National Graduate Institute for Policy Studies Global Center of Excellence Economics Working Paper No. 6 April 2009

RURAL POVERTY AND INCOME DYNAMICS IN SOUTHEAST JonnaP.Estudillo a KeijiroOtsuka b Abstract ManyruralhouseholdsinAsiahavebeenabletomoveoutofpovertyinthepresenceof increasingscarcityoffarmlandinitiallybyincreasingincomethroughtheadoptionof modernricetechnologyandgraduallydiversifyingtheirincomesourcesawayfromfarm tononfarmactivities.Increasedparticipationinnonfarmemploymenthasbeenmore pronouncedamongthemoreeducatedchildren,whoseeducationisfacilitatedbyan increaseinfarmincomebroughtaboutbythespreadofmodernricetechnology.An importantlessonforpovertyreductionistoincreaseagriculturalproductivitythroughthe developmentandadoptionofmoderntechnology,whichsubsequentlystimulatesthe developmentofthenonfarmsector,therebyprovidingemploymentopportunitiesforthe rurallaborforce.Thischapterexploresthekeyprocessesoflongtermpovertyreduction inSoutheastAsiausingtheandascasestudies. Key words GreenRevolution,poverty,nonfarmemployment,childschooling JELclassification: O12,O15,O53,Q12,Q15 aFoundationforAdvancedStudiesonInternationalDevelopment,7221Roppongi, Minatoku,1068677,,phone+81354136038,fax+81354130016, email: [email protected] . bFoundationforAdvancedStudiesonInternationalDevelopment,7221Roppongi, Minatoku,Tokyo1068677,Japan,phone+81354136035,fax+81354130016, email: [email protected] Acknowledgment:TheauthorsthankPrabhuPingaliforhiscommentsonanearlierdraft ofthispaper.Theusualcaveatapplies. RURAL POVERTY AND INCOME DYNAMICS

IN 1

1. Introduction

Therewasabelieffromthe1950stothe1970sthathighpopulationpressureonclosed landfrontierwouldresultinhighincidenceofruralpoverty,foodshortages,andeven widespreadfamineinSoutheastAsiaandSouthAsia.Highpopulationpressureto adeclineinthesizeoffarmlandandanincreaseintheincidenceoflandlessness,even thoughfarmlandisamajorsourceofincomeofruralhouseholdsintheearlystageof development(EstudilloandOtsuka,1999;HayamiandKikuchi,2000;Hazelland

Haggblade,1991;Lanjouw,2007).Indeed,theincidenceofpovertyisobservedtobe higheramongthelandpoorandlandlesshouseholdsthanamongthefarmerhouseholds

(WorldBank,2008a;Estudillo et al .,2008;Hossain et al .,2009).Thedirectimpactsof theGreenRevolution,asexemplifiedbytheadoptionofmodernricetechnology,on employmentopportunitiesforthepooragriculturallandlessandnearlandlesspopulation seemtobemodest(LiptonandLonghurst,1989).Demandforagriculturallaboris seasonalandtherehasbeenanincreasingtrendintheadoptionoflaborsaving technologies(JayasuriyaandShand,1986).ThemajordirectimpactoftheGreen

Revolutioncomesmainlythroughanincreaseinriceproduction,attributabletoyield increaseandtheshortergrowingperiodthatsignificantlyreducedriceprices,thereby increasingthewelfareofthepoorasconsumers(BarkerandHerdt,1985;Davidand

Otsuka,1994). 1ThischapterisasynthesisofSawada et al .(2009),Estudillo et al .(2009a),TakahashiandOtsuka(2009), Cherdchuchai et al. (2009),andOtsuka et al .(2009b),whichareChapters14and9ofthebook Rural Poverty and Income Dynamics in Asia and Africa (Otsukaet al .,2009a).

2 Yet,weobserveaclearandremarkablemovementofruralhouseholdsoutof povertyinSoutheastandSouthAsiainthemidstoftheunfavorablescenarioof increasingscarcityoffarmlandanddeclininglaboremploymentopportunitiesinthefarm sector.AccordingtotheAsianDevelopmentBank(2008),theproportionofpopulation livingonlessthantheAsianpovertylineofUS$1.35perdaydeclinedby5.9percentage pointsinthePhilippines,by7.8percentagepointsinThailand,by21.4percentagepoints in,andby47.7percentagepointsinVietNamfromtheearly1990stothemid

2000s.Interestingly,incomegrowthand,consequently,povertyreduction,havebecome evidentinlandscarceregionsofsubSaharanAfrica,wherelandisonceconsidereda relativelyabundantresource(Otsuka et al .,2009a).ThismayindicatethatAfricanrural householdshavebeenexperiencingthesamepatternofstructuralchangeandtracking similarpathwaysoutofpovertythatruralhouseholdsintropicalAsiahaveexperienced inthepast20–25years.Animportantissueistoidentifythestrategicprocessesby whichruralpovertyhasbeendeclininginAsia,whichservesasalesson,notonlytosub

SaharanAfrica,buttootherdevelopingcountriesaswell.

Wefoundthattheriseinnonfarmincomeisthemajordriverbehindpoverty reduction,whichwasfacilitatedbyearlierdecisionofhouseholdstoinvestinchildren’s schoolingmadepossiblebytheincreaseinfarmincomebroughtaboutbytheGreen

Revolution.Thisisthefirststudytoourknowledgethatexaminesthestructural transformationofruralfromfarmtononfarmactivitiesbyexploringthe causalmechanismsthatlinkagriculturalproductivitygrowthwithhumancapital investments,thedevelopmentofnonfarmsector,andpovertyreduction.2Weselected

2Attheaggregatelevel,RosegrantandHazell(2000)foundthatAsiancountriesthatgrewtheearliestand fastestarethosethatexperiencedrapidagriculturalgrowthintheearlystagesofgrowth.Thisgrowthwas

3 fourcountriesinSoutheastAsiathePhilippines,Thailand,Indonesia,and

Vietnamwherepovertyreductionhasbeenremarkableandthenwefocusondynamic changesinincomestructureandcompositionofrurallaborforceinselectedvillagesin thePhilippinesandThailand,wheretheGreenRevolutiontookplace.

Thischapterhasfiveremainingsections.Section2presentsanoverviewofthe structuraltransformationoftheawayfromfarminginfourcountriesin

SoutheastAsia.Section3discussestheconceptualframeworkandpostulatesbasic hypotheses.Section4describesthedatasetinthePhilippinesandThailand.Section5 identifiesthedeterminantsofhouseholdincome,investmentsinchildren’sschooling, occupationalchoiceofchildren,andnonfarmincome.Finally,Section6presentsthe summaryandconclusions.

2. Economic transformation in Southeast Asia

Ifrurallaborforceincreasesunderthescenarioofclosedlandfrontierandstagnant technology,wecanexpectadecreaseinthemarginalproductivityoflabor,whichleads toadecreaseinincomeandriseintheincidenceofpoverty.Thisisseeminglythecase inSoutheastAsia,wherethelandfrontierhadbeenclosedinthe1960sand1970sand populationgrewatanannualgrowthrateofwellmorethan2%inthesameperiod.

Unexpectedly,however,povertyincidencehasdeclinedinthesecountries,alongwiththe structuralshiftoftheeconomyawayfromagriculturetoindustryandservices,asshown broadbased,benefitingbothsmallandmediumsizedfarmsandthisgrowthwasmadepossiblebyan equitabledistributionofland.Strongagriculturalgrowthinthesecountriesisbasedonrapidgrowthon inputuseandproductivitygrowth.Themainsourcesofproductivitygrowthhavebeenpublicagricultural researchandextension,expansionofirrigatedareaandruralinfrastructure,andimprovementinhuman capital.

4 bythedeclineintheproportionofgrossdomesticproduct(GDP)comingfrom agriculture.InSoutheastAsiancountries,wefoundthattheservicesectorhasbeenthe dominantsectorinthePhilippines,whereasindustryhasbecomeimportantinThailand,

Indonesiaand(Table1).

Vietnamhasshownthemostdramaticshiftofitseconomicactivitiestoward industryandawayfromagricultureandservicesectors.Simultaneouswiththeswift transformationisthemarkeddeclineintheincidenceofpoverty,byasmuchas48 percentagepointsfrom1993to2004.Asaresult,theincidenceofpovertyin2004.has becomelowerinVietnamcomparedwiththatinthePhilippinesandIndonesia,which startedatamuchlowerincidenceofpovertyintheearly1990s.Itisbynowwellknown thatdirectparticipationinthelabormarketinthenonfarmsectorinindustryandservices isthemostimportantroutetoupwardincomemobilityandanescapefrompovertyfora largemajorityoftheruralpoor(HayamiandKikuchi,2000;Lanjouw,2007;Estudillo et al. ,2008).

EmploymentstructureinThailand,Indonesia,andVietnamshowsthatagriculture remainsthelargestemployerofbothmaleandfemalelabor.Incontrast,femalesinthe

Philippinesarelargelyemployedintheservicesector(i.e.,55%ofthefemalelaborforce intheearly1990s)andincreasinglysoinmorerecentyears(i.e.,64%inthemid2000s) whilemalesremainlargelyinagriculture(AsianDevelopmentBank,2008).Similarly,in

ThailandandIndonesia,femalesinthelaborforcehavebeenincreasinglyflockingthe servicesectorwhilemaleshavebeenmovingoutofagriculturetoindustry.Accordingto

Momsen(2004)femalesintheworldatlargehavebeenmovingoutofagriculturefaster thanmentotheindustrysector,initiallyfromthe1960stothe1980s,and,finally,tothe

5 servicesectorfromthe1990s.IncreasedinvolvementofSoutheastAsianfemalesinthe industrycoincideswiththemovementofproductionbaseoflaborintensive,low technologyproductsawayfrom,,andHongkongtoSoutheastAsia,when theseEastAsiancountriesshiftinamajorwaytomoresophisticatedproducts correspondingtotheirsharpwageincreasesinthelate1970stotheearly1980s.

