Total factor productivity and the learning mechanism by exporting case study for SMEs in vietnam

Keywords:export;totalfactorproductivity;SMEs;learningbyexporting;export destination... Frameworkforanalysis Theframeworkusedforanalyzinghowexportaffectstotalfactorproductivityatfirmlevel

PoliciesandSustainableEconomicDevelo pment |1

TotalFactorProductivity andtheLearningMechanism byExporting:CasestudyforSMEsinVietn

sizedenterprises(SMEs)inVietnamfrom2005to2013.Theresultsshowthatexporthasa positiveimpacto n totalfactorproductivityatfirmlevel.Moreover,thestudyfindsreason sforanappearanceo f learningmechanismb y exporting:

(i)exportingfirmsreceivetechnologicaltransfersfromdevelopedcountries;and(ii)expo rtingfirmshaveenoughcapabilitiestoabsorbk n o w l e d g e generatedb y exporting.Th efindingssuggestthatSMEsinVietnamshouldinvestinR&Daswellasmachineryandequi pment,andactivelyexploreandaccesstodevelopedcountries.Moreover,thegovernme ntshouldsupportformalcredits,internetservices,andbusinessenvironmentforSMEstoi mprovethecapacitytoabsorbforeignknowledgethroughexporting.

Keywords:export;totalfactorproductivity;SMEs;learningbyexporting;export

destination

1 Introduction

Themicroeconomicliteratureontherelationshipbetweenexportandproductivityatfirmlevelhasemergedrecently.Empiricalevidenceshowsthatexportersperformbetterthannon-

exportersi n selectionofthebestfirmsinexportmarketsorlearningbyexporting(Roberts&Tybout,

thetermsofproductivity.Thesefindingsmayreflecteitherself-1997;Cleridesetal.1998;Bernard&Wagner,1997;Cleridese t al.1998).Moreover,i n testinglearningbyexportinghypothesisi n developingcountriestheresultshavebeenquitemixedandsof a r inc onc lus ive (BernardandJe nse n,199 9; Cleridesetal.,1998;Awetal.,2000;Graner&Isaksson,2007;Boermans,2010).AccordingtothesurveybyCentralInstituteforEconomicManagement(CIEM)ofVietnamin2013,mostVietnameseenterprisesweremicrobusinesses(70%),whilesmall-andmedium-

sizedonesaccountedfor24%and6%respectively.Enterprises

intheformalandinformalsectormadeup70%and30%oftotalSMEsrespectively.Ownershipwasmainlyhouseholds(over70%).Fortheexportsector,theresultsshowedthatonlyabout6.27%ofthetotalnumberofSMEsengagedinexportsandanumberofexporterstotheEUwerethelargest(13.8%).Thefactthatasmallproportionoffirmsengagedinexportingwascontributingtoreducedcompetitivenessoftheeconomy.Theprimaryaimofthestudyi s totestthelearning mechanismbyexportingofSMEsinVietnam.Thistopicisanessentialthemei n thecontextofVietnam'sinternationaleconomicintegration.Thispaperhasthefollowingnewfeatures.Firstly,thestudyteststhelearningbyexportingmechanismandfindsoutwhylearningbyexportingmechanismhappened.Secondly,thestudyusesdatafrom2005-

2013ofCIEM.Thisdatasetc o n t a i n s manydetailedquestionsaboutexportactivitiesofSMEs.Thirdly,thestudyusestwo-

stepGMMdynamicpanelregressionanalysistocontroltheendogenousfactorsinthemodel

2 Literature reviewsand framework

2.1 Literatureandempiricalstudies

Thereweretwomainschoolsoftheorytoexplainwhytheexportbusinessaremoreefficient:thetheoryofself-

selectionmechanismandthetheoryoflearningbyexportingmechanism.Thetheoryo f self-

selectionmechanismshowedthatonlyhigherproductiveenterprisesparticipateinexportactivitiesandtheyhaveenoughcapacityoftheinternationalenvironment(Roberts&Tybout,1997;Cleridesetal.1998;Bernard&Wagner,1997).Meanwhile,thetheoryoflearningbyexportingarguedthatexporti s ano r i g i n ofthetotalfactorproductivitygrowth.Theeffectivenesso f learningbyexportingincludesknowledge,technology,andefficiencygainedfromtheexportingprocess.Theprocessoflearningtakesplacethroughtechnicalassistancefrominternationalbuyers(Grossman&Helpman,1991).Whenbusinessesengageinexportactivities,theywouldbeabletoabsorbknowledgefromtheirexportpartners.Itwouldhelptheseenterprisesimprovetheirproductioncapacity

