279

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

Rev. FCA UNCUYO. 2018. 50(2): 279-292. ISSN impreso 0370-4661. ISSN (en línea) 1853-8665.

PontificiaUniversidadCatólicadeValparaíso.FacultaddeCienciasEconómicasyAdministrativas.EscueladeNegociosyEconomía.AvenidaBrasil2830,piso7.Valparaíso.Chile.rodrigo.valdes@pucv.cl

Measuring market integration and pricing efficiency along regional maize-tortilla chains of Mexico

Análisis de la transmisión de precios en las cadenas regionales de comercialización de tortillas de maíz en México

RodrigoValdesSalazar

Originales: Recepción: 06/03/2017 - Aceptación: 24/10/2017

Abstract

Whenmarketperformance isanalyzed,akey issue tounderstand is theextent towhichdifferentagentsinthesupplychainrespondtochangesinpriceshocks.ThispaperexplorestheMexicanmaize–tortillachainthroughapricetransmissionanalysis.Itfocusonacountry-levelstudyoftheMexicantortillamarketbyconsideringdifferentregionsandcommercializationchains.UsingaVectorErrorCorrectionModel(VECM)theinte-grationoftheMexicanmaizetortillamarketbetweentheinter-regiontortillería,inter-region supermarket and intra-region tortillería-supermarket prices are investigated.The findings suggest that price transmission between tortillerías and supermarketswithinaregionseemstohavetriggeredashiftinsomeconsumers'placeofpurchase.Moreover,itdemonstratesthatcointegrationlevelmainlydependsofstorage,logisticsand transportation that could be the reason for the upsurge in variability of tortillapricesbetweenstates.TheseresultsappeartovalidatethatradialtransportationsysteminMexicohasbenefitedthecentralregionsattheexpenseoftheperipheryregions,thusenabling tortillerías inCentral regions to remain competitivewith supermarkets andretain their market share.

Keywordsmaizetortillaindustry•Mexico•pricetransmission•supplychain

280

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

Resumen

Cuandoseanalizaelcomportamientodelosmercados,unelementoclaveesentenderhastaquépuntolosdiferentesagentesdelacadenadecomercializaciónrespondenaloscambiosenlasperturbacionesdelosprecios.Estetrabajoestudialascadenasdecomer-cializacióndetortillasdemaízenMéxico,através,deunanálisisdetransmisióndeprecios.EsteproductoesunadelasprincipalesfuentesalimenticiasenMexico.Esteestudioanalizadiferentes regionesy cadenasde comercializaciónutilizandounModelodeCorreccióndeErroresVectoriales (VECM)para investigar la integracióndelmercadomexicanodetortillasdemaízentrediferentesagentescomercializadores.Estosagentessonlatortilleríainterregional,elsupermercadointerregionalylospreciosintra-regionalesdelarelacióntortillería-supermercado.Losresultadossugierenquelatransmisióndepreciosentrelastortilleríasylossupermercadosdentrodeunaregiónparecehaberprovocadouncambioenellugardecompradealgunosconsumidores.Loanterior,demuestraqueelniveldecointegracióndependeprincipalmentedelalmacenamiento, la logísticayel transporte,loquepodríaserlarazóndelaumentoenlavariabilidaddelospreciosdelatortillaentreestados.EstosresultadosparecenvalidarqueelsistemadetransporteradialenMéxicohabeneficiadoalasregionescentralesaexpensasdelasregionesperiféricas,permitiendoasíque lastortilleríasde lasregionescentralessigansiendotancompetitivascomolossupermercados,permitiendolaconservacióndesuscuotasdemercado.

Palabras clavetortillasdemaíz•México•transmisióndeprecios•cadenadesuministro

Introduction

Tortillais themost important Mexicanconsumer good. It represents 38%of theprotein,45%ofthecaloriesand49%ofthecalcium among the Mexican population'sdietaryintake(23).MaizeisthetraditionalsourceofgrainforthisproductanditisthemostwidelycropproducedinMexico.

Mexico produces 23 million t of maizeand ranks fourth in producing countries intheworld,aftertheUnitedStates,ChinaandBrazil,with224,146and42milliontrespec-tively (1). Themaize tortilla industry playsa significant role in the Mexican economy,accountingfor1%ofMexico'sGDP.Themaize-tortilla industry directly employs about225,000workersinmills,flourfactories,andtortillerías in theproductionofmaize flour,maizedough(nixtamal),andmaizetortillas(17).Therearemorethan78,000tortillerías (tortilla bakeries) and mills distributed

throughoutMéxico,employingover2millionagricultureworkersinthemaizeproductivesector(7,34).

There are two main processes in theproduction of maize tortilla: maize flourand nixtamal (maize dough). The maizeflourindustryishighlyconcentrated,beingtheirtransportationandstorageoperationssubsidized by the Mexican government inordertomaintaintortillapricesstable(28).Incontrast, thenixtamal industry ishighlydecentralizedandusuallynotverticallyinte-grated to the storage, logistics, and distri-bution of maize (31). In the case of millsand tortillerías they acquire maize fromcommercializing firms, who buy directlyfromproducers,distributeandstoremaizein urban places, and sell to the nixtamal industryathigherprices,thuscausingthistypeoftortillastobecostlier(35).

