Risk Management Strategy for Agricultural Production ... Num2... · Southeast Asian Journal of...

Southeast Asian Journal of Economics 4(2),July-December2016:21-45

Risk Management Strategy for Agricultural ProductionInstability: A Case Study of Paddy Production

Vanxay Sayavong

National Economic Research Institute (NERI), Vientiane Capital, Lao PDR.Email: [email protected]

Abstract WhileLaosfacesmoreintensiverainfall,frequentseverefloodsandrun-offs from theMekong river basin, its agriculture sector is likely to be increasingly vulnerable to floods leading to instability of agricultural production especially paddy production. As a result, the instability of agricultural productionwill affect theEconomicVulnerability Index (EVI)whichisregardedasoneofthecriteriaforleavingLaosoutfromthelistofLeast Development Countries (LDCs). Therefore, this research seeks to identify themost effective riskmanagement strategies for farmers againstfloodsonpaddyfarmsinthewetseasonasacasestudyinordertostabilizetheagriculturalproductionfromnaturaldisastersandtoimprovetheEconomicVulnerabilityIndex(EVI)ofLaos.Datafrom369farminghouseholdswerecollectedinvulnerablelocations(flood-proneareas)andquantitativeanalysisviaordinaryleastsquare(OLS)estimationwasused.Theresultsindicatethatcalendaradjustmentistheonlystrategythatisabletoreducetheriskfromfloods.Therefore,calendaradjustmentshouldbeencouraged.Also,thisstudysuggeststhatmoreattentionneedstobepaidtoimprovingthewaterorfloodcontrol systems such as drainage and embankments in flood prone areas. Lastly, it should be aware that the use of anti-flooded seeds should be dependedverymuchtothenatureoffloods.

Keywords: Laos,riskmanagement,floodsandpaddyfarm

JEL Classification:Q12,Q15

22 • Southeast Asian Journal of Economics 4(2), July-December 2016

1. Background Laosisvulnerabletoextremesofweather,Countriesrelyingmoreonagriculturalsectorareeasilyvulnerabletoshocksespeciallynaturaldisasters,making its production unstable. This instability in agricultural productionwould,then,directlyorindirectlyhaveanimpactonincome,livingconditionsand livelihood of people as well as economic and human capital development ingeneral.Inparticular,theannualfloodintheMekongregionhascausedanimmensedamagetolocalcommunitiesinvulnerableareasaroundtheriverbasin in terms of loss of life, properties, infrastructure, and human capitalEconomicactivities areaffectedparticularly theagricultural sector suchasfisheriesandcropsonwhichtheruralpeopledependthemost.AnestimatedannualcatastropheduetofloodsintheLowerMekongBasinregioncostsupto70milliondollars1, orwhichCambodiaandVietnambear two-thirdsoftotalcost.Globally, thenumbersoffloodshasincreasedconsiderablyfrom 50per year in the 1980s to 150 per year in the 2000swhich is driven by populationgrowthandclimatechange(IndependentEvaluation,2012).Thepopulationgrowth leads to poor urbanization, rural landuse, deforestationand land erosion facilitating the natural disasters whereas climate change resultsinanincreaseinfrequencyofnaturalhazardssuchasstorms,drought,flood and fire.According to a report of Lao PDR ResilienceAgriculture SectorClimateChange(p.22),climatechangecouldleadtoalongerannualdryseasonandmoreintensiverainfall,whileatthesametimecausingmorefrequentseveredroughtsandfloods.Whiletheacknowledgementofclimatechangeonriceproductioniswellrecognizes,desirablepolicyinterventionsarestilllimitedandsegmented(Chen.ZandDamen.B,2014,p.23).

SinceLaosischaracterizedasagriculturaleconomywheretheshareof agricultural sector accounts for more than ¼ of GDP2 and majority of population (71 per cent3) engaged in agricultural activities such as crop plantation, animal rearing, etc., it is unavoidable for Laos to expose itself totheinstabilityinagriculturalproduction,particularlyinrecenterawheretheoccurrenceofnaturaldisastershasbeenmorefrequentduetochangein

1 http://www.mrcmekong.org/topics/flood-and-drought/,viewed7thApril20152 LaoStatisticsBureau(2013).3 LaoAgriculturalCensus(2010/11)

Vanxay S., Risk Management Strategy for Agricultural Production Instability • 23

globalclimate.Especially,paddyisoneofthepromisingcashcropscontribut-ingtonationaleconomysubstantiallywhichaccountsforaquarterofGDPandoverhalfofagriculturalGDP(RIMES2011,p.1).Asaresult,thereis nodoubt thatpaddy farm is essential for economicgrowthand indeed foragriculture sector for several years, and either be an important source of incomesforruralpeopleespeciallyinCentralandSouthernpartsofLaos.

TheclimateinLaosiscategorizedintotwoseasons,oneisthewetseason influenced by the monsoon rain beginning from the mid-May to mid-Octoberandanother is thedryseasondominating therestof theyear. Asaresult,paddyfarmsinthelowlandplainsarevulnerabletofloodsduringthewetseason.AccordingtoWorldBank2011,themeanannualrainfallisprojectedtobeincreasedduringthewetseasonfor10-30%especiallyintheeasternandsouthernpartsofLaos.Yet, there isa threat fromanotherwaywheretheannualrunofffromtheMekongriverbasinisalsoanticipatedlikelytoincreaseforaround21percentby2030.ThecentralandsouthernprovincesofLaos,whichareplain-areaswithscatteredvalleys,anabundanceofriversandalongstretchofMekongRiver;areregardedasthemostvulnerabletoheavyfloods.Asaconsequence,paddyproductionisgreatlyinfluencedbytheflood.According toNationalEconomicResearch Institute (2013), damagedue to typhoon and flooding in 2013, for example, have beenwidespread coveringareasof12provincesendingupwithanestimatedeconomiclossof2,243billionkipor0.27billiondollars(11percentofagricultureGDPand3.2percentofGDP).Thelossofagriculturalproductionsuchaspaddyisthemostclaimedresultingintheinstabilityofriceproductionofhouseholds.

