marketing order suspensions and fresh lemon retail-fob … · richards et al.: fresh lemon margins...

Journal of Agricultural and Applied Economics, 28,2(December 1996):263–27701996 Southern Agricultural Economics Association

Marketing Order Suspensions and FreshLemon Retail-FOB Margins

Timothy J. Richards, Albert Kagan, Pamela Mischen, andRichard Adu-Asamoah

ABSTRACT

In August 1994, the Secretary of Agriculture announced the termination of the marketingorder and the associated flow-to-market, or prorate, controls for fresh California and Ari-zona (CA/AZ) lemons. Lemon growers and handlers have expressed concern over theimpact of this decision on retail-FOB margins. This study presents an econometric modelof fresh lemon marketing margins that tests for the presence of buyer and seller marketpower during previous periods of marketing order suspension. The results show that buyerand, to a lesser extent, seller market power cause retail-FOB margins to widen duringperiods of prorate suspension.

Key Words: conjectural variations, lemons, marketing order, monopsony power, retail-FOB margins.

The Secretary of Agriculture formally sus-pended the federal marketing order for Cali-fornia and Arizona (CA/AZ) lemons in Augustof 1994, ending the system of controlled al-locations between the fresh and processinglemon markets known as “prorates.” Prorateswere the primary mechanism through whichthe marketing order was able to control marketsupply, and thereby cause lemon prices to rise.Technically, prorates refer to the weeklyamount each handler could ship to the freshmarket as a proportion of total industry ship-ments. These amounts were set by the Secre-

Timothy J. Richards is assistant professor, and AlbertKagan is professor, School of Agribusiness and Re-source Management(SABR), Arizona State Universi-ty. Pamela Mischen is project director,Nationat Foodand Agricultural Policy Project (NFAPP), ArizonaState University.Richard Adu-Asamoahis faculty re-search associate,SABR and NI?APF!

The authors thank Gary Thompson,Sam Gao, PaulPatterson,and two anonymousreviewersfor their help-ful comments.

tary of Agriculture upon recommendation ofthe Lemon Administrative Committee (LAC)—a board of handlers, growers, and consumerrepresentatives—based on consideration of thestrength of demand, state of supply, and pro-jections of future needs. In general, the objec-tive of the marketing order was to establish an“orderly market” for lemons. Nominally, thedefinition of an orderly market is one that has“an even flow to market at stable prices, ” butthe actual purpose of the marketing order wasto raise lemon prices to the parity level (Car-man and Pick 1990).

Consumer groups lobbied for the termina-tion of the lemon prorate system, arguing thatit was merely a federally sanctioned agricul-tural cartel. Many growers, however, main-tained that price stability under the marketingorder was necessary to prevent cycles of boomand bust which cause inefficient investmentand removals. The larger cooperative handlers,not surprisingly, supported the grower posi-tion, although several independent and coop-

264 Journal of Agricultural and Applied Economics, December 1996

erative-affiliated handlers were accused ofcheating on the prorate system. Clearly, failureof the marketing order was due to forces bothinternal (independent handlers) and external(consumers) to the industry.

Consumer and producer interests also co-incide on the issue of retail-FOB margins.Such margins are often used as a measure ofmarket performance, or the ability of the mar-keting system to pass along production costsavings to consumers. Neither consumers norproducers wish to see suspension of the pro-rate system result in a greater share of the lem-on dollar going to retailer profit.

Previous studies of the effect of proratesimplicitly treat the prorate suspension decisionas exogenous to lemon supply conditions.Typically, decisions to temporarily suspendthe prorate were made only when board mem-bers felt controlling deliveries would not helpstabilize market prices or quantities. [ Failureto account for the endogeneity of the suspen-sion decision with lemon supply will producemisleading conclusions about the effect of reg-ulation on marketing margins. This study at-tempts to explain the decision to suspend andsuspension effects as simultaneous economicoutcomes.

Failure to allow for potential exercise ofbuyer or seller market power also introducesa source of bias to estimates of the effect ofregulation on margins. Margins under condi-tions of oligopsony or oligopoly respond notonly to changes in quantities allowed onto themarket, but also to regulatory factors that caninfluence a firm’s ability to use its marketpower. Rather than presume such power exists,however, this study tests for market imperfec-tions during periods when the prorate was ineffect, and periods when it was not,

Consequently, the objective of this study isto determine the effect of lemon prorate sus-pension on the retail-FOB marketing margin.While several studies have attempted to deter-

1For example, during the 1985–86 season, storageproblems caused the fresh lemon availability to be farlower than normal. Whereas flow-to-market controlcan stabilize prices and deliveries during times of over-production, prorates would have been irrelevantduringthe 1985–86 shortage.

mine the welfare implications of various mar-keting order provisions (see, for example,Powers 199 la), our investigation does not at-tempt to assign cost or benefit values to anyof the stakeholders in the lemon industry. Wedo, however, consider both the endogeneity ofprorate suspension decisions with supply andthe possibility of imperfect competitionamong lemon buyers and sellers. In the firstsection, an economic model of fresh lemon re-tail-FOB margins under imperfect competitionis presented. This model demonstrates thelikely effects of prorate suspension on the ex-ercise of buyer and seller power, and conse-quently, on margins. In the next section, wedevelop an empirical model of lemon marginsthat allows for the possibility of market powerand suspension endogeneity. The third sectionconsists of an explanation of how the empiri-cal model is estimated, including a discussionof the data, estimation methods, and empiricalresults. Finally, several implications are drawnregarding the future of lemon marketing, andother marketing-order-controlled commoditiesin California and Arizona,

Economic Model of Citrus Margins

Several studies have used data from navel or-ange suspension periods in the 1980s to pre-dict the likely impacts of marketing order ter-mination on prices, shipments, allocations, andmargins. In separate investigations by Thorand Jesse, and by Powers, Zepp, and Hoff, theauthors argue that prorate suspension causesthe price of fresh citrus to fall. Without theability to use the marketing order flow-to-mar-ket controls to price discriminate between theprice-inelastic fresh market and the price-elas-tic processing market, lemon sellers allocatemore to the fresh market. Carman and Pick’s(1988) analysis of the 1985–86 lemon proratesuspension shows that the fresh market pricenot only falls without the prorate, but also be-comes more variable. Although they do notexplicitly address the problem of retail-FOBmarketing margins, Carman and Pick suggestthat an increase in buyer market power with-out the prorate could lead to a reduction ingrower prices without a corresponding fall in

Richards et al.: Fresh Lemon Margins 265

the retail price. Such buyer power is a distinctpossibility in the lemon industry, due to thegeographic dispersion of growers and cyclicaloverproduction problems (Kinney et al.).