281

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

Themaizetortillavaluechainpresentsan important barrier to the entry fornewcompetitorsdue to thehigh levelofconcentration in imports, storage andmarketing of maize (34). Gasca (2014)suggested that government subsidiesalong the maize-tortilla value chainbenefited large maize commercializingfirms,whichused theirmarketpower tofinancespeculativeoperationsofstorage,fleet, and export. Thereby exacerbatingthe effects of the rise in the international priceofmaizeinthedomesticmarketbycausingpricedistortionsalongthemaize-tortillavaluechain.Thetortillademandisveryinelastictoachangeinpriceanditisconsideredtobehaveasanormalinferiorgood (21, 30). According to Wood et al. (2012), the income elasticity of demandfor tortilla by non-poor andpoor house-holds is not significantly different fromzero except for the urban non-poorwhohave a very small incomeelasticity, indi-catingthatchangesinincomeareunlikelyto change tortilla demand. This is areflectionoftherelevantstatusoftortillaasastaplethatthesehouseholdsconsumein the desired quantity. Therefore, lowincome household’s standard of livinghas suffered the most as a consequenceofpricepeaksevents,suchas,thetortillacrisishappenedin2007(38).

From the supply side, tortillerías andsupermarkets mainly distribute tortillas.Thefirsthaveincreaseddramaticallyafterthe deregulation of the tortilla industryin 1998 (37). However, due to the highlyinelastic demand for the consumption,its price has not decreased as would beexpected in a competitive market butinstead it has increased (23). FollowingKeleman et al. (2009),thissituationcouldbe exacerbated by the current publicprograms that incentivize the opening of new tortillerías andmillsthroughsubsidies.

The second player, supermarkets,accountedin2014forthedistributionof25%of the totalmaizetortillascommer-cializedinMexicoandtheyhaveemergedasaseriouscompetitorofthetraditionaltortillerías (27). Supermarkets have agreaterpresence innorthernandcentralparts of Mexico, having less presence insouthern Mexico and rural areas (25).Throughout the lastdecade, regular foodpurchasesatsupermarketsofmiddleandhigh-income households have increased,thisisreasontobelievethatatamedium-termsupermarketswill consolidate theirposition as the primary competing channel of tortillerías, acquiring greater marketshare of tortilla sales at the expense of them (10).

With such a broad panorama, itshouldnotbesurprisingthatfarmersandproducers behave in different followingthe socioeconomic environment of their regions, therefore influencingthemarketperformance along the supply chain ofmaizetortilla.Thisisduelargelytoasym-metries among segments of this chain,whichconditionedtheresponsesofindus-trial-scale corporations and small-scalefamilybusinesses.

Whentheperformanceofagriculturalmarkets are analyzed, a key issue is theextenttowhichthedifferentagentsinthesupplychainrespondtochangesinpriceshocks (29, 36). Therefore, the analysisof spatial price transmission is central in understanding the extent of the inte-gration of economic agents into the market process(16).Whilestudiesofpricetrans-mission in a spatially separated contextshavereceivedsubstantialattentionintheliterature [e.g. Ethiopia (12); Ghana (6)and Chile (35)], only a few studies haveexplicitlyexaminedthe impactofmarketchain structures on the transmission ofprice signals among regions.

282

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

The paper aims to fills this gap byexploring the performance of the Mexican maize-tortillachainthrougharegionalpricetransmissionanalysis.Inparticular,focusonacountry-levelstudyoftheMexicantortillamarket by considering different regionsandcommercializationchains.InnovatesbyimplementaVectorErrorCorrectionModel(VECM) to determine the spatial priceintegration level of inter-region tortillería price, inter-region supermarket price andthe intra-region tortillería-supermarketprice. Since these goods should tradefreely between the different regions inMexico,itishypothesizedthattortillapriceshould be positively cointegrated within their producing regions and betweenthe two main selling points, tortillerías andsupermarkets.

The article proceeds as follows:Sections 2 describes the data sourcesand methodology. Section 3 presentsthe results and discussion, respectively.Finally, Section 4 summarizes the mainconclusions.

Materials and methods

Data description A total of 362 weekly observations

from the January 2007 to December2013wereusedtoimplementtheVECMs(figure 1). Tortilla price series wereacquiredfromtheSNIIM(NationalInfor-mation System and Market Integration)(33) for each of the 50 urban centersspreadthroughoutallofMexico.

Figure 1.Averagepricesoftortillaforthe5regionsconsideredonthisstudy(expressedinEuros).

