Southeast Asian Journal of Economics 4(1), June 2016: 29-57 Agricultural Policy Misspecification in Thailand Nararuk Boonyanam School of Development Economics, National Institute of Development Administration, Bangkok, Thailand Email:[email protected]Abstract Severe poverty of farmers in Thailand is visible in the form of agricultural land abandonment. The causes of land abandonment will reflect the causes of poverty and the effectiveness of government agricultural policies. This paper identifies the factors that cause agricultural land abandonment and the appropriate policy measures for effectively addressing the problem. The study utilizes the optimization method based on duality production theory with 808 parcel samples. The Simultaneous Tobit estimation was used to estimate the system of nine equations. The study identified the contributing factors for land abandonment and provides alternative directions for policy makers. Key words: poverty, land abandonment, profit maximization, duality production theory 1. Introduction Thai farmers have always been in poverty. They are unable to meet their family‘s basic needs from agricultural production, mainly due to low productivity, low income and inability to manage various risks such as plummeting crop prices, increasing production input prices, shortage of cash for investment, erratic distribution of rainfall and all types of land use problems, including landlessness, land rights, land infrastructure and land fragmentation. Over the years, the Thai government has implemented several
29
Embed
Agricultural Policy Misspecification in ThailandKey words: poverty, land abandonment, profit maximization, duality production theory 1. Introduction Thai farmers have always been in
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Southeast Asian Journal of Economics 4(1),June2016:29-57
Agricultural Policy Misspecification in Thailand
Nararuk BoonyanamSchool of Development Economics,
National Institute of Development Administration, Bangkok, Thailand Email:[email protected]
Abstract
Severe poverty of farmers in Thailand is visible in the form of agriculturallandabandonment.Thecausesoflandabandonmentwillreflectthecausesofpovertyandtheeffectivenessofgovernmentagriculturalpolicies.Thispaperidentifiesthefactorsthatcauseagriculturallandabandonmentandtheappropriatepolicymeasuresforeffectivelyaddressingtheproblem.Thestudy utilizes the optimizationmethod based on duality production theorywith 808 parcel samples. The Simultaneous Tobit estimation was used to estimatethesystemofnineequations.Thestudyidentifiedthecontributingfactorsfor landabandonmentandprovidesalternativedirectionsforpolicymakers.
Key words: poverty, land abandonment, profit maximization, duality productiontheory
1. Introduction
Thaifarmershavealwaysbeeninpoverty.Theyareunabletomeettheir family‘sbasicneeds fromagriculturalproduction,mainlydue to lowproductivity, low income and inability to manage various risks such as plummetingcropprices,increasingproductioninputprices,shortageofcashfor investment, erratic distribution of rainfall and all types of land use problems, including landlessness, land rights, land infrastructure and landfragmentation.Overtheyears,theThaigovernmenthasimplementedseveral
30 • Southeast Asian Journal of Economics 4(1), June 2016
First, Several irrigation projects were constructed in the 1960s in ordertosolvethewatersupplyproblembuttheirimpactattheregionallevelremains very limited as more than 77% of the agricultural area is still underrain-fedconditions(RoyalIrrigationDepartment,2011).Withoutwatersupply, farmers in these areas are unable to cultivate and produce goods for consumption or sale. The governments’ budget constraints hinder thewidespreadconstructionofirrigationsystemsthroughoutThailand.Ontopofthis,thewidespreadsoilacidityinmanyareascreatesalabortransitionfromtheagriculturaltotheindustrialsectorwiththeabandonmentofagriculturalland, even though several attempts have beenmade by the government totacklethesoilqualityproblembyestablishinganddevelopingvolunteersoildoctorsandlearningcentersatthesub-districtlevel.
Second,Theimplementationof theLandCodeActof1954,whichcontainsproceduresforissuinglandtitledeedsrangingfromfulltousufructrightsforfarmers,theactencouragestheclearanceofforestbyrecognizingthreestepsoflandacquisition:occupancy,useandlegalownership(Cleary&Eaton,1996).Underthisact,agriculturallandsizewaslimitedto50raiperhousehold with the intention of solving the problems of land distribution andlandlessness.Onthecontrary,however,thelimitationleadtoproductioninefficiencyforcertaincropsandcreatedalandfragmentationproblem.TheLandInstituteFoundationrevealedthatmostfarmers(87%)aresmallparcelholderswithanaveragefarmsizeoflessthan5rai.Productioncostsincreasedwith fragmentation (McClosekey, 1975), because small parcel size did notallowforefficientmechanizationandlimitedchoicesoffarmingtechniques.Asaresult,smallplotswereabandonedduetothecomplicationofworkingfragmentedland(Simon,1987).Althoughthelandsizelimitwascancelled in 1960, the fragmented and abandoned land problems remain unsolved. Recently, Reform Thailand proposed that the country revert back to the maximumlandholdinglimitof50raiperhouseholdandincreaselandtaxesto 5% for abandoned land and land over themaximumholding of 50 rai.However,thereisnoevidencesubstantiatingwhythelandholdinglimitandtaxratearejustifiedat50raiand5%aswellashelptoimprovethewell-beingofthefarmers.
Nararuk B., Agricultural Policy Misspecification in Thailand • 31
Third,Inthe1970’s,severalgovernmentagencies1alsocarriedoutalandallocationprogramwithlimitedsuccess.TheAgriculturalLandReformAct of 1975 created theAgricultural Land Reform Office (ALRO) in the MinistryofAgricultureandCooperativestoimplementreformsasameansofresolvingthelandlessnessproblem.However,theprogramledtoaconcentra-tionoflandheldbytherich.ThenumberoflandlessfarmersfromtheALROprogram remains at over 800,000 families, while tenancy problems haveworsenedfrom514,717familiesin1996to678,077in2004,and1,373,145familiesin2011(Sajjanand,2013).Farmersperceivedthatthecostofproduc-tion is increased from leasing cropland.Therefore, these farmers lost theirmotivationtocontinueproductionoperationsduetoinsecurelandrightsandunprofitable operations. Fourth, Implementation of the Agricultural LandRentalActof1981aimedtoprotecttenantfarmerrights,butcreatedbarriersthat prevented agricultural landowners from leasing their land to more productiveandefficientfarmersinstead.Furthermore,theactledtoincreasedland abandonment. The minimum lease period enforced by this act was sixyears,whichistoolong.Incaseswherealandownerdesiredtoselltheirland, the act also stated that landownerswere required to provide existing tenantswiththefirstopportunitytopurchasetheland.Sincetenantfarmerswouldtypicallyofferapricebelowfairmarketvalue,thelandremainedwiththepresentownerwhoceasedagriculturaloperationswhileattemptingtosellthe land. Fifth, Civil and Commercial Code, Section 537-571, prohibits alandowner from evicting a tenant before the lease agreement has expired.While Section 1382 aims to reduce land abandonment by allowing land acquisition through adverse possession when an individual has peacefully enjoyedthelandfortenyearswithouttheexpressobjectionofthelandowner,thissectionhasproducedheatednationwidedebate.
