Figure5:AverageITAproductshareintotalexportsbytypeofITAaccession,1996-2012(percent)
vintages insteadof beingsimply mapped. . Forinstance,agiventari  linemayhavecovered alotof
ITAproducts(relativetonon-ITAproducts),whentradewasreportedintheHS1996vintage,andit
thereforewasconsideredanITAproductline.However,inlatteryears,thislinemaynotbeconsidered
anITAtarilineanymoreduetotheshifttoHS2007reporting.Thereasonischangingtradestructure:
Nowrelativelymorenon-ITAproductsmaybetradedunderthislineasaresultofsomeITAproducts
havingbecometechnologicallyobsolete.Thus,thelinesthatweconsidertobecoveredbytheITAvary
betweenvintages.WethereforerstobtainseparatelistsoftheproductlinescoveredbytheITAduring
1996-2001inHS1996,2002-06inHS2002and2007-12inHS2007.
Inanextstep,wethenmaptheHS2002andHS2007linesintoHS1996usingconversiontablesfrom
theUNStatisticsDivision(UNSD)websitetoobtainaconsistentHS1996-baseddataset. TheHS2002
lines mapintotheexact same setof lines that wealsoobtained for HS1996because the updates in
classicationmethodologywereminorbetweenthesetwovintages.ButthisisnotthecaseforHS2007,
sothatourresultingdatasetcontainsadierentnumberoftarilinesduringthetimeperiods1996-2006
and2007-12: 23(9)HS1996linesappearonlyduringtheformer(latter)period,while74HS1996lines
areincludedinallyears,therebyresultinginourabovementionedtotalof106lines.
Withthis set of ITA-related HS1996lines onhand, , we e can then obtain 6-digit HS1996 bilateral
trade owdatafor1996-2012fromUNComtrade. Weusetheimport owdataandcomplementwith
exporter-reportedmirrordata.
11
Thisgivesus3.86millionobservationsofnon-zeroITAtrade ows
covering234countries,thoughnotallobservationsareuseableinallregressionsinlightofmissingvalues
fortaris.
ThesedataontarisareobtainedfromUNTrainsinHSCombinedfortheyears1996-2012. This
11
Weapplythemirrordatawheneveracertainimport-reporterdidnotreportfortheparticularyearatall. Werestrict
themirrordatatosuchcasesonly,becauseifacountryreportsbilateraltradeintheparticularyear,butdoesn’tspecify
somelineorit is zerowhile it ispresent inthemirrordata, thenthere is not actuallyalack of reportingissue but a
dierenceinmethodologyofclassifyingproductsbetweenimporterandexporter.
10
Convert .pdf to .jpg online - Convert PDF to JPEG images in C#.net, ASP.NET MVC, WinForms, WPF project
How to convert PDF to JPEG using C#.NET PDF to JPEG conversion / converter library control SDK
convert pdf to gif or jpg; best convert pdf to jpg
Convert .pdf to .jpg online - VB.NET PDF Convert to Jpeg SDK: Convert PDF to JPEG images in vb.net, ASP.NET MVC, WinForms, WPF project
Online Tutorial for PDF to JPEG (JPG) Conversion in VB.NET Image Application
.pdf to jpg converter online; convert pdf file to jpg online
reportingin HSCombined (rather thanHS1996throughout as inComtrade) makes necessary y anad-
ditionalstep. HSCombinedgives s taris for1996-2001inHS1996,2002-06inHS2002and2007-12in
HS2007.WethereforeagainemploytheconversiontablestogeneratetarisforoursetofHS1996lines
throughoutallyears.
12
Tollinsomemissingobservations,wethenlinearlyintrapolatetarisbetween
yearsforwhichobservationsexist.
13
Asfurtherright-handsidevariables,wecollectanystandardgravityvariableswhichvaryacrosstime
withinany country orcountry-pair.
14
GDPandGDPpercapitaweretakenfromPennWorldTable
Version8.0.RTAandcurrencyunionmembershipdataaretakenfromDeSousa(2012).
15
Aremoteness
measurewascomputedanaloguetothosecommonlyusedintheliterature.
16
WTOmembershipdata
wascollectedfromtheWTOwebsite.
SturgeonandMemedovic(2010)emphasizetheimportanceofintermediategoodstounderstanding
globalvaluechains. Theydevelopanovelclassicationscheme,classifyingproductlines s for dierent
sectorsintothoseprimarilyincludingnalorintermediategoods.
17
Thisishelpfulforustoanalyzehow
ITAmembershipeectsdierbetweencountriesindierentpositionsinvaluechains{upstream(ex-
portingintermediates)anddownstream(importingintermediates/exportingnalgoods). Theauthors
providesuchaclassicationforelectronicsgoodsonHS2007basis,whichweusetosplitoursampleto
investigatehowtheITAeectsmayoperatethroughGVCs.
18
WhenconvertedtoHS1996 usingthe
UNSDconversiontables,wendthisclassicationtocover47ofour106ITAproductlines.
19
ITAproductscanbeclassiedin7broadproductcategories,asoutlinedinWorldTradeOrganization
(2012). Weresorttothesetoreducethedimensionalityofourdatasetinourrobustnesscheckforzero
trade owswhichusenon-linearPoissonestimation. Thecategoriesarethefollowing(withnumberof
6-digitHS1996linesincludedinparentheses): Computers(14),Instrumentsandapparatus(17),Parts
andaccessories(32),Semiconductormanufacturingequipment(10),Semiconductors(15),Data-storage
mediaandsoftware(9)andTelecommunicationsequipment(9).Computers,semiconductors,andparts
andaccessoriesarethemosttradedproducts,makinguparound80percentofITAproducttrade ows.