(Bernard&Jensen,1999;Wagner,

2002).Internationalconsumersandcompetitorswouldtransferknowledgeandt

echnologytodomesticexportingfirms.Thisleadstotechnologytransferf r o mthetraditionaltechnologytomoderntechnology(Rodrik,1988;Grossman&Helpman,1991;

Cleridesetal.,1998).Whenforeigndemandrequiresacertainlevelofstandards,theimportersinthedevelopedcountriesmayprovidethetechnologyforexportingfirmsindevelopingcountries.Ther e a s o n isthattheproductiontechniquesindevelopingcountriesdonotmeetthequalitystandardso f internationalexportingmarkets.Moreover,thelearningbyexportingtheoryarguesthatexportersi ndevelopingcountrieswouldgainmorebenefitthroughexportingiftheyexportmoregoodsandservicestodevelopedcountriesthantodevelopingcountries,Thisisbecausedevelopedcountrieshavehigherqualifiedtechnology,highertechnicalrequirements,andthereforehavehighlyconditionaltransferofmoderntechnologyfortheexportersinlessdevelopedcountries(DeLoecker,2004;Graner&Isaksson,2007;Eatonetal., 2008).

Themajorityof empiricalstudieso n l y f ou nd

exportingself-selectionm e c ha n i s m ra the r thanlearningbyexportingmechanism(Bernard

&Jensen,1999;Cleridesetal.,1998).However,somer e c e n t studiesf o u n d

r e s u l t s o f learningm e c h a n i s m throughexporting(Awe t al.,2000;VanBiesebroeck,2005).Inaddition,empiricalstudiesalsofoundevidenceoftheroleofexport-

destinationtototalfactorproductivityatfirm level(DeLoecker,2004;Graner&Isaksson,2007;Eaton,2008)

2.2 Otherfactorsaffecting productivityatfirmlevel

OtherfactorswhichhaveeffectsonTFPatfirmlevelincludefirmcharacteristics,suchassize,age,ownership(Barney, 1991;Wernerfelt,1984);absorption

levelo f technologyandqualityo f humanr e s o u r c e s (Barney,1991;Cohen& Levinthal,1990);entrepreneurship(Schumpeter,1947;Audretschetal.,2006);competitioninindustry(Porter,1980;N i c k e l l , 1996);andbusinessenvironment(North,1991;Lietal.,2000;Acemoglu&Johnson,2005;Cull&Xu,2005;Augier,2012;Harrisetal.2011)

2.3 Frameworkforanalysis

Theframeworkusedforanalyzinghowexportaffectstotalfactorproductivityatfirmlevelisbasedonliteratureandempiricalstudiesofneoclassicalgrowththeory,endogenoustheory,tradetheory,institutiontheory,resourcedbasedtheory,entrepreneurshiptheory,andcompetitiveadvantagetheory.Therefore,factorsaffectingtotalfactorproductivityatfirmlevelcanbedividedintoseveralgroups:exportingactivities(Group1);firm characteristicsandabsorptioncapacities(Group2);entrepreneurship

characteristics(Group3);industrycharacteristics(Group4);businessenvironmentfactors(Group5);andindustry-andregion-relatedfactors(Group6)

Characteristic Capability(age, ownership,

TFP growth

Destination

Entrepreneurs (education, skill)

Europe Russia OtheroutsideA

Fromtheexportma rket side,thepercentageofexportingfirms totheEurope

m a rke tis thehighestbutittendstodeclineovertime(27.1%and13.8%in2005comparedwith2013).ThesecondlargestexportmarketsareotherAsianandAmericancountriesbuttheytendtodecreaserelatively(otherAsiacountries:17.7%in2005comparedto11.3%in2013;US:17.7%in2005comparedwith8.8%in2013).Th

e thirdlargestexportmarketareAseancountriesandittendstoincreaseagaininr e c e n t years(9.9%in2011comparedto13.1%in2013).ThenextexportmarketsareJapan,ChinaandcountriesoutsidetheAsiaandittendstoincreaseinrecentyears(China:9.9%in2011comparedto10.6%in2013;Japan:8.6%in2011comparedto11.3%in2013;theothercountriesoutsideAsia:5 3 % in2011comparedwith8.8%in2013).Moreover,thepercentageofexportingfirmstoRussiai s thelowestandonlyaccountfor6.4%andittendstodecreaseovertime(7.9%in2011to5%in2013).Inaddition,the

researchshowsthatexporthelpstechnologytransfer;however,anumberofbusinessessupposethatlevelofexportingthathashelpedtechnologytransfertendstodecreaseovertime(76.6%versus60.3%in2005and2013respectively)