Our study extends the existing research inthree ways. First, the analysis focuses on mar-keting margins to determine the effect of ter-minating flow-to-market control on the grow-ers’ share of the retail lemon dollar. Second,the study models the decision to suspend theprorate as endogenous. Because many of thefactors that determine lemon supply also influ-ence the decision to suspend prorates, the de-cision to suspend cannot be considered as anexogenous event. Third, the model does notpresume that perfect competition prevails inthe fresh lemon market, but tests for the pos-sibility that both lemon buyers and sellers ex-ercise a degree of market power. Tests of mar-ket power employ an econometric frameworkthat has proven valuable in the study of mar-keting margins in other industries, particularlyin beef and pork packing.

Empirical analyses of imperfect agricultur-al markets often employ the “new empiricalindustrial organization” (NEIO) paradigm.Azzam and Pagoulatos; Schroeter; andSchroeter and Azzam (1990, 1991) adopt thisapproach in their study of oligopsony and ol-igopoly power in meat packing industries.Warm and Sexton apply the NEIO to the tree-fruit industry and find evidence of both buyingand selling market power by pear processors.Contrary to earlier work in the processing to-mato industry, Durham and Sexton find lim-ited support for claims of oligopsony behavior.In their analysis of the effects of regulation onmarket power, Melnick and Shalit find thatmarketing board provisions which set pur-chase quantities and minimum prices are in-sufficient to keep Israeli tomato market agentsfrom exercising oligopsony power—to thedetriment of farmers. Taylor and Kilmer em-ploy an NEIO approach to show that growersdo not use the Florida celery marketing orderto raise prices above the perfectly competitivelevel.

Others favor more traditional approaches tomargin analysis. Brester and Musick reviewseveral studies that use reduced-form price

spread models, These models consist of themarkup model, the price of marketing servicesmodel, and the relative price spread model. Ofthese, only the relative price spread model isconsistent with the predictions of a more com-plete model of margin determination. Usingthe relative price spread model to analyze theCA/AZ navel orange market, Thompson andLyon present evidence demonstrating that pe-riods of prorate suspension cause the retail-FOB margin to narrow, Powers (199 lb), onthe other hand, uses a structural model of thenavel market to find the opposite, i.e., marginsare likely to be greater during times of sus-pension because greater quantities move intothe less elastic fresh market. Neither of thesestudies considers the effect of prorate suspen-sion on the relationship between growers orhandlers and the ultimate retail buyer of lem-ons.

If it is true that the marketing order effec-tively sanctions a grower cartel, as shown byPowers (1993), then terminating the proratesystem should cause both farm and retail pric-es for fresh lemons to move toward the per-fectly competitive level. However, the mar-keting order was originally established with apurpose of allowing growers to countervailsmonopsony power by lemon buyers. Conse-quently, the alternative to monopoly sellingpower is not, in fact, perfect competition, buta degree of monopsony power on the part oflemon buyers. Therefore, rather than an ex-pectation that margins will narrow, this sug-gests that the retail-farm margin may in factwiden. Determining the relative effects of buy-er and seller power requires explicit consid-eration of lemon buyer power and the role ofcooperative pricing.

There are many reasons to expect imper-fect competition among fresh lemon buyers.First, fresh lemons share many of the char-acteristics described by Rogers and Sexton asbeing conducive to buyer power because theyare “bulky, and/or perishable, causing ship-ping costs to be high, restricting the products’geographic mobility, and limiting farmers’access to only those buyers located close tothe production site” (p. 1143). Second, Rog-ers and Sexton argue that buyer concentration


is likely to be higher than seller concentrationdue to the specialized needs of buyers. Whilefew nonagricultural inputs substitute for farmproducts in food processing, it is also truethat farm products cannot readily substitutefor other processing inputs. Third, and mostimportant for lemon growers, Rogers andSexton maintain that grower investments inspecialized, sunk assets cause a barrier to exitand, hence, inelastic product supply. Further,in 1987, the average four-firm concentrationratio (CR-4) for grocery stores among allstandard metropolitan statistical areas(SMSAS) in the U.S. was 58.3% [U.S. De-partment of Agriculture/Economic ResearchService (USDA/ERS) 1994]. A CR-4 ofgreater than 50~o is taken to be an indicatorof the potential for imperfect competition, onboth the buying and the selling side of theretail market. Beyond these structural facts,power exists in the output market simply be-cause lemons constitute a small proportion ofconsumer budgets, there are few good sub-stitutes, and they are often purchased as acomplement to another inelastically demand-ed good (alcoholic beverages, water, or forbaking). Taken together, these conditions sug-gest a strong a priori case for imperfect com-petition in the lemon market.

Recognizing this potential for lemon buyerpower, growers established a voluntary mar-keting agreement to control fresh market salesas early as 1925. The California Fruit Grow-ers’ Exchange, today’s Sunkist Growers co-operative, was the primary outgrowth of thisagreement, Beginning with a grower marketshare of over 90%, marketing conditions be-gan to deteriorate until “increasing produc-tion, lower prices, and lack of full participa-tion in the marketing agreement led toenactment of a federal marketing order in1941” (Kinney et al., p. 3). Under the mar-keting order, Sunkist managed to maintain amarket share of greater than 5090, so lemonbuyers exercise market power subject to theconstraint of dealing with a cooperative seller.z

2Lemon cooperatives are “active” in the sense ofRoyer and Bhuyan to the extent thatthey dominate the

Sunkist’s ability to sustain a high market sharewithout the marketing order is, however, ofsome question.