Figura 1.Preciospromediodetortillaparalas5regionesconsideradasenesteestudio(expresadasenEuros).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

5 W

K 20

0715

WK

2007

25 W

K 20

0735

WK

2007

45 W

K 20

072

WK

2008

12 W

K 20

0822

WK

2008

32 W

K 20

0842

WK

2008

52 W

K 20

089

WK

2009

19 W

K 20

0929

WK

2009

39 W

K 20

0949

WK

2009

7 W

K 20

1017

WK

2010

27 W

K 20

1037

WK

2010

47 W

K 20

105

WK

2011

15 W

K 20

1125

WK

2011

35 W

K 20

1145

WK

2011

3 W

K 20

1213

WK

2012

23 W

K 20

1233

WK

2012

43 W

K 20

121

WK

2012

11 W

K 20

1221

WK

2012

31 W

K 20

1241

WK

2012

51 W

K 20

12

Valu

e (E

uros

)

Period (Week)

T OCCIDENTAL

T NORTH WESTERN

T SOUTHERN

T NORTH EASTERN

T CENTRAL

283

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

In order to classify each price pairaccordingtheirsaleschannels(supermarketand tortillería),theENIGH2006andENIGH(NationalSurveyofHouseholdIncomeandExpenditure) 2012 surveys were used tocompare the percentage share of tortilla purchasesconsumersmadeateachchannel.

Price data was transformed to thehistoricEurovaluebecauseprice fluctua-tionsdistortthevalueofeconomicvaluesmeasured in nominal values (32). In thissense, since there is no daily or weeklyconsumerpriceindexmeasurement,trans-formation to Euroswas decided over thedollarorMexicanPesogiventothe lowerinflation level in the European Unionduringthetimebeingconsidered(13).

All price series were transformed tonatural logarithms to enable the coef-ficient's interpretation as elasticity. TheMexican urban centers were identifiedfollowingtheclassificationoftheMexicanSecretariat of Agriculture, Livestock,Fisheries, Rural Development, and Food(SAGARPA). These centers were sortedby regions according their distance andpopulationlevel(17).

The regions are: North West, NorthEast, Occidental, Central and SouthernDistinctvariables for tortilla (from itemsbought) and for regions (from chain

variable containing state information).The sample was filtered for tortillasthat were bought in urban centers ofmore than 100,000 habitants and in therespective region. This allowed a deeperanalysisregardingtheproportionofurbanhousehold tortilla purchased by super-markets or tortillerías for the respective regionandyear(table1).

Price transmission analysisWell integrated markets play an

essentialroleinreducingriskscausedbysupplyanddemandshocksbyfacilitatingadjustments in trade flows across space,andindoingsoreducingpricevariabilityfacedbyconsumersandproducers(4).

Market integration can be measuredthrough short and long run price trans-mission elasticity, which measures theeffectof aprice change foragood inonemarket of the price in another market (39).

Most of price transmission research in Mexico has been centered on analyzethe effect of international price shocks on Mexican maize markets and thedynamic of vertical price transmissionalong the tortilla industry. For example,Motamedet al.(2008)analyzedthepricetransmission between US yellow maizeandMexicanwhitemaize.

Table 1.Averagedistanceandpopulationforeachregionconsideredonthisstudy.Tabla 1.Distanciapromedioypoblaciónparacadaregiónconsideradaenesteestudio.

Source:INEGI(2014)/Fuente:INEGI(2014)

Av. Distance between regions

North Western

North Eastern Occidental Central Southern Population

North Western N/A 1091 1442 1776 2300 10,307,316

NorthEastern 1091 N/A 736 1136 1623 17,200,470

Occidental 1442 736 N/A 448 935 23,437,097

Central 1776 1136 448 N/A 601 38,807,720

Southern 2300 1623 935 601 N/A 22,583,935

284

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

Their findings demonstrated a weakevidence of cointegration between thesetwoprices, further suggesting that publicsubsidyprogramscouldaffecttheconver-gence of price shocks between them.Araujo-Enciso (2011) studied the verticalpricetransmissionfortheMexicanTortillaIndustry by analyzing how social capitalinfluenced price transmission betweenmaize,maizeflour,andmaizetortilla.TheworkthatmostresemblesthisanalysiswasdonebyAppendiniandQuijada(2010)whoanalyzedtheproductionandconsumptionstrategiesofstaplefood(maize-tortilla)byruralhouseholdsofMexico.

Thisworkexploresspatialpricetrans-mission of maize tortilla for 5 regions distributed throughoutMexico by consid-ering the most important urban centersandthesupermarketchannels.Toestimatethe speed of price transmission betweenregions, an approach based onEngle andGranger(1987)wasused.TwopriceseriesƤt

A andƤtB from regions A and B respec-

tively are cointegrated if each is a non-stationaryprocess(integratedoforder1orI(1))andiftheresidualsεtofthelong-runrelationship between these prices arestationary(integratedoforder0orI(0)).

(1)

Before the price transmission modelcan be modeled, stationarity, cointe-gration level and long run cointegrationneedtobeverified.TheLeybourne(1995)andOteroandSmith(2012)criticalvalueswere applied to testthe stationarity foreach price series. When the existence of unitrootsareconfirmed, it ispossibleto

formulateatestforcointegrationvectors,that is, the long run equilibrium rela-tions.Ingeneralterms,inthepresenceoftwovariablesytandxt which are I (1) the residualswillbeI(1).

However, if this variables are I(0), yt andxt arecalledcointegrated.Thecointe-gration level between serieswas rankedaccording the Johansen trace method(18)and theoptimal lagwasdecidedbytheAkaikeCriterionandFinalPredictionError(21).Thetestresultsarepresentedin table 2 (page 285). Assuming that Ƥt

A andƤt

B are indeedcointegrated,avectorerrorcorrectionmodel (VECM) for thesetwopricesareestimated.TheVECMtakesthefollowinggeneralformequation2*.