InvestmentExpenditureofSMEs

Investment sinotherente rprises

Otheri nvestme

;Levinsohn&Petrin,2003).Therefore,theresultofT F P estimationisbias.TosolvetheproblemforTFPestimation,theestimationtechniquessuchasGMM(Hansen,1982) , OP(Olley&Pakes,1996)andLP(Levinsohn& Petrin,2003)areusedt

estimateTFPbyLevinsohnandPetrin(2003)toestimatethetotalfactorproductivityatfirmlevel

4.2 Modelfortestingthelearningbyexportingmechanism

Duetothenatureofsmallandmediumenterprisesin2005-2013thedatasetfeaturesdynamicpaneldataandtherem a y bea c o r r e l a t i o nbetweenindependentvariablesi n themodelwhenestimatingoftherelationshipbetweenexportingandTFP.ThatmeansifweestimatemodelbyOLSregression,itwillcausebiastotheresults.Toovercometheproblemofendogeneity,thisstudyusesGMMregressiontechniquesthatweredevelopedbyHasen(1982),GrilichesandHausman(1986),andArellanoandBond(1991)

m m

n n

a 3 ijtJapan 4 ijtAsean 5 ijtAsia 6 ijtAmerican

tXK ijtAsian:exportof

firmitoAseanatindustryjattimetX K ijtAsia:expor

toffirm itoAsiaat industryjattimet

XK ijtAmerican :exportoffirmi toAmericanat

industryjattimetXK ijtEupore :exportoffirmitoEuropeati

ndustryjattimetXK ijtRussian : exportoffirmitoRussiaati

ndustryjattimetX K ijtOtherAsia :exportoffirmitootherAs

iaatindustryjattimet

Formsofownershipareclassifiedintofivegroups:householdowner ship,privateenterprises,cooperatives,limitedliability,jointstockc Group2:Variablesrelatedtoentrepreneurs

EducationlevelofownershipEducationisdividedintothreelevels:primarylevel;secondaryl

evel,andhighs c h o o l level,encodedintotwodummyvariab les.Variableprimaryeducationisc h o s e n forcomparison.

variablenotechnicalexpertiseischosenasthebasevariable.

s

Lawknowledgedividedinto3levelsofgood,average,lessknownande ncodedintotwodummyvariables.hbldnisselectedascomparativev ariables.

Destinationofexport

Destinationofexportisdividedinto9levels:exporttoChina,exporttoJapan,

exporttoSoutheastAsia,exporttoAsia,exporttotheUS,exportt otheEU,exporttoRussia,andotherexportstoAsia,andnoexpor tdestinationandw a s encodedinto8dummyvariables.Noexp ortdestinationischosenasthebasevariable.

companydoesnotreceiveanysupportfromthestate.

Diversityofnetworkismeasuredbythetotalgroup(organization)which

firmscontactregularly.Thereare5maingroups:samesector,ot hersectors,b a n k i n g , government,others)anddiversityin dexrangesfrom0to5.

Qualityofnetwork

Totalnumberoftimesthatfirmreceiveshelpfromtheregularcontactsinthe

corporatenet work 5.3 Hardinfrastructure

Conditionsofcarriage

Conditionsofcarriageisadummyvariable;=1iffirm’slocationisnearmajor

roadsorrailwaysorports;=0iffirm’slocationisnotnearmajorroads, railwaysorports.

Location

Locationisadummyvariable.Locationequals1iffirmislocatedinin dustrial

parks,exportprocessingzones,ortechnologyparks;locationequ als0ifitislocatednearresidentialareas.

Internet

Internetisadummyvariable.Internet=1iffirmusesInternet;interne t=0ifit

doesnot.

Group6:Variablesrelatedtothesectorsandregions

Industry

PoliciesandSustainableEconomicDevelo pment |14

Industryisdividedinto10sectorsandencodedinto9dummiesa ndgarmentindustryischosenasthebasevariable.