Royer and Bhuyan develop a model ofoutput and price determination under variousrelationships between farmer cooperativesand farm-product buyers. Adopting the per-spective of the fresh lemon retailer (or whole-saler), the current model assumes the mostgeneral case where the retailer can exerciseno market power (perfect competitor) or canuse full monopoly or monopsony marketpower. Assume retailers receive a price P,(q)

from consumers in market i and pay the co-operative handler j a price P~(q) for q,~boxesof lemons, and assume that the cost of trans-porting and retailing lemons is a constant k

per unit. Therefore, the retailers’ profit max-imization problem is

(1) max [T,] = max [P,qti – Pgqd – kqo].q,, qq

Growers, through their handlers, supply up tothe point where the price is equal to the mar-ginal costs of production, so their inverse sup-ply function is written: Pg = A4Cg. Substitutingthis expression into (1), the first-order condi-tion for the retailer becomes

Multiplying the first term by (q/q) and(P,/Pr), and the second by (q/q) and(&fC,/MCz), and summing over all handlers. .,and markets provides the market markup func-tion in general elasticity form:

(3) ~=‘r(’+wl’+:)-k=o

Lemon Administrative Committee, which recommendsthe weekly flow-to-market, but are “passive” in thesense that they cannot control the amount of output oftheir member-growers. As shown by Royer and Bhuy-an, however, this distinction is immaterial to the equi-librium price and quantity when the buyer is dominant.


where T-Iis the own-price elasticity of demandfor lemons, and ~ is the elasticity of supply oflemons. In equation (3), the ratio of 0, the con-jectural elasticity of demand, to q shows thedegree of market power exercised by the re-tailer in the output market, while the ratio of~, the conjectural elasticity of supply, to ●

shows the degree of market power exercisedby the retailer as a lemon buyer. For each ofthe conjectural elasticities, a value of zero in-dicates perfect competition, while a value ofone indicates monopoly or monopsony, re-spectively.

Equation (3) provides an expression for theretail-FOB margin in terms of the supply anddemand elasticities, and the conjectural elas-ticities in the input and output markets. Thisexpression allows a comparison of the relativesize of the margin under the prorate and in anunregulated market. Rewriting (3) with theproportionate retail-FOB margin on the left-hand side gives

P, – MC, $(4) M=

() ()

.g; – !? + k*,P, ~

where g = A4C~1P, and k% = klPr, and M isthe retail-FOB margin. Under the prorate, as-sume that lemon sellers are able to exerciseoutput market power, but cooperative controlover lemon supplies prevents them from usingany input market power. In the extreme case,this implies that 6 = 1 and ~ = O, so themargin is equal to

(5) M, = –(l/q) + k*,

or the elasticity of demand plus marketingcosts.3 However, if the loss of the marketingorder represents a loss of handler control overpricing, then retailers maximize profit wheretheir marginal revenue is equal to the marginaloutlay on lemons, or 6 = 1 and @ = 1. Con-sistent with Carman and Pick’s (1988) obser-vations of the suspension of the marketing or-

3Where q is left as a negative value, this marginis positive.

der in 1985–86, this exercise of buyer powerleads to wider retail-FOB margin:

(6) Mz = (l/e) – (l/q) + k*.

Of course, cases of monopsony, monopoly, orperfect competition are all extremes. In actu-ality, the values of 6 and @ are more likely tolie somewhere between zero and one. In thefollowing section, an econometric model is de-veloped that estimates these parameters usingmonthly fresh lemon pricing data.

Empirical Model of Prorate Suspension

Retailers, through handlers or packers, choosethe amount of lemons to purchase from grow-ers and decide the spatial allocation betweenretail markets.4 In doing so, a retailer’s profitis equal to the sum of the revenue from eachretail market’s sales, less the amount paid forlemons, less the cost of marketing. The profitmaximization problem is written above asequation (1).

Following Schroeter and Azzam (1991),the margin equation (4) is rewritten such thatthe retail-FOB margin is a function of mar-keting costs, exogenous variables, and outputand input market power parameters:

(7) M,, =‘r’’~r~=(%)(w:)

+ c(w) + ~

P ‘“r,,1

where Mit is the proportional retail-FOB mar-gin in region i and time period t; retailingcosts, C, are a general function of labor andtransport costs, w; and Zil is a vector of ex-ogenous variables that influence the propor-tional margin. However, as the discussionabove indicates, the locus of power in the mar-keting channel is likely to change depending

4Note that, because the problem is cast from theperspective of lemon buyers, the empirical model neednot consider the optimal allocation between the freshand processed markets. Studies such as Powers (1993)deal adequatelywith this issue.


upon whether the prorate restrictions are ef-fective or suspended.

Econometric estimates of the effect of mar-ket power on the retail-FOB margin take intoaccount the different bargaining power rela-tionships between the prorate and non-prorateperiods. In terms of the empirical margin mod-el, a binary variable, Dt, differentiates betweenthe conjectural elasticities during eachprorate regimes:

of the

(8) M,, =(-)( )

P,,,, – MC, _ MC, @— —

+<%$):’(:;

()-D, ! +~+Z,,,

T ,,,1

where D, = 1 during prorate periods, and D,= O during suspensions. With this model, testsfor the effect of prorate suspension on the ex-ercise of market power, and hence marketingmargin behavior, amount to tests of the equal-ity of the conjectural elasticities between thesuspension and non-suspension regimes.5

Identifying the conjectural elasticity valuesrequires a multi-stage estimation procedure.Details are provided in several studies (e.g.,Warm and Sexton; Schroeter and Azzam 199 1;and Azzam and Pagoulatos, among others) onthe framework for estimating marketing mar-gins equations under imperfect competition.Specifically, the first stage estimates time-varying elasticities of demand and supply.Substituting these estimates into equation (8)identifies the conjectural elasticities of demandand supply. In the second stage, the marketinginput cost function, C(w), is specified as a dualretailer cost function to account for the effectof input price variation on the retail-FOB mar-gin. The following discussion examines eachof these steps in turn.

Empirical specification of the demand

5Note that Schroeter and Azzam’s (1991) assump-tion of firm homogeneity applies equally in this case,so each of the conjectural elasticities in (6) is a weight-ed average over all of the firms in the industry.

function closely follows Carman and Pick’s(1988) model, This model treats weekly retailprice as a function of per capita consumption,lagged real prices, income, grade, and seasonaldummy variables. Modifying their model toaccount for monthly data, the U.S. demand forfresh lemons is written in inverse form withthe national average retail lemon price a func-tion of the per capita quantity delivered (Q),the ratio of quantity to time (Q/T), per capitaincome (Z), the lagged retail price (F’., ), adummy variable that takes a value of one dur-ing prorate periods (MO), the average size oflemons shipped during the marketing year(QUAL), and dummy variables for each monthof the year, excluding December—which actsas the base, or reference month (MN~). Al-though monthly price data are less likely toreflect lagged price adjustment, preliminarytests find strong support for the inclusion oflagged prices in these data as well. With thisspecification, the slope of the demand functionvaries with time as follows:

(9) P, = aO + alQ + az~ + a31

+ abPc,. 1, + a5M0 + aGQUAL

II+ ~ a#fN~ + e].

k=1

Carman and Pick’s model of lemon supplyincludes each of the factors the Lemon Ad-ministrative Committee considers when rec-ommending weekly deliveries. Attempting toforecast the strength of demand in the currentweek, the LAC looks at recent price and salestrends, storage levels, seasonal demand pat-terns, and indications of futurb demand com-mitments. Because the current model usesmonthly data, the set of explanatory variablesdiffers from the weekly model of Carman andPick, but reflects many of the same economicinfluences on supply. Namely, lemon deliver-ies are a function of the equivalent-on-treeprice per box in Southern California (P,) mul-tiplied by a time trend, the level of lemonstocks (S), lagged sales (~., ), and the fittedvalue of the prorate suspension variable (~).Multiplying the lemon price by a time trend

Richards et al.: Fresh timon Margins 269

creates the necessary time variation in supplyelasticity.G Finally, the supply model incorpo-rates a binary variable to account for the pro-rate suspension period. Given these consider-ations, the supply function appears as

(10) (2 = b, + b,Pg,,T + bJ,

+ b3@- , + b4~, + U,.