The adjustment parameters αpAB and

αpBA measure the speed with whicp price

transmission takes place. Hence, they arethe parameters of interest in the following stages of this analysis. Error correction,and thus cointegration between Ƥt

A andƤt

B, requires that αpAB<0 and αp

BA>0 . For example, ifƤt

A is too high with respect to Ƥt

B then ἒt*willbepositive.Inthiscase,and

inequation(2)ensurethatp Afallsandp

B increases in the next period, thus drivingthese prices closer to the long-run rela-tionship.Overall,thelong-runpricetrans-Overall,thelong-runpricetrans-mission elasticity (ß1)andshorttermerrorcorrections terms (α1 & α2) for tortillería prices between regions (3 pairs), super-market prices between regions (2 pairs)andfinally,betweensupermarketandtorti-llería prices with in regions (2 pairs) were estimated.Therefore,7VECM'swereimple-mentedbyusingpricepairscorrespondingto each tortillería-supermarketandregionalcentersinteractions(table3,page285).

*

285

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

Results

Observing table 4 (page 286) theaverage sale prices of tortillas at super-markets are considerably lower thantortillerías inallregions.Whenbothmeanprices are compared, it was found that

supermarket prices between regions aresmaller than tortillería prices, furthersuggesting higher volatility due to theatomization of these last sale points.

Table 2.CointegrationranksaccordingtheJohansenTraceMethod(18).Tabla 2.RankingsdecointegracióndeacuerdoconelMétododeJohansen(18).

Region Rank Eigen value Trace Test P-valueT.NorthWestern-T.NorthEastern

0 0.1318 57.741 0.00001 0.0183 6.689 0.3886

T.NorthWestern-T.Southern0 0.1251 55.981 0.00001 0.0211 7.738 0.2825

T.NorthEastern-T.Southern0 0.1387 60.342 0.00001 0.0176 6.433 0.4179

S.Occidental-S.Southern0 0.1055 45.259 0.00001 0.0141 5.118 0.5871

S.Occidental-S.Central0 0.1540 66.266 0.00001 0.0161 5.8813 0.4858

T.Central-S.Central0 0.0376 14.890 0.01721 0.0028 1.0174 0.3647

T.Occidental-S.Occidental0 0.0493 18.004 0.01891 0.0008 0.2873 0.5919

Table 3. VECMResultsfortheBivariateCointegration.Tabla 3.ResultadosdelModeloVectorialdeCorrecciondeErrores.

Interactions Variables ß1 α1 p-value α2 p-valueDurbin Watson

Test

Dist. Reg.

(Km)

Inter-regional Supermarket Cointegration

ΔLogP Central / ΔlogP Occidental

-0.8626 -0.1258 0.0024*** 0.1552 0.0002*** 1.8698 448

ΔLogP Southern / ΔlogP Occidenal

-0.8995 -0.0361 0.3982 0.1679 0.0001*** 1.8213 935

Inter-regional Tortillería Cointegration

ΔLogP Southern / ΔLogP NorthEastern

-0.9178 -0.1956 0.0009*** -0.0776 0.2198 1.8932 1623

ΔLogP Southern / ΔLogP NorthWestern

-0.9867 -0.2178 0.0001*** -0.0867 0.1756 1.8390 2300

ΔLogP NorthWestern / ΔlogP NorthEastern

-0.9156 -0.1810 0.0075*** 0.0687 0.2287 1.8817 1091

Intra-regional Tortillería-SupermarketCointegration

ΔLogP T.Occidental / ΔlogPS.Occidental

-0.4976 -0.0886 0.0003*** -0.0390 0.2561 1.8516 N/A

ΔLogP T.Central/ ΔLogP S.Central

-0.5567 -0.1056 0.0002*** -0.0701 0.0886 1.9984 N/A

286

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

TheLeybourne(1995)andOteroandSmith (2012) critical values confirmedthat12ofthe25pairsand15of25pairsare cointegrated of order 1, respectively.For all price pairs, the Durbin Watsonstatistics are both close to 2, meaningthat there is low auto-correlation ofthe residuals, all results are availableupon request. These findings confirm along-term bivariate cointegration rela-tionship in the following pairs: Tortillería Central-SupermarketCentral, Tortillería Occidental-SupermarketOccidental,Torti-llería North Eastern-Tortillería Southern,Tortillería North Western-Tortillería Southern,Tortillería North Western-Torti-lleríaNorthEastern,SupermarketCentral-Supermarket Occidental, SupermarketOccidental-Supermarket Southern. TheJohansen trace test method confirmed abivariatecointegrationrank1for7pricepairs. These results are presented ontable2(page285).

With respect to the VECM estimationresults, the cointegration parametersfor each pair of regional interactions are presentedintable4.

Discussion

Inter-Regional SupermarketFortheOccidental-Centralpricepair,the

long-term elasticity (ß1)was0.86,thatis,a1%incrementinthepriceofCentralregionwould generate an average increment of0.86%inthepricesoftheOccidentalregion.Both short-term error correction termsindicateaproper(negative)andsignificantadjustment toward equilibrium repre-sented by α1 (-0.13) and α2 (0.15). Theseresults demonstrate that prices betweenOccidentalandCentralregionsfolloweachother, being the Central region’s α1 the dominantpricemaker,sinceitcorrectslessthantheOccidentalregion.