Region

Regionisdividedinto3groups:North,Central,South,andencod edintotwodummyvariables.Northisabasevariable.

Year

Years2005,2007,2009,2011,and2013;encodedintofourdum mies;2005isthebaseyear

4.4 Hypothesis

Thetheoryoflearningbyexportingarguesthatexportisanoriginofthetotalfactorproductivitygrowth(Grossman&Helpman,1991;Bernard&Jensen,1999;Wagner,2002;Rodrik,1988;Grossman& Helpman,1991;Cleridese t al.,1998).Thisstudyteststhelearningmechanismbyexportingsothefirsthypothesisis:

H1:Exportingfirmshavehigherproductivitythannon-exporitngfirms

Moreover,thelearningbyexportingtheorymaintainsthatexportersi n developingcountrieswouldgainmorebenefitthroughexportingiftheyexportgoodsandservicestodevelopedcountriesm o r e thantodevelopingcountries(DeLoecker,2004;Graner&Isaksson,2007;Eatonetal.,2008).Thus,thesecond hypothesisis:

H2:Firmsexportingtodevelopedcountrieshavehigherproductivitythanfir

msexportingtodevelopingcountries

4.5 Data

DatawascollectedbytheCentralInstituteforEconomicManagement(CIEM)ofPlanningandInvestment,theInstituteofLaborScienceandSocialAffairs(ILSSA)oftheMinistryofLabor,InvalidsandSocialAffairs(MOLISA),andtheFacultyofEconomics(DoE)oftheUniversityofCopenhagentogetherwiththeEmbassyofDenmarkinVietnamin2005,2007,2009,2011,and2013.Theobjectso f thissurveywereSMEsmanufacturingin10provincesandcities

includingHanoi,HaiPhong,HoChiMinhcity(HCMC),HaTay1(earlier),PhuTho,NgheAn,QuangNam,KhanhHoa,LamDong,andLongAn.One of theu n i q u e featureswasthatthedatai nc lu de d householdenterprisesandunregistered(informal).Thus,thesurveydataincludedallinformalenterprisesoperatinginparallelw i t

)

(1.7 4)

4*

0.470

* (1.93

)

(1.75)

-0.14 (-

0.91)

1.47)

4

0.244 (1.03

(2.02)

**

(2.50 )

(2.35)

7

0.203 (1.65

(2.38)

**

value>0.5,andfromHansenandAR(2)wecouldconcludethattheregressionresultsinmodel2stepGMMsystemarereliable

ThestudyresultsshowthatexporthasarelationshipwithTFPoffirms.Accordingly,productivityo f exportingfirmsis0.792times higherthan thatnon-exportingfirmsat1%statisticalsignificance(seemodel1).Theresultsof thisstudyareconsistentwiththeempiricalstudiesonlearningbyexportingindevelopingcountrieslikeIndonesia(Blalock&Gerler,2004),China(Girma,2004;Parke t al.,2010).Next,thes t u d y fi n d thee v i d e n c e thatthedestinationso f exportals

oa ff e c t theproductivityoffirms.Specifically, exports tothe ChinaandEuropehaveimpactontheTFPoffirmsw i t h statisticalsignificancelevelsof5%and10%respectively.Theresultsofthisstudyareconsistentw i t h thestudyofGranerandIsaksson(2007)andEatonetal.(2008)

Therearetworeasonsfortheexistenceoflearningbyexporting.Firstly,exportingfirms havee n o u g h absorptivecapacitytoabsorbknowledgegeneratedbyexportactivities.Thisargumentisr e i n f o r c e d bytheresultsofthestudythatshowthatprofessionalqualificationsofentrepreneurs,investmentspendingonmachineryequipment,theratequalityofhumanresources,andtheeaseofaccessingformalcredithavepositiveimpactsonproductivityatthestatisticalsignificanceof5%

(seeresultsinmodel1andmodel2).Secondly,exportingtodevelopedcountrieshasapositiveimpactonproductivitybecausewhenfirms exporttodevelopedcountries,itwouldmotivateenterprisestoi m p r o v e theirproductioncapacityandt h e y wouldr e c e i v e thetransfero f technologyf r o m thedevelopedcountry