Equation [10) includes the fitted value ofthe prorate suspension binary variable becausethe decision to suspend is hypothesized to beendogenous to lemon supply conditions.7 For-mally, the null hypothesis is that the suspen-sion decision is exogenous, so rejecting thenull supports the endogeneity of the suspen-sion dummy. Kennedy’s omitted variables ver-sion of Hausman’s general specification testdetermines whether or not the prorate suspen-sion decisions are endogenous. In order tosimplify the description of the Hausman test,the explanatory variables of (10) are separatedinto sets of exogenous (X) and possibly en-dogenous (Y) variables. With this definition, Yconsists of the prorate suspension dummyvariable. In general notation, the supply equa-tion simplifies to

(11) Q,, = X@, + Yy + Y*T, + W,,

GSupply response studies for citrus crops find thatthe most important determinants of supply includelong-run expected prices, current bearing acreage, theage of the existing grove, and proxies for urbanization,technological change, farm labor availability, and taxlaws (Carman). Besides the price variable, most ofthese factors are either unobservable or unavailable ona monthly basis; therefore, they are usually proxied bya simple time trend.

7As pointed out by an anonymous reviewer, pro-rate suspensions only occur in periods of unusually lowsupply. Because the primary purpose of the proratewas to hold market deliveries below their unregulatedlevels, prorates were simply not required when sup-plies were limited due to poor harvests, or storageproblems, Therefore, to accurately estimate the supplyresponse of lemon shippers, the empirical model dif-ferentiates between prorate and non-prorate regimes.Further,by including those factors considered by theLAC in suspending the prorate in a separateprobit re-gression, this analysis models, as well as possible, thecharacteristics of supply response in suspension peri-ods as different from those under non-suspension pe-riods.

where Y* is the vector of fitted prorate sus-pension values. A Wald statistic tests the jointnull hypotheses that all elements of T, are

equal to zero. Rejection of the null indicatesthat lemon prorate suspension decisions are in-deed endogenous. If the suspension dummy isendogenous, then estimating (10) with D[ in-stead of ~ with ordinary least squares will re-sult in biased and inconsistent parameter es-timates as the binary variable is correlatedwith the equation residuals. Clearly, policyrecommendations that result from inconsistentparameter estimates have the potential to beseriously in error.

To correct for this possible source of bias,the model incorporates Heckman’s two-stageendogenous dummy-variable method. In thismethod, a first-stage probit model specifies thesuspension binary variable as a function ofcurrent deliveries, prices in the previousmonth, and the level of lemon stocks, each asdefined above. In the second stage, the fittedvalues from the probit regression are used asinstruments for the binary variable in the sup-ply equation. The resulting parameter esti-mates are consistent, and are used to calculatethe price elasticity of supply. In addition to themarket power measures, the margin equationalso includes various measures of marketingcost.

Estimates of retailers’ operating costs aredependent upon labor (W) and transportationcosts (TR) for delivery to each region. Giventhat the actual form of the cost function is notknown a priori, a generalized Leontief func-tional form is used which approximates an ar-bitrary cost function. Schroeter and Azzam(1991 ) detail the conditions and assumptionsmade in the estimation of a single cost func-tion as an aggregation over all retailers. Withthese assumptions, the empirical specificationof the cost function for each region j is

(12) C,,,(w,,,) = CJOW,+ cj(TR,

+ C,j(W,TR,)”2 + V,.

Substituting the cost function (12) into equa-tion (8), and adding a set of monthly binaryvariables completes the derivation of the mar-gin model.


In the margin specification, the prorate sus-pension dummy acts as a slope shifter on themarket power measures. Estimating both thesuspension and non-suspension conjecturalelasticities provides a separate test for marketpower in each regimes Including the retailcost function, monthly dummy variables, anda random error term, the retail-FOB marginfor each region j (with the time subscript omit-ted) is given by the following:

(13)(P, - MC8) = ~

P, 1

“’W+4?)()+ ~ * W,TR,) 112

1J

+($+;(-)

A 6,

‘O,

( )()

MC,,,+— q

P, q

( )()

MC,, &+L—

P, q

where Pj is normally distributed with zeromean and with a variance-covariance matrixgiven by

(14) E(NjPj’) = ~]]IM,

U WJJ+,‘) = U,.,zn,, a,,,, # 0,

and MT~ is the kth monthly dummy (July–Oc-tober). Preliminary specifications of the modelshow that lagged margin values, incorporatedto test for the slow adjustment of margins re-ported by Pick, Karrenbrock, and Carman, areinsignificant determinants of current margins.Tests of the non-diagonality of the variance-covariance matrix indicate that the sample dis-turbances are contemporaneously correlated,

8An example of a similar application of this pro-cedure is found in Chang,

so the set of equations is estimated with aseemingly unrelated regressions (SUR) pro-cedure.

With this model, tests of the equality of theconjectural elasticity parameters between theprorate and non-prorate regimes determine theeffect of lemon prorate suspension on retail-FOB margins. Specifically, if the output mar-ket conjectural elasticity is greater withoutprorates than with, then the retail-FOB marginwill be greater without the prorate system.Similarly, for buyer market power, if the inputmarket conjectural elasticity is greater withoutprorates, then margins widen in the absence ofregulation.