Becausetheinstabilityinagriculturalproductionaffectedbynaturaldisastersbecomesoneofsub-indicativeindexeswithinEconomicVulnerabilityIndex(EVI)thatLaosneedstoimproveinordertograduatefromthelistofLeastDevelopedCountries(LDCs)alongwithothercriteriasuchasHAIandGNI(UN2008).Therefore,improvingEVIthroughenhancinghousehold’sagricultural production more resilient to natural disasters is necessary.At the same time, the prevention or riskmanagement strategy against naturaldisastersisconsideredasacost-effectiveapproach.AccordingtoIndependentEvaluation 2012, disaster risk reduction is regarded as a cost-effective approach rather than post-disaster coping strategies shows that one dollar


invested in the risk reduction could save money of four dollars for the cost of a disaster recovery. Therefore, this research concentrates on the study of disasterprevention/riskmanagement strategies as themain researchobjec-tive.To achieve the objective, two research questions can be identified as (1)Identifyingriskmanagementorpreventivemeasuresappliedbyhouseholdsinthefloodedproneareas.(2)Linkingthepreventivemeasuresandhouseholdcharacteristicswiththedegreeoftheimpactandidentifythemosteffectivemeasures lessening the impact.Paddyfarmin thewetseason isgivenasacasestudy.Paddyreferstotheunpolishedrice.

Hence,thestudyisstructuredinto7sections.Backgroundandidentifiedproblemsarepresentedinthefirstsection.InformationofnaturaldisastersinLaosandliteraturereviewaredescribedinsector2whereasthemethodologyanddataarepresented insection3and4.Thestatisticsfromthesurvey issummarizedinsection5andthequantitativeanalysis(econometricsresult)isinsection6.Therefore,lastsectionistheconclusion.

2. Natural Disaster and Literature Review It is likely to be that a number of natural disasters has increased substantially during the last two decades. Since 1990 to 2012, Laos had sufferedapproximately2,331naturaldisastersnationwide;ofwhichoverahalfwasflooding.As shown inFigure 1, a number offloods had covered 56.5percentoftotalnumberdisastersovertheperiodof1990to2012whichhad caused around 601.9millionUS dollars or 65.9 percent of total cost.Storm and drought become the second and the third ranking in term of numbersanddamages(value).Overtheperiod,thedamagesincludenotonlyahuglossofproperties,economicactivitiesandinfrastructuresuchashouses,assets,crops,livestock,roads,waterandelectricityinavalueof938.2millionUSdollarsoraround40millionUSdollarsannually;butalsoinvolveadeathof 191 people, injured 46,896 people and left 38 people still missing. In particular,thestormofHaimaorEgayandNock-tenorJuaningthathitLaosin2011hadcost297.3millionUSdollarsor0.45percentofGDP.ThestormofHaimahadgonethroughthenorthernandcentralprovinces(Sayaboury,Xiengkhouang, Vientiane and Bolikhamsay) in the end of June whereas Nock-tenswipedthecentralandsouthernprovinces(Vientiane,Bolikhamxay,


Khammuan,SavannakhetandChampassak)inthelateJuly.Thecombinationcostofbothstormsincludesthedamagesofvehicles,buildings,bridges,watersupply systems, electricity networks, and infrastructure. Moreover, over63,000 hectares of farmland and 400 irrigations had been damaged or destroyed. At the same time, 25 villages and 15 schools were remained underwaterfortwodaysafterthestorms.

Figure 1.TypesandcostofnaturaldisastersinLaoPDRduring1990to2012

Numbersofnaturaldisasters Damages(value)bynaturaldisasters byshare(%) asashare(%)

Source:DisasterInformationManagementSystem(Desinventar),

http://www.desinventar.net/DesInventar/profiletab.jsp?countrycode=lao,viewed15thMay2015.

Regarding to the flood, though the numbers of floods had been fluctuatedgreatlybutitstrendhadbeeninanupwardasthesameasthelossin values. For instance, a number of floodswas over 100 times occurring annuallysince2008whichjumpedfrom12timesin1995beforemovingdownto32in2012.Intermsofdamages,thevalueoflosswasmorethan200timesin thecurrentprice.Under this circumstance, thecapital cityofVientiane,Xayaboury,KhammuaneandChampasackprovincewerethemostsufferersintermsofnumbersandlosseswithanapproximately18millionUSdollarsand41.2thousandhectaresoffarmlandperyear.In2014,accordingtothelatestreportfromtheDepartmentofDisasterManagementandClimateChange


underMinistryofNaturalResourcesandEnvironment(MoNRE),morethan90,000residentsinseveralprovinces(Phongsaly,Luangnamtha,Luangprabang,Hauphane,Xayaboury,Xiengkhuane,Khammoun, Savanhnakhet, Saravan,Xekong,Champasack andAttapue province)were affected by 3 executivefloodsduetostormsgeneratedbysoon-moonseasonduringJunetoOctoberresultinginanextremelylossof12millionUSdollars.Thedamagesincludefarmland,fisheries,infrastructure(road,waterandelectricity),andschools.

Basedontheliterature,arangeofhousehold’scopingstrategiesafterthe shocksofnaturaldisasters canbe identified.For instance,WorldFoodProgramme (2007) shows that agricultural households in Laos apply the copingstrategiesofconsumptionadjustment,borrowingandaskinghelpfromrelativesandfriendsand,toextension;theirsavingisusedagainstshocksoffloodingandlandslides.Beegle.K,Dehejia.H.RandGatti.R(2006)foundthatchildlaborswereusedascopingstrategiesagainstunexpectedcroplossamong agricultural households in Tanzania. Moreover, Mukherjee. S &Nayyar.S(2011)identifiesthatconsumptionreductiononfoodsandeducationarethecopingstrategiesofhouseholdsinthecaseofPhilippinesandKenya.Lastly,Chen.ZandDamen.B(2014)showveryusefulinformationofclimateadaptationandriskmanagementcarriedoutbyneighbouringcountries.GivenanexampleofChina,adjustingcroppingpatternstoexpandricecultivation,breedingnewricevarietieswithlongergrowingperiodsthataretolerantto awider range of temperatures, improving irrigation and drainage systems,especiallypromotingalternativewetting-and-dry(AWD)irrigation;pestanddiseasecontrol;fertilizerapplicationandmechanizationarethetoppriorities.Unlike,Vietnamgivesadifferentapproachemphasizingmoreonparticipatorycommunityplanninganddecision-makingonadaptationstrategiesandactionplans,developmentofadaptivelivelihoodforsmallholderricefarmersinthecoastalareas suchascombined rice-fish, rice-duck farmingandalternativerice-vegetable farming, breeding and adoption of early-maturing and salt-tolerantricevarieties;implementationofintegrateddisasterriskmanagementprojectsandimplementationofirrigationmodernizationprojects.