Meanwhile,agriculturalproductivityinthesecountriesbeganrisingbeforethe structuraltransformationoftheentireeconomy,asexemplifiedinriceyieldincrease, attributedtothedevelopmentandadoptionofmodernricevarieties(MV)inthe1970s and1980s(Figure1).InthePhilippines,MVadoptionandthesubsequentyieldincrease canbeobservedfromtheearly1970stothemid1980sduringwhichstructural transformationhadbeenonlyslowlytakingplace.Thesamepatternholdstruein

Indonesia,wherericeyieldishigherthaninthePhilippinesbecauseofitsfavorable agroclimaticconditions(largelyfreefromtyphoons).Amoredramatictransformationin

Thailand’seconomytookplaceeventhoughtheadoptionofMVsandriceyieldgrowth arelower.TraditionalThairicehasbettergrainqualityandcommandsahigherpricein theinternationalmarket,thusitoccupiesalargershareofthecountry’sricearea.MVs wereintroducedinVietnaminthemid1970sandMVadoptionquicklyreached80%of thetotalriceareaintheearly1990s.Dramaticyieldincreasehasbeenobservedsincethe early1980s,reachingabout5tonsperhainthemid2000s.Vietnamhasbecomethe world’ssecondlargestriceexporter,nextonlytoThailand,andmuchoftherural populacedependsonriceproduction.Itisreasonabletoassumethattheboomingrice

6 sectorisoneofthemajorpropellingforcesinincomegrowthandpovertyreductionin ruralVietnam. 3

WhileitisclearthatstructuraltransformationinSoutheastAsiancountrieshas subsequentlyledtopovertyreduction,thestrategicprocessesbywhichagricultural growthinearlieryearshastriggeredthesubsequenttransformationoftheruraleconomies havenotbeenidentified.Wefocusonnewtechnologyinthericesectorbecausericeis largelyproducedinownercultivatedfarms,whichcompriseabout80%ofthetotal numberoffarmsinAsia(Otsuka,2007). 4Sincetheproductionpossibilityfrontiershifts outwardwithmodernagriculturaltechnology,thedevelopmentofthenonfarmsectorwill bestimulated,giventherisingincomeandhighincomeelasticityofdemandfornonfarm products.Thedevelopmentofagricultureislikelytostimulatethedevelopmentofthe ruralnonfarmsectorthroughtheconsumptionandproductionlinkages(Haggblade et al. ,

2007).Empiricalevidenceonthemagnitudeoftheseeffects,however,remainsscanty.

We,therefore,postulatethefollowingsequelofeventsinthetransformation:Green

Revolution higherfarmincome largerinvestmentsinschoolingofchildren  supplyofeducatedlaborforcetothenonfarmsector higherincomeofchildrenand povertyreduction furtherdevelopmentofthenonfarmsector.

3HouseholdlevelsurveysinmajorriceproducingareasinnorthernandsouthernVietnamrevealthat modernricetechnologysignificantlyincreasesriceproductionincomeandtotalhouseholdincomeinareas withwelldevelopedirrigationsystem,floodcontrolanddrainage,andtransportationandcommunication facilities,asruralinfrastructureisanecessaryconditionforanefficientricemarketingsystem(Ut et al., 2000) 4Hayami(2001)tracesthetrajectoriesofdevelopmentperformanceofIndonesia,thePhilippines,and Thailandthroughthecountry’secologicalconditionsandcolonialhistory.InIndonesia,ruralcommunities werebifurcatedintoricefarmingpeasantproprietorsandlargeplantationsfortropicalexportcropsbased onhiredlaborduringtheDutchcolonialperiodwhenlargescaleexploitationoftropicalrainforeststook place.InthePhilippines,exploitationofthesameresourcebaseunderSpanishruleresultedinpervasive landlessness.InThailand,landowningpeasantscontinuedtodominatebecausethedeltaplainsthatformed theresourcebasefordevelopmentweremainlysuitableforriceproduction.

7

3. Conceptual framework and hypotheses

ClassicstudiesonthedualeconomicmodelpioneeredbyLewis(1954)anddeveloped furtherbyRanisandFe(1961),Jorgenson(1961),andHarrisandTodaro(1970)focuson theprocessoflaborreallocationawayfromthelowproductivityfarmsectortoahigh productivitynonfarmsector.Theyarguethatcapitalinvestmentinmodernurbansector istheleadingfactorthatpromotesthedevelopmentofoveralleconomies,withoutmuch regardtotheroleoftechnologicalchangeinagriculture(HayamiandGodo,2005).In reality,however,significanttechnologicalchangehasbeentakingplaceinAsian agriculture.Wearguethatthemainfactorbehindthestructuraltransformationisthe

GreenRevolutionthatincreasesthemarginalproductsoflaborandpurchasedinputs, suchaschemicalfertilizer,inagriculture.AsillustratedinFigure1,wehypothesizethat theGreenRevolutionsignificantlyincreasesfarmincomethroughhigherriceyieldand highercroppingintensityattributedtoshortergrowingperiodandnonphotoperiod sensitivityofMVs.Farmhouseholdincomeisdeterminedlargelybyagricultural technologyandhouseholdaccesstofarmland,whilethecontributionofhumancapital maynotbelargeunlessamodernanddynamictechnologyissuccessivelyintroduced, whichischaracterizedbyachangingandcomplexoptimuminputmixandmanagement technique(Schultz,1975).Wepostulatethefirsthypothesis:

8 Hypothesis 1: In the early stage of economic development, when farming is a

dominant source of income, access to land and agricultural technology are the

major determinants of farm household income.

Astheeconomydevelops,theavailabilityofjobsintheruralnonfarmsector,urbanlabor markets,andoverseasmarketsincreasespartlybecauseofthedeclineinfoodprices, whichareconsideredwagegoodsfortheurbanworkers(HayamiandGodo,2005),and partlybytheproductionandconsumptionlinkagesbroughtaboutbytheGreen

Revolution(Figure1).Increaseddemandforlaborleadstotheriseinreturnstoboth

“quality”and“quantity”ofhumancapital,whichmeansthatthewagesofboththe educatedanduneducatedlaborforcehaverisen.Thus,wepredicttheincreasing importanceofhumancapitalandthedecreasingimportanceoflandassourcesof householdincomeofruralhouseholds.

Hypothesis 2: As the economy develops, the availability of nonfarm jobs increases

so as to increase the returns to both the “quality” and “quantity” of human

capital. Consequently, the development of the nonfarm sector leads to an increase

in nonfarm income of rural hoiuseholds and a major reduction in rural poverty.

Povertytendstobehigherinareascharacterizedbyunfavorableagriculturalconditions susceptibletodroughtsandfloodingwithpooraccesstomarkets.Increasingavailability ofnonfarmjobsisthesinglemostimportantfactorinincreasingruralhouseholdincome insuchunfavorableareassolongasthenonfarmlabormarketsareregionallyintegrated.

9 Thelandlesshouseholds,whichdependedprimarilyonagriculturallaboremployment, alsobenefitfromtheexpandednonfarmemploymentopportunities.Totheextentthat theincomeoflandlesshouseholdsincreasesfasterthanthatoflandedhouseholdsandthe incomeinunfavorableareasgrowfasterthaninfavorableareas,wecanreasonably assumethatthestructuraltransformationofruraleconomiesispropoor.

Hypothesis 3: While in the early stage of economic development, income was

lower and poverty incidence was higher in unfavorable areas, household income,

particularly nonfarm income, grew more rapidly and the incidence of poverty

declined more sharply in unfavorable areas than in favorable areas.

Withthedevelopmentofthenonfarmsectorandthesubsequentincreasesinreturnsto schooling,ruralhouseholdstendtoinvestinschoolingofchildrenbecausethiscan generateincreasednonfarmincomeandremittancesinthelongrun.Weexpectthatfarm incomeisthemoreimportantdeterminantofinvestmentsinschoolingintheearlystage ofdevelopmentbecausetheparents,whoareinterestedininvestinginchildren,earn substantialportionoftheirincomefromfarming.Thus,wepostulatethefollowing hypotheses:

Hypothesis 4: In the early stage of economic development, when farming is a

dominant source of income, farm income is a major determinant of schooling

investments in children of rural households.

10 Hypothesis 5: Educated workers tend to find lucrative jobs in the nonfarm sector,

where returns to schooling are higher.

Ifhypotheses4and5areempiricallysupported,inalllikelihood,theGreenRevolution contributedtopovertyreductionbyinducinginvestmentsinchildren’sschoolingand increasingtheavailabilityofeducatedlaborforcetothenonfarmsector.

4. Description of the data sets

Toexaminetheempiricalvalidityofhypotheses1to5,weanalyzethedetailed householdpaneldatacollectedinthePhilippinesandThailandinthelastfewdecades.

Thissectiondescribesthebasiccharacteristicsofthedatasets.

TherearetwodatasetsfromthePhilippines.Thefirstonecomesfromasurvey ofhouseholdslocatedalongaloopofthenationalhighwaysintheCentralLuzonregion

(henceforthreferredtoas“CentralLuzon”).ThesamplesetinCentralLuzonconsistsof essentiallyidentical126householdsinboth1979and2004,inadditiontoasurveyof499 grownupchildrenin2004.Thissamplesetconsistsoffarminghouseholdswiththe exclusionoflandlesshouseholds.Farmerhouseholdsarethosewhichoperatefarmland, includingownercultivators,leaseholdtenants,andsharetenants.Landlesshouseholds arethosewhichdonothaveafarmtooperate,includingtheagriculturalworkers,who ekeoutalivingoncasualfarmwork,andthenonagriculturalhouseholdswhose livelihooddependspurelyonnonagriculturalwork.ThesamplesetinCentralLuzon representstherelativelyhighincomefarminghouseholdsthathaveeasyaccessto

11 marketsandagriculturalextensionservicesowingtotheirproximitytomajorhighways connectingCentralLuzontoMetro.

Theseconddatasetcomesfromthesurveysof447householdsconsistingofboth farmerandlandlesshouseholdsthatwererandomlyselectedintwovillageseachin

NuevaEcijaintheCentralLuzonregionandIloiloProvinceinPanayIslandinsouthern

Philippines(henceforthreferredtoas“NuevaEcijaandIloilo”),whichwereinterviewed repeatedlyin1985,1992,1997,2001,and2004.Thesevillageswereselectedfroman extensivesurveyof50villagesrepresentingirrigatedandrainfedlowlandriceproduction environmentsinnorthern,central,andsouthernLuzon(DavidandOtsuka,1994).This datasetinthePhilippinescapturestheintricaciesoflandtenuresystem,includingthe subtenancyarrangementsthatemergedbecauseoftheregulationsofthelandreformlaws.

InThailand,295householdsinthreevillageseachintheCentralPlain,whichis relativelyaffluentandlocatednear,andnortheastprovinces,whichareknown tobemuchpoorerandlocatednearthelocalcityofKhonKaen,weresurveyedin1987 and2004.Onevillageineachprovincerepresentsoneofthethreetypicalrice productionenvironmentsinThailand.Twovillagesrepresentthefavorableenvironment owingtotheexistenceofthegravityirrigationsystem,twovillagesrepresentthe favorablerainfedenvironment,andanothertworepresenttheunfavorableenvironment, includingafloodproneintheCentralPlainandadroughtproneareainthenortheast.