Inmanyofourregressions,weusecontrolsectorshelpusassesshowITAtradehasperformedrelative
12
WeusetheconversiontableforconversionofHS2007andHS2002toHS1996.IftherearemultipleHS2007orHS2002
linescorrespondingtoaHS1996lineinourlist,wetakeasimpleaverageacrossthesHS2007orHS2002linestoobtain
thetarifortheHS1996line.
13
FurthermorewehadtotakeintoaccountthattheEUispresentedasasinglecountryinTRAINS.Thusweappended
thedatasettoincludeallitsmembersinvariousyearstoachieveconsistentcoverageofactivesignatoriesthroughoutthe
sampleperiod.
14
Non-timevariantvariablessuchasdistancearecontrolledforbycountry-pair(-product)xedeectsinallourspeci-
cations.
15
DeSousa(2012)dataonlycover currency unionrelationships upto2009. . To o extendthe data,we addedEstonia
joiningtheEuroin2011. Aswearenotawareofanyothercountriesjoiningorexitingacurrencyunionafter2009and
before2013,weassumethatnofurtherchangesincurrencyunionmembershipoccurredafterthistime. LiketheGlick
andRose(2002)currencyuniondenition,oursisalsotransitive,i.e. ifcountry-pairsx{y,andx{zareincurrencyunions,
theny{zisacurrencyunion. ThereforewithbothElSalvadorandEcuadorhavingadoptedtheU.S.Dollar,theywould
bothbeconsideredtobeinacurrencyunionwiththeUnitedStatesaswellaseachother.
16Our remoteness measure is computedfor importers andexporters using g the standardformula, weighting bilateral
distancesbytradingpartnersharesinworldGDP(seee.g. UNCTADandWTO(2012)). . Toobtainasingleremoteness
measuresforanybilateralpairintheinterestofparsimony,importerandexporterremotenessarethenmultipliedbefore
takingthenaturallogarithm.
17
Their classicationcouldbecome part of arevisedBEC C classication, whichwill distinguish betweencustomized
intermediategoods(typicallyrelatingtotradewithinglobalvaluechains)andotherintermediategoods.
18
Thesedata onHS2007basis s werekindly providedtousby theauthors. . SturgeonandMemedovic c (2010)include
analogsonSITCandISICbasis.
19
WhentheICT(machinery)controlsectorisadded102of202(165of995)linesarecoveredbytheclassication.
11
Online Convert Jpeg to PDF file. Best free online export Jpg image
Online JPEG to PDF Converter. Download Free Trial. Convert a JPG to PDF. You can drag and drop your JPG file in the box, and then start
convert pdf pictures to jpg; convert pdf to high quality jpg
Online Convert PDF to Jpeg images. Best free online PDF JPEG
Online PDF to JPEG Converter. Download Free Trial. Convert a PDF File to JPG. Drag and drop your PDF in the box above and we'll convert the files for you.
change format from pdf to jpg; change pdf to jpg
tothatofcomparablegoodspostITAaccession. Weusetwoofsuchcontrolsectors:otherinformation
andcommunicationstechnology(ICT)goods,notcoveredbytheITA,andmachinerygoods.
For ICTgoods, , theOECD D provides adenitionwhichcovers a totalof193 productlines inthe
6-digitHS1996classication.
20
Ofthese193lines,77arealsocoveredbytheITAunderourdenition
of106lines. Thusnon-ITAICTgoods{thecontrolsector{comprise116linesandaddanother3.70
millionobservationstothedataset.
21
Meanwhile,29linesarecoveredbytheITAthatarenotconsidered
ICTgoodsbytheOECD.
22
Finally,wealsoconstructabroadmachinerycontrolsector.WeselectHSsections84,85,87,and90.
Thesecompriseelectricalandnon-electricalmachinery,roadvehiclesandoptical/photographic/precision
instrumentsandwerechosenbecausethesesectorsalsotendtobequiteintegratedinGVCs.Thisbroad
machinerysectorcomprisesallITAandICTtarilines.
23
.Itsinclusionbringsourdatasettoatotalof
28.36millionobservations.
4. EmpiricalStrategy
Thepaper’sestimationstrategyusesthebasicstructuralgravitymodeldevelopedbyAndersonand
Van Wincoop(2003), adaptedfor a panel approach and varying trade costs across goods. . Already
AndersonandvanWincoop(2004)discusstheadvantagesofusingdisaggregate,product-level,datato
accountforvaryingtrade costs andelasticities.
24
Inaddition,pair-specicobserableorunobservable
tradedeterminants mayvaryacrossdierentproducts. . Usingproduct-leveldata,asthisstudy y does,
thereforehastheadvantageofminimizingaggregationbias.
25
Ourproduct-levelpanelversionofAndersonandVanWincoop’s(2003)structuralgravityequation
is
x
ijkt
=
y
ikt
x
jkt
y
kt
(
T
ijkt
ikt
P
jkt
)
1 
k
(1)
wherethetsubscriptdenotesyears. Thevariablex
ijkt
denotesimportsofcountryj fromcountry
iofgoodk;y
ikt
isthetotalproductionofgoodkbycountryi;x
jkt
istotalexpenditureforgoodkin
country j;y
kt
is globalproductionofgoodk; ; andT
ijkt
stands for bilateraltrade costs. . Multilateral
resistance,orgeneralequilibriumeects,arerepresentedby
ikt
,theoutwardtradebarriersfacedby
countryi,andP
jkt
,theinwardtradebarriersofcountryj.Ifsuchoveralltradecostsfacedbyacountry
arehigh, itwillbe e expectedtotrademorewithany bilateralpartner at agivenbilateraltrade cost
(thanacountryfacinglowoveralltradecosts).Notably,multilateralresistancevariesovertime.