Nevertheless,previousstudiesontheissueofpreventivemeasuresorriskmanagementagainstthefloodarerarelyfoundespeciallyinthecaseofLaos.However, there isadocumentofprojectproposal4on improving the

4 MoredetailspleaseseeUNDP(2006)inthereference.


resilience of agricultural sector to climate change, which is regarded as a usefulliteratureforthisstudy,theevaluationofsuchprojectontheimpactofriskmanagementprogramsorpreventivemeasuresonagriculturalproductionstabilityinparticularpaddyproductionagainstnaturaldisastersrevealsthat,byprovidingflood/drought tolerant ricevarietiessuchasTDK1,TDK1/1,TDK8andTDK11for110hectaresinthetargetedareas,theaverageyield is higher than the local seedusedbefore theproject for aboutone tonperhectare5.

3. Methodology3.1 Theoretical Concept

ThisstudyisbasedonthetheoreticalconceptofStochasticProductionFrontiersAnalysis(SPF)initiatedbyAigner,Lovell&Schmidt(1977)andMeeusen&vanderBroeck(1977).Despitethemainobjectiveofstochasticproduction frontier is to estimate the technical

inefficiency(internalshocks)groundedonthemaximumlikelihoodestimate(non-linear estimation), the external shocks can also be investigated. TheoverallconceptoftheStochasticProductionFrontierisillustratedthroughtheFigure2.

Figure 2.TheStochasticFrontierProductionFunction

Source:editedCoelli.T,Rao.PandBattese.G,2003,p.186

5 Someachievementsandotherprojectinterventionscanbefoundonthewebsiteofhttp://www.la.undp.org/content/lao_pdr/en/home/operations/projects/environment_and_energy/IRAS/.


Figure2demonstratestheconceptofproductionfrontierwherethehorizonaxis-x representstheinputs;theverticalaxis-y is denoted as output and where i and j refer to the observed farm household i and household j respectively.Theproductionfunctionofequation(1)isassumedasdiminishingreturn toscale.Household i,as thesameashousehold j,uses inputs(x) toproduce outputs (y**)atthefrontierlevelasintheequationasfollows:

Y**it = f (Xit ,β) evit -uit (1)

Where,i isthenumberoffarmhouseholds(i=1,2,……,n),t is time (t=1, 2,….,T),Y**

it is theoptimaloutput of paddy,Xit is a (1xk) vector of inputssuchascapital, labor,seedsandfertilizerandβ is a (kx1) vector ofparameters tobe estimated.However,with the existingof the technicalinefficiencyandgiventheamountofinputs,farmhouseholdcouldproducethe output only at y*.Atthesametime,farmhouseholdscanproduceatyonlywhentheexternalshockssuchasfloodinghaveanegativeimpact.Technicalinterpretation on the term of uit and vit are as follows:

The error term uit captures farms-specific technical inefficiency inproduction,specifiedby

uit = z it δ + wit (2)

Where, zit is a (1xm)vectorof explanatoryvariables,δ is a (mx1) vector of unknown coefficients andwit is a random variable whereas uit is obtainedbyanon-negativetruncationofN~(zitδ,σ2

u)orN~(μ,σ2u).When

uit ≥0inequation(1),itisconditionaltoensureallobservationsfallunder the stochastic production frontier. Input variables can be included in bothequation (1) and (2) as long as technical inefficiency effects are stochastic(Battese and Coelli 1995). According to Farrell (1995), the efficiency offarm’s production consists of two components (1) technical efficiency and (2)allocativeefficiency.Thetechnicalefficiencyreferstotheabilityofafarmtogaintheoptimal(maximal)outputgivenasetofinputsandtheallocativeefficiencydemonstratestheabilityofafarmtousetheinputsintheoptimalproportionsgiventheirrespectiveprices.Therefore,thetechnicalinefficiencyissourcedfromtheuseofinputsinefficientlyandnotintherightproportion(anissueofallocation).


Normally, the error term vit is assumed to be independently and identicallydistributedasN~(0,σ2

v) and captures random variation in output due to factorsbeyond thecontrolof the farmhouseholds suchasweather,markets,priceandgovernmentpolicies.However,sincethisstudyconcentratesontheimpactoffloodingonpaddyfarmsinflood-proneareasasanexternalshock,theerrortermvit is now assumed to capture the observed variation in outputdue to the impactoffloodsas theexternalshock.Fortunately, theremightbesomeresponsiveactionsagainstthefloodpracticedbyfarmersaswellaspreventiveprogramsinitiatedbythegovernmentandnon-governmentorganizationswhichcanmitigatethepotentialimpact.Asaresult,equation3investigatesthepreventiveprogram/responsiveactionthatcanminimizetheeffectoffloodsstatistically.

vit = γFit+ Ɛit (3)

Where,vit isa(1xm)vectorofimpactfromexternalshocks(flood)representedbyoutput losses.Fit is a (1xm) vector of explanatory variables representingtheresponsiveactionsorriskmanagementagainstfloodssuchasthericespeciesoffloodingresistant,landrelocation,adjustmentofcalendarplantation and building drainage and embankment (water or flood control system),γ isa(mx1)vectorofunknowncoefficientsandƐit isanerrorterm.

3.2 Econometric specification

Since this study seeks to examine the role of risk management mitigatingtheimpactoffloods(outputlosses)onpaddyfarms,theattentionispaidonlyonequation(3)whichisunderthetheoreticalconceptofstochasticproductionfrontieranalysis.Thelossesofpaddyrefertotheimpactoffloods.The data of paddy losses is collected from farming households in floodedproneareasbythesurvey.Moredetailsonthedatacollectionandthesurveyaredescribedinthesection3.2.

Equation(3)intheprevioussectioncanbeseeninamorespecificeconometricsofequation(4).Alistofriskmanagementstrategiesorpreventiveprogramsareexaminedinequation(4)inordertoidentifytheonethatareable tomitigate thepaddy lossesdue tofloods.Riskmanagement includesspeciesoffloodingresistant,farmlandrelocation,andreductionoffarmland


inthefloodedareas,calendarplantationadjustmentandwater/floodcontrolsystems (building drainage and embankment). The differences between locations, infrastructuresandhouseholds’characteristicsarealsotakenintoaccountascontrolvariables.Duetoasmallnumberofsamples,onlyrelevantcontrolvariablesareselectedtobeincludedintothemodel.Forinstance,adummyvariableofProvinceisusedtocontrolthelocationeffectsbetweenSavannakhetandChampasackprovince.Similarly,thevariablesofMainRoadandIrrigationaccessareconsideredasinfrastructureeffectswhereasavariableofHouse-ownership represents thewealthofeach farminghousehold.ThetreatmenteffectsforeachhouseholdarecapturedbythevariablesofUnderDevelopmentProjectsandRiceBankMembershiprespectively.