Data

The U.S. Department of Labor/Bureau of La-bor Statistics (USDL/BLS) consumer price in-dex database provides monthly retail lemonprice data for the four markets of New York,Chicago, Atlanta, and Los Angeles from Au-gust 1984 through December 1993. Regionalpersonal disposable income data are alsobased upon BLS information. The data on re-gional monthly lemon arrivals in each of theabove cities are found in the USDA’s annualFresh Fruit and Vegetable Arrivals in Western

(Eastern) Cities publications. Grower pricesare monthly average FOB prices to Californiapacking houses from the USDA/ERS publi-cation Fruit and Tree Nuts: Situation and Out-

look Yearbook, which also provides the lemonstock data. Total monthly shipments and av-erage quality values are taken from various is-sues of the Annual Report of the Lemon Ad-

ministrative Committee, which, since thedissolution of the committee, are held by theUSDA/Agricultural Marketing Service (USDA/AMS).

Proxies for the marketing cost variables in-clude average hourly wages in food retailing(USDL/13LS) for each of the four regions, andthe cost of transporting produce by truck fromSouthern California to each of the terminalmarkets. The latter data are available in theFruit and Vegetable Truck Rate and Cost

Summary, published by the Federal-State Mar-ket News Service of the USDA. Prorate sus-


Table 1. Lemon Summary Statistics

Variable Units Mean Variance Minimum Maximum

Pc

P:

P:

P;

Q.Q,,Q.Q,,MC

M,

M.

M,v

Wcw,wew.

$/lb.@/lb.$tnb,$/lb.000s Cwt000s Cwt000s Cwt000s Cwt

$/hr.$/hr.$/hr.$Ihr.

38.1837.9637.1937.7768.4942.09

108.7657.33

0.770.780.770.76

10.989.569.86

11,13

37.5140,0347.0052.18

368.1481.63

1,554.30154.49

0.0170.0160!0170.0150.310.820.480.41

28.7327.5125,9224.7830.0024.0038.0035.00

0.450.460.430.498,897,268.279.50

51,7655,3755.9757.46

128.0067.00

220.0092.00

0.970.960.960.95

11.6911.7810,9612.21

Note: Variablesare defined as follows: H is the retailprice in market i, Q, is the amount of lemon deliveries to market

i, M, is the retail-FOB margin in market i, and W, is the food-retailing wage rate in market i.

pension periods are given in various issues ofthe Annual Report of the Lemon Administra-

tive Committee. Table 1 provides the descrip-tive statistics for the regional fresh lemonquantities, prices, margins, and packing costs.

Results and Discussion

The multi-stage empirical procedure begins byestimating the elasticities of supply and de-mand for lemons. In this section, we presentand interpret the results from these equationsfirst. A discussion of the marketing margin es-timates follows.

Before estimating the supply curve, how-ever, the results of the Hausman endogeneitytest [equation (1 1)] are used to determine theform of the binary prorate suspension variable.The omitted variable test accepts the null hy-pothesis of erogeneity if the estimated param-eter on the fitted value of the suspension dum-my (~) is not significantly different from zero.In fact, estimates of equation (11) show a t-

statistic of 2.542 on the suspension parameter,so the test rejects erogeneity. As a result, es-timates of the supply equation use Heckman’sendogenous dummy variable procedure.

Stage one of Heckman’s method requiresestimating a probit model of the prorate sus-

pension binary variable. Specifically, the prob-it model provides estimates of the regressionmodel given by

(15) Prob(D, = 1) = @(aO+ alQ,

+ a2P,_ , + %s,),

where @ is the normal distribution function,Q, is the level of deliveries in month t,P,. ~ isthe FOB lemon price in the previous month,and St is the level of stocks in the currentmonth,

The parameter estimates, marginal effectsof each variable, and goodness-of-fit statisticsresulting from this procedure are shown in ta-ble 2. Although not all of the explanatory vari-ables contribute significantly to the decision tosuspend prorates, the overall prediction suc-cess rate of 76.42’%oindicates the model effec-tively predicts administrative committee be-havior. At a 5?Z0level, only the total level ofshipments and the level of lemon stocks aresignificant. Taken as a group, however, a like-lihood ratio test rejects the null hypothesis thatall the parameters of the probit model arejointly equal to zero. Specifically, with threedegrees of freedom and a 5% significance lev-el, the critical chi-square value is 7,815,


Table 2. Prorate Suspension Probit Model:CA/AZ Fresh Lemons (monthly, 1983–94)

Variable Estimate’ Marginalh t-Ratio

Constant –6.0302 –2.3395 5.305**

output 0.0055 0.0021 –1.861*FOB Price,., 0.2683 0.1041 –1.401Stocks 0.0002 0.0008 –4.718**

# 46.934

Note.r: Single and double asterisks (*) indicate signifi-

cance at the 5’70 and 1‘Yolevels, respectively. The chi-

square test of all parameters = O at the 5% level with

three degrees of freedom and a critical value of 7.815.

0 Parameter estimates are maximum likelihood obtained

with the probit function in LIMDEI?t, The marginal effects are giVen by: i~E(Y)/&x = +( ~ ‘x)x,

where y is the dummy suspension variable, (3 is the pa-

rameter vector, and x is the vector of explanatory vari-

ables.

whereas the likelihood ratio test statistic is46.934.

The marginal effect of each explanatoryvariable gives the expected change in theprobability of observing a prorate suspensionfor a one-unit change in the value of the vari-able in question. For example, an increase inlemon shipments of 1,000 cwt causes theprobability of suspension to rise by 0.2%. Al-though this result appears to be contrary toexpectations, the model already accounts for

the effect of a short supply on the suspensiondecision through the lagged FOB price. If theFOB price rises by one dollar per carton, thenthe probability of observing a suspension risesby 10.41%. The price effect also appears tooutweigh the effect of lemon stocks. Com-monly, low lemon stocks are felt to heavilyinfluence the LAC decision to suspend theprorate, but the probit results suggest thatprice movements are more important. As ex-plained in the following discussion, low de-mand elasticities contribute to the volatility oflemon prices.

Because the price and quantity are jointlydetermined in the demand equation (9), esti-mation uses a two-stage least squares proce-dure. The set of instruments consists of severalvariables that are felt to be correlated withboth lemon shipments and retail price, but areexogenous to lemon demand, such as laggedvalues of lemon stocks, lagged lemon retailprices, average seasonal lemon size, the priceof fresh oranges, the average U.S. retail wagerate, the average U.S. truck rate, per capitaincome, U.S. population, and a time trend. Theaggregate fresh lemon inverse demand curveestimates are presented in table 3, At the sam-ple mean price and quantity, these results im-ply a demand elasticity of –O. 139, suggestinga highly price-inelastic demand for lemons. It

Table 3. Fresh Lemon Retail Demand Function: U.S. Per Capita Monthly Consumption(1984-93)

Variable Estimate t-Ratio Variable Estimate t-Ratio

Constant –37.702 –1.502 March 1.329 0.964Quantity –0.019 –2.118* April 4.698 2.986**@’Time –0.003 –0.370 May 6.644 2.267*Prorate’ 1.904 2.088* June 7.293 2.896**Income 1.883 2.593** July 7.807 3.707**Ret. Price,_, 0.465 3.682** August 3.346 2.739**Quality” 0,368 2.063+ September 0.033 0.020January –2.062 –1.802* October 0.721 0.576February –4.058 –3.545* November –1.452 –1.090

R2 0.860D.w. 2.097

Nate: Single and double asterisks (*) indicate significance at the 5~. and 1% levels, respectively.