20
ThiscoverageresultswhenwecombinetheOrganizationforEconomicCooperationandDevelopment(2003)andthe
updatedOrganizationforEconomicCooperationandDevelopment(2011)denitionstoachieveabroaddenitionofICT
goodsacrosstime.
21
Inaddition,productlinesthatarecoveredbytheITAforinstanceonlyin2007-12areconsideredcontrolsectorlines
during1996-2006,ifcoveredbytheOECDICTdenition.
22
These29linescovermanifoldproducts,mainlyprintingmachinery,electrictypewritersandopticalphoto-copiers;laser
discsandmagnetictapes;electricandpowercapacitators;equipmentformeasuringliquidorgas;andpartsofaccessories
ofaforementionedproducts.
23Tobeexact,twoITAtarilines(HS381800: Chemicalelement/compoundwafersdopedforelectronics;HS950410:
VideogamesusedwiththeTVreceiver)arenotcoveredbythefourHSsections,butremaininthedatasetthroughout.
24
OtherstudieshighlightingthisincludeClausing(2001),AndersonandYotov(2010b),AndersonandYotov(2010a).
25
Wendthat suchaggregationbiascanbeconsiderablewhenreplicatingour analysisonaggregateITAtrade ows
(AppendixA).
12
C# Image Convert: How to Convert Adobe PDF to Jpeg, Png, Bmp, &
This demo code just converts first page to jpeg image. String inputFilePath = @"C:\input.pdf"; String outputFilePath = @"C:\output.jpg"; // Convert PDF to jpg.
convert pdf file to jpg; convert pdf into jpg online
C# Image Convert: How to Convert Tiff Image to Jpeg, Png, Bmp, &
RasterEdge.XDoc.PDF.dll. This demo code will convert first page to jpeg image. C:\input.tif"; String outputFilePath = @"C:\output.jpg"; // Convert tiff to jpg.
change pdf to jpg online; batch pdf to jpg converter
Thegravityrelationisconvenientlylinearinlogarithmicform,whichallowsestimationbyOLS.The
relationshipispresentedthenas:
lnx
ijkt
=lny
ikt
+lnx
jkt
lny
kt
+(1 
k
)lnT
ijkt
(1 
k
)ln
ikt
(1 
k
)lnP
jkt
:
(2)
Thistransformationdoesnotallowtoincorporatezerotrade ows.Theexclusionsofthezerotrade
owscanleadtobiaswheneverthenon-positivetrade owsarenotrandom(Helpmanetal.,2008).To
avoidthebias,wewillalsoestimatethemultiplicativeversionofequation(1)intherobustnesssection.
Asdataontotalproductionandexpenditureonspecicproductsarenotavailableandmultilateral
resistance terms arehardto measure,our model l relies s on xed eects analog to Olivero and Yotov
(2012):
lnx
ijkt
=
ikt
+
jkt
+
kt
+(1 
k
)lnT
ijkt
(3)
where
ikt
=lny
ikt
(1 
k
)ln
ikt
,
jkt
=lnx
jkt
(1 
k
)lnP
jkt
and
kt
=lny
kt
.
Next we setoutthe trade cost equation. . Bilateraltrade e costs,T
ijkt
, are e acomposite aectedby
variouspolicyandnon-policyvariables,allofwhichneedtobeaccountedforinordertoobtainaccurate
estimatesofourITAimpactsofinterest. Wepositthattheremaybevariouschannelsthroughwhich
theITAaectsbilateraltradecostsT
ijkt
andtherebyultimatelytrade ows.Werefertothesevarious
channelsaslayers. Twooftheselayersrelatedirectlytotaris. Astariscanhardlybequantiedin
aggregatedata,thisprovidesanotherimportantmotivationforusingproduct-leveldata.Furthermore,
wepositthattheITAmayhavenon-tarirelatedimpactsonboththeimportandexportside. This
resultsinthefollowingtradecostequation:
T
ijkt
=(1+t
ikt
)exp(
1
t0
ikt
+
2
imita
ijt
+
3
exita
jt
+
4
I
ijt
+
ijk
)
(4)
Thetaricostsarebrokendownintotheappliedbilateraltari(t
ijkt
)andabinarycostforhavinga
positivetari(t0
ijkt
).First,theITAmayboostimportsbyreducingtheappliedbilateraltari,ourrst
layer.Introducingtarisdirectlyasanexplanatoryvariableintheestimationidentiesthisimpact.
Thesecondlayerquantiesthatreducingtaristozeromayhaveanadditionalimpactonimports
beyondtari reduction. . Eliminatingthetari  completelymightimply adecrease intransactionand
administrativecostsrelatedtoclearingcustomsasnotari is tobepaid. . The e intuitionhereisthat
reducingtarisfrom2to0percentcouldhaveabiggerimpactthanreducingthemfrom4to2percent.
Agrowing"timeintrade"literaturehighlightsthatreducingbureaucratichurdlesincustoms,therefore
curbingassociatedclearingtimes,isexpectedtohaveasubstantialimpactontrade.
26
Therelationship
couldalsogotheotherway. First,evenwithzerotari,thecustomsprocedures s mightnotbemuch
reducedastheVAThastobepaid
27
.Second,therecouldbeextraadministrativeburdenfromtheneed
toprovetothecustomsocialsthatthegoodshouldbeclassiedunderthelinethatiscoveredbythe
agreement(HuntandHunt(2014)). Ourempiricalndingssuggestthattradebarrierreducingeects
dominate.
Thenon-taricostsarealsodissectedintotwodierentchannels: thenon-tariimpactoftheITA
onimportsandonexports.
26
FreundandPierola(2012)showthatcustomsclearingstimeshaveabigimpactontrade.
27
e.g.seeOceoftheRevenueCommissioners(2013)forEUcustomsprocedures.