Thesignsof thecoefficientsareallowedtobepositiveornegativebased on the effectiveness of risk management strategies or preventive programs. For instance, if water or flood control system (drainage and embankment)inthefloodedproneareasisnotwelldesignedforcontrollingtheflowoffloods,paddylossesasanimpactofthefloodscouldbehigherthantheareawithawelldesignwaterorfloodcontrolsystem.Similarly,iftheriskmanagementofcalendaradjustmentispracticedinappropriately,theoutcomecouldbecostlyandviceversa.

Risk Management Model

Vit = con + θRMit + γXit + Фit (4)

Where,

• Vit ispaddylossforfarminghouseholdi attimet,t=2014.

• RMit capturesallriskmanagement(dummyvariables)practicedbyfarminghousehold i. Riskmanagementare(1)Adjustingtoearlycalendarplanta-tion,(2)Adjustingtolatelycalendarplantation,(3)Farmlandrelocation,(4)Reducingplantedareas,(5)Applyinganti-floodseed,(6)Watermanage-mentsystem.

• Xit arecontrolvariablescapturinghouseholds’characteristics,developmentprojects and geographies. Control variables are (1) House ownership, (2)Accesstomainroad,(3)Ethnic,(4)Off-farmactivity(familybusiness),


(5)Accesstoirrigation,(6)Underdevelopmentprojects,(7)Membershipofricebank,and(8)Province.

• con is constant term

• γ and θ arecoefficients

• Фit is errorterm,assumingasnormaldistribution

4. Data Source Household survey was carried out in the flood-prone areas wherethereisplentyofpaddyfarms.AccordingtostatisticsfromMAF,ChampasackprovinceandSavannakhetprovince(Figure3)areconsideredasthelargestpaddy producers sharing 14 percent and 22 percent of total paddy output respectively in2013.Towhatextent, theLaoAgriculturalCensus2010/11indicatesthatSukhumaandPathoumphonedistrictsinChampasackprovince,andChamphoneDistrictinSavannakhetprovinceareamongthemostpaddyplantationintensiveareas.Atthesametime,thesethreedistrictssufferatleasttwotimesoffloodingperyear.Asaresult,theyareselectedasthetargetedlocationsforthesurvey.Around368householdsintentargetedvillages6 in each district inChampasack province and twenty villages inChamphonnedistrict inSavannakhetprovinceare randomly selectedas a sampleby theprovincialstatisticscenters.ThesurveywasundertakenaftertheharvestinearlyDecember2014.

Farming households are asked to give a bundle of information particularlyrelatedtopaddyplantationandfloodssuchasoutput,paddylossesduetofloods,durationandheightoftheflood,inputsusedforplantationanddifferent strategies of risk management adopted by farmers. Household’scharacteristicssuchassize,gender,age,education,assets,income,consump-tion, saving, and infrastructure ofmain road, electricity,water, school andhealthcenterarealsotakenintoaccount.Asummaryofthesurveyisreportedinthefollowingsection.

6 Five targeted villages in Pathoumphone distruct areBanKhonenoi, BanTomo, BanMoungkhoumtarnpiew,BanHouybangliengandBanNamsaitha.Fivevillagesin Sukhuma district are Ban Korknongboua, Ban Bork, Ban Kodthanou and BanHouyphai.TenvillagesinChamphonedistrictareBanLamthean,BanLambong,BanSakeunneua,BanThameung,BanNonsithan,BanKeangphoun,BanThamoung,BanPheiykar,BanDongmeung,andBanDongnoi.


Figure 3.SavannakhetandChampasackprovinceinthemapofLaos

Source: https://www.google.la/search

5. Survey Locations Thespecificinformationofsamplinghouseholds’characteristicsandtheireconomicstructureincludingoccupationsalongwithlocalinfrastructureandgeographiesinthestudylocationsarepresentedthissectiontocapturethedifferencesbetweenhouseholds.NotethatSavannakhetprovinceisreferredto thesamplein thesurveylocationofChamphonedistrict inSavannakhetprovince and Champasack province refers to the sample in the survey locationsofPhatumphoneandSukhumadistrictinChampasackprovince.

5.1 Household’s characteristics

Itseemstobethatsamplinghouseholdsaremainlyoriginatedbeingborninthetargetedvillageratherthanrecentlymigratedfromotherpartsofthe countryand foreigncountries.The survey reports that their familyhasbeenlivinginthevillageaslongas41.6yearsasanaverage.Theaveragesize


of family’smembers is 7 persons per household implying a relative large familycomparingtothenationalleveland,either;theaverageageoffamily’smembersisaround27yearsoldillustratingasayoungfamilyaveragely.Moreimportant,theeducationofsamplinghouseholdsismostlyundertheprimarylevel. Only a very small proportion of 1.79 percent of total sample gains highereducation.Furthermore,allmostofsamplinghouseholdsareLao-ethicandBuddhismdominantlylivinginruralareawithoutmainroadaccess.

It is apparently seen that the main occupations of the sample are farmersandworkinginThailand.Thereisasmallproportionofsampleengagedasfarmlabors,governmentofficials,traders,andworkersinmanufacturingfactories.Accordingtotheinterview,ashortageoflaborsisoneofthemainproblemsforpaddyfarmswhichpusheswagesgettinghigher.Itisnodoubtfulthatthesourcesofincomearemainlyfromfarming(livestockandcrops)andmoney transfer. The survey shows that approximately 35 percent of total income is from paddy rice and crop production, about 9 percent is from livestock activity. Money transfer is not only considered as an importantsourcewhichcontributes17percentoftotalincomebutfunctionsasacreditforfarmactivitiesbesidesborrowingloansfromthecommercialbanks.

