242

F.CANOVAANDA.SCHLAEPFER

itisimportanttomeasuretheimpactoftheEuro-MediterraneanpartnershiponMediterraneanbusinesscyclesandtheexternalitiesthatEUpoliciesmayhaveforbusinesscyclesinnon-EUcountries.

Howshouldcyclical uctuationsofasmalleconomychangewhenitbecomesmoreinterconnectedwithalargeeconomy?Theinteractionbetweentradeopenness, nancialintegrationandcyclicaldy-namicsiscomplex(seeImbs,2010,forareview).Increasedbilateraltradeisexpectedtobeaccompa-niedbyhighercyclicalsynchronization(see,forexample,CanovaandDellas,1993;FrankelandRose,1998)buttheimpactof nancialintegrationisunclear.Limitedabilitytoborrowandlendinternation-allymayhamperthetransferofresourcesacrosscountriesinresponsetochangesinproductivitygrowthandthusleadtohighercyclicalcomovements(seeHeathcoteandPerri,2004).Ontheotherhand,ifagentshavelimitedinformationorfaceliquidityconstraints,limitingcapital owsmayleadtolowercyclicalcomovements,asinvestorsmaybeunabletosimultaneouslywithdrawcapitalfrommanydestinations(see,forexample,CalvoandMendoza,2000).Specializationisalsoanimportantdimensiontoconsider,sincecountrieswithsimilarproductionstructurewilltendtodisplayhighcyclicalsynchronization,evenwhentradeand nancial owsarelimited.However,sincetradeand nancialopennessmayleadtoincreasedspecialization,morecross-borderinterdependenciesmayin-directlymakecyclicalcomovementslower(seeKraayandVentura,2007).SincesomeMediterraneancountriespossessnaturalresourcesandothershavecomparativeadvantageinlabor-intensivesectors,thislatterhypothesismayberelevantfortheregion.Ingeneral,toevaluatetheeffectsoftheEuro-Mediterraneanpartnership,onemustknowwhatcyclical uctuationsintheMediterraneanbasinlooklike,whethertheproductionstructure,thelevelofdevelopment,themonetaryregimeandtheinstitu-tionalframeworkmayexplainregionalheterogeneities,andunderstandwhatcontributestotransmit uctuations,bothwithintheregionandfromEUtonon-EUcountries.CanovaandCiccarelli(2012)providea rstlookatbusinesscyclesintheareabutdonotstudythestructureofcyclical uctuationsnorthetransmissionmechanismacrosscountries.Thispaper llsthisgap.

Wecontributetotheliteratureinthreeways.Forstudentsofbusinesscycles,weprovideaturningpointclassi cation,previouslyunavailableformanycountriesintheregionandfortheMediterraneanasawhole,andasetofstylizedfacts,summarizingfeaturesofthe uctuationsinthearea.Thesefactsaresomewhatatoddswiththewisdomrepresentedbythecurrentwaveofbusinesscyclemodelsandshednewlightonrelativedifferencesinthecyclicalcharacteristicsofdevelopedanddevelopingecon-omies.Forinternationalmacroeconomists,weprovidenovelevidenceontherelationshipbetweentrade, nanceandcyclicalsynchronizationwhichquestionsthewaycross-countrytransmissionofbusinesscyclesisthoughttotakeplace.Forpolicymakers,weprovideacharacterizationofthecyclicalheterogeneitiespresentintheregion,anecessarysteptoformulatepoliciesbetterachievingintegrationgoals,andaninterimevaluationoftheEuro-Mediterraneanpartnershipprogram.

Weuseupto vequarterlyseries(realgrossdomesticproduct(GDP),unemployment,industrialproduc-tion,realincomeandrealsales)for19Mediterraneancountriesinourexercises.Sincenotallseriesareavailablefortheentiretimespan,andsincestartingdatesareoftenirregularlydistributed,weuseaweightingschemetoaggregatevariablesspeci cturningpointsinareferencecycle.Theschemereproduceswellexistingturningpointdates,generates uctuationswithconventionalfeatures,andminimizesspuriouspatternsinthefrequencyofthecyclesandintherelativedurationofthephasesformajorEuropeancountriesofthearea.Mediterraneancyclesaredifferentfromthoseofotherregionsoftheworldanddisplayconsiderableheterogeneitiesintermsofpersistence,volatilityandcomovementsacrosscountriesandsubregions.We nd,forexample,thatcyclicalupturnsanddownturnsarenotwellsynchronized,andwhilecomovementsincreaseinthe2008–09recession,theirabsolutelevelisconsiderablybelowtheonereportedinAsiaorLatinAmerica.Inaddition,amplitudesanddurationsareverymuchcountryspeci candthecross-sectionaldistributionofoutputlossesinrecessionsisquitespreadout.Interest-ingly,theseheterogeneitiesareunrelatedtothestandardwealth,monetaryandindustrialindicatorsthatthetheoryhassingledoutascrucialtoexplainthecross-sectionaldistributionofbusinesscycle uctuations.Theyarealsohardtorelatetothelevelofdevelopmentortogeography.