13
C# Create PDF from images Library to convert Jpeg, png images to
Batch convert PDF documents from multiple image formats, including Jpg, Png, Bmp, Gif, Tiff, Bitmap, .NET Graphics, and REImage.
best pdf to jpg converter online; change pdf file to jpg
VB.NET PDF Convert to Images SDK: Convert PDF to png, gif images
Resize converted image files in VB.NET. Convert PDF to Jpg, Png, Bmp, Gif, Tiff and Bitmap in ASP.NET. Embed PDF to image converter in viewer.
convert pdf to jpg for; convert pdf file to jpg file
Theimportsidenon-tariimpact,imita
ijt
,isthusthethirdlayer.ThisITAimporterdummyonly
takesavalueofoneonlywhentheexporterisaWTOmember;thisisbecausetheconcessionsofITA
areonlyguaranteedtoWTOmembers.TheintuitionforthischannelisthattheITAmayhaveafurther
positivetradeimpactapartfromtarireductions.Thisissuggestedbytheliteraturehighlightingthat
tradepolicycertaintyhasanimpactoninvestmentandentrydecisionsofrms,includingthroughrm
location(HandleyandLim~ao,2012, 2013). . Therefore,we e alsorefer tothesetypes ofeects as ITA
commitmenteects.
Thefourthandnallayeristheexport-sideanalogue,exita
jt
,tothethirdlayer. Thenotionhere
is that location andsourcingdecisions ofrmsincreasingly depend on importavailability inlightof
globalproductionsharing. Thus,aliberalandcertaintradepolicymayfosterexports,particularlyin
theITsector,whichhasbeendocumentedtobeoneofthesectorsmoststronglycharacterizedbyglobal
productionsharing(seeMilbergandWinkler(2013)chapter2foranalysisofUSeconomy).Anecdotal
evidencesuggestsforinstancethat,inshort-listingpotentialproductionlocations,multinationalenter-
prisesmayonlyconsidersuchlocationsthatarepartofcertaininternationalagreementsorhavecertain
institutionalfeaturesthatsuggestthatrisksofbusinessoperations,includingimportingandexporting,
beingdisruptedarelow. The e exporterITAmembershipdummy exita
jt
equalsonesimply when n the
exporteris anITAmembervis-a-visanytradingpartnerashigherexports,materializinge.g. . dueto
moretechnologytransfer,shouldnotnecessarilyonlygotoWTO(orotherITA)members.
The I
ijt
matrix rst contains two variables capturing the impact ofthe ITAand d WTO on non-
members.
28
TherstcapturestheITA’simpactonotherWTOimportersthatdidnotjoin.Thisdummy
takesthevalueofonefornon-ITAWTOimporters,whentheexporterisalsoaWTOmember. When
itscoecientisnegativeitindicateslessITAproductsbeingtradedwithnon-ITAWTOimporters. If
positive,itimpliesthathigherITAgoodsimportswerecommonalsoamongWTOmembersingeneral.
Inthatcase,theadditionalimpactthatjoiningtheITAwouldhaveforaWTOmemberislower. The
seconddummyvariableissimilar. IttakesthevalueofoneifonlyonepartnerisaWTOmemberand
is often referred toas WTO trade diversion. . Ifitscoecient t takes anegativevalue,it implies that
countries,afteraccedingtotheWTO,reducetheirtradewithcountriesthatarenotmembers.
TheothervariablesintheI
ijt
matrixarecommonlyusedgravityregressorsthatvarysimultaneously
across the importer, , exporter r and time dimensions.They include dummies for joint currency union
membershipandforacommonpreferentialorregionaltradeagreement(RTA).Relatingtopreferential
trade agreements,weintroduceanother two variables. . Wenotethatmany y countries that joinedthe
ITA"passively"wereEUaccessioncountries. WithITAandEUaccessionoftenhappeningaroundthe
sametime,itispossiblethattheITApassiveexportereectcouldcapturesomeEUtradecreation,in
caseitexceedsthatofotherRTAs.ToforestallthispossibilityweincludeanEUinternaltradedummy
forexportsofcountriesthatjoinedtheEUafter1997.
29
Likewiseweaddadummyforexportstothe
USofcountriesthatjoinedanFTAwiththeUSafter1997.
Finally,equation(4)includes axedeectterm 
ijk
whichaccountsfor allcountry-pair-product
specictradedeterminants{whetherobservableorunobservable.Givenlargeheterogeneityinbilateral
traderelationships,suchaxedeecttermhasbeenstronglyadvocatedbyvariousauthors(e.g.Baldwin
28
Thistypeofeectisoftencalledwithin-WTOtradediversionoftheITAbyBoraandLiu(2010)but{asopposed
tointhe caseof preferentialtrade agreements { wendthat theterm mightbe misleading incase of theITA as its
preferences areextendedonannon-discriminatorybasistoallotherWTO members dejure andde facto oftentimesto
allothercountries.
29WhilesomeoftheEUimpactshouldbepickedupbytheRTAdummy,someauthors havepointedout,thattrade
creationmayvarysubstantiallyacrossspecicagreementsandmaybeparticularlystrongfortheEU(Eicher,Henn,and
Papageorgiou,2012).
14
VB.NET PDF - Convert PDF with VB.NET WPF PDF Viewer
Jpeg, VB.NET compress PDF, VB.NET print PDF, VB.NET merge PDF files, VB.NET view PDF online, VB.NET Convert PDF to image formats, such as PNG, JPG, BMP and
convert pdf into jpg online; best way to convert pdf to jpg
VB.NET Create PDF from images Library to convert Jpeg, png images
Components to batch convert PDF documents in Visual Basic .NET class. Support create PDF from multiple image formats in VB.NET, including Jpg, Png, Bmp, Gif
change pdf to jpg file; convert pdf page to jpg
andTaglioni(2007)).