‘ Tbe Prorate variable is a binary variable that is equal to one during prorate periods, and is equal to zero during

suspension periods.i, ~ualj~Y is indicated by the average size of lemons shipped during the season,


is not surprising that lemon demand is inelas-tic, because of the relative unimportance offresh lemons in consumer budgets and the vir-tual absence of viable substitutes.

Contrary to the navel orange results ofPowers (199 lb), retail prices are considerablyhigher during suspension periods than whenthe prorate is effective. The parameter esti-mate on the lagged lemon price is consistentwith the finding of Pick, Karrenbrock, andCarman, who show that the retail price doesnot adjust instantaneously. Furthermore, thepositive coefficient on the QUALITY variablesuggests that smaller lemons (those with ahigher count value) have higher retail prices.Just as the vector of time-varying demandelasticities identifies the conjectural elasticityof demand, estimates of the supply elasticityidentify the degree of monopsony power.

Once again, the simultaneity of price andquantity in the supply equation requires an in-strumental variables estimation method. Ap-plying two-stage least squares to equation (10)provides estimates of the supply response ofCA/AZ lemon growers. The assumption thatone supply equation adequately describes lem-on supply is appropriate given that Californiaand Arizona supply over 90% of total domes-tic lemon consumption. With monthly data,lemon supply is defined in terms of growerdeliveries rather than production. Productiondecisions are made at the time of planting,some five to seven years prior to harvest anddelivery. The inclusion of the prorate suspen-sion dummy models the LAC’s tendency torecommend greater deliveries in high supplyyears in order to prevent stock buildups or theabandonment of fruit. The least squares esti-mates of the lemon supply equation are pro-vided in table 4. These results show thatlagged production, the ratio of the FOB priceto time, and the level of stocks are all signif-icant determinants of lemon supply. Althoughthe elasticity varies over time, the averageelasticity for the entire period is 0.46.

Substitution of the time-varying supply anddemand elasticities into the margin equationidentifies the conjectural elasticity parameters.In this context, conjectural elasticities showthe percentage change in aggregate supply or

Table 4. Farm Supply Function: CA/AZFresh Lemons (monthly, 1983–94)

Variable Estimate t-Ratio

Constant 312.72 4.202**

Output,-l 0,543 7.979**

FOB Price/ Time 0.081 2.345*

Prorate’ 23.966 0.573

Stocks 0.076 7.611**

R2 0.748

D.W. 2.045

Note: Single and double asterisks (*) indicate significance

at the 5 ?ZO and 1% levels, respectively.“ See table 3 footnote for definition of the Prorate vnri-

able.

demand for a given percentage change in re-tailer purchases or sales. Besides the marketpower parameters, the margin equation esti-mates the effect of marketing costs, the sea-sonality of margins, and the effect of proratesuspensions. Table 5 shows the margin equa-tion parameters for each of the four geograph-ic regions.

Tests for autocorrelation produce Durbin-Watson test statistic values for the North-Cen-tral, South, East, and West regions of 1.267,1.420, 1.295, and 1.494, respectively. At the590 level of significance, all but the West sta-tistic lie below the lower-bound d-value of1.434. Therefore, the parameter estimates intable 5 use the Prais and Winsten autocorre-lation correction algorithm described inGreene.

Preliminary model estimates reject the par-tial-adjustment specification used by Pick,Karrenbrock, and Carmam Although marginsappear to adjust slowly in their weekly data,the adjustment period is evidently less thanone month, as such an adjustment becomesundetectable in the monthly data of this study.Several other factors do contribute to variationin the retail-FOB margin. Namely, table 5shows that retail wages and transportationcosts are significant components of the mar-gin, and that margins tend to narrow duringthe months of July, August, September, andOctober compared to the excluded months.

The conjectural elasticity estimates showonly limited evidence of selling power during

2’74 Journal of Agricultural and Applied Economics, December 1996

Tab1e5. SUREstimates of Regional Lemon Retail-FOB Mmgins (monthly, 1983–94)

Variable North-Central South East West

w –1.135 – 1.463 –1.204 –1.005(–17.973)** (–18.083)** (–18.791)** (– 19.940)**

TR –0.582 –5.705 –0.691 N.A.a(-0.068) (– 10.686)** (-1.369)

(w. I’-l/)(‘“) 0,086 0.895 0.113 N.A,(0,580) (9.404)** (1.103)

0’ 0.013 0.013 0.011 0.011(4.277)** (4.450)** (3.879)** (3.739)**

+’ 0.002 –0.209 –0.018 0.049(0.017) (-0.239) (-0.144) (0.396)

0 0.133 0.096 0.132 0,134(14.225)** (15.423)** (16.487)** (30.918)**

+ 0.239 0.249 0.326 0.240(2.135)* (2,000)* (3.067)** (1.919)*

July –0.052 –0.043 –0.047 –0.044

(-3.294) (–3.025)** (-3.264) (–3.202)**

August –0.065 –0.058 –0.060 –0.059

(–3.575)** (–3.503)”” (-3.566) (–3.562)**

September –0.066 –0.060 –0.060 –0.052

(–3.604)** (-3.662) (-3.568) (–3.211)**

October –0.043 –0.040 –0.037 –0.032

-------- --- ‘–2’8!9?-:-- ‘:?980)** ------!:?:62:)-----------!:?!-18?-----

R2 0.805 0.738 0.632 0.759

N 97 97 97 97

Notes: Numbers in parentheses are t-ratios. Single and double asterisks (*) indicate significance at the 5% and 1%

levels, respectwely. The conjectural elasticity of demand is denoted by 6, and + is the conjectural elasticity of supply.

7R is the level of transportation costs from Southern California to each market, while W is the wage rate in food

packhvg specific to the given region.