In contrast with ß parameters, theaveragelowercorrectiontimetowardequi-libriumofCentral-Occidentalisconsistentwiththenegativeimpactofdistance.

Despite its shorter avg. distance(435km), theCentral-Occidentalpairhasthe lower ß1 (0.86) with respect to the Occidental-Southern pair, which presentsa ß1 of0.90andadistancebetween themof 935 km. This is not consistent with

Table 4. Timeseriesavg.priceinEurosforperiodJanuary2007-December2013.Tabla 4.PreciopromedioenEurosparaelperiodoenero2007-diciembre2013.

Source:SNIIM(2014)/Fuente:SNIIM(2014)

Region & Point of Sale Avg. Price in Euros Standard Deviation

Tortillería_Occidental 0.615 0.077Tortillería_NorthWestern 0.733 0.105Tortillería_Southern 0.664 0.105Tortillería_NorthEastern 0.652 0.098Tortillería_Central 0.608 0.072Supermarket_Occidental 0.438 0.101Supermarket_NorthWestern 0.456 0.100Supermarket_Southern 0.435 0.099Supermarket_NorthEastern 0.459 0.112Supermarket_Central 0.429 0.095

287

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

the concept that distance should have anegative impact on price transmission. According toASERCA (3), theway super-marketsdeterminetheirpricescouldbeafactorintheasymmetricrelationbetweenprice transmission and distance. Forexample,inmostofMexicansupermarkets,theconsumerpriceisdeterminedbytheirregional headquarters, which define anaverageof40%marginindependentofthelocalization of each one (28). This couldgenerate a relative market segmentation,whichseemtohavetriggeredashiftintheconsumerpreferredplaceofpurchaseandalackofcointegrationwithintheSouthern,NorthWesternandNorthEasternregions.

Following Araujo-Enciso (2011),Mexican supermarkets have seen theirmarket share increased, while cornerstoresandstreetvendorshaveseentheirjoint market share reduced. The factorsunderlyingthissituationarediverse.First,they show an increasing consumption oftortillaamongurbanareas.Intheseareas,the supermarkets concetrate the salesoftortilla, thusgeneratinga trade-off effectwith respect to the rest of sale´s points (14).Second,regardingpricesoftortillas,maizeparticipationinthecoststructureofthedevelopmentofsatellite tortillamassdough is over 68% as opposed tomaizemealpathway(5).Asaresult,theincreasein the price of maize flour or dough isreflected directly in the price of tortillassoldintortillerías andsupermarkets.Dueto the greater bargaining power of thesupermarkets,theyobtainmoreattractivecustomer prices in relation to cornerstoresandstreetvendors.

On the other hand, Keleman et al(2009) demonstrated that supermarketswouldacquiregreatermarketshareastheoverall tortilleríapricesremainedhighandsupermarket prices remained low. Of thefindings, it is expected that supermarket

increases their market power allowing a more intense participation of the overall maize-flour industry on these sale points.Althoughnotallsupermarketpricespresentevidenceofcointegrationbetweenregions,the difference in regional prices is smallwhen compared to the range of averagetortillería prices. This can be attributed tothe vertical integration of flour companiesinbuyinggraindirectlyfromproducersoncontract,grainstorage,andlogisticaldistri-bution.Theyhavemanyplantsstrategicallydistributed throughout Mexico and buydirectlyfromproducers,storeanddistributegrains to their own facilities and also sellmaizeatapremiumtothenixtamalindustry(27). In addition, supermarkets have thefinancial resources towholesale purchase,havegreatstoragecapacity,andbargainingpower. Therefore, they can negotiate flourprices forall their stores throughoutalloftheregionsanddefinetheirpricestohavelessvariabilitywithinregions(31).

Inter-Regional Tortillería Cointegration

Itwasfoundthattortillería prices in the peripheral regions of Mexico presentedevidenceofcointegrationwitheachotherwhilethecentralregionsdidnotpresentsignificant evidence of such behavior.Thelackofcointegrationbetweencentraland peripheral regions is an interestingfinding that could be attributed to theradial transportation systemand storagere-distributionactivitiesinMexico,whichiscenteredintheCentralregion(22,29).

When long-term elasticities are analyzed, it was found that all coeffi-cients of the inter-region tortillería pairs are near to one. Among them, theNorthWestern-North Eastern pair has theshortestaveragedistance(1091km),thelargest ß1 (0.92)andthelowestα1 (-0.18) of the three relations.

288

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

Further confirming the assumptionthatdistancehasanegativeeffectonpricetransmission.

When Southern-North Eastern andSouthern-North Western price pairs arecompared,itwasfoundthatlongtermelas-ticities and short error correction termsare larger for the first one than for thesecond.FollowingtoMcMichel(2009),netproductionandconsumptionamongtheseregions could be a factor in the determi-nation of these parameters. When terms of production are compared, North Easternregionsanetconsumerofwhitemaizeand90%oftheirproductionisconcentratedintheSpring-Summercycle.