NotethatalthoughMVadoptionratesaregenerallylowinThailand(Figure1),theyare relativelyhighinourstudysites(i.e.,54%oftheareaplantedintheCentralPlainand

91%inthenortheastin2004).Landlesshouseholdswerenotincludedin1987because therewereveryfewofthem,althoughitsshareincreasedto13%ofthetotalhousehold

12 populationin2004.Aninterestingphenomenoninthesevillagesisthetendencyof householdstousepumpirrigationforricefarmingandtoventureintohighvaluecrop productioninrecentyears.Moreimportantly,householdshavebecomeincreasingly mobiletolocalcitiesaswellasBangkoktoengageinregularnonfarmwork,whichhave becomemorenumerousin2004.

Allthethreedatasetsprovidedinvaluableinformationonthechangingsourcesof householdincomeandthefactorsaffectingrice,nonrice,andnonfarmincomeaswellas thedeterminantsofhouseholds’investmentsinschoolingofchildren.TheCentralLuzon andThaidatasetswereusedfurthertoexaminethefactorsaffectingoccupationalchoice ofgrownupchildrenandtheirnonfarmincome.BecauseofthepanelnatureofNueva

EcijaandIloiloandThaidatasets,whichincludedbothfarmerandlandlesshouseholds, wemeasuredtheextentoflandlessnessandtracedthemovementsofhouseholdsinand outofpoverty.Overall,theselongtermdatasetsservedasthebasefromwhichwewere abletosystematicallyanalyzethelongtermprocessesofpovertyreductioninthecontext ofSoutheastAsia.

5. Household income, children’s schooling, and occupational choice

Totracethelongtermprocessesofpovertyreduction,wedescribethechangingsources ofhouseholdincome,patterninprogressofchildreninschool,andchoiceofoccupation ofchildren.

5.1 Farm size and landlessness

13

Table2showsalargereductioninthesizeofoperationallandholdingsoffarm householdsandanincreaseintheproportionoflandlesshouseholdsfromthelate1980s totheearly2000s.InthePhilippines,theaverageoperationalfarmsizeofsample householdswas1.0hain1985,butitdecreasedto0.76hain2004.Theproportionof landlesshouseholdsrosefrom22%to44%inthesameperiod.Amoredrasticreduction infarmsizewasfoundinThailandandtheproportionoflandlesshouseholdsrosefrom nilin1987to13%in2004.Surprisingly,aswewillshowlater,therewasamovementof householdsawayfrompovertyinthemidstofincreasingscarcityoffarmlandand declineinlaboremploymentopportunitiesinricefarmingduetomechanizationand wideradoptionoflaborsavingdirectseedingmethodreplacinglaborusingtransplanting

(Otsuka,Asano,andGascon,1994).

5.2 Household income sources and poverty

Wedividedthevillagesintofavorable(orirrigatedareas)andunfavorablerainfedareas susceptibletodroughtorflooding.Unfavorableareasinthemid1980saregenerally characterizedbylowerincomeandhigherincidenceofpovertyowingtopoorproduction environmentandpooraccesstomarkets.Householdsderivetheirincomefrom(1) agriculturalwages,(2)ricefarming,(3)propagationoflivestockandpoultryand productionofnonricecrops,and(4)nonfarmactivitiesincludingnonfarmwages, remittancesandgifts,andincomefromoperatingownbusiness.Wehaveincludedthe imputedvalueofhomeconsumedproductsinbothriceandnonriceincome.

14 WehaveafewimportantobservationsfromTable2,whichshowsthechanging sourcesofhouseholdincomeinthe1980sandearly2000s.First,theimportanceoffarm incomewasmuchhigherinthe1980sthanintheearly2000s.Riceincomewas particularlyimportantinfavorableareasinthe1980s,yetitscontributiondeclinedinlater yearsimportantlybecauseofthedeclineinricepricescoupledwithonlyamodest increaseinyield.InthePhilippines,riceincomecomprised37%ofthetotalhousehold incomebutdeclinedto12%;inThailand,itwas66%butdeclinedto26%.Thisindicates thedeclineintheimportanceoffarmlandasasourceofhouseholdincomeofrural households.

Second,theproportionofincomefromagriculturalwagehasdeclined(exceptin thefavorableareasinThailand,whereitsshareisrelativelysmall)andthedeclinewas spectacularintheunfavorableareas.Thisisexplainedtoalargeextentbythedeclinein labordemandinricefarming,broughtaboutbytheaccelerationintheuseoflaborsaving technologies,decreaseinriceprices,andstagnantproductivity.Inasmuchashiredlabor wassuppliedbythelandlesshouseholdsandmarginalfarmers,wecanfairlypresume thatrelianceonagriculturallabormarketaloneisnotsufficienttopromoteincome growthandpovertyreduction(OtsukaandYamano,2006).

Third,thecontributionofnonriceincomehasriseninunfavorableareasinthe

Philippines,indicatingthattheagriculturalsystemsintheseareashavemovedawayfrom thedominanceofriceproductiontoamorediversifiedsystemwithincreasedimportance ofhighvaluecropsandlivestock.Yet,theincreaseinthenonricefarmincomewas hardlythedriverofincomegrowthbecauseofitsrelativelysmallshareintotalhousehold income.

15 Fourth,theshareofnonfarmincomehasincreaseddramaticallyinallareasalong withasharpriseinpercapitaincome.Themoredrasticchangewasobservedinthe unfavorableareasinThailand,wherethenonfarmincomesharerosefrom21%in1987 to74%in2004andpercapitaincomeroseby2.7times.Thisremarkabletransformation wasfacilitatedbytheavailabilityofnonfarmwageemploymentinlocaldistrictsof

BangkokandKhonKaen.Traditionally,farmersinthisregionmigratetowestern regionstoengageinlowwageemploymentinsugarcanecuttingbut,inmorerecentyears, theselowwagejobswerereplacedbyhighwagejobsinurbanareas.InthePhilippines, theincreaseinnonfarmincomesharewaspartlybroughtaboutbytheinfluxofdomestic andforeignremittances,theincomeshareofwhichrosefrom11%in1985to17%in

2004.SawadaandEstudillo(2008)foundthattransferincomefromabroadservesasan importanttransmissionmechanismthroughwhichinternationalemigrationpositively contributestopovertyreduction.InboththePhilippinesandThailand,thenonfarm incomeaccountsfor,byfar,thelargestshareoftotalruralhouseholdincomein2004.

Fifth,whilepercapitaincomeroseinboththefavorableandunfavorableareas, theincomeincreasewasmoredramaticinthelatter,wheretheinitialincomewaslower andpovertyincidencewashigher(i.e.,66%inthePhilippinesin1985and70%in

Thailandin1987).Thisgrowthofincomewasgenerallybroughtaboutbytherisein nonfarmincomeinallareasbutmorevisiblyinunfavorableareas,whichresultedina muchlargerreductioninpoverty,therebysupportinghypothesis3onthecatchupof unfavorableareaswithfavorableareaswithrespecttoincomegrowthandpoverty reduction.

16 Acomparisonbetweenthegroupoflandless,nearlandless(0–1ha),andsmall farmer(1–2ha)householdsandthegroupofbigfarmers(2haandabove)demonstrates thattheformergrouptendstoearnalargerpercentageofincomefromnonfarmsources, particularlythoselocatedinunfavorableareas(Otsuka et al .,2009b).Acomparisonof percapitaincomewithinthegroupoflandless,nearlandless,andsmallfarmer householdsshowednoconsiderabledifferences,eventhoughthepercapitaincomeofthe largefarmerwasmuchhigher.Largefarmerhouseholdsincludedistinctivelylarge farmersandthosereceivinglargeremittancesfromtheirwelleducatedchildrenworking outsidethecitiesandoverseas.Yetthefactremainsthattherecentdevelopmentofthe nonfarmsectorfavorsthelandpoorhouseholds,whichwouldotherwisehavebeenmuch poorerinthemidstoflowagriculturalwagesanddecliningemploymentopportunitiesin thefarmsector.Otsuka et al .(2009b)foundamoreremarkablemovementoutofpoverty forthelandlesspoorandadeclineintheincomegapbetweenthegroupoflandless,near landless,andsmallfarmerandlargefarmerhouseholds.

Overall,itseemsclearthatthedevelopmentofthenonfarmsectorandincreased accessofagriculturalhouseholdstononfarmlabormarketshaveproventobethemajor drivingforcebehindthereductioninpovertyinruralvillagesinthePhilippinesand

Thailand.Thetransformationofruraleconomiestowardnonfarmactivitiesisclearly propoor,asitincreasesthedemandforlabor,therebydecreasingtheincidenceof unemploymentandunderemployment.

5.3 Determinants of household income

17 Itisimportanttoexplorerigorouslythesignificantfactorsaffectingsourcesofhousehold income.Wedividedhouseholdincomeintothreemajorcomponents:(I)rice,(II)nonrice, and(III)nonfarmincome.Riceincomeincludesincomefromagriculturalwagesand fromricefarming.

Wereporttheimportantfactorsaffectinghouseholdincomeinthetwocountries usingtheregressionresultsreportedinTakahashiandOtsuka(2009)inCentralLuzon andEstudillo et al .(2009a)inNuevaEcijaandIloilointhePhilippines,and

Cherdchuchai et al .(2009)inCentralPlainandnortheastThailand(Tables3and4).

Therearefiveimportantcategoriesoffactors:(I)sizeofcultivatedarea,showninterms ofowned,leasehold,andsharetenantlands;(II)adoptionofMVsandavailabilityof irrigation;(III)“quantity”characteristicsofhumancapital,representedbythenumberof adulthouseholdmembersandnumberofmembersfallingintodifferentagecategories;

(IV)“quality”characteristics,representedbytheratioofadultswithsecondary(orlower secondary)andtertiary(orpostsecondary)schooling;and(V)marketaccessrepresented byvillagedummies. 5

Theestimationresultsofreducedformfarmincomeandriceincomefunctions showthatthesizeofcultivatedarea,particularlyownedandirrigatedareasplantedwith modernricevarieties,wasthemostimportantdeterminantoffarmincomeintheearly years,whichclearlysupportshypothesis1ontheimportanceofaccesstolandandfarm technologyforfarmincome.Therelativecontributionofriceincometototalhousehold income,however,hasdeclined,owingtothedeclineinricepricesandstagnantyield growth.Accordingly,whilethesizeofcultivatedarearemainsasignificantfactor

5Adultsmeansworkingagemembersbetween22and65yearsoldinthePhilippinesand23and65years oldinThailand.