5.2 Infrastructure Access

The infrastructure access such as health and education by the sample isvaried.Thesurveyrevealsthattheprimaryeducationaccessislikelytobepreferableandaccessibleasindicatedbyashortdistanceof0.8Kmaveragelybetween the house and the school.However, the distance to access highereducationismuchlongerwithapproximately3.5Kmwhichshouldadvocatesomedifficultiesaccessingtoeducationserviceespeciallyforthepoor.Sincethe longer distance to education service means higher cost to access the service,thismightbeoneofthefactorsresultinginloweducationlevelofthesample.Also,thesurveydemonstratesthataround25.3percentofsampleislocatedlessthan2Kmfarfromlower-uppersecondaryschool.Thedistancetolocalhealthcenter,hospitalandadministrativeofficeiseitherlong.Thereisonly19.2and2.6percentoftotalsamplelocatedwithadistanceoflessthan2 Km from local health center and administrative office respectively. In addition,itisshownthatlessthanaquarterofsamplehasaccessedtopipewater and just over a quarter has accessed to electricity. Lastly, a half of


samplehastoilets.Itconcludesthatamajorityofsampleshaveloweducationsandmayhavedifficultiesaccesstopublicservicessuchashighereducation,healthcenterandadministrativeoffice.

5.3 Flood and Paddy Loss

As the main objective research of this study is to identify the most effectivestrategiesofriskmanagementforpaddyfarmerstoberesilientagainstflooding,Table1givesimportantinformationrelatedtopaddyproductionandtheflood.It illustrates thatalmostahalfofplantedareas(1.5hectares)forbothsurveylocationswasfloodedresultinginalossofalmost3tonsofpaddyas an averageper household. It can be seen that a farminghousehold had suffered thefloodfor two times in2014.At thesametime, theflood takesabout 13 days per time with the maximum case of 25 days. Heavy rain, immediatefloodandstormwerethesourcesoffloods.Moreimportant,paddyfieldswerefloodedinthestagesofpost-plantationandpaddy-grownmainlybetween July and September in 2014.Also, it is interesting to know that farmersinthesurveylocationinChampasackprovinceseemtosuffermoredifficultiesoffloodcontrolbecausetheleveloffloodsisashighas2.5meters.

Accordingtothereportonnaturaldisastersin2014fromtheDepart-mentofDisasterManagementandClimateChangeunderMinistryofNaturalResources and Environment (MoNRE), there were three natural disastersstriking Laos in 2014.The first storm had hit Laos betweenMarch 20 to April20resulting inheavyrainwhile theseconddisasterhadranbetweenAugust1-25,2014duetothenoon-soonseasonandstormleadingtofloods.ThelastonehadvisitedinLaosduringSeptember16-22.Withthisregard,thedurationoffloodsareexpectedtobemorethan20daysalthoughitmightbevariesindifferentlocations.Asaresult,thepaddythatcan’tbetolerantduringthefloodwouldbedamaged.Despitethereisanapproximated17percentofsamplespurchasingandreceivinganti-floodseedsfromaresilienceproject, itisdoubtfulforthevalidityofanti-floodseedsbecausethedurationoftheanti-floodingseedstakesabout twoweeksor13dayswhichis likelytobeshorter thanthedurationof theflood.However, theuseofanti-floodseedswillbetestedstatisticallyinsectionofeconometrics.AroundtwoquartersofthesampleinSavannakhetprovinceandaquarterofthesampleinChampasackprovincearetreatedbydevelopmentprojects.Theassistancefromtheprojects


includesprovidinganti-floodseeds,naturaldisasterpreparationtrainingsuchaspreparingtomigratepropertiesandlivestock,preparingboatsformigration,providingwaters andbasic foods such as packs of noodle, and training inagriculturesuchasreducingtheimpactfromdiseases,encouragingtheuseofnon-chemicalfertilizer.

Table 1:Paddyproductioninthewetseason(perhousehold)

List Savannakhet Std. Dev Champasack Std. Dev

PlannedareaActualOutputDamagedpaddyfieldPaddylossduetofloodsNumberofFloodsDurationofafloodLevelofafloodFarmersunderdevelopmentprojects

1.6hectare1.4Ton1.43hectare2.7Ton2.6times13.38days1.41meters34%

(1.13)(1.4)(1.17)(6.5)(3.04)(16.8)(0.7)

1.63hectare1.6Ton1.59hectare3.04Ton1.8times12.19days2.5meters16.8%

(1.1)(2.08)(1.32)(3.87)(0.9)(8.25)(5.9)

Source:Farminghouseholdsurvey2014,NERI

5.4 Early Warning

Since an early warning system is considered as one of effective diagnoses in order to enable farm households formulating the effective strategiesofriskmanagement,thesurveyshowsthatoverahalfofthesamplereceivingearlywarningsystemsbeforetheappearanceoffloodsdominantlythroughTVandRadio.Table2reportsvarioustypesofearlywarningsystemstakenby farminghouseholds in two study locations. It is clearly seen thathouseholds were warned before the flood for 4-6 days. The warning via TVnews is themostaccessible taking for58.19percentof the sample inSavannakhetprovinceand40.36percentinChampasackprovince.However,thesampleinChampasacktendstoreceivemoreearlywarningsfromvillage’sannouncementandlocalknowledgethanthesampleinSavannakhetprovince.Earlywarningsthroughradio,village’sannouncement,localknowledgeandothersourcesshare18percent,11percent,8.7percentand5.9percentofthesample inbothprovinces respectively.Nevertheless, this sta tistics impliesthatalargeproportionofhouseholdsisinaccessibletothesystemsofearly


warning.Yet,thetimeofreceivingearlywarningsseemstobeshortbeforetheflood.Asaconsequence,farmersmightnotbeabletofindanappropriatedstrategyandtakeanactionontimetoreducetheriskoffloods.

Table 2:Earlywarningsystems

List Savannakhet Champasack Days

EarlywarningviaRadioEarlywarningviaTVEarlywarningviaSMSEarlywarningviaVillage’sannouncementEarlywarningviaLocalknowledgeEarlywarningviaOthers

20.34%58.19%0%9.6%4.52%6.76%

16.87%40.36%0%

13.25%11.45%6.02%

4405.55.864.86

Note:Earlywarningthroughlocalknowledgeincludeslocalfortunetellersgroundedontraditionalhoroscopecalendar,observingthelevelofMekongriver,observingtherain(heavyrainingwouldleadtoflooding),observingthebehavior of animal such as chicken and past experience. Others include districtannouncement,callingfromsons,friendsanddaughters.