32 • Southeast Asian Journal of Economics 4(1), June 2016
sub-districtissetforaspecificcropproductionandfarmerswhoparticipate in the programwill receive a guaranteed selling price of the crop throughsubsidizedprograms.Althoughtheprogramcandictatecroptypeforspecificareas, it cannot control the yield or output; crop yield depends on many uncontrollablefactorssuchasrainfall,soilquality,temperature,pestactivityandactsofnature, etc. Inaddition, thecropprice ina competitivemarket cannotbecontrolledeitherdomesticallyorglobally.Successfulagro-zoninginothercountriescomesfromadifferentsetofobjectiveswheretheprimarygoalforagro-zoningistocontroltheexpansionofurbanareasintoagriculturalareas. Therefore, the zones are use-based, not crop-based as in Thailand. Thusfar,theresultsofapilotprojectforagro-zoningin30provincesremaininconclusiveandthecroppricesareuncontrollableaspreviouslyexpected,whichdiscouragesfarmersfromstayingintheoccupation.
Seventh,TheThai government implemented theDebtMoratoriumProgram(DMP)in2001toalleviatepovertyamongagriculturalhouseholdsand use a period of relief to encourage structural changes in farmers’ operations. Tambunlertchai (2004) indicated that the DMP fell short of itsobjectives; furthermore,after twoyears, theprogramhadnosignificantimpact on its participants.While the goal of the DMP was to relieve its participants’creditconstraints, itsparticipantsfacedloanrestrictions in theformoftakingoutnewloansduringtheperiodofthemoratorium.Therefore,these farmers used new loans from illegal lenders to finance agricultural productionandendupinalargerpovertycycle.Eighth,Thericemortgageschemewas introduced in 2011with the objective of assisting farmers bymanipulatingriceprices,butledtonationwidecorruptionandpricearbitrage.Theschemefailedduetoenormoussubsidycosts,hugeamountsofriceingovernment storageasattempts tomanipulate the internationalcommoditymarketfailed,andamajorityofthebenefitsgoingtothemillersandlargestfarmerswhosellthemost.Itwasproofthatpoorfarmersreceivedverylittlebenefitfromthescheme(Poapongsakorn,Pantakao,Nanthajit,Arunkong,andJanepeungporn,2014).
Ninth,ThecurrentThaigovernmentisconsideringtheCommunityLandTitleAct to solve landlessnessand landabandonmentproblemswith nosupportingevidenceonhowtoputtheactintoactionandwhetherornot
Nararuk B., Agricultural Policy Misspecification in Thailand • 33
it is the right measure to be adopted and Tenth. The government is also consideringareformoftheLandandPropertyTaxActtoresolvetheissues of landlessness, landownership concentration and land abandonment. The initiative refers to theworkofLaovakul (2010)whoproposedvarious taxrates for each categoryof landusebasedon the assumptionof the annual average increase rate for landprices. It shouldbenoted thatvariousflaws existwhenresearchreliesheavilyonassumptions.
Theaforementionedseriesofprogramsandpolicieswereimplementedaccording to the farmers’ trouble trends in each period of time, which is not the right way to solve the problem andmay increase economic costs associatedwithpolicymisspecificationas theprogramswerenotbasedonrigorous,comprehensivequantificationstudy.Obviously,thefarmers’povertyfactors theThaigovernmenthasbeentryingtosolveare thecausalfactorsbehindagriculturallandabandonment;watersupplyshortagesandinappro-priate soil quality create a transitionof labor into the industrial sector and thelandabandonment.Fragmentedparcelshaveledtotheabandonmentofagriculturalland.Leasingofcroplandsincreasethesenseofinsecurityandunprofitablefarmingoperationswhicheventuallyledtolandabandonment.Protectionoftenantfarmerrights ledtothelandabandonmentofagriculturallandowners. Shortage of cash for investment and low crop prices led to discouragement about staying in the farmingoccupation and abandonment of agricultural land. Therefore, solving the core issue of agricultural landabandonmentisthewaytosolvethefarmers’povertyproblem.
AgriculturallandabandonmentinThailandhasdrawnalotofatten-tionoverthepastdecadedueprimarilytotheover-exaggeratedreportingontheamountofabandoned land.Reports routinely reference theestimateof 48millionraiofabandonedlandinThailand.Further,reportstypicallystatethattheamountofabandonedlandisincreasingeachyear(Makkarapirom,2011),and(Sutthiwatthananiti,2012).TheOfficeofAgriculturalEconomics(OAE) defines abandoned agricultural land as agricultural land that was previouslyusedinagriculture,buthasbeenleftidleforfiveormoreyears.Thailandowns320millionraiofland.Asof2011,agriculturallandrepresented48.51%of the total land area and abandoned agricultural land representedonly1.89%or2,879,642rai.Further,OAEdatashowsthatatacountrylevel,
34 • Southeast Asian Journal of Economics 4(1), June 2016
Nararuk B., Agricultural Policy Misspecification in Thailand • 35
theamountofabandonedlandhassteadilydeclinedduringtheperiod1986through2011aspresentedinTable1.However,whenwetookacloserlook at the data from the OAE, we found that there was a drastic increase in agriculturallanduseandabandonmentduring2001-2002.Figure1presentsthe graphs thatwere plotted from the data inTable 1. Since land use and landabandonmenthaveahugevariationinscale,aseparategraphshowingabandonedlandisprovidedtotheright.Attheprovinciallevel,threepatternsoflandabandonmentwerefound;landabandonmentwaseitherincreasing,decreasing,orremainingconstant.Oftheseventy-sixprovincesinThailand,onlythecapitalcity,Bangkok,hadaconstantpatternofnoabandonedland.Fourteenprovincesappeartohaveadeclininglandabandonmenttrend,whilethebalanceofsixty-oneprovinceshaveanincreasinglandabandonmenttrend.