Plugginginthetradecostequation(4)intothegravityequation(3)yieldsourestimationequation:
lnx
ijkt
= 
it
+
jt
+
kt
+
ijk
+(1 
k
)ln(1+t
ijkt
)
+ 
1
t0
ikt
+
2
imita
ijt
+
3
exita
jt
+
4
I
ijt
+
ijkt
(5)
where
l
=
l
(1 
k
)forl=1;:::;4.
Wedropthe productdimensionofthemultilateralresistance proxies 
it
and
jt
inequation(5).
Thisimpliesthatwedonotcontrolforasymmetricchangesofmultilateralresistanceacrossproducts.
Wedeemtheriskofsuchchangestobelow,giventherelativehomogeneityoftheproductsweconsider
(includinginthecontrolsectors).Moreover,theremainingsetofxedeectsremainsratherrichtocap-
turetheimportantoftheunobservableeconomicimpactsbyaccountingforimporter-exporter-product
andproduct-timeeectsinadditiontothethesemultilateralresistanceterms. Droppingtheproduct
dimensionfromthemultilateralresistancetermsisneededtoallowustoidentifyourITAcommitment
eectsofinterest,whoseidenticationreliesonvariationbetweenITAandnon-ITAproducts.Itisim-
portanttohighlightthatequation(5)canthereforeonlybeestimatedonadatasetincludingacontrol
sectorinadditiontoITAproducts. Wepresenttwoofthesecontrolsectors: other,i.e. . non-ITA,ICT
productsandmachinery. TheITAcommitmenteectestimatesresultingfromthisequationillustrate
howITAtrade owsperformedrelativetothoseofthecontrolsectorafterITAaccession.
Wealsoemploythefollowingsimpliedequationinsomeestimations:
lnx
ijkt
=
kt
+
ijk
+(1 
k
)ln(1+t
ijkt
)+
1
t0
ikt
+
2
imita
ijt
+
3
exita
jt
+
4
~
I
ijt
+
ijkt
(6)
Ithastheadvantagethatitcanallowsustofocusonthevariationwithinacountry-pairovertime.
Thiscanaddressthequestion{oftenposedbypolicymakers{howjoiningITAhasaectedacountry’s
tradecomparedtothatbeforeaccessionandgivesanotionoftheabsolutegainsinimportsorexports.
The
it
and 
jt
terms have beendroppedandonlythe country-pair-product andproduct-timexed
eectsareretained.Thisismethodogicallyrisky,becausewetherebyloseourcontrolsforthevariation
ofmultilateralresistanceovertime. However,average e multilateralresistanceduring1996-2012remains
contolledfor via the country-pair-product xedeects and we expand I
ijt
to include four additional
variablesinthissimpliedspecication.Theseadditionalvariablesare(logsof)importerandexporter
GDP,adistance-basedremotenessdummy{oftenusedasaproxyformultilateralresistanceinearly
literature {anda dummy takingthe value of one for allcountries that have joined the WTO after
1997.
30
This lastvariable isincludedtodisentangleITAimpactsfromthosecausedby generaltrade
enhancing-reformsinthewakeofWTOaccession.ThereisevidencethatrecentWTOaccessionshave
beencharacterizedbymoreonerousreformrequirements(TangandWei,2009),whichcouldalsoaect
theITAsector. Nonetheless,wearefortunatethat t omissionofthemultilateralresistancetermsdoes
havelargeimpactsonestimates.
31
Thisstrengthensourcondenceinthevalidityofsuchasimplied
regressionthatonlyexploitsITAproductdataandreliesonvariationacrosstimeforidentication.
Equation(5) andits moresimpied versionare estimatedrst for allgoods andthenseparately
for intermediategoodsandnalgoods toobtainadditionalinsightsonvaluechainstructures. . Com-
putationally, estimationof equation (5) ) has s been n challenging until recently y because it t includes four
30
IncludingGDPsasexplanatoryvariableswillpartlycapturesomeofthevariationattributabletomultilateralresistance
giventhatthelatteriscorrelatedwithcountrysize(AndersonandYotov,2010a).
31
AppendixTableA5presentsestimatesforequation(5),usingITAandcontrolsectordata,butdroppingthemulti-
lateralresistanceterms. EstimatesremainverysimilartoourTable2baseline,suggestingthatmultilateralresistancedid
notvaryhugelyduringoursampleperiod.AppendixTableA5isfurtherdiscussedintherobustnesssection.
15
sets of high-dimensionalxed eects. . Becauseofourpanelbeingunbalanced,traditionalestimation
wouldrequirethatthreesetsofxedeectdummiesbeheldinmemory. Aseachdummywouldcon-
tainthan5millionobservations,computer memoryconstraints bind. . Traditionallythese e constraints
impliedthatonlyonehigh-dimensionalxedeectcouldbeconsideredbytransformingtheestimation
equation(Greene,2003).
32
.Laboreconomistshavedevisedsolutionstothechallengesofmultiplehigh-
dimensionalxedeects,startingwithapproximationsinAbowdetal.(1999).GuimaraesandPortugal
(2010)provideaniterativetechniquetoobtainexactestimatesofequationswithtwohigh-dimensional
xedeectsinacomputationallymanageableway,whichhasrecentlybeengeneralizedtoanunlimited
numberofxedeects.
33
Finally,itisimportanttoconsiderendogeneityinourestimationequations.Thereisampleempirical
evidencethatsectorscharacterizedbyhigherlevelsofimportpenetrationreceivegreaterprotection(e.g.
Tre er(1993),LeeandSwagel(1997)),whichisinlinewiththepredictionsofpoliticaleconomymodels
oftradeprotection.