5.5 Risk Management

It is crucial to know that the sample practiced different types of strategiestoreducetheimpactsorriskoffloodssuchascalendaradjustmentfor plantation, relocating the planted areas and using anti-flooding seeds providedbyresilienceprojectsandself-purchase.Table3indicatesthatlessthan10percentofthesamplehavecarriedoutvariousstrategiesexcludingtheuseofanti-floodseedsagainst theflood.Thissuggests that therearemanyfarmersinthesurveylocationshavingnoideasofanystrategiesorpreventionstodealwith thefloodon theirpaddy farms.Under thiscircumstance, theyhave potentially their farms fallen under the risk of beingflooded and theconsequenceisthecostofpaddylosses.However,thesurveyshowsthatthereis a litterhigherproportionof the sample (over aquarter)usinganti-floodseedsagainsttheflood.Thus,itisinterestingtoseewhethertheuseofanti-floodseedscouldhelpfarmersreducingtheriskorimpactsoftheflood.Regarding


tothewater/floodcontrol/managementsystems,thesampleinChampasackseemtoapplywatercontrolsystemsontheirfarmssuchasbuildingdrainageand embankment more than the samples in Savannakhet. In contrast, the sample in Savannakhet province seem to prefer the strategies of calendar adjustmentandplantedareasreductionasriskmanagement.Around5.7percentand 4.3 percent of the sample in both provinces had built drainage and embankmentonsurroundtheirfarmswhereasabout4.6percentofthesamplehadreducedplantedareas in thefloodedprobeareas inorder tosoften the riskfrombeingflooded.Thestrategiesofcalendaradjustmentandfarmlandrelocationwerepracticebyasmallproportionofsample.

Fromtheviewoffarmers,itturnsoutthatwater/floodcontrolsystemsviaembankmentanddrainagearethemostdesirablestrategytolessentheriskfrom floods on their paddy plantation in the wet season. They claim that water/floodcontrolsystemscancontrol theflowoffloodandprevent theirfarms from being flooded.Also, an expansion of the irrigation systems is requested by farmers.They stress that such an expansion of the irrigationwouldallowfarmers toplantmore in thedryseasonby increasingplantedareas.Withthisregard,farmerscanavoidtheriskandthelossofinvestmentthatmight be caused by the flood in thewet season.Therefore, as so far, it concludes thatwater or flood control systems such as embankment anddrainage,andanincreaseofirrigatedlandinthedryseasonareconsideredasthemosteffectivestrategiestoreducetherisk.

Table 3:Riskmanagementstrategies

List Savannakhet Champasack Days

AdjustingtoearlycalendarplantationAdjustingtolatelycalendarplantationRelocatingpaddyfarmlandReducingplannedareasApplinganti-floodseedsBuildingdrainageBuildingblock(embankment)

1.6%3.39%1.13%5.08%24.29%5.8%0.8%

1.81%1.81%1.81%3.61%10.24%7.23%7.06%

24171.811.54

13(6.55)--



6. Econometrics Result This section presents the result of the hypothesis specified in theeconometricspecificationmodel(equation4).Themainpurposeistoexaminewhether there are any riskmanagement practicedby farmhouseholds thatwouldbeabletomitigatetheeffectoffloodsinordertostabilizethepaddyproduction statistically.Other factors includinghousehold’s characteristics,geographicandprojectdevelopmentarecontrolledtocapturethedifferencesbetweenhouseholdsthatmightinfluencetheexpectedoutcomes(paddyloss)indirectly.

Themethodofordinaryleastsquare(OLS)estimationisemployedandtheeconometricresultcanbeseeninTable4.Firstly,themodelshowsthatitisoverallsignificantwithahighF-valueof7.47.Thetotalobservationis340,whicharehopefullysufficientforthestatisticalestimation.Theresultindicatesthatthesignofdependentvariablesespeciallythestrategiesofriskmanagementareexpected.Althoughthevariablesofriskmanagementhaveexpectedsigns,mostofthemareinsignificantwhichimpliesthattheseriskmanagement practiced by farmers are not effective sufficiently to deal withthefloodinthesurveylocations.Forinstance,theresultindicatesthatwatermanagementsystemsuchasdrainageandembankmentisinsignificantstatisticallytocontrolthefloodoutfromthepaddyfieldsalthoughthesignisnegativeasexpected.Thismayexplainthatthewatermanagementsystemsarenotbewell designedenough incapable tohandlewith theflood in thesurveywheretheheightanddurationoffloodsishighandtakeslongdays.

Similarly,thevariableofanti-floodseedshasalsoanexpectednegativesignbutareinsignificant.Sincemanyfarmersaboutonefifthofthesample in the survey locations had used the anti-flood seeds as a strategy of risk management,itisimportanttodiscusstheuseofanti-floodseeds.Thefailureofanti-floodseedscanbeexplainedbythedurationandheightoftheseedthatcan’t be comparable to theduration andheight of theflood.Table1&3 in section5.3and5.5showthatthedurationofanti-floodseedtakesabout13dayswhile thedurationof theflood isalsoaround13dayswithamaximumof 24daysinsomeareas.Atthesametime,theleveloffloodsisalsohighas highasabout2metersonaveragewhichishigherthantheheightofpaddy(around 1.50meters).Hence, the anti-flood seeds can’t survive during the


flood.Thissuggeststhattheuseofanti-floodseedsshouldbeconditionaltothe characteristics of floods such as duration and height in flooded prone areas. Inotherwords, theseedsshouldbegiven to farmers in the locationwherethedurationandheightoffloodsareshort(lessthantwoweeks)andlow(lessthan1.50meters).Otherwise,itsuggeststodeveloptheanti-floodseedstobemoretolerantwithlongerdaysandtallerthantheheightoffloods.

Furthermore,theresultalsoshowsthattheriskmanagementoffarmrelocation is insignificant.This reflects that farmlandrelocationbyfarmersmightnotbefarawayenoughfromtheexpectedfloods.Asaconsequence,farmsarestillfloodedandpaddiesarelostasthesameasotherfarmerswhodidn’tadoptthismethod.Towhatextent,itcanbesaidthatthefloodislikelyunpredictableforfarmerstopreciselyanticipatethecoverageanddegreeoffloods.Accordingtothesurvey,itreportsthatasmallproportionoffarmershad relocated their farms far away from the vulnerable areas (expected to be flooded)for1.8kilometers.Thisadvisesthatfarmersinthesurveylocationsshouldhaverelocatetheirfarmsafurtherdistance.