18 affectingriceincome,itsimpactontotalhouseholdincomehasdeclined,owingtothe decliningshareofriceincome.InNuevaEcijaandIloilovillages,wefoundthat leaseholdlandunderrainfedconditionpositivelyandsignificantlyaffectednonrice incomeintheearly2000s,reflectingthediversificationofrainfedagricultureto productionofhighvaluecrops,whichmayhavebeenfacilitatedbythespreadofportable irrigationpumps,whichallowedtheproductionofanonricecropduringthedryseason.

Thisargumentisalsosupportedbythepositiveandsignificantcoefficientofirrigation ratioinnonriceincomefunctionintheearly2000s,whichagainrenderssupportto hypothesis1,whichexplainstheimportanceofagriculturaltechnologyingenerating farmincome.IntheCentralPlainofThailand,householdstendedtoretreatfromfarm worktogetinvolvedmoreheavilyinnonfarmactivitiesbyleasingoutland.

InNuevaEcijaandIloilo,thenumberofhouseholdmembersfallingintovarious agecategories,ingeneral,hadpositiveandsignificanteffectsonriceandnonricefarm incomeinthelate1980s,indicatingthatadditionallaborresourcesofhouseholdswere mainlyallocatedtofarming.Suchpositiveandsignificantimpactdisappearedinthe early2000s,indicatingashiftoflaborresourcesawayfromfarmingtononfarmactivities.

Indeed,thecoefficientsofthefouragecategoriesaboveage31wereallsignificantly positiveinnonfarmincomeregressionintheearly2000s.Thistendencywasalso confirmedintheCentralPlainofThailand,wheremalesbetween23and40yearsoldand between41and60wereactivelyinvolvedinnonfarmworkintheearly2000s.

Interestingly,Filipinofemalestendedtobemoreactiveinnonfarmjobsinearlieryears thanmenbecausetheyhavethecomparativeadvantageinnonfarmworkastheyhave moreeducationthanmales(Quisumbing et al .,2004).Yetintheearly2000s,the

19 expandedlaboremploymentopportunitiesinthenonfarmsectorgaverelativelyequal employmentopportunitiesforallworkingmembersregardlessofsex,whichwas observedinboththePhilippinesandThailand.Thisobservationindicatestheabsenceof genderdiscriminationinthenonfarmlabormarektsinthesecountries.

ThecoefficientsofsecondaryandtertiaryschoolingvariablesinthePhilippines andpostsecondaryschoolingintheCentralPlainandnortheastofThailandwerepositive andsignificantinthenonfarmincomeequationintheearly2000s,butnotinearlieryears, whichrenderssupporttohypothesis2ontheimportanceofqualityofhumancapitalin nonfarmincome.Thismeansthatthemoreeducatedlaborforcebecamemoreactively involvedinnonfarmworkinlateryearsperhapsbecausereturnstoeducationbecame higher,whichsupportshypothesis5onthepreferenceofeducatedhouseholdmembersto workinthenonfarmsector.Itisalsointerestingtonotethatriceincomeandnonrice farmincome,asawhole,werenotaffectedbyschooling,whichindicatesthatschooling didnothaveasignificanteffectontheefficiencyoffarmmanagementatleastinthe

PhilippinesandThailand.FosterandRosenzweig(1996),however,foundthatthe adoptionandsubsequentprofitabilityofnewseedswerehighlydependentonschooling ofadultmembersintheearly1970s,whentheGreenRevolutionhadjustbeganin.

Indeed,theclassicworkofSchultz(1975)indicatedthatschoolingincreasesproductivity byenhancingtheabilitytodealwithdisequilibriainresourceallocationbroughtaboutby newagriculturaltechnologyamongotherthings.Sincemodernricetechnologywasno longernewinSoutheastAsiaintheearly1980s,therewasnoinconsistencybetweenour findingsandthoseofFosterandRosenzweig(1996).

20 Theestimationresultsofincomefunctionscapturetheincreasingimportanceof boththe“quantity”and“quality”ofhumancapitalandthedecreasingimportanceof farmlandingeneratingruralhouseholdincomeinthecourseofeconomicdevelopmentof ruralsocietiesinAsia,renderingsupporttohypothesis2.Estudillo et al .(2008),using theOaxacadecompositionmethod,foundthattheriseinreturnstothenumberof workingagemembershasaccountedformuchofthegrowthofpercapitaincomeinrural

Philippines.Thismeansthatthepoorhouseholds,whichdidnotinvestinschooling becausetheycannotaffordtodoso,wereabletoimprovetheirincomepositionby participatinginthenonfarmlabormarket,wheretheybenefitedfromtherisingwagesof uneducatedandunskilledlabor.Overall,itappearsthatthedevelopmentofthenonfarm sectorispropoor,asthepoorhouseholdsareabletoutilizetheirprimaryasset,whichis unskilledlabor.

5.4 Determinants of children’s schooling

Table5comparestheschoolingattainmentofparentsandchildreninthesamplevillages.

Thereareafewimportantobservations.First,averageschoolingattainmentofboth parentsandchildrenwashigherinthePhilippinethaninThaivillagesbyabout3–4years, thankstoanextensivepublicschoolsystemthathaditsrootssincetheAmericancolonial periodinthe1900s.Second,childrenhaveattainedsignificantlyhigherlevelsof schoolingthantheirparents.Filipinoparentscompleted5–8yearsofschoolingonly, whiletheirchildrencompleted7–11years.InThailand,parentscompleted3–5yearsof schoolingonly,whiletheirchildrencompleted5–9years.Investinginchildren’s

21 schoolingseemstobethemainformofintergenerationaltransfersofforrural householdswhenthesizeoffarmlandhasbecomesmaller.Andthird,theproportionof adultworkingmemberswithsecondaryschoolinghasincreasedinbothcountriesand, moreremarkablyinthePhilippines,indicatingthat,inthiscountry,returnstohigher levelsofschoolinghaverisenparticularlyrapidly.Thehighlyeducatedlaborforcemust havebeenabsorbedbythenonfarmsectorinasmuchasschoolingdoesnotaffectfarm managementefficiency.

WeusedincrementalyearsinschoolofadultchildreninthePhilippinesand completedyearsofschoolinginThailandasmeasuresofparentalinvestmentsin schooling.Theseadultchildrenarethosewhowereinschoolageatthetimeofthebase yearsurveys. 6Table6showstheestimationresultsofthedeterminantsofschooling investments.Wedividedourexplanatoryvariablesintothefollowingcategories:(I)size offarmlandandtenureinthebaseyear,(II)farmandnonfarmincomesinthebaseyear,

(III)modernagriculturaltechnologyrepresentedbytheadoptionofMVsandirrigation ratio,(IV)completedyearsinschooloffatherandmother,(V)characteristicsofthechild suchasageandgender,and(VI)supplysidefactorssuchastheavailabilityofschools androadsanditsqualityascapturedbythevillagedummies.InTable6,weshowthe coefficientsof(I),(II),(III)and(IV)only,whiletheimpactsof(V)and(VI)are discussedinthetext.Model1inCentralLuzonhasatwostageprocedurethatuses predictedvaluesoffarmandnonfarmincomesfromthefirststageincomefunctions.

Model2usesreducedformregressionmodelonthepresumptionthattheimpactsof

6InNuevaEcijaandIloilovillages,oursamplewasthegroupofchildrenwhowere620 yearsoldatthetimeofthebaseyearsurveysin1985and2002.Incrementalyearsmeans increasesinschoolingyearsbetween1979and2003inCentralLuzonandbetween1985 and1989and2002and2004inNuevaEcijaandIloilo.

22 accesstolandandagriculturaltechnologyontotalhouseholdincomecomeindirectly throughitseffectonhouseholdfarmincome.Reducedformregressionwasalsousedin

CentralPlainandnortheastThailand.

Itisimportanttomentionthatalandreformprogramwasimplementedinthe

Philippines,whichwasthemajorcauseofanincometransferfromthelandlordtoshare tenants,therebyallowingthelattertoinvestinchildren’sschooling.ThePhilippineland reformconsistsoftwomajorprograms:(1)atenancyreform(OperationLeasehold), whichconvertssharetenantsintoleaseholdtenants;and(2)alandredistributionprogram, whichconvertssharetenantsintoamortizingowners(holdersofaCertificateofLand

Transfer[CLT]).Leaseholdrentandamortizationfeeswerefixedat25%oftheaverage riceyieldforthreenormalcropyearsprecedingthelandreformimplementationin1972.

RiceyieldsroseinthevillagesbecauseofthediffusionofMVssothatadivergence betweenthereturnstolandandfixedleaseholdrentsandamortizationfeesprescribedby lawwascreated.Thisdivergenceledtotheemergenceofasubtenancyarrangementin theformoflandpawning.

Underthepawningarrangement,themoneylenderadvancescashtothefarmer andtakesoverthecultivationofthelandwhiletheindebtedfarmercommonlyremainsin possessionofthecultivationrightofthelandasasharecropper(Andersen,1962;

McLennan,1969;HayamiandKikuchi,2000).Themoneylenderpocketsthedifference

(amountingtoabout15%ofthegrossoutput)betweenhisshareofoutputandthefixed leaseholdrentmandatedbylaw.Thepawningarrangementenablesbeneficiariesofland reform,whowereformersharetenants,toinvestinhumancapital,includingschooling andmigrationofchildren,andselfemployednonfarmactivities.

23 InNuevaEcijaandIloilo,thesizeofpawnedoutlandhaspositivelyaffected progressioninschoolingofchildrenin2002–04.Thisindicatesthatthedevelopmentof pawningmarketforfarmlandhasenabledhouseholdstoraisefundsforchildren’s schooling.Infact,pawningrevenueshavebeenusedalsotofinanceoverseasmigration andtoventureintononagriculturalbusinesses,bothofwhichentailhighfixedcostsbut nonethelessopenupopportunitiesforruralhouseholdstostepfurtherupintheincome ladder(Estudillo et al .,2009b).

InNuevaEcijaandIloilo,wealsofoundthatthedifferenceintheprogressof childrenthroughschool,measuredasthedifferenceinyearsofschoolingcompletedin

1985–89and2002–04,betweenthelandlessandfarmerhouseholdswaslargein1985–89, butthendeclinedin2002–04,evenforthosehouseholdswithchildrenintertiaryschool age.Thisimpliesthat,landlessparentshavebeenabletoaffordtosendtheirchildren eventotertiaryschools,inrecentyears,asmuchastheparentsoflandedhouseholds.