I Transportation costs for CA/AZ lemons are assumed to be zero for the West region, as Los Angeles prices are used.

prorate periods. Contrary to the navel orangeresults of Powers (1993), the Eastern,Southern, and North-Central lemon marketsdiffer only slightly from perfect competitionat the 5% level. The output market conjecturalelasticities for these markets are small in anabsolute sense, averaging slightly over 1% foreach market, but are statistically quite signif-icant. None of the regions provide evidence ofinput market, or buying power during the pro-rate regime. Under prorate suspension, how-ever, all fresh lemon markets are imperfect onboth the buying and selling side. With respectto the output market conjectural elasticities,estimates vary from 0.096 in the South to0.134 in the West region. Input market powerrepresents the largest influence on margins un-

der prorate suspension, as expected based onboth the arguments of Carrnan and Pick(1988), and the theoretical model of this study.Whereas the input market conjectural elastic-ities are not significantly different from zerounder the prorate regime, they are significantin each market under prorate suspension. Es-timates of these parameters vary from 0.239in the North-Central region to 0.326 in theEast. Given that their allowable range is fromO to 1, these results suggest that margins riseby an average of almost 3070 without the flow-to-market regulation of prorates. Clearly, pro-rate suspension has its greatest impact on mar-gins through the exercise of monopsonybuying power on the part of lemon distributorsor retailers.

Richards et al.: Fresh Lemon Margins

Likelihood ratio tests formally compare theoutput and input conjectural elasticities be-tween the suspension and non-suspension re-gimes. The first test restricts the output marketconjectural elasticities to be equal between thetwo regimes. Comparing the values of the log-likelihood function from this and an unrestrict-ed model using a likelihood ratio test providesa chi-square distributed test statistic. Withthree degrees of freedom, the critical chi-square value is 7.815 at the 5Yolevel of sig-nificance. The calculated likelihood ratio teststatistic is 220.97, so the test clearly rejectsthe null hypothesis of parameter equality. Asimilar test of the equality of the input marketconjectural elasticities produces a likelihoodratio value of 14.212. Again, this result indi-cates that lemon buyers exercise a greater de-gree of market power without the prorates ineffect.

Although the lemon and orange marketsare somewhat different, these market powerresults contradict those of Powers (199 lb);Powers, Zepp, and Hoffi and Thor and Jessefor navel oranges, and the results of Carmanand Pick (1988) for lemons. Each of thesestudies argues that suspension of the prorateleads to a larger allocation to the fresh marketand lower retail prices. Wider margins underprorate suspension due to the exercise of mo-nopsony and monopoly power mean that re-tailers sell fewer fresh lemons at higher retailprices. When the prorate is in effect, retailersshow little success in taking advantage of theinherent inelasticity of demand because theamounts they sell are controlled by the LAC.However, in an unregulated market, retailersare free to sell any amount they choose. Theseresults indicate it is in retailers’ interests tolimit the amount they sell and increase retailprices, irrespective of their ability to force theFOB price down. If retailers are unable to ex-ercise buyer power due to the existence of acooperative with countervailing market power,their only alternative is to increase profitsthrough higher retail prices. This result woulddetermine changes in the retail-FOB marginwithout prorates if cooperatives are able tocontinue in their present role.

However, these tindings also show that,

275

without the prorate, cooperatives are less ableto influence the price paid at the packer level.As a result, lemon buyers are able to restricttheir purchases and force the FOB pricedown—the standard monopsony-power result.Monopsony power becomes important be-cause cooperatives in the lemon industry haveopen-membership policies; that is, they haveno ability to control the amount that growersbring to the market. As such control is a nec-essary condition for the ability to countervailsthe use of buyer power, they are left somewhatpowerless without their control of the LACand its ability to regulate deliveries.

Conclusions

This study argues that the deregulation offresh lemon sales resulting from the CA/AZcitrus marketing order termination is likely tosignificantly widen the retail-FOB margin.Whereas the prorate system enables lemonmarketing cooperatives to limit the exercise ofinput and output market power by retailers,this power disappears without the prorate. Theexercise of significant monopsony and, to alesser extent, monopoly power by lemon re-tailers under prorate suspension results in arise of the retail price and a decline in the FOBprice. Clearly, this leads to wider retail-FOBmargins-a result that benefits neither con-sumer nor producer interests.

Empirical estimates of a structural modelof marketing margins support these conclu-sions, Regional, monthly fresh market margindata from 1984 through 1993 show the retail-FOB margin to be significantly wider duringprorate suspension periods than when the mar-keting order controls fresh market allocations.In fact, input market power parameter esti-mates, which show the response of margins tochanges in the elasticity of supply, increase bya factor of up to 10 times their non-suspensionvalues.

Several factors influence the level of freshlemon retail-FOB margins that have been pre-viously ignored in the literature. First, themodel treats the decision to suspend the pro-rate as endogenous to supply conditions in thelemon market. Heckman’s endogenous dum-


my variable method corrects for the resultingsimultaneity bias in estimating the supply elas-ticity when this consideration is not taken intoaccount. Second, allowing buyer and sellermarket power to be determined by the policyregime represents a new way of looking at theeffects of government-sanctioned marketingorders in produce markets. Contrary to muchof the previous research in this area, the resultsof this study demonstrate that buyer marketpower does increase in the absence of the mar-keting order, and, in fact, becomes the domi-nant influence on margins.

There are many potential implications forthe lemon industry. Cyclical overproductionand characteristic cycles of boom and bust inall citrus markets led to the establishment ofthe marketing order. Without the ability to reg-ulate market flows, producers claim they willnot be as willing to make long-term invest-ments in new orchards. Such investment islikely to fall as a result of growers’ smallershare of the retail dollar and the likely increasein the variability of lemon prices-an issue onwhich this study is silent. Any rise in pricevolatility will likely reinforce the reductions inorchard investment caused by lower margins.A reduction in new plantings, or an increasein orchard removals, will result in a supplyreduction at a time of bright future prospectsfor international sales of both fresh and pro-cessing lemons.

Some argue that the long-run reduction insupply will lead, ultimately, to higher lemonprices. However, if the industry is to attain anequilibrium in the long run, a lower level ofproduction will only be consistent with lowerprices as the industry moves down its long-run aggregate supply curve.

These results also suggest that the cooper-atives themselves may be harmed. One of thereasons given for the termination of the mar-keting order concerns allegations of “cheat-ing” on the prorate regulations by both inde-pendent and some cooperative handlers.Without the prorate to enforce orderly mar-keting, the temporal and geographic dispersionof lemon deliveries results in ample opportu-nisty for the cheaters to undercut the coopera-tive price openly and effectively. It is hoped

that this study will serve to reopen the inves-tigation of the effect of marketing orders inthis and related produce markets.