Fortunately, the crucial finding reveals that the strategy of lately calendar plantation adopted by farmers is valid and regarded as the most effectiveriskmanagementstrategywhichshinesahopeofawaytodealwiththefloodonpaddyfarms.Ithighlightsthatfarmers,whoadjustthecalendarplantationinalatertimeafterfloods,havefewerpaddylossesfor50.1percentthanwhodonotadoptthisstrategy.Basically,alifecircleofpaddyplantationinthewetseasonstartsfromMaytoNovemberandDecemberforsomeareaswhichisverymuchdependedonthebeginningandthenatureoftherainfallinfed lowlands.Inaddition, it takesabout90daysor3monthsbefore theharvest. Following the survey, paddy fields were flooded in the stages of post-plantationespeciallythematurityofpaddybetweenJulyandSeptembersubjectedtodifferentlocations.SincesomefarmersplantedthepaddyafterfloodsinAugustorSeptemberrespecttodifferentlocations,theirfarmscanbeavoidedofbeingflooded.Then,theyharvestedpaddiesduringNovemberandDecemberwheretherewerenoeventsoftheflood.Thisevidencesuggeststhat a strategy of calendar adjustment after the flood can help farmers to reducetheriskoffloods.Onthecontrary,farmerswhostartedplantationearlyinMayorJunewouldprobablyharvestinAugustandSeptemberwhichwas


theperiodoffloodsresultinginpaddylosses.Similarly,farmerswhomadeanadjustmentofearlycalendarplantationhadtheirfarmsfloodedaswell.Asaresult,thevariableofearlycalendarplantationistestednotsignificantly.

It is surprised to know that a strategyof farmland reductionhas apositivesignsignificantlymeaningthatfarmerswhoimplementsthisstrategyofriskmanagementhavealossfor33.9percentlargerthanwhoarewithout.Becausefarmersareafraid to lose their investment if theystillkeeppaddyplantationinthefloodedareas.Thence,byscarifyingacertainareaofpaddyfarmthatareexpectedtobefloodedwouldhelpfarmerstosavesomeoftheinvestment.However,theresultturnsoutthatthesefarmershavelossesmorethanother farmerswho still keepplanting in thefloodedprone areas.Thereasonisthatfarmerswhokeepplantingmayhavesomesurvivalofpaddies.Therefore,astrategyofreducingacertainareaofpaddyfarminthefloodedproneareaswouldhavea larger loss.Withthisregard, itsuggests that thisstrategyshouldnotbeintroducedtofarmersasriskmanagementagainsttheflood.

Forother factorsbesides the riskmanagement strategies, it canbeseen that farmlandswith irrigationaccessaremoreflooded thanfarmlandswithoutirrigationaccess.Asaresult,farmerswhohavetheirfarmlandswithirrigationaccesshavealargerlossfor22.6percentthanwhosearewithout.Thisreflectstothecharacteristicsoffloods(longdurationandhighheight)inthesurveylocation,theexistingirrigationshouldnothelppreventorrelievethefloods.Onthecontrary,theirrigationsystemsextendthelossesduetotheland slope of irrigated areas comparing to non-irrigated areas where the farmland is inhigher slope. In the sameway,Farmers arewithmain road access and under development projects tend to have their paddy farms more vulnerable thanwhoarewithout. It isdue to the fact that farmerswhoareliving at villageswithmain road access especially in Sukhuma district inChampasackprovincehavetheirfarmsarelocatedalongthemainroadwherethedrainageispoor.Becauseofpoordrainage,thefloodcan’tbehandledandconsequently flows into paddy farms resulting in paddy losses. Since the developmentprojectstendtoselectthevulnerablefarmersastargetedgroups,farmerswhoareundertheprojectsaremorevulnerablecomparingtofarmerswhoarewithout.Finally,theresultshowsthatthemajorityofethnicgroup


hasmore capabilities or opportunities access to public services of dealingwiththefloodthantheminority.

Table 4:OLSregression

Numberofobservation:340

Paddy loss Coefficient Robust Std. Err.

Adjustingtoearlycalendarplantation# 0.1034368 0.2640207

Adjustingtolatelycalendarplantation# -0.5016118*** 0.3015363

Farmlandrelocation# -0.0120293 0.1969084

Reducingplantedareas# 0.3394376** 0.1620558

Applyinganti-floodseed# -0.1798099 0.130771

Watermanagementsystem# -0.3119612 0.3500978

Access to main road# 0.8939148* 0.1282629

Ethnics -0.6060914** 0.2539453

Accesstoirrigation 0.2260881* 0.062467

Underdevelopmentproject# 0.3872217* 0.1071151

Constant 8.781043 0.3335458

F(14,340) 7.47

Prob>F 0.0000

R-squared 0.1362

Note: *issignificantat1percent,**meanssignificantat5percentand*** is significantat10percent.#referstodummyvariable.Onlyinterestedvariablesandsignificantcontrolledvariablesarepresented.

Source:Author’sestimations


7. Conclusion Laosisconsideredasanagriculturaleconomywhichhasamajorityofpopulationlivinginruralareas.Themainsourcesofincomearepredominatelyfrom agricultural activities especially paddy production. However, due tonatural disasters particularly the flood, the agricultural production such aspaddy is easily unstable affecting the stability of farmers’ income and theEconomicVulnerabilityIndex(EVI)whichisoneofthecriteriaforleavingout from the list ofLDCs.Therefore, this research attempts to search andidentifythemosteffectiveriskmanagementtomitigatetheriskfromfloodsinordertostabilizethepaddyproductioninthewetseasonortoenablefarmersberesilienttothefloodasacasestudythroughthequantitativeresearchbasedonordinaryleastsquare(OLS)estimation.Dataiscollectedintheflood-proneareasaftertheharvestperiodwithatotalsampleof369farminghouseholds.The targeted flooded locations as a case study are Champhone district in Savannakhetprovice;andSukhumaandPatumphonedistrictinChampasackprovince.

The statistics from the survey shows that a large proportion of samplinghouseholdsdidn’treceivethesystemsofearlywarning.Atthesametime,italsohighlightsthatonlyasmallproportionofsampleshadcarriedoutriskmanagementsuchaschangingthecalendarplantationadjustment,usinganti-floodseeds,buildingdrainageandembankments,andfarmlandrelocation.The econometric result indicates that the lately calendar adjustment is the only riskmanagement that enables farmers tobe resilient to theflood.Byadjusting thecalendar latelyafter theflood, the lossesofpaddydue to thefloodcanbeeasedrelativelytofarmerswhodon’tadjustthecalendar.Sincetheresultshowsthattheriskmanagementoffarmlandreductionissignificantwithapositivesigncoefficient,thismeasureshouldnotbeencouraged.Also,the result shows that using anti-flood seeds is insignificant in the survey locations.Thismightbeduetotheshortlifeofanti-floodseedscomparedtothedurationoffloods.Inaddition,accesstoirrigationsystemsandmainroadcanallowpaddyfarmsbeingfloodedduetolowslopeofirrigatedlandandpoordrainage.Lastly, theresult indicates thatavariableofethnicgroup issignificant.