Sincethelandlesshouseholdsdidnothavepawningrevenues,itseemsclearthatthe increasingdominanceofnonfarmincomewasthemajorforcebehindtherisein schoolingattainmentofthechildrenoflandlesshouseholdsvisàvisthechildrenof farmerhouseholds.Inasmuchassecondaryandtertiaryschoolinghaspositiveand significanteffectsonnonfarmincome,wecanreasonablyconjecturethatthe improvementintheincomepositionofthelandlessvisavisthelandedhouseholdswas broughtaboutbytheacquisitionofhigherlevelsofschoolingoftheirchildren,who eventuallybecameactiveparticipantsinthenonfarmlabormarket.

InCentralLuzon,bothfarmandnonfarmincomeshavepositiveandsignificant effectsoncompletedyearsinschoolofadultchildren,whileinNuevaEcijaandIloilo,

24 farmincomeappearedtobethesinglemostimportantsourceoffundstofinance additionalyearsinschoolin1985–89.InCentralLuzon,thesizeofownedlandand irrigationratioin1979andintheCentralPlainofThailand,irrigatedareaplantedwith traditionalvarieties(TVs)havesignificantlyandpositivelyaffectedcompletedyearsin schoolofadultchildren.Thesefindingsindicatethataccesstolandandagricultural technologywascriticalinhumancapitalaccumulationinearlieryears,whenfarmingwas adominanteconomicactivity,renderingsupporttohypothesis4.

Completedyearsinschoolofparents,ingeneral,hasimprovedchildren’s schoolingattainmentbecausethemoreeducatedparentscanperceiveincreasesinreturns toschoolingsothattheytendtoinvestmoreinschoolingofchildreninanticipationof theirchildrenjoiningthenonfarmlabormarketuponcompletingschool.Itisinteresting tonotethatmother’sschoolinghasahigherimpactonchildren’sschoolingthanthe father’sinNuevaEcijaandIloiloandinCentralPlainofThailand,perhapsbecause womeninthoseareastendtobemoreactiveinnonfarmemployment,importantlyinlight manufacturingindustries(Momsen,2004).

Birthyeardummiesshowedthatschoolinginvestmentsweremadeinfavorof youngercohort.MaledummywasnegativeandsignificantinthePhilippines,evenafter controllingforothereffects,implyingthatparentstendtofavorfemaleswheninvesting inchildschooling.Thispreferencewasparticularlysignificantforeldestdaughters,who areexpectedtohelpfinancetheschoolingofheryoungersiblingsuponfinishingschool andenteringthejobmarket.InThailand,themaledummywasnotsignificant,which meansthatThaiparentsinvestequallyonschoolingofchildren,irrespectiveofgender.

25 Tosummarize,ourregressionresultsdemonstratethataccesstolandand agriculturaltechnology,throughitspositiveimpactonfarmincome,arebyfarthemost importantfactorsthathaveinducedinvestmentsinschoolingofchildreninthe1980s.In thecaseofthePhilippines,theGreenRevolutionandlandreformimplementation stimulatedthedevelopmentofthelandpawningmarketand,subsequently,investments inschoolingoftheyounggenerations.Itisworthemphasizingthat,despitethelackof accesstofarmland,theFilipinolandlesshouseholds,whowerethepoorermembersof ruralsocieties,wereabletoinvestinchildren’sschoolinginearly2000sasmuchasthe farmerhouseholds,mostlikelybecauseoftheincreaseintheirnonfarmincome.

TheresultsofouranalysesofthePhilippinesandThailandareconsistentwith thoseinSouthAsia.In,Hossain et al. (2009)reportedthattheadoptionof improvedagriculturaltechnologies,cropdiversification,andoccupationalmobilityfrom farmtononfarmactivities,suchastrade,business,andservices,aretheimportant pathwaysoutofpovertytoasignificantproportionofthepoorhouseholds.Theshiftin thestructureofhouseholdincomeinfavorofnonfarmactivitieshasbeenfacilitatedby thedecisionofhouseholdstoinvestinschoolingofchildren,wholaterjoinedthe nonfarmsector.Farmincomeandeducationofadultworkersarebyfarthemost importantdeterminantsofchildren’sschoolenrollment,pointingtotheimportanceofthe adoptionofnewricetechnology.Highernonfarmincomeisthemajorfactorbehindthe dramaticriseinhouseholdincomeandthereductioninpoverty.InTamilNadu(India),

KajisaandPalanichamy(2009)reportedthatchildrenfromhouseholdswithhigherfarm incomeareabletoattendschoolbeyondthecompulsorylevel,pointingtotheimportant roleofnewtechnologyinrice,nonricecrops,andlivestockproductionininducing

26 children’sattendanceinschool.UnlikeinthecaseofthePhilippines,nonfarmincomeis notasignificantfactor,which,accordingKajisaandPalanichamy(2009,p.138),maybe duetothefactthatnonfarmearningsofchildrenarenotspenttofinanceschooling investmentsoftheiryoungersiblings.

5.5 Occupational choice and nonfarm income

Themainoccupationofthehouseholdheads,whoarepredominantlymaleinCentral

Luzon,wasricefarmingbecauseearlysurveysweretailoredtolookcloselyoneconomic activitiesofricefarminghouseholds.InNuevaEcijaandIloilo,oursamplesetincluded householdsheadedbybothfarmersandlandlessworkers.Becauseofthedeclinein inheritablesizeoffarmlandandlandreformrestrictionsonlandtransfer,therewasan increaseinthenumberoflandlessagriculturalworkersinthechildren’sgeneration.

Clearly,withouttheincreasingavailabilityofnonfarmjobs,therurallowincome populationisboundtoincrease.

Whiletheparentswereengagedinagriculture,theiradultchildrenoccupied highlydiversifiedjobsinthevillage,localtowns,andcities.Occupationsinthevillage andlocaltownswerepredominantlyunskilled,includingjobsintheinformalsectorin transportation,commerce,domesticwork,andskilledartisanwork,reflectingthe increasingdemandfortheseservicesintheruralareas.Incontrast,manufacturingjobs areseldomavailable.Theseobservationsseemtosuggestthatthegrowthlinkageeffects workforthedevelopmentofservicesectors,whoseproductsarenontradablegoods,but notformanufacturingsectors,whoseproductsaretradable.Skilledjobswereheldbythe

27 moreeducatedchildrenlivinginthecitiesandoverseas,andmanyofthemwere professionalsincludingnurses,doctors,teachers,andengineers.Professionaljobsand overseasworkrequireearlierinvestmentsofhouseholdsinschooling,whichwas facilitatedbyanincreaseinfarmincomeinearlieryears,aswehaveshownearlier.We alsofoundanincreasingtendencyfortheinternationallabormarkettoacceptunskilled workerssuchaswomenindomesticandmeninconstructionworksinlateryears.These workerswerecommonlyhighschoolgraduateswithonly10yearsofschoolingoreven less.Theymusthavecomefromthelowerincomegroup,yettheywereabletoventure intotheinternationallabormarketpartlybecausejobplacementfeeshavebecome affordableasinternationallabormarketshavebecomemorecompetitive.

InCentralPlain,Thailand,urbanwageemploymentispopularbecausechildrenin thisregioncaneasilymigratetoworkinfactorieslocatedinBangkokandnearby industrializedareas.InnortheastThailand,ruralnonfarmworkconsistingofcasualand regularsalaryworkandselfemploymentdominatesamongyounggenerations.

Distributionofchildren’snonfarmoccupationswasfairlysimilarformalesandfemales intheCentralPlain.Femalesinthenortheastaremorelikelytoworkinruralcasual nonfarmjobs,whilemenmigrateouttoobtainregularnonfarmjobsinBangkok.

Table7showsthedeterminantsofcurrentoccupationalchoiceofchildren,shown asthemarginaleffectsoftheregressorsevaluatedattheirmeans,ontheprobabilityof choosingruralnonfarm,urban,andoverseaswork.Filipinofemalestendedtoengage moreinruralnonfarmactivitiesthanmales,reflectingthecomparativeadvantageof femalesinnonfarmjobsandmalesinfarmjobs(Quisumbing et al .,2004).Filipino femalesalsohadhigherpropensitytoventureintooverseasmigration,reflectingthe

28 increasingcontributionoffemalesinhouseholdincomegenerationasaresultofthe integrationoftherurallabormarketwiththeinternationalmarkets.Educationwas positivelyassociatedwiththeprobabilityofparticipationinbothruralandurbannonfarm activities,whichsupportshypothesis5ontherelationshipbetweenchildschoolingand occupationalchoice.Themarginalimpactofeducationonoutmigrationwashigherthan inparticipationinruralnonfarmemploymentinruralareas.Educationdoesnotseemto beasignificantfactorinoverseasmigrationperhapsbecauseofthelargenumberof overseasworkersinunskilledjobssuchasdomesticandconstructionwork,especiallyin theMiddleEastandEastAsia.

InbothCentralPlainandnortheastThailand,wefoundthatthemarginaleffects ofhighereducation,shownasdummyforpostlowersecondaryschooling,suggesta patterninwhichtheprobabilityofjoiningthenonfarmlabormarketsincreasesatthe expenseoffarmandselfemploymentaseducationlevelgoesup.Childrenwithmore educatedmothersarelesslikelytobeinvolvedinfarmingandmorelikelytoworkin nonfarmjobsintheCentralPlainandBangkok. 7Agriculturaltechnology,shownasthe interactiontermbetweentheratioofirrigationandratioofareaplantedwithMVs, increasedtheprobabilityoffarmingintheCentralPlain,whichindicatesthatfarmingis profitablewiththeadoptionofmodernricetechnology.Childrenwithmoreeducated mothersaremorelikelytoparticipateinnonfarmemploymentandlesslikelytobe involvedinfarmingintheCentralPlain.

Tosummarize,ourregressionresultsrevealthatinboththePhilippinesand

Thailand,thereisaclearshiftofoccupationalchoiceawayfromfarmtononfarmjobs, 7InstudyvillagesinThailand,nonfarmactivitiesreferalmostexclusivelytourbannonfarmactivitiesin nearbycitieswhereasinPhilippinevillages,nonfarmactivitiesincludebothruralandurbannonfarm activities.

29 withthemoreeducatedchildrenventuringintothemorelucrativenonfarmjobs,inwhich returnstoschoolingareexpectedtobehigher,renderingsupporttohypothesis5onthe relationshipbetweenoccupationalchoiceandschoolingattainment.Animportantpoint ofinquiryistowhatextenteducationhasaffectednonfarmincome,astheincreasein nonfarmincomeisexpectedtobethemajordrivingforcebehindpovertyreduction.

Weshowstatisticalresultsofthedeterminantsofindividualnonfarmincome drawnfromTakahashiandOtsuka(2009)forthePhilippinesandCherdchuchai et al .

