!
!
ABSTRACT!
!
This!thesis!examines!possible!determinants!of!the!difference!in!stock!market!return! correlations!for!the!financial!crisis!that!flared!up!in!August!2007!for!20!countries.!Using!a! Wald!test!it!is!shown!that!stock!market!correlations!have!significantly!increased!during!the! financial!crisis.!Using!two!models!it!it!shown!that!financial!integration,!economic!integration,! exchange!rate!volatilities!and!an!interaction!term!of!the!volatility!of!monetary!rate!and! economic!integration!are!significant!determinants!of!stock!market!correlations.!The!results! obtained!in!this!paper!are!important!for!investors!that!are!interested!in!investing!across! borders!that!can!now!take!into!account!more!factors!to!motivate!investment!descisions.!
!WHAT!DETERMINED!STOCK!MARKET!
CORRELATIONS!ACROSS!THE!WORLD!FOR!
THE!2007!FINANCIAL!CRISIS?!
MILIAN!BACHEM!/!SUPERVISOR:!DORINTH!VAN!DIJK!
JULY!27,!2015!
!
Table!of!Contents
! ! ! Acknowledgements!...!3! I.! Introduction!...!4! I.! Literature!Review!...!6! A.! The!Markowitz!model!of!portfolio!selection!...!6! B.! Previous!research!on!world!wide!correlations!during!bear!markets!...!8! C.! Possible!factors!that!lead!to!varying!degrees!of!correlations!in!stock!market!returns!across! countries!...!9! i.! Monetary+environment+...+9! ii.! Level+of+economic+integration+...+10! iii.! Rate+of+financial+integration+...+11! iv.! Exchange+rate+risk+...+11! v.! Dominant+position+of+the+United+States,+EMU+effects+and+other+monetary+union+effects+...+13! III.! Methodology!...!14! A.! Model!...!14! B.! Variable!description!...!14! IV.! Data!description!and!hypotheses!...!16! A.! Data!Description!...!16! B.! Hypotheses!...!18! V.! Results!...!20! A.! Increase!in!correlations!...!20! B.! OLS!factors!regression!...!21! C.! Robustness!analysis!...!23! VI.! Conclusion!and!discussion!...!25! VII.!Appendix!...!27! A.! Table!...!27! B.! Test!for!normality!of!error!terms!and!serial!autocorrelation!for!model!2!...!27! VIII.!References!...!29! ! ! ! ! !Acknowledgements!
! A!number!of!people!have!been!instrumental!in!the!formation!of!this!paper!and!therefore! deserve!a!special!word!of!thanks.!! ! Firstly,!I!would!like!thank!my!supervisor!Dorinth!van!Dijk.!Without!his!time,!effort!and! dedication!this!thesis!would!simply!not!have!become!what!it!is!now.!This!has!ranged!from! the!numerous!times!that!he!has!been!willing!to!reread!segments!of!the!article!to!helping!me! formulate!the!very!outset!of!this!research.! ! Secondly,!I!would!like!to!express!thanks!to!my!parents,!who!have!supported!me!during! obstacles!in!the!research!and!have!added!interesting!ideas.!I!am!sure!their!suggestions!have! made!this!thesis!more!understandable!and!engaging.! !!
I.! Introduction!
!