The Southern region is slightly anet importer and it has cyclical maizeproduction with deficits and surplusesand produces nearly 80% of its maizein the spring-summer cycle. The NorthWesternregionisthebiggestnetexporterofmaizeinMexicoandaccountsforabout80%oftheautumn-winterproduction(9).

Considering these seasonalities, sincethe North Western region is the only net producer among them, the likelihood oftrade is definitely greater between theNorthWesternandtheSouthernregions.Therefore,itisclearthatlargertradeflowscanbeaccountingforincreasinglevelsofpricetransmissionandfasteradjustmentstowardequilibriumfromnetproducingtonetconsumingcountries.

Despitetheirhigherpricesandlookingattortillaplaceofpurchase,thetortillerías ofCentralregioncouldbeenabletoretaintheir market share (76%). This couldbe explained by an inelastic consumerbehavior and greater competitivenessof tortillerías in these regions that were previouslyconfirmedbyRetes-MatillaandReyes (2010).

Intra-Regional Tortillería-Supermarket Cointegration

It was found that Central regionpresentedthehighestlong-termelasticity(ß1=0.50)andthefastesterrorcorrectionterm (α1=-0.11) when compared withthe Occidental region, which present along-term elasticity of 0.49 and an errorcorrection term of -0.08.

The fact that elasticitiesdiffer signifi-cantly from 1 could be attributed tothe higher costs of production (30%higher than supermarkets) and qualitydifferencesoftortillerías for which certain consumers are willing to pay a higherprice (10). Nevertheless, these regionsaretheonlypresentingevidenceof long-termprice cointegration between super-marketsandtortillerías.

As demonstrated by Espejel-García(2012), due to they low prices, they arethe only two regions where tortillerías have been able to retain their largermarket share. It canbe attributed to thegreater importance placed on the priceof tortillaby the lowandmiddle-incomehouseholds.

From this results, infers that themost affected by tortilla’s high price arethe lower income consumers, who haveneglected their quality preference of thenixtamaltortillabygivinggreaterimpor-tance to the price and quantity in orderacquire sufficient tortillas to meet theirbasic dietary needs. In this sense, sinceconsumer preferences have shifted lowpriceasapriority inordertomeetbasicconsumption needs, in seeking for thelowest tortilla prices, they have shiftedtheir place of purchase preference forthe cheaper channels, which are: super-markets, corner stores, and streetvendors(15).

289

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

Overall, it is very likely that costsstructureismuchhigherfortortillerías in regionswheresupermarketandtortillería pricesarenotcointegrated.Theextracostis likely to emanate from higher transpor-tationcosts,whichresultinhighermaizeprices. Therefore, infers thath tortillerías inperipheralregionsarenotabletoreactto lowerpricesofsupermarkets,as theirmarketsharehaseroded.

In contrast, tortillerías located in theCentral region, which are cointegratedwith supermarkets, havemaintained theirmarket share due to the combination ofgreaterpricetransmissionandaninelasticdemandfortortillaelaboratedattortillerías.

Thesefindingsappeartovalidatethatradial transportation system in Mexicohas benefited the central regions at theexpenseoftheperipheryregions,enablingtortillerías in Central regions to remain competitivewithsupermarketsandretainmarketshare.Giventhatcommoditieslikemaize have relative low prices, they arevery sensitive to transportation, ware-housing and transfer cost. Therefore, itisreasonabletobelievethatwhitemaizecostscouldbeuptodoublethepriceintheextreme areas of Mexico.

From this results, infers that thepriceofmaizepaid toproducers isnotdirectlyrelated to the big increases in tortillaprices or tortilla price high variability.Instead it attributes theirprice formationto distribution costs and intermediaries.From a policy perspective, the Mexicangovernment must revise the subsidypolicies that primarily benefit graincommercializing companies and largefirms. The lack of competition in storageand distribution of maize has adverselyaffected the end consumer, tortillerías locatedintheperiphery,andnixtamalmillslocatedintheperiphery.

The World Bank (2007) recommendthat the government allow the bankingsector to provide better financial mecha-nisms to maize producers, like allowingthemtousetheir inventoriesascollateralto acquire loans. This would empowerproducersandwouldgive themagreaterflexibilityincommercializingtheirproduct,potentially benefiting the nearest tortilla consumermarkets.

Furthermore, subsidies to tortillerías should not encourage new players in analready saturated market, but rather theyshould encourage the efficient productionandthepropereconomyofscalesrequiredto produce tortillas at a competitive price. Additionally, FIRA (2014) recommendsthat tortillerías use marketing to promoteconsumptionoftheirbetterqualitytortillas,so that they are able to maintain marketshare even when their price is above thesupermarketprice.

Nevertheless,alowerthecoststructurefor tortillerías wouldbeneededinordertoretain the tortillería market share necessary for the tortillerías currently in business.Achieving proper production economiesofscaleanddirectlinkagesbetweenmaizeproducers and tortillerías is essential to reductionofcostsrequiredtomaintaintheircompetitiveness.

Conclusions

The aim of this research was tode-termine the spatial price transmission level of the Mexican maize tortilla market across different market chains andregions.Theresultswereheterogeneous,demonstrating some evidence of pricecointegrationbetweenregions forsuper-market and tortillería prices and withinregionsbetweentheplacesofpurchase.