AlthoughthepercentageofagriculturallandabandonmentinThailandis considered low, most of the area in Thailand has an increasing land abandonment trend and it creates a substantial economic loss each year. In1998,itwasestimatedthataneconomicvalueaddedof26,355.32millionbahtwasgeneratedfromtheadditional3%increaseinagriculturallanduse.(LandInstituteFoundation,2001).
36 • Southeast Asian Journal of Economics 4(1), June 2016
Moststudiesonagriculturallandabandonmenthavefocusedontherelation-shipbetweeninstitutionalchangesonlanduseandabandonmentduringthetransitionfromthestate-commandtomarket-driveneconomies,particularlyinpost-SovietEasternEurope.Prishchepov,Radeloff,Baumann,Kuemmerle,andMuller(2012)indicatedthatstronginstitutionandlandusepoliciescanreducetherateofagriculturallandabandonment.Baumannetal.(2011)foundthe most important predictors explaining substantial spatial variations inabandonment rates to be topography, soil type, population, institutional factorsandeconomicshock.InAlbania,MullerandMunroe(2008)indicatedthatmost cropland abandonment at the onset of the transition periodwas concentratedinlessdenselypopulatedareasbutdependedoneconomicreturnsafter the transition period. Evidence from Romania, (Muller, Kuemmerle,Rusu, and Griffiths, 2009) showed that isolated crop lands were more abandoned thanmorehomogenous lands, andhillyareasweremore likely to be abandoned than the plains.Some researchers have indicated that theabandonmentofagriculturallandismostlydrivenbysocio-economicfactorssuchasimmigrationintoareaswhereneweconomicopportunitiesareofferedto rural people, and ecological factors such as elevation and soil erosion(Benayas,Martin,Nicolau,andSchulz,2007).However,resultsfromSpain,(Zaragozi et al., 2012) showed only geographic factors as the important drivingfactoroffarmlandabandonment.Thefactorsareirrigation,vegetationindex,topographicwetnessindexandclimaticindex.NotmuchresearchhasbeenconductedonagriculturallandabandonmentoutsideEurope.InJapan,(Nishihara,2012)foundagriculturallandzoningpolicytohavegreatimpactonlandvalueincreaseanddiscouragethelandownersfromcultivatingland.Asmallgapin theanticipationof landpolicytoconvertagriculturalzonesinto residential development zones prevents landowners from selling or leasinglandtomoreefficientfarmers.EvidencefromThailand(Gine,2005)indicatedthatpartialrightstitlescreateadistortionofthelandrentalmarketbytriggeringasenseofinsecurityforlandowners.Thelandownersaremorelikelytoleaseorabandonsecuredplotsandcultivateunsecuredplotsinordertoavoidexpropriationrisks.
Nararuk B., Agricultural Policy Misspecification in Thailand • 37
andpoliciesshouldbeadoptedbytheThaigovernmenttodealwithagriculturelandabandonment?Thispapercoversalltengovernmentpoliciespreviouslydiscussed, namely, land infrastructure development, limited land holding,landallocationprogram,landrentallaw,adversepossessionlaw,agricultural-zoning,debtmoratoriumprogram,pricesubsidyprogram,community landtitleactandreformoftheLandandPropertyTaxAct.Thefollowingsectionwill discuss the research scope, followedby themethodology used in thisstudy,data,resultsandconclusionwithadiscussiononpolicy
Thefactorsaffectingagriculturallandabandonmenttobemeasuredwere farm gate prices for the three major crops of Khon-Kaen (cassava, rice,andsugarcane);thepriceofinputusedinagriculturalproduction(cashon hand, fertilizer, household labor, machine labor, pesticides, capital or agriculturalmachineryandloaninterestrates).Thephysicalandinstitutionalfactorsareparcelsize,typeofpropertyrights,mainsourceofwatersupply,soilquality,distancetocity,farmingtechnology,landtax,perceptionoflandrentalandadversepossessionlaws.Landsharesforthethreemajorcropsandlandshareforabandonmentwerealsotobemeasured.Thenumbersofyearsinwhichagriculturalmachineryhasbeenusedwerecollectedtocalculatethepresentvalueofeachagriculturalmachinepriceby the following formula: PV=C1/(1+r)nwhereC1iscashflowinPeriod1,ristherateofreturnequaltotheinflationrate,andnisthenumberofperiods.
38 • Southeast Asian Journal of Economics 4(1), June 2016
2.2 Methodology
In this paper, a profit optimization method based on the duality productiontheorywillbeapplied.Thisisthefirstattempttoapplythelanduse model with the land abandonment problem and take the presence of censoredvariablesintoaccount.Withoutrecognizingthesecensoredvariables,theparameterestimatewillbeinconsistency.Theprofitoptimizationmethodis the most widely used method for investigating agricultural land use determinantsasitcancapturethedecision-makingprocessesthatdrivelandmanagement.Themethodhasnolimitationsintermsoffactortypeandisnotlimited to an aggregate level analysis. In addition, it does not require a stackofeconomicdatawhichisfrequentlyunavailableorofpoorqualityindevelopingcountries.