One!of!the!strongest!global!downturns!in!recent!history,!resulted!from!the!financial!crisis! that!flared!up!in!August!2007!in!the!US.!A!consensus!has!been!reached!that!the!origin!of!the! financial!crisis!was!the!subprime!housing!finance!market,!which!had!been!showing!signs!of! weakness!throughout!2007.!Many!considered!the!financial!crisis!in!the!US!to!be!a!largely! domestic!problem,!since!the!subprime!market!only!represented!15%!of!the!total!US! residential!mortgages.!After!the!collapse!of!Lehman!Brothers!in!September!2008!however,! the!crisis!spread!rapidly!across!borders.!The!resulting!global!crisis!has!been!estimated!to!be! one!of!the!broadest,!deepest!and!most!complex!crises!since!the!Great!Depression.!! Research!by!Markowitz!(1952)!suggests!that!investors!should!be!interested!in! international!diversification!due!to!the!lower!correlations!in!equity!returns!across!borders.! Some!studies,!like!that!of!Longin!and!Solnik!(1995),!have!suggested!that!correlations!in! country!portfolio!returns!increase!during!periods!of!turbulence!in!capital!markets.!If!this! were!the!case,!then!benefits!from!portfolio!diversification!would!shrink!exactly!when!they! are!needed!most.!A!study!by!Roll!(1988)!shows!that!all!23ecountry!indices!studied!of!a!crash! period!in!October!1987.!He!finds!that!a!common!factor!underlies!the!movements!of!stocks! around!the!world.!His!model!predicts!that!a!macroeconomic!shock!would!affect!all!countries! and!that!diversification!can!only!mitigate!countryespecific!events.!More!recent!research!by! Didier,!Love!and!Perìa!(2012)!found!a!similar!effect!for!the!2007!financial!crisis.!Specifically,! they!find!that!all!the!stock!market!correlations!in!their!sample!increased!with!respect!to!the! US!for!the!period!after!the!collapse!of!Lehman!Brothers.!Furthermore!they!find!the!main! determinant!of!the!increase!in!stock!market!correlations!to!be!financial.! The!question!that!remains!is,!if!the!stock!market!correlations!between!other!countries! also!increased!during!the!financial!crisis.!It!is!important!to!answer!this!question!as!noneUS! investors!interested!in!international!portfolio!divesification!could!have!avoided!the!US!stock! market!in!the!crisis!period.!Furthermore,!it!is!important!to!consider!some!other! determinants!of!stock!market!correlations!that!have!not!been!considered!in!the!research!of! Didier,!Love!and!Perìa!(2012).!This!is!due!to!the!wealth!of!literature,!that!suggests!that!there! are!other!nonefinancial!determinans!of!stock!market!correlations.!It!is!important!to!find! these!factors!so!investors!that!are!interested!in!investing!across!borders!can!take!into! account!more!factors!to!motivate!investment!decisions.! This!paper!investigates!seven!factors!that!could!determine!the!difference!in!stock! market!correlations!for!the!precrisis!to!crisis!periods.!For!this!research!20!stock!market! indices!have!been!used!to!create!190!different!stock!market!correlation!differences.!! The!seven!factors!that!have!been!considered!are:!the!rate!of!financial!integration,!the!rate! of!economic!integration,!the!exchange!rate!volatility,!the!monetary!rate!volatility,!a!EMU! dummy!variable,!a!US!dummy!variable.!In!the!second!model!an!interaction!term!between! the!monetary!rate!volatility!and!the!level!of!economic!integration!during!the!crisis!is!added! as!suggested!by!Chor!and!Manova!(2012).!! A!significant!increase!in!the!stock!market!correlation!in!the!precrisis!to!the!crisis! period!using!a!Wald!test.!Using!OLS!regression!two!different!models!are!estimated.!A! number!of!important!conclusions!are!drawn.!Specifically,!for!several!factors!the!first!model! confirms!previous!research!that!has!been!conducted.!Conformations!have!been!found!for! the!rate!of!financial!integration!and!differences!in!exchange!rate!volatility.!It!is!shown!that! differences!in!stock!market!correlations!increased!during!the!financial!crisis!if!!differences!in!financial!integration!increased!during!the!financial!crisis.!For!exchange!rate!volatility!it!is! shown!that!differences!in!stock!market!correlations!decreased!during!the!crisis!if!the! volatility!of!the!exchange!rate!between!the!country!pairs!increased!during!the!crisis.! Conformations!have!not!been!found!for!the!other!determinants.!Specifically,!for!the! difference!in!volatility!of!monetary!rates!and!levels!of!economic!integration.!! In!the!second!model,!the!interaction!term!is!shown!to!have!a!significant!effect!on! stock!market!correlations.!Using!this!model,!a!significant!effect!is!also!found!for!the!level!of! economic!integration,!contrary!for!the!first!model.!Specifically,!the!results!from!the!second! model!show!that!a!large!difference!in!economic!integration!during!the!financial!crisis! increased!the!difference!in!stock!market!correlations!for!the!country!pairs.!The!interaction! term!deminished!this!effect!if!the!difference!in!monetary!rates!of!the!country!pair!was!high! during!the!crisis.!This!model!has!shown!a!possible!justification!for!the!reason!that!Didier,! Love!and!Perìa!(2012)!failed!to!find!an!effect!for!difference!in!the!rate!of!economic! integraton!on!stock!market!correlations!during!the!financial!crisis.!A!similar!effect!for!the! volatility!in!monetary!rates!has!not!been!found.!The!two!dummy!variables!have!been!shown! to!be!insignificant!in!determining!stock!market!correlations!beyond!the!other!factors! included!in!the!model.!! Using!an!alternative!approach!of!considering!increased!and!decreased!correlations! separatly!the!results!of!the!regressions!show!significantly!different!results.!Specifically,!for! country!pairs!with!increasing!stock!market!correlations,!the!rates!of!financial!integration,! economic!integration!and!the!dummy!variables!have!been!shown!to!be!more!significant!in! determining!stock!market!correlations.!For!country!pair!that!experienced!decreasing! correlations!rates!of!economic!integration,!exchange!rate!volatility,!monetary!rate!volatility! and!the!interaction!term!were!significant!in!determining!stock!market!correlations.!Both! these!results!are!relative!to!when!all!correlations!are!considered!together.! The!remainder!of!the!paper!is!structered!as!follows:!Section!II!presents!the!literature! review!justifying!the!factors!used!in!the!models.!Section!III!presents!the!methodology.! Section!IV!presents!the!data.!Section!V!dicussed!the!results!from!the!statistical!tests.!Section! VI!concludes.!! ! ! ! ! ! ! !