290

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

The findings of this paper were thanviewedinlightoftheENIGHsurveyresults,whichdemonstratedchangesintheplaceof purchase of tortillas between theyears2006(priortothecrisis)and2014(after the crisis). These surveys reveal anotice able reduction in consumption atthe tortillería channel and an increasingconsumptionatsupermarketorresellersofsupermarkettortilla,likecornerstoresor street vendors. The regions with thegreatest consumer reaction seem tobe the regions with the lower levels ofprice transmissionbetween thedifferentmarket chains.

The tortillerías of Central and Occi-dental regions enjoy some of the lowesttortilla prices in Mexico, which can be

attributedtothecheaperlogisticcostsofsuchregions.Theseregionshaveastrongsupermarketcointegrationbetweenthemandunliketheotherthreeregions; thesearetheonlytwothathaveevidenceoflong-termcointegrationbetweensupermarketprice and tortillería price. Accordingly, itis concluded that maize producers andtortilla producers would benefit from adirect transaction through the reductionofcommercializationmarkupofmaize.

Further research is recommended tounderstandthecausesoftheregionalvari-ability.Forexample,atechniqueofdimen-sionality reduction could be performed,tomeasure the impactof factors suchasinsecurity,fuelpricesandstoragecapacityin the spatial price transmission of tortilla.

References

1.Appendini,K.;Quijada,M.2010.ConsumptionstrategiesinMexicanruralhouseholds:pursuingfoodsecuritywithquality.AgricultureandHumanValues.4:1-16.

2.Araujo-Enciso,S.2011.Análisisdetransmisióndepreciosentrelosmercadosdemaízmexicanosy el mercado estadounidense: Métodos lineales y no lineales. Revista Española deEstudiosAgrosocialesyPesqueros.5:39-78.

3.ASERCA(AgenciadeServiciosalaComercializaciónyDesarrollodeMercadosAgropecuarios).2014.Laformacióndelospreciosenlacadenadevalormaiz-grano-tortilla.155p.

4.Barrett,C.2005.Spatialmarketintegration.Thenewpalgravedictionaryofeconomics.2:1-7.5. Caspi, C. E.; Pelletier, J. E.;Harnack, L. J.; Erickson,D. J.; Lenk,K.; Laska,M.N. 2017.Pricing

of staple foods at supermarkets versus small food stores. International journal ofenvironmentalresearchandpublichealth.14(8):915.

6.Cudjoe,G.;Breisinger,C.;Diao,X.2010.Localimpactsofaglobalcrisis:foodpricetransmission,consumer.FoodPolicy.Elsevier.35(4):294-302.

7.DeIta,A.2012.CatorceañosdeTLCANylacrisisdelatortilla.ProgramadelasAméricasReporteEspecial.33:1-25.

8.Engle,R.;GrangerC.1987.Co-Integrationanderrorcorrection:representation,estimationandtesting.Econometrica.55:251–276.

9.Espejel-Garcia,M.2012.CaracterizacióndelconsumodetortillaenelEstadodeMéxico.TesisDoctoradoenEconomía.ColegiodePostgrados.103p.

10.FIRA(FondoMexicanodeGarantiayFomento).2014.AnálisisdelmercadodelatortillaEnMéxico.112p.

11. Gasca, J. 2014. Corporate control of food distribution in Mexico. Revista Problemas delDesarrollo.176:45-55.

12. Getnet, K.; Verbeke, W.; Viaene, J. 2005. Modeling spatial price transmission in the grainmarketsofEthiopiawithanapplicationofARDLapproachtowhiteteff.AgriculturalEconomics.33:491-502.

13.González-Rivera,G.;Helfand,S.2001.Economicdevelopmentandthedeterminantsofspatialintegration in agricultural markets. Working Papers. Department of Economics.UniversityofCalifornia.Riverside.33p.

291

Market integration along the Mexican maize-tortilla chains

Tomo 50 • N° 2 • 2018

14.González,H.;Martinez,D.2016.EfectodetransmisióndepreciodelmercadodelmaízalmercadodelatortillaenMéxico.RevistaMexicanadeCienciasAgrícolas.6(6):1149-1162.

15.Guzmán,G.2007.Incrementoenlospreciosdelmaízylatortilla.RevistaLatinoamericanadeEconomía.38:151-167.

16. Holst, C.; Von Cramon-Taubudel, S. 2013. Trade, market integration and spatial pricetransmissiononEUporkmarkets following eastern enlargement.DAREDiscussionPapersGeorg-AugustUniversityofGöttingen.DepartmentofAgriculturalEconomicsandRuralDevelopment.1307:1-35.

17. InstitutoNacionaldeEstadísticayGeografía(INEGI).2010.CensodePoblaciónyVivienda.Available in http://www.inegi.org.mx/est/contenidos/proyectos/ccpv/cpv2010/Default.aspx

18. Johansen, S. 1991. Estimation and hypothesis testing of cointegration vectors in Gaussianvectorautoregressivemodels.Econométrica.59:1551-1580.

19.Keleman,A.;García,H.;Hellin, J.2009.Maizediversity,poverty,andmarketaccess: lessonsfromMexico.DevelopmentinPractice.19:187-199.