(2009)forThailand.Table8showsa“Minceriantypeincomefunction”usingpersonal characteristicsoftheworker(i.e.,schooling,workexperience,age,andgender)as explanatoryvariablesoflogfarmearnings(i.e.,dailylaborearningsinthePhilippines andannualearningsinThailand).TheMincerianfunctionwasestimatedusingatwo stageproceduretocontrolforsampleselectionbias,asonlythosechildrenwhoworkin ruralnonfarmsectorandManilawereincludedinthesamplesetwiththeexclusionof childrenworkinginthefarmsector,suchasfarmersandagriculturalworkers.

Nonetheless,theselectivitycorrectiontermwasnotsignificantforthetwocountry regressions.

Education,whichisspecifiedinbothyearsofschoolingcompletedandasdummy variablesforsecondaryandtertiaryschoolingcompleted,haspositivelyandsignificantly affectednonfarmincomeofchildreninCentralLuzon.Theaverageratesofreturnsto educationweresubstantiallyhigherinManilathaninruralareas,judgingfromthelarge differenceinthemagnitudeofthecoefficients.Moreover,tertiaryschooling(butnot secondaryschooling)haspositivelyaffectedruralnonfarmincome,whereasbothtertiary andsecondaryschoolingweresignificantfactorsexplainingearningsinManila.These

30 findingssuggestthatthedisadvantageofhavingprimaryschoolingonlyisrelatively smallinruralnonfarmjobs,wheretheinformalservicesectorremainsdominant.In otherwords,thedevelopmentoftheruralnonfarmsectorislikelytobeespeciallymore propoorthanthatoftheurbannonfarmsector.

Itisinterestingtoobservethatthecoefficientsofexperienceanditssquaredterm aresignificantonlyinruralareas,butnotinManila,suggestingtheimportanceof accumulationofspecifichumancapitalthatisspecifictotheavailablejobsinruralareas.

Itimpliesthateventhelowlyeducatedworkerscanincreasetheirnonfarmearningsin ruralareasbysimplyaccumulatingworkexperience.ThisstorydoesnotholdinManila, wheretertiaryschoolingisparticularlyimportanttoincreaseearning,perhapsbecause jobsinManilarequiregeneralskills,whichcanonlybeobtainedthroughformaltraining inschools.

InThailand,ourregressionresultsalsoshowthesignificantandpositiveimpact ofeducationonannualearnings,whichagainsupportshypothesis5thatthemore educatedworkerstendtoobtainemploymentinthemorelucrativenonfarmsector,where theycanfullymaximizethereturnstotheirschooling.Educationdidnothaveanimpact onselfemployment,whichprovidesmainlyinformalservice,suggestingthatformal schoolingisnotarequisitetoincreaseincomeinthissector.Thus,ruralnonfarmjobsin servicesectorstendtoprovideemploymentopportunitiesforthepoor,whoareless educatedthantherich.Nonetheless,thisimportantfunctionoftheruralnonfarmlabor marketsdidnotreceivemuchattentionintheliteratureonrurallnonfarmeconomies

(Haggbladeetal.,2007).

31 Inbothcountries,wefoundthatgenderdidnotaffectearnings,exceptintherural nonfarmsectorinCentralLuzon,wherepresumablydomesticwork,whicharemore appropriateforwomen,aremorecommonamongthesamplerespondents.IntheCentral

PlainofThailand,childrenofthemoreeducatedparentstendedtoearnsignificantly higherincomeinthenonfarmsector,reflectingthetendencyofeducatedparentstorelay informationtochildrenonthemorelucrativenonfarmjobs.Overall,ourregression resultsshowtheprimeimportanceofacquiringhighereducationasastrategytoincrease individualincome,thatneitherworkexperiencenorgenderaffectsnonfarmearnings, especiallyinthecity,andthatruralnonfarmjobsareparticularlypropoorastheyseldom requireshighereducation.

6. Summary and conclusions

ThischapterexploresthestrategicprocessesbywhichruralhouseholdsinSoutheastAsia wereabletochangetheirsourcesofhouseholdincomeandhowpoorhouseholdswere abletomoveoutofpovertyusinglongtermpaneldatasetsinvillagesinthePhilippines andThailand.Ruralhouseholdsareabletomoveoutofpovertyinthepresenceof increasingscarcityoffarmlandanddeclininglaboremploymentopportunitiesin agriculturebydiversifyingtheirincomesourcesawayfromricetononricecropsand, moreimportantly,byengaginginnonfarmactivities.Theriseinnonfarmincomeisthe mostdecisivefactordirectlyresponsibleforpovertyreductioninruralAsia.

Weobservedthattheyoungerandmoreeducatedchildrenarethosewhoaremore activelyinvolvedinnonfarmjobs.TheGreenRevolutionisthemajordrivingforce

32 behindtheriseininvestmentsinchildren’sschoolingthroughtheincreaseinfarm income,therebycontributingtopovertyreductionnotonlyintheshortrunbutinthe longerrunaswell.Thesefindingssuggestasequenceoflongtermchangesfromthe

GreenRevolutiontowardincreasedfarmincome,increasedinvestmentinchildren’s schooling,andthechoiceoflucrativenonfarmoccupationsbytheyoungerandeducated laborforce,whichcontributedtopovertyreductionandthedevelopmentofthenonfarm sector.

Yet,incomegrowthandpovertyreductionwereobserved,eveninareaswherethe

GreenRevolutiondidnottakeplace.Thiswasfacilitatedbytheincreasedavailabilityof nonfarmjobs,evenfortheunskilledlabor,whichisthemajorassetofthepoor,who cannotinvestinschoolingbecausetheycannotaffordtodosointheabsenceofefficient creditmarkets.Inalllikelihood,themajordrivingforcebehindthemovementofthe pooroutofpovertyinareas,wheretheGreenRevolutiondidnottakeplace,istherisein jobopportunitiesbroughtaboutbythedevelopmentofthenonfarmsector,includingthe urbansector,whichwouldhavebeeninducedtodeveloppartlybytheGreenRevolution.

AmajorresearchagendaistoseehowtheGreenRevolutionhasstimulatedthe growthoflaborintensiveruralindustriesandservicesthroughtheproductionand consumptionlinkages,whichhasprovidedgreateremploymentopportunitiesfor unskilledlabor.Thegrowthlinkageeffects,however,donotnecessarilyworklocally.

Increaseddemandfornonfarmtradablecommoditiesmayfacilitatethedevelopmentof urbannonfarmsectors,particularlyinindustrialclusterswhereagglomerationeconomies tocostandproductionadvantages(SonobeandOtsuka,2006).

33 Themajorpolicyimplicationisthat,inordertostimulatethedevelopmentofthe entireeconomy,itissensibletodevelopagriculturefirst,whenthatsectordominatesthe economy.Thus,itiscriticallyimportanttodevelopimprovedagriculturaltechnologies andtodiffusetheproductionofhighvaluecropsinpoorareaswhereagricultureisthe dominantsourceofruralhouseholdincome.Agriculturaldevelopmentcantriggera subsequenttransformationofruraleconomiestowardincreasednonfarmactivitiesby stimulatinginvestmentsinschoolingofyoungerchildren,whosubsequentlycontributeto povertyreductionandfurtherdevelopmentofthenonfarmsector.Inthiscontext,another majorresearchagendaistoexplorethestrategytodevelopruralnonfarmsectorswith dueconsiderationofmarketfailuresthathinderthedevelopmentoftheunskilledand uneducatedlaborintensivesegmentoftheeconomy.

References

Andersen, J. (1962). “Some aspects of land and society in a Pangasinan community”. PhilippineSociologicalReview10(1),4158. Asian Development Bank (2008). Key Indicators for Asia and the Pacific. Manila, Philippines. Barker, R., Herdt, R. (1985). The Rice Economy of Asia. Resources for the Future, Washington,DC. Cherdchuchai, S., Otsuka, K., Estudillo, J.P. (2009). “Income dynamics, schooling investment,andpovertyreductioninPhilippinevillages,19852004”.In:Otsuka, K.,Estudillo,J.P.,Sawada,Y.(Eds.)RuralPovertyandIncomeDynamicsinAsia andAfrica,Routledge,London.

34 David,C.C.,Otsuka, K.(1994).ModernRice Technology and IncomeDistributionin Asia.LynneRienner,Boulder,CO. Estudillo, J.P., Otsuka, K. (1999). “Green revolution, human capital, and offfarm employment: changing sources of income among farm households in Central Luzon,196694”.EconomicDevelopmentandCulturalChange47(3),497523. Estudillo, J.P., Sawada, Y., Otsuka, K. (2008). “Poverty and income dynamics in Philippinevillages,19852004”.ReviewofDevelopmentEconomics12(4),877 890. Estudillo,J.P.,Sawada,Y.,Otsuka,K.(2009a).“Incomedynamics,schoolinginvestment, and poverty reduction in Philippine villages, 19852004”. In: Otsuka, K., Estudillo, J.P., Sawada, Y. (Eds.) Rural Poverty and Income Dynamics in Asia andAfrica,Routledge,London. Estudillo,J.P.,Sawada,Y.,Otsuka,K.(2009b).“Thechangingdeterminantsofschooling investments: evidence from villages in the Philippines, 198589 and 200204”. JournalofDevelopmentStudies45(3),391411. Foster,A.D.,Rosenzweig,M.(1996).“Technicalchangeandhumancapitalreturnsand investments:evidencefromtheGreenRevolution”.AmericanEconomicsReview 86(4),93153. Haggblade, S., Hazell, P., Reardon, T. (2007). “Sectoral growth linkages between agricultureandtheruralnonfarmeconomy”.In:Haggblade,S.,Hazell,P.,Reardon, T.(Eds.)TransformingtheRuralNonfarmEconomy:OpportunitiesandThreatsin theDevelopingWorld,TheJohnsHopkinsUniversityPress,Baltimore,MD. Harris,J.R.,Todaro,M.P.(1970).“Migration,unemploymentanddevelopment:atwo sectoranalysis”.AmericanEconomicReview60(1),12642. Hayami, Y. (2001). “Ecology, history, and development: a perspective from rural SoutheastAsia”.TheWorldBankResearchObserver16(2),169198. Hayami,Y.,Godo,Y.(2006).DevelopmentEconomics:FromthePovertytotheWealth ofNations,3 rd ed.OxfordUniversityPress,Oxford. Hayami, Y., Kikuchi, M. (2000). A Rice Village Saga: Three Decade of Green RevolutioninthePhilippines.MacmillanPress,London.

