EPTD DISCUSSION PAPER NO. 46 DETERMINANTS OF LAND

DETERMINANTSOFLANDUSECHANGE:EVIDENCEFROMACOMMUNITYSTUDYINHONDURASGillesBergeronandJohnPenderEPTDDISCUSSIONPAPERNO.46EnvironmentandProductionTechnologyDivisionInternationalFoodPolicyResearchInstitute2033KStreet,N.W.Washington,D.C.20006U.S.A.July1999EPTDDiscussionPaperscontainpreliminarymaterialandresearchresults,andarecirculatedpriortoafullpeerreviewinordertostimulatediscussionandcriticalcomment.ItisexpectedthatmostDiscussionPaperswilleventuallybepublishedinsomeotherform,andthattheircontentmayalsoberevised.ABSTRACTThisstudyinvestigatesthemicro-determinantsoflandusechangeusingcommunity,householdandplothistories,anethnographicmethodthatconstructspaneldatafromsystematicoralrecalls.A20-yearhistoricaltimeline(1975-1995)isconstructedforthevillageofLaLimaincentralHonduras,basedonarandomsampleof97plots.Changesinlanduseareexaminedusingtransitionanalysisandmultinomiallogitanalysis.Transitionanalysisshowsthatlandusetransitionswererelativelyinfrequentinareasunderextensivecultivation,butmoresoinareasofintensivecultivation;andthatmostchangesfavoredintensification.Econometricanalysissuggeststhatlanduseintensificationwasinfluencedbyplotlevelvariables(especiallyaltitude,slope,distancetoaroadandtenure),farmlevelvariables(humancapital,farmsize,andownershipofproductiveimplements),andbycommunityvariables(especiallypresenceoftechnicalassistanceprograms).Totheextenttheseresultsarefoundtobemorebroadlyrepresentative,theysuggestthattheremaybegoodpotentialtopromoteincome-enhancinghorticulturaldevelopmentthroughinvestmentsintechnicalassistanceandeducationinsimilarcommunitieselsewhereinHonduras.Thestudyconcludesthattheplothistoryapproachisapotentiallyvaluabletoolforinvestigatingtheunderlyingcausesofchangeinlanduseatthemicro-level.Themethodisparticularlywelladaptedtosituationswheretheavailabilityofdataispoor.Itisalsosuggestedthattheapproachwouldhaveadditionalbenefitswhenreplicatedoveralargenumberofsitesasthiswouldallowintegrationofhigherorderdeterminants(e.g.nationalpoliciesandmarketincentives)whileexpandingtheapplicabilityandrepresentativityoffindings.iCONTENTSAbstract.............................................................11.Introduction.......................................................32.ConceptualFrameworkandHypotheses..................................5CommunityFactors...............................................7FarmCharacteristics..............................................9PlotLevelCharacteristics..........................................103.Methods.........................................................13DataCollection.................................................13Sampling......................................................15AnalyticalMethods..............................................16TransitionAnalysis.........................................16EconometricModel.........................................17EconometricProblems.......................................224.Results..........................................................24CommunityLevelChanges.........................................24HouseholdLevelChanges.........................................25PlotLevelChangesandTransitionAnalysis............................29EconometricAnalysis.............................................36RegressionResults..........................................36Diagnostics...............................................38DiscussionofResults........................................425.Conclusion.......................................................50SummaryofFindings.............................................51PolicyImplications...............................................51ResearchImplications............................................54References..........................................................56Wedistinguishlandusefromlandmanagement,asthelatterreferstothe1practicesrealizedwithinaparticulartype(e.g.,whichtypeofcropisgrowninannualcropland).DETERMINANTSOFLANDUSECHANGE:EVIDENCEFROMACOMMUNITYSTUDYINHONDURASGillesBergeronandJohnPender1.INTRODUCTIONThisstudyinvestigatesthedeterminantsoflandusechangeinonecommunityinHonduras.Landuseisdefinedhereasagenerictypeoflandcover—forests,pasturelands,andannualcroplandsaresuchtypes.Changeinlandusecanhavefarreaching1consequencesforfarmers’welfareaswellastheenvironment.Forinstance,conversionofaforestorpastureintoirrigatedcroplandmayincreasefarmers’incomes,butmayalsoincreasesoilerosion,reduceplantbio-diversity,orleadtoenvironmentalpollution.Ontheotherhand,intensificationoflanduse(e.g.,throughterracing)maybeassociatedwithinvestmentsinlandimprovementthatrestorefertilitytodepletedlands(ScherrandHazell1994;Pender1998).Givensuchwiderangingimpactsitisimportanttounderstandhowlandmanagersarriveattheirlandusedecisions,sothatthepotentialbenefitscanbeexploitedwhileminimizingthenegativeconsequencesassociatedwithsuchchange.Landusedecisionsaregenerallyviewedasafunctionofbothmacro-andmicro-levelprocesses(LandUseCoverandChange[LUCC]WorkingGroup1996).Theeffectsofmacro-levelfactors(definedheretoin