Forothervariables,theresultsalsoshowthatsmallandmediumenterprisesbusinesseshavehigherproductivitythanthatofmicroenterprises—

0.288and0.47timesrespectivelyat1%statisticalsignificant(model2).Formsofownershiphaveimpactonproductivityoffirms,especiallylimitedliabilitycompaniesandjointstockcompanies.Thestudyresultsalsoshowthattheownershavinghigherqualificationswouldleadtheirfirmstobehigherinproductivity.Forinstitutionalvariables,thestudydoesn o t fi n d evidenceo f theinfluenceo f f o r m a l institutionsandi n f o r m a l o n theproductivityofSMEsinVietnam,whichsuggeststhattheinstitutionsdonotimpactdirectlyontheproductivity.I t m a y bethatinstitutionscouldi m p a c t o n productivityindirectlythroughotherc h a n n e l s F o r thegroupo f variablesrelatedtohardinfrastructureandsoftinfrastructure,ther e s e a r c h showsthatthelocationofthebusinessaffectstheproductivitywithstatisticallysignificance

of0.1%;however,thesignofthisvariableisnotasexpected.Softinfrastructurevariables(includingaccessingtoformalcreditandaccessingtointernet)affectproductivitywithastatisticalsignificancelevelof5%.Forvariablesrelatedtotheindustrygroup,thestudyshowsproductivityinthemetalsectorthathasa higherproductivitythanproductivityi n thegarmentindustryw i t h statisticalsignificanceatlevelof1%,whilethebeverageindustryhaslowerproductivitythanproductivityinthegarmentindustrywithsignificancelevelatp<0.1.Forrelatedvariableregions,productivityoffi r m s insouthernregionis0.38timesthanthatoffirms intheNorth.Moreover,productivityoffi r m s intheCentralis0.4timeslowerthanthatoffirmsintheNorthatsignificancelevelof5%.

5.2 Conclusion

Thestudyfindstheimportantroleoflearningbyexportingmechanism(exportinghasimpactonproductivity),andexporttoChinaandEuropehaveapositiveeffectontheproductivityofSMEs.Therearetworeasonsforthis:

(i)firmshaveenoughcapacitiestoabsorbknowledgegeneratedbyexporting;and(ii)developedcountriestransfertechnologyf o r exportingfi r m s i n developingcountries.Next,theelementso f c h a r a c t e r i s t i c s suchassize,ownership,andqualifiedbusinesso w n e r s haveimpactsontheproductivityofbusinesses.Theinstitutionalvariablesdonothavedirecti m p a c t o n thep r o d u c t i v i t y o f SMEs.Hardi n f r a s t r u c t u r e doesn o t a ff e c t productivitywhilesoftinfrastructure(accessinginternetandformalcredit)haspositiveimpactonproductivity.Finally,i n d u s t r y andregionhaveimpactsonproductivity

5.3 Policyrecommendations

ExportplaysanimportantroleinincreasingtheproductivityofSMEsinVietnaminthecontexto f internationaleconomicintegration.However,therearetwoconditionsfortheappearanceo f learningbyexportingmechanism:

(i)enterpriseshavetoi n v e s t i n R&Dandm a c h i n e r y andequipmenttoimprovethecapacitytoabsorbknowledgefromoutsideintotheprocessproduction;and(ii)firmsactivelyexploreandaccesstodevelopedcountriestotakeadvantageoftechnologicaltransferf r o m thesecountries.F r o m thegovernment,i t i s necessarytoc r e a t e conditionsf o r enterprisestoimprovethecapacitytoabsorbknowledgebyenhancingaccesstoformalcredit,accesstointernetservicesforSMEsaswellasafavorablebusinessenvironmentandotherkindsofsupport

5.4 Limitationofthestudy

Likeanyotherempiricalstudies,thestudycannotavoidsomelimitationsinanalysis.Oneofthemc o m e s fromthenatureofthedatasetconductedbyCIEM,DoE,andILSSA.Thisdatasetonlyfocuseso n SMEsinmanufacturingfirmsandonlycoversonly10provinces.ThesecondlimitationisthatthestudyonlyusesonemethodtoestimateTFPbasedontheCobb-

DouglasproductionfunctionofSolow(1957).Fromtheaboveweakpoints,itmaybebettertoimprovetheestimationresultsbyusingotherdatathatcoverallofsmallan

methodstoestimateTFPforcheckingtherobustnessoftheresults