34
Totheextentthathighimport ows(ourdependentvariable)coincidewithhigh
importpenetration,wethusneedtosuspectthattheycausehighertarilevels,reducethelikelihood
ofzerotarisaswellasthatofjoiningtheITA.Inabsenceofappropriateinstrumentalvariables,the
standardinthe literature e which h wealsofollowhere e has s become to rely oncountry-pair(-product)
xed eects (Baier r and Bergstrand, 2007). . This s mitigates the e issue e as long g as s import t penetration
doesnotdramaticallychangeoverthesampleperiod.Ifimportpenetrationdoeschange,butrelatively
homogeneouslyacrosssectorswithinacountry(sayduetochangesinmacroeconomicconditionssuchas
exchangerates),thenourimporter-timexedeectsconstituteasecondlineofdefense.Anyremaining
endogeneitywouldbiasourestimatestowardzero. Thus,totheextentthatsomeendogeneityremains
despitetheextensivexedeectcontrols,ourestimateswouldneedtobeinterpretedaslowerbounds.
5. Results
Our results highlight thatthe ITA’s impact on trade consists offour layers. . Moreover,many y of
these impacts s vary depending on n whether r a country acceded d \passively". . Previous s literature only
allowedforasingleimport-sideimpactoftheITAandusedaggregateITAtrade. Thediscussionhere
focusses exclusively on regressions which account for allthese layered and heterogeneous impacts in
product-leveldata. Toallowcomparabilitytopreviousliterature,weaddtheselayersgraduallyinthe
appendix, moving from m aggregate toproduct-leveldata inthe process. . Coecientson n the standard
gravityvariablesarealsodiscussedthere.
35
Finally,thediscussionhighlightsth+atusingproduct-level
dataiscrucialtoallowforproduct-specicunobservabledeterminantsincountrypairs.
36
32Inabalancedpanel,twosetsofxedeectscouldbestrippedalgebraically.
33StatacommandreghdfebySergioCorreiaofDukeUniversityimplementsthisgeneralization. Itreliesonthenotion
thatthematricesfor thecomputationof the coecient estimates aresparseandonly identiesnon-zeroentries. . This
reducesmemoryconstraintsatthecostofhighercomputationtime.Forregressionswithmorethantwoxedeects,no
exactstandarderrorscanbederivedbytheroutine,becausetheexactnumberofabsorbedxedeectsishardtocalculate
toobtainanexactnumberofdegreesoffreedom. Sowerelyontheroutine’sconservativeboundstandarderrors,which
weclusterbycountry-pair-pruductcombinations.
34
WhiletheseminalmodelofGrossmanandHelpman(1994)predicts thathigherimport penetrationwouldactually
lead to lower r levels of protection, , Maggi i andRodrguez-Clare (2000) ) showthat t this s predictionis s reversed d whenthe
assumptions arerelaxedthat(i)tradetaxesarethegovernment’s onlypolicy instrumentandthat(ii)governmenthas
accesstonon-distortionarytaxation.
35
WesuppresstheseforreasonsofspacefromTable2,whichcontainsourmainresults. ForTable2regressions,these
coecientsremainsimilartothoseofregression24inAppendixTableA2.
36
Most notably eects s on exports may y be overstatedinaggregate data, as s comparison of f regressions s 25 and23 in
AppendixTableA2andAppendixTableA1shows.
16
Table2presents ourresults. . Goingacross,it t is dividedintothreesets ofcolumns analysingITA
impactsonall,intermediateandnalgoods.Asnotallproductlinescanbeclassiedintointermediate
andnalgoods,sampleselectioneectsimplythattheallgoodsestimatesdonotnecessarilyliebetween
theformertwo. Goingdown,itisdividedbyhorizontaldashedlinesintosectionsseparatingthefour
dierentlayersofITAtradeeects,WTOtradediversionandotherselectedcontrolvariables.
37
Table2alsoincorporatesdirectlytwotypesofrobustnesschecks.First,inadditiontothesimplied
specicationonlybasedontheITAproductdata(regressions1,4,7),weallowforourdierentcontrol
sectors ICT(regressions s 2, , 5,8)andMachinery y (regressions 3,6,9). . Thesecondrobustness s check
relatestoChina.Oneofthemainnoveltiesofourdatasetisthatitcoversasubstantialperiodoftime
afterChina’sITAaccession.ThisallowsustoanalysetowhichextentChina’sperformancehasdiered
from that ofother ITA members. . As s section 2 already y highlighted, China’s s importance in trade of
ITAproductshasincreasedimmenselyontheexportside,and,toalesserextent,ontheimportside.
Interestingly,however,allcoecientsofinterestremainbroadly unchangedwhenChina’sexportsare
excluded.
38
TheonlyexceptionisthepassiveITAexportereectandwethereforeincludeinTable2(in
greyshading)itscoecientfromtheanalogueregressionexcludingChineseexports.
39
Thishighlights
that, while e China is not very y distinct t in its s import t patternfrom other r countries, it t has s become e an
exceptionalcaseofexportsuccessinITAproductssinceitsaccessiontotheagreement.
37
AllTable2regressionsincludeataminimumcountry-pair-productaswellasproduct-timexedeects.Product-time
xedeectsensurethatestimatesarenotaectedbyglobalsupplyordemandshocksinspecicproducts.F-Testsreject
timexedeectsintheirfavourthroughoutourregressions,althoughcoecientshardlychangebetweenthetwotypesof
specications.
38
WealsoexploredwhetherresultschangefurtherwhenexcludingChineseimports. Thisisnotthecase.
39ThesecoecientsaretakenfromAppendixTableA3,whichrepresentsthecompleteanaloguetabletoTable2but
excludingChineseexports.