35 Hazell,P.,Haggblade,S.(1991).“RuralurbangrowthlinkagesinIndia”.IndianJournal ofAgriculturalEconomics46(4),515529. Hossain, M., Rahman, M., Estudillo, J.P. (2009). “Income dynamics, schooling investments,andpovertyreductioninBangladesh”.In:Otsuka,K.,Estudillo,J.P., Sawada, Y. (Eds.) Rural Poverty and Income Dynamics in Asia and Africa, Routledge,London. Jayasurija, S.K., Shand, R.T. (1986). “Technical change and labor absorption in Asian agriculture:someemergingtrends”.WorldDevelopment14(3),415428. Jorgenson,D.W.(1961). “Thedevelopmentofadualeconomy”.EconomicJournal71 (282),30934. Kajisa,K.,Palanichamy,N.V.(2009).“Incomedynamicsandschoolinginvestmentsin Tamil Nadu, India, 19712003: changing roles of land and human capital”. In: Otsuka,K.Estudillo,J.P.,Sawada,Y.(Eds.)RuralPovertyandIncomeDynamics inAsiaandAfrica,Routledge,London. Lanjouw,P.(2007).“Doestheruralnonfarmeconomycontributetopovertyreduction?”. In:Haggblade,S.,Hazell,P.,Reardon,T.(Eds.)TransformingtheRuralNonfarm Economy:OpportunitiesandThreatsintheDevelopingWorld,TheJohnsHopkins UniversityPress,Baltimore,MD. Lewis, W.A. (1954). “Economic development with unlimited supplies of labor”. ManchesterSchoolofEconomicandSocialStudies22(1),13991. Lipton,M.,Longhurst,R.(1989).NewSeedsandPoorPeople. UnwinHyman,London. McLennan,M.(1969).“LandandtenancyintheCentralLuzonplain”.PhilippineStudies 17(4),651682. Momsen,J.H.(2004).GenderandDevelopment.Routledge,London. Otsuka, K. (2007). “Efficiency and equity effects of land markets”. In: Evenson, R., Pingali,P.,HandbookofAgriculturalEconomics,Vol.3.,Elsevier,Amsterdam. Otsuka,K.,Asano,S.,Gascon,F.(1994).“’SecondGeneration’MVsandtheEvolution

oftheGreenRevolution:TheCaseofCentralLuzon,196690.”Agricultural

Economics10(3),28395.

36 Otsuka,K.,Estudillo,J.P.,Sawada,Y.(2009a).RuralPovertyandIncomeDynamicsin AsiaandAfrica.Routledge,London. Otsuka, K.,Estudillo,J.P.,Sawada,Y. (2009b).“Towardanewparadigmoffarmand nonfarm linkages in economic development”. In: Otsuka, K., Estudillo, J.P., Sawada, Y. (Eds.) Rural Poverty and Income Dynamics in Asia and Africa, Routledge,London. Otsuka,K.,Yamano,T.(2006).“Introductiontothespecialissueontheroleofnonfarm income in poverty reduction: evidence from Asia and East Africa”. Agricultural EconomicsSupplementtoIssue35.3. Quisumbing,A.R.,Estudillo,J.P.,Otsuka,K.(2004).LandandSchooling:Transferring WealthAcrossGenerations.JohnsHopkinsUniversityPress,Baltimore. Ranis,G.,Fei,J.C.H.(1961).“Atheoryofeconomicdevelopment”.AmericanEconomic Review51(4),53358. Rosegrant, M.W., Hazell, P.B. (2000). Transforming the Rural Asian Economy: The UnfinishedRevolution.OxfordUniversityPress,Oxford. Sawada,Y.,Estudillo,J.P.,Otsuka,K.(2009).“Introduction:anoverviewandconceptual framework”.In:Otsuka,K.,Estudillo,J.P.,Sawada,Y.(Eds.)RuralPovertyand IncomeDynamicsinAsiaandAfrica,Routledge,London. Sawada, Y., Estudillo, J.P. (2008). “Trade, migration and poverty reduction in the globalizingeconomy:thecaseofthePhilippines”.In:Nissanke,M.,Thorbecke,E. (Eds.) Globalization and the Poor in Asia: Can Shared Growth Be Sustained? PalgraveMacmillan,NewYork. Schultz, T.W. (1975). “The value of the ability to deal with disequilibria”. Journal of EconomicLiterature13(3),82746. Sonobe, T., Otsuka, K. (2006). ClusterBased Industrial Development: An East Asian Model.PalgraveMacmillan,Hampshire,UK. Takahashi,K.,Otsuka,K.(2009).“Humancapitalinvestmentandpovertyreductionover generations: a case from the rural Philippines, 19792003”. In: Otsuka, K., Estudillo, J.P., Sawada, Y. (Eds.) Rural Poverty and Income Dynamics in Asia andAfrica,Routledge,London.

37 Ut, T., Hossain, M., Janaiah, A. (2000). “Income distribution and poverty in Asia: insightsfromvillagestudies”.EconomicandPoliticalWeekly35(5253),2542. World Bank (2008a). World Development Report 2008: Agriculture for Development. Washington,DC. WorldBank(2008b).WorldDevelopmentIndicators.Washington,DC. WorldBank(2009).WorldDevelopmentReport2009:ReshapingEconomic. Washington,DC.

38 Table1 IndicatorsofstructuraltransformationinselectedcountriesinSoutheastAsia Descriptionandcountry 1995 2006 Philippines Grossnationalincomepercapita(US$)a 1,050 1,390 %agriculture b 22 14 %industry(%manufacturing) b 32(23) 32((23) %services b 46 54 Thailand Grossnationalincomepercapita(US$) 2,740 3,050 %agriculture 10 11 %industry(%manufacturing) 41(30) 45(35) %services 50 45 Indonesia Grossnationalincomepercapita(US$) 980 1,420 %agriculture 17 13 %industry(%manufacturing) 42(24) 47(28) %services 41 40 Vietnam Grossnationalincomepercapita(US$) 240 700 %agriculture 27 20 %industry(%manufacturing) 29(15) 42(21) %services 44 38 % of household below $1.35 per day c Initialyear Finalyear Philippines(19942006) 32.9 27.0 Thailand(19922002) 7.9 0.1 Indonesia(19932005) 60.7 39.2 Vietnam(19932004) 73.3 25.6 aTakenfromWorldBank(1997,2008b,Table1.1).Referstogrossnationalproductin 1995andgrossnationalincomein2006. bTakenfromWorldBank(2008b,Tables4.2and6.1). cTakenfromAsianDevelopment(2008,Table6.1).ThepovertylineistheAsian povertyline,whichisthe2005purchasingpowerparitybasedonconsumption.

39 Table2 Farmsizeandsourcesofhouseholdincomeofsamplehouseholds inthePhilippinesandThailand Descriptionandcountry Year Philippines Late1980s a Early2000s a Averagefarmsize b(ha) 1.00 0.76 Landlesshouseholds(%) b 22 44 Compositionofhouseholdincome c Favorableareas Percapitaincome(PPP$) 1,065 2.364 Agriculturalwage(%) 13 11 Rice(%) 37 12 Nonricefarmincome(%) 5 7 Nonfarmincome(%) 45 70 Unfavorableareas Percapitaincome(PPP$) 386 1.119 Agriculturalwage(%) 30 7 Rice(%) 20 9 Nonricefarmincome(%) 13 24 Nonfarmincome(%) 36 60 Povertyincidence(headcountratio) d Favorableareas 40 23 Unfavorableareas 66 42 Thailand Averagefarmsize e(ha) 4.24 2.42 Landlesshouseholds(%) e 0 13 Compositionofhouseholdincome c Favorableareas Percapitaincome(PPP$) 2,014 4,617 Agriculturalwage(%) 4 6 Rice(%) 66 26 Nonricefarmincome(%) 21 22 Nonfarmincome(%) 10 47 Unfavorableareas Percapitaincome(PPP$) 959 2,543 Agriculturalwage(%) 12 5 Rice(%) 54 7 Nonricefarmincome(%) 13 14 Nonfarmincome(%) 21 74 Povertyincidence(headcountratio) d Favorableareas 51 12 Unfavorableareas 70 21 aLate1980srefersto1985inthePhilippinesand1987forThailandwhileearly2000srefersto2004. bTakenfromEstudillo et al .(2009a,Table2.1). cTakenfromOtsuka et al .(2009b,Table9.2). dTakenfromOtsuka et al .(2009b,Table9.3). eTakenfromCherdchuchaietal.(2009,Table4.4).

40 Table3 DeterminantsofhouseholdincomeinsamplevillagesinthePhilippines Variable CentralLuzon a NuevaEcijaandIloilo b Late1970s Late1980s Early2000s Farm Nonfarm Rice Nonrice Nonfarm Rice Nonrice Nonfarm Ownedland(ha) 4.23** 4.42** Ownedlandirrigated(ha) 5.17** 0.63 3.11** 73.42** 3.04** 1.74 Ownedlandrainfed(ha) 1.44 0.05 1.05 5.85 2.92 7.29 Leaseholdland(ha) 2.52** 0.75 Leaseholdlandirrigated(ha) 0.03 0.17 0.89 12.05 0.54 4.16 Leaseholdlandrainfed(ha) 1.84 0.36 0.65 0.13 12.65** 8.64 Sharetenantland(ha) 1.40** 2.17** Sharetenantlandirrigated (ha) 31.0 2.06 14.97 Sharetenantlandrainfed(ha) 2.94 1.11 0.36 7.37 3.67 8.83 Irrigationratio 10.50** 0.68 1.26 0.71 2.76 3.49* 13.13* Numberofadults 0.97** 2.76** Numberofhousehold members 2230yearsold 1.58 1.99** 1.24 8.61 1.74 1.56 3140yearsold 3.26** 2.11** 0.18 5.87 0.83 10.32** 4150yearsold 4.45** 2.07** 1.16 6.14 2.97* 17.85** 5160yearsold 2.31 1.95** 2.29* 3.56 1.19 25.28** 61yearsoldandabove 4.15** 1.60** 0.78 1.62 1.15 10.43** Ratioofadultswith secondaryschooling 0.373 3.50 2.37 0.57 2.53 20.54 1.34 36.80** Ratioofadultswithtertiary schooling 2.27 12.26** 0.40 1.49 4.53 20.30 12.15** 89.64** *=significantat5%level,**=significantat1%level. aTakenfromTakahashiandOtsuka(2009,Table3.6).Incomereferstopercapitaincomeofparents.Late1970srefersto1979. bTakenEstudillo et al .(2009a,Table2.8).Incomereferstohouseholdincome.Late1980srefersto1985and early2000srefersto2001.