References

Azzam, A., and E. Pagoulatos. “Testing Oligopo-listic and Oligopsonistic Behavior: An Appli-cation to the U.S. Meat Packing Industry. ” J.

Agr. Econ. 41(1990):362–70.Brester, G.W., and D.C. Musick. “The Effect of

Market Concentration on Lamb Marketing Mar-gins.” J. Agr. and Appl. Econ. 27(1995): 172–83.

Carman, H. “Income Tax Reform and CaliforniaOrchard Development. ” West. J. Agr. Econ.6(1981):165–80.

Carman, H.l?, and D.H. Pick. “Marketing Califor-nia-Arizona Lemons Without Marketing OrderShipment Controls.” Agribusiness 4(1 988):

245-59.—. “Orderly Marketing for Lemons: Who

Benefits?” Amer. J. Agr. Econ. 72(1990):346-

57.

Chang, C.C. “Asset Fixity, Non-Reversibilities,and Dynamic Structure of Production: A Micro-Application to Pennsylvania Dairy Farms.” Un-published Ph.D. dissertation, Pennsylvania StateUniversity, 1987.

Durham, C.A., and R.J. Sexton. “Oligopsony Po-tential in Agriculture: Residual Supply Estima-tion in California’s Processing Tomato Market. ”Amer. J. Agr. Econ. 74(1992):962–72.

Greene, W. LIMDEP, Version 6.0: User’s Manualand Reference Guide. Bellport NY Economet-ric Software, Inc., 1992.

Hausman, J.A. “Specification Tests in Economet-rics, ” Econometrics 46(1978): 125 1–71.

Heckman, J. “Dumtny Endogenous Variables in aSimultaneous Equation System. ” Econometrics46(1978):931–51.

Kennedy, P. A Guide to Econometrics. CambridgeMA: The MIT Press, 1985.

Kinney, W., H. Carman, R. Green, and J. O’Con-nell. “An Analysis of Economic Adjustments inthe California-Arizona Lemon Industry.” Rep.No. 337, Giannini Foundation of AgriculturalEconomics Research, University of Californiaat Davis, 1987.

Lemon Adrninistrative Committee. Annual Report

of the Lemon Administrative Committee. Vari-ous years, 1984–93. On file, USDA/Agricultur-al Marketing Service, Washington DC.

Melnick, R., and H. Shalit, “Estimating the Market


for Tomatoes.’’ Amer. J. Agr. Econ. 67(1985):573-82.

Pick, D.H,, J. Karrenbrock, and H.l? Carman.“Price Asymmetry and Marketing Margin Be-havior: An Example for California-Arizona Cit-rus.” Agribusiness 6(1990):75–84.

Powers, N.J. “Effects of Marketing Order ProrateSuspensions on California-Arizona Navel Or-anges.” Agribusiness 7( 1991 a):203–29.

—. “Evaluating Orange Growers’ Exercise ofMarket Power with Marketing Order VolumeControl Regulations. ” J. Agr. Econ. Res.

44( 1993):33-42.

—. “Marketing Order Impacts on Farm-RetailPrice Spreads.” Amer. J. Agr. Econ. 73(1991 b):507–10.

Powers, N.J., G.A. Zepp, and EL. Hoff. “Assess-ment of a Marketing Order Prorate Suspension:A Study of California-Arizona Navel Oranges.”Pub. No. AER-557, USDA/Economic ResearchService, Washington DC, 1986.

Rogers, R.T., and R.J. Sexton. “Assessing the Im-portance of Oligopsony Power in AgriculturalMarkets.” Amer. J. Agr. Econ. 76(1994): 1143–50.

Royer, J.S., and S. Bhuyan. “Forward Integrationby Farmer Cooperatives: Comparative Incen-tives and Impacts.” J. Cooperatives 10(1995):

33–48.Schroeter, J.R. “Estimating the Degree of Market

Power in the Beef Packing Industry. ” Rev.

Econ. and Statis. 70(1988): 158–62.Schroeter, J.R., and A. Azzam. “Marketing Mar-

gins, Market Power, and Price Uncertainty. ”Amer. J. Agr, Econ. 73(199 1):990–99.

—. “Measuring Market Power in Multi-Prod-uct Oligopolies: The U.S. Meat Industry. ” Appl.

Econ. 22(1990): 1365–76.

Taylor, T.G., and R.L, Kilmer, “An Analysis ofMarket Structure and Pricing in the Florida Cel-ery Industry,” S. J. Agr. Econ. 20(1988):35–43.

Thompson, G. D., and C.C. Lyon. “Marketing Or-der Impacts on Farm-Retail Price Spreads: TheSuspension of Prorates on California-ArizonaNavel Oranges.” Amer. J. Agr. Econ. 71(1989):

647–60.Thor, I?K., and E,V. Jesse. “Economic Effects of

Terminating Federal Marketing Orders for Cal-ifornia-Arizona Oranges. ” Tech. Bull. No.1664, USDA/Economic Research Service,Washington DC, 1981.

U.S. Department of Agriculture, Agricultural Mar-keting Service. Fresh Fruit and Vegetable Ar-rivals in Western Cities: By Commodities,

States, and Months. USDA/AMS, Fruit andVegetable Division, Market News Branch,Washington DC. Various issues, 1984–93.

—. Fruit and Vegetable Truck Rate and Cost

Summary. USDA/AMS, Fruit and Vegetable Di-vision, Market News Branch, Washington DC.Various issues, 1984–93.

U.S. Department of Agriculture, Economic Re-search Service. Food Marketing Review, 1992–93. Agr. Econ. Rep. No. 678, USDA/ERS,Washington DC, April 1994.

—. Fruit and Tree Nuts: Situation and Out-look Yearbook. USDNI!3RS, Commodity Eco-nomics Division, Washington DC. Various is-sues, 1984–93.

U.S. Department of Labor, Bureau of Labor Statis-tics. Consumer Price Index Detailed Report.USDL/BLS, Washington DC. Various issues,1984–93.

Warm, J.J., and R.J. Sexton. “Imperfect Competi-tion in Multiproduct Food Industries with Ap-plication to Pear Processing. ” Amer. J. Agr.Econ. 74( 1992):980–90.

marketing order suspensions and fresh lemon retail-fob … · richards et al.: fresh lemon margins...

Documents