Forpolicyrecommendations,theresultsuggeststhattheimplicationshouldbevariedaccordingtothecharacteristicsoffloodsinfloodedproneareas. However, based on this study, a number of areas should be paid attentionsuchas(1)datacollectionandinformationofthefloodinfloodedprone areas should be given up-to-date, (2) based on the evidence of thisstudy,earlywarningsystemsshouldbeenhancedparticularlyfromconcernedorganizationssuchastheDepartmentofMeteorologyandHydrologyinordertomakesurethatmorefarmersinthefloodedproneareasareaccessibletoweather forecast and earlywarnings throughvarious channels such asTV,radioandvillage’sannouncement,(3)thefeasibilitystudyofcalendarplanta-tion adjustment should be carried out in order to identify the appropriated calendarforpaddyplantationconditionaltothenatureoffloodsindifferentfloodedproneareas.Thisstudyshouldinvolvetheparticipatoryfromrelevantsectors such as farmer, local and central government agency, developmentproject and agricultural expert. (4) The design of floods or water control systemssuchasdrainagesystemandembankmentinthefloodedproneareasincludingin theareasof irrigatedlandandmainroadshouldbe improved.Thewater/floodcontrolsystemsshouldbewelldesignedtobeabletohandlewiththespecificcharacteristicsoffloodsinfloodedproneareasatpresentandinthefuture.Lastly,(5)Theresultofthisstudyalsosuggeststhatapplyingtheanti-floodseedswouldbeeffectiveinthefloodedproneareaswherethenatureof floods (duration and height)matchedwith the application of anti-floodseeds.As a result, using the anti-flood seeds should base on the nature offloodsinthefloodedproneareas.

Acknowledgements The author has to thank theNationalHumanDevelopmentReport(NHDR)project toprovidefundsfor this researchandallcommentsat2nd NAFRI/IRASConference2015on22nd July2015andNHDRconsultationworkshopon23rdJuly2015.Also,theauthorwasindebtedtoDr.SthabandithInsisienmay, Deputy General Director at NERI, for his supervisory and Mr.SoulaxayBounthidethandMs.OukhamInthavong,researchassistantatNERI,fortheirassistanceoncarryingoutthesurvey.


ReferencesAigner,D.,Lovell,C.,&Schmidt,P.(1977).Formulationandestimationof

stochasticfrontierproductionfunctionmodels.Journal of econometrics,6(1),21-37.

Battese,G.E.,&Coelli,T.J.(1995).Amodeloftechnicalinefficiencyeffectsinastochasticfrontierproductionfunctionforpaneldata.Empirical Economics,20,325-332.

Beegle,K.,Dehejia,H.R.,&Gatti,R.(2006).Childlaborandagriculturalshocks.Journal of Development Economics,81(1),80-96.

Coelli,T.,Rao,P.,&Battese,G. (2003).An introduction to efficiency and productivity analysis (pp.183-192).Boston:Kluwer.

Chen,Z.,&Damen,B.(2014).Rice,climatechangeandadaptationoptions.Rice in the shadow of skyscrapers: policy choices in a dynamic East and Southeast Asian Setting, FAO, The World Bank & IRRI.http://223.27.200.5/www.fao.org/3/a-i4238e.pdf

DepartmentofDisasterManagementandClimateChange.(2014).The report on the natural disaster in 2014. MinistryofNaturalResourcesandEnvironment(MoNRE).(Laoversion)

Ferrell,M.J.(1995).ThemeasurementofProductiveEfficiency.Journal of the Royal Statiscal Society,120(3),253-290.

Heltberg,R.,Oviedo,M.A.,&Talukdar,F.(2013).World development report: what are the sources of risk and how do people cope? Insights from household surveys in 16 countries.Backgroundpaper,TheWorldBank.http://siteresources.worldbank.org/EXTNWDR2013/Resources/ 8258024-1352909193861/8936935-1356011448215/8986901-1380568255405/WDR15_bp_What_are_the_Sources_of_Risk_Oviedo.pdf.

IndependentEvaluation.(2012).Special Evaluation Study: ADB’s response to natural disasters and disaster risks.ADB.

LaoNationalEconomicResearchInstitute(NERI).(2014).Macroeconomy in 2013 and the outlook in 2014.LaoNationalEconomicResearchInstitute(NERI),Vientiane.http://www.neri.gov.la/download/ann2013_eng.pdf.


Mukherjee, S.,&Nayyar, S. (2011).Monitoring household coping duringshocks:evidencefromthePhilippinesandKenya(draft).http://www.uneca.org/sites/default/files/page_attachments/mukherjee_and_nayyar_monitoring_household_coping_during_shocks_evidence_from_the_philippines_and_kenya_1.pdf.

RIMES.(2011).Technical Report: managing climate change risks for food security in Lao PDR.RegionalIntegratedMulti-HazardEarlyWarningSystem(RIMES),FAO,Bangkok.

Rasabud,S.(2011).Agriculture sector’s adaptation to climate change in Lao PDR.WorkshoponClimateChangeand Its ImpactonAgriculture(Presentation).

The World Bank. (2011). Climate Risk and adaptation country profile: Vulnerability, Risk Reduction and Adaptation to Climate Change ‐Lao PDR.TheWorldBank.

UNDP.(2006).Improving the resilience of the agricultural sector in Laos PDR to climate change impacts.ProjectDocument.http://www.la.undp.org/content/dam/laopdr/docs/Project%20Documents/Environment%20and%20Energy/UNDP_LA_IRAS_Pro%20Doc.pdf.

UnitedNations.(2008).Handbook on the Least Developed Country Category: Inclusion, Graduation and Special Support Measures.UnitedNations.

UNDP,WREAandGEF.(2009).National Adaptation Programme of Action to Climate Change.http://www.la.undp.org/content/dam/laopdr/docs/Reports%20and%20publications/UNDP_LA_NAPA%20Final%20Report%20Apr%2009.pdf.

World Food Programme. (2007). Lao PDR: comprehensive food security & vulnerability analysis (CFSVA).World FoodProgramme. http://documents.wfp.org/stellent/groups/public/documents/manual_guide_proced/wfp203208.pdf.

Risk Management Strategy for Agricultural Production ... Num2... · Southeast Asian Journal of...

Documents

Transcript of Risk Management Strategy for Agricultural Production ... Num2... · Southeast Asian Journal of...