Thedualityproductiontheorywasusedtoobtainproductsupplyandfactordemandequationsforapricetakingfirmbypartialdifferentiationofprofitfunction(Beattie,Taylor&Watts,1985,p.258).Theagriculturalland-ownermakesadecisionwhether touseorabandontheir land tomaximizetheirlong-termprofit.ThetheoreticalmodeladoptedforthispaperwasbuiltupontheChambersandJust(1989)farmprofitmaximizationproblem.Themodel further developed by Fezzi and Bateman (2011) is specified in theequation(1)
Π=f(p,w,z,l1,…,lh), li L1i
h
==
/ (1)
Where Π is the maximum profit associated with the vector of competitiveoutputpricesp,thevectorofcompetitiveinputpricesw,andthevectorofquasi-fixinputzwhichisphysicalandinstitutionalcharacteristicswith l thevectorofh landuseallocationandL is the total landavailable. Assuming the profit function to be positive, linearly homogeneous and convexin the inputandoutputprices, the inputdemandandoutputsupplyequations and the optimal land use share equations can be derived using Hoteling’sLemma.Indicatingwithsthehlandusesharescorrespondingtothe land use allocation, the farm profit function for analysing land use is expressedasfollows:
Nararuk B., Agricultural Policy Misspecification in Thailand • 39
ΠL(p,w,z,L)= ( , , , , , ..., ):max p w z L s s Si 1,...,s sh
hi
h
11
1
P ==
) 3/ (2)
WhereΠL(.)istheprofitperunitoflandandsisthelanduseshareinwhich all shares sum to one.According toHotelling Lemma, the optimalquantityofithinputandthejthoutputarespecifiedasEquations(3)and(4),respectively.
Empirical Specification: The normalized quadratic profit functionwas applied tomulti-input andmulti-output production processes andwasproof outperforming all other forms used in the study (Villezca-Becerra and Shumway, 1992 and Fare, Martins-Filho and Vardanyan, 2009). Theproperties of the normalized quadratic profit function are that it is locally flexibleandself-dualwithagloballyconvexHessianmatrix.Definingx1asthenumerairegoodandvectorofnormalizednetputprices representwith p’=(p/p1,w/p1).Thenormalizedquadraticprofitfunctioncanbeexpressedasfollows:
Π*’= a a p s z p si ii
m n
i ii
h
k kk
n
ij i jj
h
i
m n
0
2 1
1
1 1
1
2
b c d+ + + +=
+
=
-
= =
-
=
+
l l/ / / //
p z s z p p2
1ij i l
l
n
ij i ll
n
ij i jj
m n
i
m n
i
h
i
m n
1 1 221
1
2
{ aU+ + += = =
+
=
+
=
-
=
+
l l le/ / ////
s s z zij i jj
h
i
h
kl k ll
n
k
n
1
1
1
1
11b c+ +
=
-
=
-
==
o// // (6)
40 • Southeast Asian Journal of Economics 4(1), June 2016
WhereΠ*’ is theprofitdividedby thepriceofnetput1,pi isnet putprice, pil is thenormalizedprice (pi/p1, i=2,…m+n), zk is avectorof n quasi-fix input which includes physical and institutional characteristic variables andai,bi,ck, δij, ɸij, φij,aij,bij, andckl are the parameters to be estimated. Linear homogeneity was imposed on the quadratic productionfunction by normalization. Conditions for symmetry are imposed on the modelswiththeconstraintsaij=aji, bij=bjiand cij=cji.Thispaperusesoutputprice;cassavapriceasnumeraire.Theoutputpricewas,onaverage,abetterpredictorofoutputthanthemodelsbasedoninputprices(Paudel&Mcintosh,2005).EvenpreviousworkbyShumway,SaezandGottret(1998),Shumwayand Alexander (1998), and Shumway and Gottret (1991) reported that numeraire was arbitrarily chosen to be the material input. By Hoteling Lemma,inputdemandandoutputfunctionisspecifiedas(7)whereqiisnetputwithpositiveoutputquantityandnegativeinputquantities.
qi= p s zi ij ji
m n
ij jj
h
ij ll
n
2 1
1
1
a a d U+ + +=
+
=
-
=
l/ / / (7)
Optimallandusesharewasderivedfromsolvingthesystemofh-1equations in (5)with s 1j
j
n
1=
=
/ .At the end, therewill be h reduced form equationsasin(8)
A farmermay not plant certain crops in a particular year, thereby resulting in a zero value for the corresponding land shares and output observation. Some landowners leave their land totally abandoned which leadstoazerovalueinbothinputandoutputyields.Byapplyingatraditionalthree-stageleastsquaretothesystemofequationconsistingofthesecensoredvariables leads to inconsistent parameter estimates (Amemiya, 1973). Therefore,theTobitregression(Tobin,1958)wasusedtoestimatethesystem
Nararuk B., Agricultural Policy Misspecification in Thailand • 41
ofequation(7)and(8)tofindfactorscausingagriculturallandabandonment,cropsupplyandinputuse.Becauseofcensoring,thedependentvariableyisthe incompletely observed value of the latent dependent variable y*. ThestructuralequationintheTobitmodelis:yi*=Xi’β+ewhenthedependentvariableiscensoredfrombelowatzeroasfollows:
yi=* * 0
0 * 0
if y
if y
y 1
#) (9)
Afterthesystemofequationshasbeenestimated,themarginaleffectsof each factor will be calculated to find the effect size of each factor on the dependent variable. The following three possible marginal effects are possible: the marginal effect on the latent dependent variable; y*, the marginaleffectontheexpectedvaluefory:E[y]foruncensoredobservationsandthemarginaleffectontheE[y]forcensoredanduncensoredobservations.Wooldridge(2002,p.520)andGreene(2003,p.764)sidewithE[y]asthemostusefulcomponent.Therefore,forthispaper,themarginaleffectontheE[y]forcensoredanduncensoredobservationswillbedeterminedby:∂E[y]/∂xk=ɸ(xiβ/σ)βk.Thispaperestimatesonesystemofnineequations:thetwooutputsupplyequations for riceandsugarcane (note: theprice forcassavawassetasnumerairebecauseitsfarmgatepricewasthemoststableamongthestudycrop),sixinputdemandequations(cashonhand,fertilizer,householdlabor, machine labor, pesticides and capital), and one land share equation for abandonment. Based on the theoretical model proposed by Fezzi and Bateman(2011),thelandshareforeachoutputmustbejointlyestimatedwiththesystemofsupplyanddemandequations.However,theaboveauthorsdidnotestimatethefullmodel;instead,theyestimatedthelandshareequationsseparately because of the convergence and parameter instability problem. Incontrast, thepresent studyestimates the fullmodelbyusingSeeminglyUnrelatedRegressionanalysis.Accordingtothefindings,however,alltypesof land shares cannot be incorporated into the system due to computationburden. Therefore, the researcher followed the suggestion ofArnade andKelch(2007)inconstructingthesimplestmodelpossiblebyincludingonlylandsharesofinterestwhichisthelandshareforabandonment.