41 Table4 DeterminantsofhouseholdincomeinsamplevillagesinThailand a Variable Rice Nonrice Nonfarm Late Early Late Early Late Early 1980s 2000s 1980s 2000s 1980s 2000 Central Plain Owned land (ha) 1.58* 0.84 -0.40 1.03 -3.35* -1.84 Leasehold land (ha) 0.86 3.25** 0.56 0.42 -0.36 -3.10** Share tenant land (ha) 0.56 1.22 0.13 -0.48 -2.36 -3.35 Ratio of irrigation times ratio of area planted with MVb 32.36** 22.20** 6.09 10.93 5.00 4.76 Ratio of irrigation times ratio of area planted with TV c 17.24** 2.07 9.03 Number of household members Male 23-40 years old -0.68 -3.61 -4.98 -4.28 3.32 11.37** Male 41-60 years old -0.65 -2.49 9.21 -3.21 6.58 11.94* Male over 60 years old -1.65 -5.90 12.12* 1.87 -5.58 3.35 Female 23-40 years old -0.70 -7.68 -4.12 3.37 7.04 8.50 Female 41-60 years old -2.24 4.33 -2.81 1.40 -4.18 1.60 Female over 60 years old -6.75* 4.63 -1.00 0.13 -7.48 -6.38 Ratio of adults with lower secondary schooling -5.49 7.61 9.49 5.36 9.21 18.15 Ratio of adults with postlower secondary schooling -21.06 -0.26 16.40 -9.78 -19.45 23.56*

42 Table4(continued) Variable Rice Nonrice Nonfarm Late Early Late Early Late Early 1980s 2000s 1980s 2000s 1980s 2000s Northeast Owned land (ha) 1.18** 1.43* 0.73** 3.31* 1.26 -3.76 Leasehold land (ha) 0.58 -1.53 0.38 1.20 -2.22 -24.76 Share tenant land (ha) 1.21** 0.92* 0.21 0.89 -3.60 -4.45 Ratio of irrigation times ratio of area planted with MV 10.54 -0.39 -0.71 -12.17** -26.87 -1.24 Ratio of irrigation times ratio of area planted with TV 9.03 7.58 -3.64 0.03 -17.44 3.27 Number of household members Male 23-40 years old 0.14 -0.91 -0.49 -2.29 -4.01* 2.62 Male 41-60 years old -1.63* -1.65 -0.03 -2.19 0.04 0.26 Male over 60 years old 0.48 -0.53 0.38 1.71 -4.09 -8.45 Female 23-40 years old -1.01 -1.36 0.14 -3.58 -1.71 -1.80 Female 41-60 years old -0.43 -0.79 -0.26 -0.24 -6.90* 7.58 Female over 60 years old -0.14 -1.80* -0.32 -3.04 -8.72* -7.10 Ratio of adults with lower secondary schooling -0.22 -2.97 1.38 -3.20 -18.51* 3.81 Ratio of adults with postlower secondary schooling -7.28 -1.47 -3.88 -1.53 23.93 26.01*

*=significantat5%level,**=significantat1%level. aTakenfromCherdchuchai et al. (2009,Table4.8).‘Late1980s’refersto1987and‘early2000s’refersto2004.Allincomesare calibratedinpercapitahouseholdincome. bModernvarietiesofrice. cTraditionalvarietiesofrice.

43 Table5 Schoolingattainmentofparentsandchildreninsamplevillages inthePhilippinesandThailand Description Philippines CentralLuzon a NuevaEcijaandIloilo b Late Early Late Early2000s 1970s 2000s 1980s Yearsinschooling Parents 7.0 8.1 6.2 7.7 Adultchildren c 9.2 10.7 7.2 9.6 %ofadultworkingmemberswith Secondaryschooling d 25 36 23 37 Tertiaryschooling e 13 31 8 20 Thailand f CentralPlain Northeast Late Early Late Early 1980s 2000s 1980s 2000s Yearsinschooling Parents 3.1 4.6 4.0 4.7 Adultchildren 5.7 8.9 5.7 9.1 %adultworkingmemberswith Lowersecondaryschooling g 4 9 6 11 Postlowersecondaryschooling h 4 11 2 11 aCalculatedfromthedatabaseofTakahashiandOtsuka(2009).Late1970srefersto 1979andearly2000srefersto2003. bTakenfromEstudillo et al .(2009a,Table2.3).Late1980srefersto1985andearly 2000srefersto2004. cChildrenwhoare22yearsoldandover. dRefersto710yearsofschooling. eRefersto11yearsofschoolingandover. fTakenfromCherdchuchai et al .(2009,Table4.3).Late1980srefersto1987andearly 2000srefersto2004. gRefersto79yearsofschooling. hRefersto10yearsofschoolingandover.

44 Table6 DeterminantsofschoolinginvestmentsinsamplevillagesinthePhilippinesandThailand Variable Philippines Thailand c NuevaEcija Central CentralLuzon a andIloilo b Plain Northeast 19792003 198589 200204 2004 Model1 Model2 Sizedofpawnedoutland(ha) 0.19 1.02** Farmincomeinthebaseyear(1985or2002) 0.14** 0.00 Nonfarmincomeinthebaseyear(1985or2002) 0.05 0.02** Predictedfarmincomein1979 0.05* Predictednonfarmincome1979 0.04* Ownedlandin1979(ha) 1.15** Leaseholdlandin1979(ha) 0.04 Sharetenantlandin1979(ha) 0.28 Irrigationratioin1979 0.54* Cultivatedlandin1987(ha) 0.03 0.11 Proportionofownedlandin1987 2.81 1.46 Proportionofleaseholdlandin1987 4.34 0.16 RatioofirrigationtimesratioofareaplantedwithMV 1 0.29 0.43 RatioofirrigationtimesratioofareaplantedwithTV 2 3.48* 1.25 Completedyearsinschooloffather 0.14* 0.15* 0.00 0.01 0.23 0.05 Completedyearsinschoolofmother 0.10 0.08 0.03** 0.01 0.33* 0.13 **=significantat1%level,*=significantat5%level. aTakenfromTakahashiandOtsuka(2009,Table3.7).Dependentvariableisincrementalyearsinschoolbetweentwosurveyyears. bTakenfromEstudillo et al .(2009a,Table2.9).Dependentvariableisincrementalyearsinschoolbetweentwosurveyyears. cTakenfromCherdchuchai(2009,Table4.9).Dependentvariableiscompletedyearsinschool.

45 Table7 Determinantsofcurrentoccupationalchoiceofchildreninsamplevillages inthePhilippinesandThailand(marginaleffects) Variable Central Luzon, Philippines, in early 2000s a Rural nonfarm Migrate Abroad Year of birth 0.01 0.10 0.03 Year of birth squared 0.00 -0.00 0.00 Female dummy 0.15** 0.10 0.23** Education 0.01** 0.04** 0.02 Owned land in 1979 (ha) -0.22 -0.29 0.68 Leasehold land in 1979 (ha) 0.00 -0.21 0.05 Share-tenant land in 1979 (ha) -0.04 -0.27 -1.14 Irrigation ratio 0.01 0.02 0.08 Father's education (years) -0.00 0.00 0.00 Mother's education (years) 0.00 -0.01 -0.00 Central Plain, Thailand in early 2000s b Farm Nonfarm Self-employed Age (years) -0.03 -0.13 0.16 Age squared/100 0.04 0.19 -0.24 Female dummy -0.01 0.03 -0.01 Dummy for lower secondary schooling -0.02 0.06 -0.04 Dummy for postlower secondary schooling -0.21* 0.35** -0.14** Cultivated area in 1987 0.01 -0.01 0.00 Ratio of owned land in 1987 -0.02 0.00 0.02 Ratio of irrigation times ratio of area under MVs c 0.50* -0.29 -0.21 Ratio of irrigation times ratio of area under TVs d 0.48* -0.31 -0.18 Father's education (years) 0.00 -0.03 0.02 Mother's education (years) -0.04* 0.05** -0.01 Northeast, Thailand in early 2000s b Farm Nonfarm Self-employed Age (years) -0.09 0.03 0.06 Age squared/100 0.15 -0.07 -0.07 Female dummy 0.03 -0.04 0.02 Dummy for lower secondary schooling -0.02 -0.09 0.11 Dummy for postlower secondary schooling -0.17** 0.12 0.04 Cultivated area in 1987 0.01 -0.03 0.02 Ratio of owned land in 1987 -0.07 0.16 -0.09 Ratio of irrigation times ratio of area under MVs -1.00 -1.37 2.37 Ratio of irrigation times ratio of area under TVs -1.58 0.79 2.37 Father's education (years) 0.02 -0.01 -0.01 Mother's education (years) -0.03 0.03 0.00 aTaken from Takahashi and Otsuka (2009, Table 3.8). Early 2000s refers to 2003. bTaken from Cherdchuchai et al . (2009, Table 4.10). Early 2000s refers to 2004. cRefer to improved varieties. dRefer traditional varieties. Table8 Determinantsofnonfarmearningsinthesamplevillages inthePhilippinesandThailand Variable Central Luzon, Philippines in early 2000s a Rural Manila Model 1 Model 2 Model 1 Model 2 Education (years) 0.06** 0.13** Dummy for secondary schooling -0.12 0.66** Dummy for tertiary schooling 0.43** 0.73** Experience (years) 0.07** 0.07** 0.02 0.04 Experience-squared -0.002* -0.002* -0.00 -0.00 Female dummy 0.25 0.27* 0.05 0.28

Thailand in early 2000s b Central Plain Northeast Nonfarm Self-employed Nonfarm Self-employed Dummy for lower secondary schooling 0.27 0.00 0.03 0.21 Dummy for postlower secondary schooling 0.27* 0.53 0.69** 0.50 Age (years) 0.15 -0.24 0.25* 0.64 Age squared/100 -0.17 0.43 -0.38 -0.97 Female dummy -0.17 -0.37 -0.12 -0.34 aTaken from Takahashi and Otsuka (2009, Table 3.9). Early 2000s refers to 2003. bTaken from Cherdchuchai et al . (2009, Table 4.11). Early 2000s refers to 2004.

47 Figure1 AdoptionofmodernriceandriceyieldinselectedAsiancountries,19602007

Adoption of modern varieties of rice (% area) 100 Philippines Thailand 80 Indonesia Vietnam

60

40

20

0 1960 1970 1980 1990 2000

Rice yield (tons per ha) 5.00 Philippines Thailand 4.00 Indonesia Vietnam 3.00

2.00

1.00

0.00 1960 1970 1980 1990 2000 Data source: World Rice Statistics online .

48 Figure2 Anillustrationofthelinkagesbetweenthefarmandnonfarmsectors Developmentofthenonfarmsector Nonfarmincome Humancapital Investment GreenRevolution Farmincome Accesstoland Source: Otsuka et al., (2009b, Figure 1.1).

49