20.Khim,V.;Sen,L.2004.Whichlaglengthselectioncriteriashouldweemploy?.UniversitiPutraMalaysia:EconomicsBulletin.3:1-9.Print.

21.Kjöllerström,M.2004.Liberalizacióncomericalagrícolaconcostosdetransporteytransacciónelevados:evidenciaparaAméricaLatina.CEPALDesarrolloProductivo.45p.

22.Leybourne,S.;Newbold,P.;Vougas,D.1998.Unitrootsandsmoothtransitions.JournalofTimeSeries Analysis. 19: 83-97.

23.Mantilla,R.;Fernando,R.;TorresSanabria,G.;GarridoRoldán,S.2014.UnmodeloeconométricodelademandadetortillademaízenMéxico,1996-2008.Estudiossociales(Hermosillo,Son.). 22: 37-59.

24.McMichael,P.2009.Afoodregimeanalysisofthe‘WorldFoodCrisis’.AgricultureandHumanValues.26:281-295.

25. Motamed, M.; Foster, K.; Tyner, W. 2008. Applying cointegration and error correction tomeasure trade linkages:Maize prices in theUnited States andMexico. AgriculturalEconomics.39:29-39.

26.Otero,J.;Smith,J.2012.ResponsesurfacemodelsfortheLeybourneunitroottestsandLagOrderDependence.ComputationalStatistics.27:473-486.

27.Retes-Mantilla, J.;Reyes, F.2010.Demandade tortillademaízenMexico,1996-2008.TesisdeDoctoradoenCiencias,especialistaenEconomía.ColegiodePostgraduado.Mexico.124p.

28.Rogel,F.;Torres,G.;Garrido,S.2014.UnmodeloeconométricodelademandadetortillademaízenMéxico,1996-2008.EstudiosSociales.22:37-59.

29.SAGARPA(SecretaríadeAgricultura,GanaderíaDesarrolloRural,PescayAlimentación).2014.SituaciónActualyPerspectivas1996-2010.111p.

30. Sánchez-Toledano, B. I.; Kallas, Z.; Gil, J. M. 2017. Importancia de los objetivos sociales,ambientales y económicosde los agricultores en la adopcióndemaízmejorado enChiapas,México.RevistadelaFacultaddeCienciasAgrarias.UniversidadNacionaldeCuyo.Mendoza.Argentina.49(2):269-287

31.Schmitt-Grohé,S.;Uribe,M.2013.DownwardnominalwagerigidityandthecasefortemporaryinactionintheEurozon.TheJournalofEconomicPerspectives.27(3):193-211.

32.Shields,D.2011.Mexicansupermarketsspurnewproducedistributionsystem.USDAEconomicReports. 41p.

33.SistemaNacionaldeInformacióneIntegracióndeMercados(SNIIM).2014.Informeestadísticodemercadosagrícolas.Availablein:http://www.economia-sniim.gob.mx/nuevo/mapa.asp?br=Mozilla/5.0%20(Macintosh;%20Intel%20Mac%20OS%20X%2010_13_4)%20AppleWebKit/605.1.15%20(KHTML,%20like%20Gecko)%20Version/11.1%20Safari/605.1.15

34.Torres,F.2014.TransformationsinfoodsupplyinMexicanCities.VoicesofMexico.90:19-23.35.Valdes,R.;VonCramon-Taubadel,S.;Díaz-Osorio,J.;Engler,A.2011.TheChileanwheatmarket

anditspricesupportmechanism,aspatialmarketintegrationanalysis.RevistadelaFacultad de Ciencias Agrarias. Universidad Nacional de Cuyo. Mendoza. Argentina.43: 47-58.

292

R.V.Salazar

Revista de la Facultad de Ciencias Agrarias

36.Valdes,R.;DiazOsorio,J.2015.TheBrazilianbeefmeatsectorintoadomesticandinternationalcontext: a Supply Chain Management (SCM) approach. Revista de la Facultad deCienciasAgrarias.UniversidadNacionaldeCuyo.Mendoza.Argentina.47(1):233-239.

37.WorldBank. 2007.EstadosUnidosMexicanos Integracióndelmercadonorteamericanodemaíz:implicacionesparalosproductoresyconsumidoresmexicanos.DepartamentodeDesarrolloSostenibleRegióndeAméricaLatinayelCaribe.21p.

38.Yunez,A.2008.ElTLCAN,laspolíticaspúblicasalsectorruralyelmaíz.rumborural.7:1-10.39.VonCramon-Taubadel,S.;Loy,J.;Meyer,J.2006.Anexperimentwithgermanfoodprices.16th

InternationalAssociationofAgriculturalEconomistConference.GoldCoast.Australia.40.Wood, B. D.;Nelson, C.H.;Nogueira, L. 2012. Poverty effects of food price escalation: The

importanceofsubstitutioneffectsinMexicanhouseholds.FoodPolicy.37(1):77-85.

AcknowledgmentFundingforthisresearchcomesfromtheDeutscherAkademischerAustauschDienst(DAAD).IamgreatlyindebtedtoDavidQuinteroforhisconscientiousandableassistanceduringthis

research. Any remaining errors are my own.