42 • Southeast Asian Journal of Economics 4(1), June 2016
2.3 Sampling procedure
The population consists of a number of agricultural parcels in Khon-Kaen. The total number of agricultural parcels in Khon-Kaen are729,375,whilethereare350,100agriculturallandowners(OfficialInformationCenter,Khon-KaenGovernor’sOffice,2012).Multi-stagecluster samplingusingprobabilityproportionaltosize(PPS)wasemployedinthisstudy.Thetotalnumberofagriculturalparcelsampleswasdeterminedbythefollowingformula: n = (deff * Z2pq)/d2 (Cochran, 1977). The United Nations Children’s Fund (UNICEF) recommends design effects equal to two for sub-nationalestimates.Theproportion,p,is0.48(350,100÷729,375).Themarginoferrorisplusorminus5percentagepoint.Witha95%confidenceinterval,α=0.05,Z=1.96,totalsamplesizeis767samples.Thetotalnumberof clusters is determined by the cluster size, which typically ranges from 20-40 households with 15-20 clusters in each province, state or district (UNICEF,1995).Duetobudgetconstraints,thisresearchusedtheclustersizeof52with15clusters(767/15=52)whichisslightlyhigherthantheUNICEFrecommendation. Ineachcluster, thesamplingunitswereselected throughPPSofmulti-stagesampling.Thefollowing4stageswereusedforsampling:district,sub-district,villageandparcel.TheparcellistwasattainedfromtheofficeoftheSub-districtAdministrativeOrganization(OfficeoftheSAO)taxdatabase.Thailawrequiresapropertydeclarationformeveryyearlandtaxesarecollected.TheformcalledPor.Bor.Tor.5provides the followingusefullandinformation:landownername,address,contactnumber,typeoflanduse,landsize,cropsgrownandlocationoftheparcel.Simplerandomsamplingwas used to select the 52 parcels in each village. In total, therewere 780samplesfrom15villages.Itshouldbenotedthattherewereinstanceswherelandowners owned multiple agricultural parcels. However, the random sampleprocessselectedonlyoneparcelofthemultipleparcelsforthesurvey.Whileconductingthefieldsurveyontheselectedsamples,landownerswerealsointerviewedfordataontheotheragriculturalparcels.Therefore,thetotalsampleobtained from the surveywas808 samples.Of these808 samples,thereareabandonedparcelsfor118samplesandtherest690samplesareusedparcels.
Nararuk B., Agricultural Policy Misspecification in Thailand • 43
Anaverageparcel size is around11 rai and the standarddeviation is 9.5, which indicates a wide gap between large and small parcel sizes. Theminimumparcel size is 0.38 rai or 152 square-wa, and themaximum is 100 rai. Cash on hand, or the agricultural landowners’ average savings13,163.77bahtwithaminimumof3,200baht,which isquite low.Under-standably,Thaifarmersarealwaysinpoverty.
Nararuk B., Agricultural Policy Misspecification in Thailand • 45
In addition, the average loan interest rate was 7.82%, while the maximumratewasashighas14%.Thelandownerspossessedanaverageoftwoagriculturalmachinesinwhichthemaximumwassixwhiletheminimumwaszero.Thefrequencyanalysis,whichisnotpresentedhere,showedaveragesoilqualitytobemoderateat56.93%andhighat25.25%,whichindicatedthat80%of theparcelsampleshadmediumtohighsoilquality.Themainwatersourcewasfromrainat81.56percent,followedbyirrigationatonly8.29%,whichindicatedthatmostofthefarmersreliedonnature.
AnestimationoftheoutputsupplyequationsinTable3showsthatdifferent factorsaffectedeachcropsupply.Allsignificant factorsandsignswereasexpected.Thehigherricesupplycamefromareductioninothertypesof land allocation, an increase in thewater supply from irrigation system, alargerparcelsizeandadecreaseinthepropertyright.ThecoefficientoftheFULL_RIGHTandSEMI_RIGHTvariablesuggeststhatthoselandownerswhohavefullandsemirightofpropertyrightonaveragetendtosupplylessofricethanthoselandownerswhohavenopropertyright.ThisisagreeswithGine (2005) in a study indicating that a partial property right inThailand triggersasenseofinsecurityamonglandowners.So,theyprefertoabandonthe securedparcelandcultivate theunsecuredplots toavoidexpropriationrisks. The higher sugarcane supply came from increases in agricultural machinepricesthatbelongtothelandowner,landallocationforcassavaandsugarcaneanddecreasesinpropertyrightsandloaninterestrates.Themorecapitalresourcesthelandownerpossessed,thegreaterthelikelihoodadvancedagriculturalmachinerywasutilizedinsugarcaneproduction,whichresultedinhigheryields.Landsharesforcassavaalsoaffectedthesupplyofsugarcane,butnotlandsharesforabandonment.Itshowedthatlandownersstilldidnotconsiderallocatingabandoned landforsugarcaneproduction,even if therewasavailable land,which indicates that theabandonedplotsmusthaveaninfrastructure problem that is too difficult to correctwith the resources onhand. In contrast, greater land shares for cassava resulted in additional suppliesofsugarcanebecausebothcropsgrowwellinpoorsoilqualitywithlow rainfall and they share the same input such as agriculturalmachinery,productiontechnology,fertilizerandpesticides.