17
Table2:ThelayersofITAtradeeects
Typeofgoods
All
Intermediate
Final
4Fixedeects
ijk&kt
ijk,kt,it&jt
ijk&kt
ijk,kt,it&jt
ijk&kt
ijk,kt,it&jt
5Sample
ITA
ITA&ICT
Machinery
ITA
ITA&ICT
Machinery
ITA
ITA&ICT
Machinery
ExplanatoryVariables
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Layer1:
ln(1+tari)forITAgoods
-0.347***
-0.389***
-0.293***
-0.117
-0.592**
-0.327*
-0.677***
-0.194
-0.118
Tari
(-4.26)
(-4.54)
(-3.89)
(-0.72)
(-3.01)
(-2.00)
(-4.56)
(-1.24)
(-0.78)
reduction
ln(1+tari)forothergoods
-0.677***
-0.614***
-0.928***
-0.562***
-0.577***
-0.528***
eect
(-9.52)
(-22.10)
(-4.13)
(-6.24)
(-4.93)
(-6.10)
Layer2:
ZerotariforITAgoods
0.101***
0.123***
0.099***
0.130***
0.190***
0.139***
0.071***
0.090***
0.092***
Tari
(20.45)
(21.33)
(20.11)
(13.99)
(15.72)
(13.92)
(8.53)
(9.34)
(9.89)
elimination
Zerotariforothergoods
0.030***
0.008**
0.051**
0.014
0.021*
0.034***
eect
(4.81)
(3.23)
(3.05)
(1.66)
(2.07)
(3.89)
1ActiveITAImporter
0.336***
0.155***
0.159***
0.405***
0.284***
0.192***
0.339***
0.078**
0.097***
Layer3:
(11.63)
(9.18)
(11.47)
(7.34)
(7.02)
(6.54)
(6.69)
(2.77)
(3.64)
Non-Tari
1PassiveITAImporter
0.349***
0.113***
0.122***
0.478***
0.368***
0.218***
0.347***
0.023
0.061*
eectfor
(14.32)
(6.49)
(8.70)
(10.40)
(8.74)
(7.27)
(8.23)
(0.82)
(2.28)
Imports
1Non-ITAWTOImporter
0.325***
0.102***
0.094***
0.446***
0.361***
0.194***
0.265***
-0.020
0.013
(12.70)
(6.57)
(7.48)
(9.14)
(9.64)
(7.27)
(5.93)
(-0.78)
(0.52)
7TotalimpactofjoiningITA(Lin.combinationsofwithin-WTOITAtradecreationanddiversion):
Indiv.ITAIm.
0.010
0.053***
0.065***
-0.041
-0.077*
-0.002
0.074**
0.097***
0.084***
-non-ITAWTOIm.
(0.63)
(3.50)
(5.12)
(-1.36)
(-2.22)
(-0.08)
(2.57)
(4.06)
(3.66)
Exog.ITAIm.
0.023
0.011
0.029*
0.032
0.007
0.024
0.082**
0.043
0.048*
-non-ITAWTOIm.
(1.49)
(0.68)
(2.28)
(1.08)
(0.18)
(0.90)
(3.08)
(1.70)
(2.02)
ActiveITAExporter
-0.074***
0.167***
0.086***
-0.025
0.091
-0.014
-0.109**
0.028
0.018
Layer4:
(-3.57)
(8.18)
(5.11)
(-0.64)
(1.61)
(-0.35)
(-3.21)
(0.90)
(0.61)
Non-tari
PassiveITAExporter
0.429***
0.177***
0.143***
0.393***
0.195***
0.051
0.507***
0.381***
0.419***
eectfor
(25.74)
(9.09)
(9.55)
(13.20)
(3.87)
(1.58)
(17.52)
(11.48)
(13.45)
Exports
Exog.ITAExporters
0.018
0.125***
0.111***
0.060
0.085
0.071
0.081*
0.309***
0.253***
8otherthanChina
(0.83)
(4.63)
(5.58)
(1.61)
(1.13)
(1.64)
(2.20)
(6.95)
(6.20)
OneinWTO
0.070**
-0.009**
-0.049***
0.066
-0.105***
-0.080***
0.181***
0.036
0.053*
(3.12)
(-0.62)
(-6.06)
(1.61)
(-3.48)
(-3.86)
(4.45)
(1.40)
(2.43)
2ExporterlateWTOsignatory
-0.036
-0.047
-0.027
(-1.45)
(-1.02)
(-0.63)
3ExporterlateEUsignatory
0.480***
0.083***
0.063***
0.376***
-0.006
0.190***
0.823***
0.188***
0.249***
(18.36)
(3.34)
(5.65)
(7.93)
(-0.10)
(5.73)
(18.01)
(4.46)
(6.36)
3ExporterlateUS-FTAsignatory
-0.116
0.080
0.127***
-0.069
0.140
0.135
0.018
0.123
0.107
(-1.73)
(1.62)
(5.06)
(-0.58)
(1.47)
(1.93)
(0.16)
(1.50)
(1.45)
Observations
2477294
5632921
21813553
680728
1165824
2530265
825203
1970737
2397118
2R
0.8050
0.7978
0.7984
0.8238
0.8161
0.8268
0.7839
0.7936
0.8003
2AdjustedR
0.8049
0.7974
0.7982
0.8237
0.8150
0.8262
0.7838
0.7928
0.7997
Notes:Basedonrobuststandarderrorsclusteredbycountry-pair-productcombinations.Allregressionsincludeall(non-collinear)standardgravityvariablesasinAppendixTablesA1andA2(coecientsnot
reported).Thenumberofobservationsishigherinit/jtspecications,becausethesexedeectsleadtoGDPsandremoteness{whichcontainsomemissingvalues{tobedropped.
1ITAimportervariablesonlytakethevalueofoneifexporterisaWTOmemberandiftheproductinquestionisanITAproduct.
