price arise
from information collected and given an interpretation implied by relevant
economic theory.7 For instance, if the Exchange's price forecasts had
some systematic bias it would pay either producers or buyers to take
advantage of this bias by altering shipping or buying patterns (DeCanio).
No evidence of this type of activity was discovered, however.
These findings together indicate that the Exchange provides a
valuable information collection and interpretation system which probably
yields stable prices at or close to short-run competitive levels and
substantially reduces the buyers' search costs.
1. For an overview of the Florida celery industry see Rose.
2. The Florida celery marketing order does limit acreage planted, but
on a weekly basis the crop base is considered predetermined.
3. Their model has been extended by Hartley to one where supply is
exogenous, and an application to an agricultural market has been
presented in Goldfeld and Quandt.
4. The importance of unsold celery stocks on price formation is noted
by Riggan and Brooke.
5. For a discussion of non-price determinants of demand see Godwin and
6. Kirman has suggested that imperfectly informed firms may attain
"false" equilibria. Salop and Grossman and Stiglitz have both
recently discussed some aspects of information and market structure.
7. For example, see Muth or Pashigian.
Bohall, Robert. "Pricing Performance in Marketing Fresh Winter Tomatoes."
Marketing Research Report No. 977, ERS, USDA, November 1972.
Breimyer, Harold F. Economics of the Product Markets of Agriculture. Iowa
State University Press, 1976.
Brooke, D. L. and G. H. Jung. "Market Organization and Operation of the
Florida Celery Industry." Bulletin 709, Institute of Food and
Agricultural Sciences, University of Florida, April 1966.
DeCanio, Stephen J. "Rational Expectations and Learning from Experience."
quarterly Journal of Economics, February 1979, 47-56.
Fair, Ray C. and Dwight Jaffee. "Methods of Estimation for Markets in
Disequilibrium." Econometrica, May 1972, 497-514.
Garbade, Kenneth D. et al. "On the Information Content of Prices."
American Economic Review, March 1979, 50-59.
Goldfeld, Stephen M. and Richard Quandt. "Estimation in a Disequilibrium
Model and the Value of Information." Journal of Econometrics, 3(1975),
Godwin, Marshall R. and William T. Manley. "Customer Preference Aspects
of Competition between Florida and California Celery." Bulletin
648, Florida Agricultural Experiment Station, June 1962.
Grossman, Sanford J, and Joseph Stiglitz. "Information and Competitive
Price Systems." American Economic Review, May 1976, 246-253.
Hartley, Michael J. "The Estimation of Markets in Disequilibrium: The
Fized Supply Case." International Economic Review, October 1976,
Kirman, Alan P. "Learning by firms about Demand Conditions." in
Adaptive Economic Models, Richard H. Day and Theordore Groves, eds.
Academic Press, 1975.
Mathis, Kary and Robert L. Degner. "Marketing Florida Celery: A Whole-
sale and Retail Analysis." Florida Agricultural Market Research
Center, Food and Resource Economics Department, Institute of Food
and Agricultural Sciences, University of Florida, August 1976.
Muth, J. F. "Rational Expectations and the Theory of Price Movements."
Econometrica, July 1961, 315-335.
Pashigian, B. Peter. "Rational Expectations and the Cobweb Theory."
Journal of Political Economy, May 1979, 338-352.
Pratt, John W. et al. "Price Differences in Almost Competitive Markets."
Quarterly Journal of Economics, May 1979.
Riggan, Wilson B. and Donald L. Brooke. "Predicting the Price of Florida
Celery." Proceedings of the Florida State Horticultural Society,
Rose, G. Norman. "Celery Production in Florida--A Historic Data Series."
Economics Report 69, Food and Resource Economics Department, Institute
of Food and Agricultural Sciences, University of Florida, May 1975.
Salop, Steve. "Information and Monopolistic Competition." American
Economic Review, May 1976, 240-245.
Wu, S. Y. "An Essay on Monopoly Power and Stable Price Policy." American
Economic Review, March 1979, 62-72.