46 • Southeast Asian Journal of Economics 4(1), June 2016Ta
ble
3Parameterestimationofinputdem
and,outputsupply,andabandonedlandsh
areequations
Vari
able
Out
put S
uppl
y E
quat
ions
Inpu
t Dem
and
Equ
atio
nsS_
AB
AN
DO
NY
_RIC
EY
_SU
GA
RC
AN
EC
ASH
FER
TL
AB
OR
MA
CH
INE
PEST
CA
PITA
L
Intercept
PP_R
ICE
PP_SUGARCANE
PP_INTE
RES
T
PP_FER
T
PP_LABOR
PP_M
ACHINE
PP_PES
T
PP_C
APITA
L
S_CASSAV
A
S_SU
GARCANE
S_ABANDON
SIZE
1.194***
(0.407)
0.009
(0.009)
-0.043
(0.144)
0.015
(0.013)
-0.038
(0.009)
0.868
(0.062)
0.141
(0.009)
-1.720
(0.011)
19.724
(0.002)
-1.202***
(0.214)
-1.104***
(0.087)
-0.903***
(0.103)
0.009***
(0.002)
-39.544
(6.115)
0.071
(0.414)
1.987
(4.193)
-0.712*
(0.410)
-0.926
(0.288)
-23.990
(1.291)
15.973
(0.184)
-50.670
(0.315)
1140.013***
(0.038)
10.076***
(2.849)
28.582***
(2.256)
1.953
(3.569)
-0.073
0.049
-1.011
(0.813)
-0.001
(0.018)
0.151
(0.264)
0.007
(0.025)
-0.310*
(0.017)
2.412
(0.103)
2.175
(0.015)
5.393***
(0.021)
-49.884*
(0.003)
-0.060
(0.152)
-0.186**
(0.082)
-0.026
(0.197)
-0.005
(0.003)
-62.322*
(32.364)
-2.203***
(0.741)
-1.577
(10.628)
-1.84*
(1.001)
13.530**
(0.689)
-295.857
(4.146)
56.289
(0.614)
191.861**
(0.839)
-1496.134
(0.119)
-2.997
(5.960)
-17.045***
(3.212)
90.431***
(7.848)
0.872***
(0.130)
-176.908***
(41.317)
-2.005**
(0.948)
-7.879
(13.561)
-1.097
(1.284)
11.242
(0.882)
-77.426
(5.322)
150.873*
(0.785)
82.900
(1.072)
-2637.352*
(0.152)
-10.949
(7.767)
-12.433***
(4.159)
81.978***
(9.735)
1.294***
(0.167)
-0.941
(10.260)
0.357
(0.242)
-0.161
(3.383)
-0.305
(0.323)
-5.140**
(0.224)
-58.764
(1.362)
-8.093
(0.196)
-2.942
(0.269)
1516.601***
(0.042)
-0.347
(1.935)
-0.012
(1.051)
20.102***
(2.293)
-0.006
(0.042)
-0.635
(1.775)
0.050
(0.047)
0.538
(0.654)
-0.033
(0.063)
-0.810*
(0.045)
18.625
(0.266)
2.144
(0.038)
-5.017
(0.052)
122.864
(0.008)
-0.394
(0.379)
-0.047
(0.206)
1.640***
(0.483)
0.011
(0.008)
7.040
(669.793)
-0.103***
(0.027)
0.327
(0.375)
0.022
(0.036)
0.120
(0.026)
3.367
(0.150)
6.175***
(0.020)
2.050
(0.030)
-340.594***
(0.004)
-0.229
(0.218)
0.033
(0.113)
0.417
(0.280)
0.008*
(0.004)
-10.167
(0.910)
0.010
(0.040)
-1.457***
(0.559)
0.069
(0.051)
-0.212
(0.035)
9.652
(0.227)
2.496
(0.034)
11.088**
(0.047)
128.762**
(0.006)
- - --0.026***
(0.008)
Nararuk B., Agricultural Policy Misspecification in Thailand • 47Ta
An estimation result of the input demand equations shows that pesticide price affects the quantity of fertilizer used; while the fertilizer pricealsoaffectsthequantityofpesticideuse,whichindicatesthatfarmerstraded-offbetweenthesetwotypesofinput.Asexpected,thehigherthepriceforrice,themorethelandownerswouldinvestinfertilizer,householdlaborand agricultural machine. Technology affects all input use with negative effects,excepthouseholdlabor,whichindicatesthatbettertechnologyhelps
Nararuk B., Agricultural Policy Misspecification in Thailand • 49
todecreaseallinputuseincropproduction.Theperceptionoflandrentallawhasapositiveeffectonfertilizerandcapitalusewhichimpliesthattheparcelthat being used is encouraged by the law for both short and long-term investments.Landownersareinclinedtocultivatethelandratherthanleaseittoothersforproduction.
The estimation result from the land abandonment share equationshows the factors with the greatest effect on land abandonment to be agricultural machinery prices, followed by semi-right of property right, pesticideprice,fullrightofpropertyright,sugarcaneprices,landrentallaws,water,soilquality,parcelsize,anddistancetocity.Thedegreeofeffectfromthemarginal effect calculation presented in Table 4.5 indicates that if the capital price of the landowner is increased by one baht, there will be an additional9.151or9rai151square-wa2inlandabandonment.Thisindicatesthatthericherlandownerstendtoabandontheirlandmorequicklythanthepoorerlandowners.Duringthefieldinterviewsurvey,theselandownerswerefoundtohavetheequipmentnecessarytoaccomplishallofthetasksfortheirenterprises,buttopreferprovidingacustomhiringserviceforothersinsteadofcultivatingtheland.Theservicesrangefromcustomplanting,harvestingor hauling. The custom hiring business not only made better use of theirequipment,labor,andmanagementresourcesthroughouttheyearbutitalsoincreasedtheirincomeandprofitsfortheyearand,inmostcases,madetheaforementionedconsiderablyhigherthancropproduction.Thecoefficientofthe SEMI_RIGHT and FULL_RIGHT variable suggests that those land ownerswho have full and semi right of property right on average tend toabandonmoreoftheirplotthanthoselandownerswhohavenopropertyrightfor0.163and0.997or398.8and65.2squarewarespectively.This isalsoagreeswithGine(2005) thatapartialproperty right inThailand led to theabandonment of secured parcel. If the price of pesticides increases by one baht, there will be an additional 0.788 or 315.2 square-wa in land abandonment,whileareductioninsugarcanepricebyonebahtwillresultinanadditional0.104or41.6square-wainlandabandonment.Allofthesepricefactorsareconsistentwitheconomictheory.Relativetolandownerperception
50 • Southeast Asian Journal of Economics 4(1), June 2016
oflandrentallaws,therewillbeanadditional0.027or10.8square-wainlandabandonmentifthelawsareperceived.Relativetoparcelsize,iftheparcelsizeissmallerbyonerai,therewillbeanadditional0.002or3.2square-metersin land abandonment. This is due to the production inefficiency of small parcelsasreportedbySimon(1987)andMcClosekey(1975).ThecoefficientoftheIRRIGATIONvariablesuggeststhatthoselandownerswhohavethemainwatersourcesupplyfromirrigationsystemonaveragetendtoabandonlessoftheirplotthanthoselandownerswhosemainsourceofwatersupplyarerainfor0.022or8.8square-wawhilethoselandownerswhohavethehighqualityofsoilonaveragetendtoabandonlessoftheirplotthanthoselandownerswhose soil quality aremedium for 0.018 or 7.2 square-wa.Closerdistancetoacitybyonekilometerwillgenerateanadditional0.002or3.2squaremetersinlandabandonment.Thisoccursbecausetheeconomicrentandproductivityoftheparcelsnearurbanareasishigherfornon-agriculturalactivities.ThesummaryofsignandsizeeffectofsignificantvariablesonlandabandonmentsharearereportedinTable5.