2TakesthevalueofoneforallexportersthatacceededtoWTOafter1997.
3Takesthevalueforexportsof\1"forintra-EUtrade(afteraccession)ofallcountriesthatjoinedtheEUafter1997.AnalogouslyforUSFTA.
4Setsofxedeects:kt=product-time;ijk=country-pair-product;it=importer-time;jt=exporter-time.
5ICTgooddenitionbasedonunionofOECD(2003)andOECD(2011)denitions.
6Broadsectorincludingelectricalandnon-electricalmachineryandequipment(HSsections84-85),vehicles(HSsection87),andoptical/cinematographic/precisioninstruments(HSsection90).Thesealso
compriseallbuttwo6-digitITAproductlines.
7Thedierenceofthesetwovariables{ITAtradecreationandITAtradediversionwithintheWTO{expresseshowmuchmoreITAimportersimportcomparedtonon-ITAWTOmembers.Inotherwords,this
wouldbetheamountthatacountry,whichisalreadyaWTOmember,couldexpecttoimportmorefromotherWTOmembersbyjoiningtheITA.
8ThiscoecientisobtainedfromanexactanalogregressionthatexcludesChina’sexportsfromthesample.ThePassiveITAexportercoecientistheonlyonetosubstantiallyvaryasaresultofsuchasample
modication.ThefullregressionresultsfromthisrestrictedsamplearereportedinAppendixTableA3.
18
5.1. Layer1: : Tarireductioneect
TherstofthreelayersofimpactthattheITAmayhaveonimportsisrelatedtotarireductions.
TheeectmeasuresimpactsoftarireductionswhetherornottheywererelatedtoITAaccession.
40
Theresultssuggestthataonepercentreductionoftaris onITAproductswouldcausea0.3to0.4
percentincreaseinimports.Inthecontrolsectorregressions,weallowtheseimportdemandelasticies
todierforotherICTandmachineryproductsandobtainhigherelasticitesfortheseinabsoluteterms
(-0.6to-0.7).
As expected, , these tari  elasticities are lower than most import demand elasticities s reported d in
theliteratureandderivedbasedontotaltrade,whichalsoincludesmanyhomogenousproducts. For
instance,Keeetal.(2008)andTokarick (2014)estimatesuchelasticitiesformanydierent countries
andcomeupwithaveragesintherangeof-1.1to-1.2,althoughanearlierstudybySenhadji(1998)is
relativelyclosetoourvalue,at-0.32. However,thatourvaluesarelowerismainlyduetousallowing
fornon-linearimpactsoftariliberalizationbyincludingtheseparatetarieliminationdummyinthe
secondlayer.Whenthissecondlayerisdroppedandtherebycomparabilitytothesestudiesisachieved,
our ITA import demandelasticity increasesto the-0.7to -0.8range(see regression 22 in Appendix
TableA2),whichseemsintuitivegiventhatITAproductsarefairlydierentiated.
41
Ourresultsfromthecontrolsectorregressionssuggestthattarireductioneectsonintermediate
goodsarehigherthanfornalgoods,bothforITAandcontrolsectorgoods. ForITAgoods,however,
this cannot be conrmedinthemore basicspecication without multilateralresistance controls (re-
gressions4and7). Thismaybearesultofthemuchsmallersamplesforwhichtheintermediate/nal
gooddistinctionisavailable,whichbecomesaccentuatedinthesmallerITAonlysampleusedinthese
regressions.
5.2. Layer2: : Tarieliminationeect
ReducingtaristozerohasalargeimpactonimportsofITAproducts,aboveandbeyondthoseof
tarireductions.Thatthereisanadditionalimpactofeliminatingtarisseemsintuitive,becausezero
tarisreduceborderformalitiesconsiderably. Ourestimatessuggestthattarielimination{whether
donebecauseofITAaccessionorinanothercontext{willboostITAimportsbyabout10-13percent
across allgoods.
42
Thus,makingthelasteortto reducesmalltaris,say from 1to0percent,will
achieveamuchhigherimpactthanreducingahightaribyseveralpercentagepointswithoutreaching
zero.
TarieliminationisespeciallyimportantforITAgoodsimports,moresothanforotherICTgoods
or the e broader r machinery y sector r for which we obtain tari  elimination impacts s of 3 3 and 1 1 percent
tradeincreases,respectively. ThesedierenceslikelycomeagainstthebackgroundoftheITAsector’s
especiallyhighintegrationinto valuechains,so that burdensome border formalities imply highcosts
whichgetsre ectedinlowertradevalues. Theestimatesforintermediateandnalgoodssupportthis
conclusion.Forintermediategoodstheimpactoftarieliminationishigher{inthe14-20percentrange
{likelybecausethesegoodsareparticularlyimportantinvaluechains.TarieliminationfornalITA
goodsisexpectedtoincreasesuchimportsby7-10percent.
40Inlightofourcontrolvariablesforotherlayers,thereisnoreasontobelievethattheirimpactsshouldvarydepending
onwhethertarireductionswererelatedtoITAaccessionornot.
41
Moreover,addinglaggedtarisinourrobustnessanalysis(Table3)increasesthetarireductioneectconsiderably.
42
Regressioncoecientsondummyvariables,suchasourITAmembershipvariablesofinterest,expressimpactsinlog
units,whichareverysimilartopercentagechangesforvaluesclosetozero. Theexactpercentagechangeimpliedbyany
coecient bcanbe calculatedas exp(b)-1. . The10-13 3 percent t rangementionedhere isobtainedfrom the highest and
lowestcoecientson\ZerotariforITAgoods"inregressions1-3: exp(0.099)-1=10.4%;exp(0.123)-1=13.1%.
19
Documents you may be interested
Documents you may be interested