Nararuk B., Agricultural Policy Misspecifi cation in Thailand • 51
3.2 Marginal eff ect scenario test
Themarginaleffectscenariotestingonthestructuralmodelprovidesanexpectedvalueforthedependentvariablewhichislandshareforabandon-mentfordifferentvaluesofsignificantvariables.Thepurposeofthistestwastofindaturningpointoflandsharesforabandonment,orazerovalueoflandsharesforabandonment.Thepresentstudytestedthefollowingtwovariables:distancetocityandlandsize.Thescenariovaluesforlandsizearesetbasedontheproposedmeasuretolimittheholdingsizeofagriculturallandat50rai.Thescenariovalues for thedistance tocityvariableweresetbasedon themaximumdistanceof theparcelsamplefromtheKhon-Kaenmunicipality,whichis144km.Therefore,thescenariovaluesforbothvariablesweresetas:landsize:0,25,50,75,100,125anddistancetocity:0,25,50,75,100,125,150.
Themarginaleffectscenariotestingondistancetocityandparcelsizeispresented inFigure2.The testshows thatagricultural landwillbefullyutilized at approximately 125 kilometers from municipality and the mosteffective land size for agricultural production at approximately 100 rai atwhichpointtherewillbenoagriculturallandabandonment.
Figure 2:MarginalEffectTesting
52 • Southeast Asian Journal of Economics 4(1), June 2016
3.3 Model Validity Test
Theproblemsofheteroscedasticityandnon-normality,whichleadtoinefficiency andunbiasedparameter estimates, are critical inTobit setting.Bothproblemsweretestedinthepresentstudy.TheoverallsystemHeterosce-dasticitytestwiththeBreusch-PaganLMTestindicatednosystemheterosce-dasticity.Doornik-Hansenwasusedfornon-normalitytestingunderthenullof non-normality and failed to accept the null, thereby indicating that thestructuresystemisunbiased.
Inadditiontotheestimationapproach,adirectquestiononwhytheland owners decide to abandon their land was asked directly. The resultsshowedthatwater,soilquality,andagriculturalmachineareamongthehighestfrequency answer which is complying with the result from our structural estimation.Laborshortageisalsooneoftheimportantfactorsforagriculturallandownertoabandontheirparcelfor13.72%.Interestingly,12.83%percentofthesamplesintendtokeeptheparcelforspeculation,thereforetheyceasethecropproductionwhilewaitingforitspricetorise.Wefoundthat0.66%haveotherreasonstoabandontheirlandthatisnotlistedinourchoices.Theyspecifythattheparcelbelongtothemaresaturatedallyear;thereforetheyareunabletocultivatetheirland.Thisisanotherproblemrelatedtophysicalandinfrastructureofparcel.
4. Discussion and policy implication
This study challenges key agricultural policies in Thailand. It has successfully constructed and applied an economic model to help answer agricultural policy issues. The outcome showed that the factors with the greatest effect on land abandonment are agricultural machinery prices, followedbysemi-rightofpropertyright,pesticideprice,fullrightofpropertyright,sugarcaneprices,landrentallaws,water,soilquality,parcelsize,anddistancetocity.Inlightoftheaforementionedresults,thisstudymakesthefollowingpolicyimplications:
First, Relative to land infrastructure development policies such as irrigationsystemandsoilimprovementprograms,bothpoliciesaretherightmeasure for theThai government to adopt.Although, financial limitations
Nararuk B., Agricultural Policy Misspecification in Thailand • 53
havehinderedthewidespreadconstructionofirrigationsystemsthroughoutThailand, the shortage can be solved by introducing a value capture tax. Landownerswhoenjoyandbenefitfromagovernment-builtirrigationsystemshould return a portion of their increased profits to the government to maintainthesystemandbuildadditionalirrigationsystems.Relativetothesoilimprovementprogram:Thegovernmenthasdoneagoodjobbyestablishinganddevelopingvolunteersoildoctorswithlearningcentersatthesub-districtlevel,whichcanbeaccomplishedinacost-effectivemanner.However,theyneedtobepromotedmorebecausemanyoftheagriculturalareascontinuetofacesoilqualityissues.
Third,With the landallocationprogram, the fullandsemi rightofpropertyrightfactorispositivelysignificant,whichsuggeststhatthoselandownerswho have full and semi right of property right on average tend toabandonmore of their plot than those land ownerswho have no propertyright.Theresultimpliedthattheeffectivenessofthelandallocationprogramisinconclusive.
Fourth,Withthelandrentallaw,theresultsuggestedreformationoftheAgriculturalLandRentalActof1981.Theaforementionedlawfavorsthetenantsandtherebyencourageslandabandonment.Itwaseffectiveinthepastunder a different social context, but is currently out-of-date and shouldbe revised.Revisingthelawwillnotonlyreducethelandabandonmentproblem,but also encourage land use at its maximum capacity through the leasing process.Fifth,Withtheadversepossessionlaw,theperceptionofthelawisinsignificant,whichsuggeststhatthelawisaninconclusivepolicy.
Sixth,Agricultural-zoningis therightpolicyfor thegovernment toadopt,butshouldbebaseduponause-based,notcrop-basedformat.Aradiusof 125 kilometers from municipality is proposed for the province of Khon-Kaen.Theparcelsnearthecityshouldbeconvertedtootherusesmore
54 • Southeast Asian Journal of Economics 4(1), June 2016
Eighth, With the price subsidy program, although the significant factors related to this policy are pesticide and sugarcane prices. Only the inputprice subsidy;pesticideprices canbe implemented,butwith cautionbecausegovernment interventiononoutputpricedestroys themarketpricemechanismandithasbeenproveninotherpartsoftheworldthatinterveningoutput price exacerbates the situation. Therefore, the output price subsidyshouldbecancelled.
Arnade, C., & Kelch, D. (2007). Estimation of area elasticities from a standardprofitfunction.American Journal of Agricultural Economics,89(3),727-737.
Baumann,M.,Kuemmerle,T.,Elbakidze,M.,Ozdogan,M.,Radeloff,V.C.,Keuler,N. S., ... & Hostert, P. (2011). Patterns and drivers of post-socialistfarmlandabandonmentinWesternUkraine.Land Use Policy,28(3),552-562.
Beattie, B. R., Taylor, C. R., & Watts, M. J. (1985). The economics of production(No.338.5B369).NewYork:Wiley.
Nararuk B., Agricultural Policy Misspecification in Thailand • 55
Benayas,J.R.,Martins,A.,Nicolau,J.M.,&Schulz,J.J.(2007).Abandon-mentofagriculturalland:anoverviewofdriversandconsequences.CAB reviews: perspectives in agriculture, veterinary science, nutrition and natural resources,2(057),1-14.
Chambers,R.G.,&Just,R.E.(1989).Estimatingmultioutputtechnologies.American Journal of Agricultural Economics,71(4),980-995.
Cleary,M.,&Eaton,P.(1996).Tradition and reform: land tenure and rural development in South-East Asia(pp.80-83).KualaLumpur:OxfordUniversityPress.
Cochran,W.G.(1977).Samplingtechniques-3.Färe, R.,Martins-Filho, C., &Vardanyan,M. (2010). On functional form
representation of multi-output production technologies. Journal of Productivity Analysis,33(2),81-96.
Fezzi,C.,&Bateman,I.J.(2011).Structuralagriculturallandusemodelingforspatialagro-environmentalpolicyanalysis.American Journal of Agricultural Economics,93(4),1168-1188.
Gine,X.(2005).Land security in rural thailand: Evidence from a property rights reform (Vol. 3734). World Bank, Development ResearchGroup,FinanceTeam.
Green,W.(2003).H.Econometricanalysis.LandInstituteFoundation.(2001).Theprojecttostudylandholdings, land
utilization, economicmechanisms and laws for optimally efficientlanduse.Bangkok,LandInstituteFoundation.
Makkarapirom, P. (2011, March). Land Reform for the Equality. Paper presentedat theseminaron landreformfor theequality,Bangkok,ChulabhornResearchInstitute.
McCloskey, D. N. (1975). The persistence of the English common fields.Princeton:PrincetonUniversityPress.
Müller,D.,&Munroe,D.K.(2008).ChangingrurallandscapesinAlbania:Cropland abandonment and forest clearing in the postsocialist transition.Annals of the Association of American Geographers,98(4),855-876.
56 • Southeast Asian Journal of Economics 4(1), June 2016
Müller,D.,Kuemmerle,T.,Rusu,M.,&Griffiths,P.(2009).Lostintransition:determinants of post-socialist cropland abandonment in Romania.Journal of Land Use Science,4(1-2),109-129.
Nishihara,M.(2012).Realoptionvaluationofabandonedfarmland.Review of Financial Economics,21(4),188-192.
Prishchepov,A.V.,Radeloff,V.C.,Baumann,M.,Kuemmerle,T.,&Muller,D.(2012).Effectsof institutionalchangeson landuse:agriculturalland abandonment during the transition from state-command to market-driven economies in post-Soviet Eastern Europe. Environ-mental Research Letter,7(2),024021.
RoyalIrrigationDepartment.(2011).IrrigationStrategicPlanningyear2013-2016. Retrieved August 31, 2015, from water.rid.go.th/hydhome/document/2557/Plan_2557-2558.pdf.
Shumway,C.R.&Alexander,W.P. (1998).Agricultural product supplies and input demands: regional comparisons. American Journal of Agricultural Economics,70(1),153-161.
Shumway,C.R.,Saez,R.R.,&Gottret,P.E.(1988).Multiproductsupplyandinput demand inUSagriculture.American Journal of Agricultural Economics,70(2),330-337.
Nararuk B., Agricultural Policy Misspecification in Thailand • 57
Simon,S.(1987).Land fragmentation and consolidation: a theoretical model of land configuration with an empirical analysis of fragmentation in Thailand.
The official information center, Khon-Kaen Governer ‘s office. (2012). Khon-Kaen ‘s Agricultural land information. Retrieved February 25,2013, fromwww.khonkaenpoc.com/khonkaen/admin/.../file/tot_land_47.xls
Tambunlertchai, S. (2004). The Government’s Helping Hand: A Study of Thailand’s Agricultural Debt Moratorium.
Tobin,J.(1958).Estimationofrelationshipsforlimiteddependentvariables.Econometrica:Journal of the Econometric Society,24-36.
UNICEF. (1995).MonitoringProgress toward thegoalsof theWorld food Summit for Children: a practical handbook for multiple indicator surveys.New York.
Villezca-Becerra,P.A.,&Shumway,C.R.(1992).Multiple-outputproductionmodeled with three functional forms. Journal of Agricultural and Resorce Economics,13-28.
Wooldridge,J.M.(2002).Econometric analysis of cross section and panel data.MITPress.
Zaragozí, B., Rabasa,A., Rodríguez-Sala, J. J., Navarro, J. T., Belda,A., & Ramón,A. (2012). Modelling farmland abandonment:A study combiningGISanddataminingtechniques.Agriculture, ecosystems & environment,155,124-132.