I.! Literature!Review!!
! A.! The!Markowitz!model!of!portfolio!selection! ! One!of!the!most!influential!discoveries!in!investment!and!portfolio!theory!was!made!by! Markowitz!(1952).!In!Financial!Economics!his!theory!has!become!known!as!the!Markowitz! model!of!Portfolio!selection.!His!results!will!be!of!fundamental!importance!to!the!research! conducted!in!this!thesis!and!are!thus!important!to!consider.!In!his!research,!he!suggested! that!investors!should!not!only!be!interested!in!maximizing!expected!returns,!but!should!also! be!aware!of!the!risk!of!those!returns.!! Assuming!no!short!sales,!Markowitz!suggested!that!a!line!can!be!constructed!on!a! graph!with!returns!on!the!vertical!axes!and!risk!or!volatility!on!the!horizontal!axes,!where!at! every!expected!return!the!volatility!is!minimized.!The!shape!of!the!line!can!be!described! simply!in!a!portfolio!containing!two!securities.!The!line!is!constructed!by!considering!all!the! different!allocations!that!one!can!hold!the!two!securities!in!a!portfolio.!! In!the!case!that!the!returns!are!perfectly!positively!correlated!and!thus!ρ = 1,!the! line!will!be!perfectly!linear!showing!that!there!will!be!no!benefit!from!diversification.!In!the! other!extreme!case,!where!the!securities!are!perfectly!negatively!correlated!ρ = −1.!This! means!a!portfolio!can!be!created!in!which!no!risk!is!involved!and!a!guaranteed!return!can!be! achieved.!This!line!will!still!be!linear!but!it!will!be!skewed!towards!the!optimal!hedging! opportunity.!With!all!other!correlations!between!the!two!securities!the!curves!will!be!none linear!and!skewed!to!the!point!of!the!perfect!hedging!opportunity.!In!all!the!cases!that! stocks!a!not!perfectly!positively!correlated,!all!allocations!on!the!line!above!the!point!of! minimum!variance!will!be!preferred!to!the!portfolio!weights!below!the!curve!as!they!offer! the!same!expected!return!but!different!volatilities.!The!choice!of!the!optimal!portfolio!will! be!a!matter!of!risk!aversion;!investors!with!greater!risk!aversion!will!choose!allocations! closer!to!the!minimal!variance!portfolio!with!lower!expected!returns!but!lower!volatility.! When!the!amount!of!securities!is!increased!one!can!create!ever!greater!diversification! benefits!through!differing!correlations!of!returns.!It!must!be!noted!that!the!points!on!the! graph!of!only!holding!one!stock!in!a!portfolio!will!always!be!on!Markowitz’s!efficient!frontier! or!will!be!riskier.!This!led!to!an!important!result!that!shows!that!increasing!returns!can!be! received!with!lower!risk!than!would!have!been!the!case!with!just!one!stock!when!stocks!are! not!perfectly!correlated.!An!example!of!the!top!half!of!such!an!efficiency!frontier!is! presented!in!figure!1.!! Figure+1:+The+top+half+of+an+efficiency+frontier+using+the+S&P+500+index+and+an+emerging+market+index+ ! ! ! The!power!of!diversification!can!also!be!shown!analytically!with!some!simple!algebra! and!statistics.!The!expected!return!and!variance!of!the!portfolio!are:! !! ! % &' = * ()% )+, (&))!! ! ! ! ! ! (2.1)! ! ! ! /'0 = *)+, *1+,()(1234(&), &1)! ! ! ! ! (2.2)! ! Formula!(2.2)!can!be!rewritten!to:! ! ! /'0 = *)+, *1+,()(16)1∗/) ∗ /1!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!(2.3)! ! This!makes!the!relationship!between!low!correlations!between!stock!returns!and!low! portfolio!variance!formal.! !
B.! Previous!research!on!world!wide!correlations!during!bear!markets! ! The!effect!of!bear!markets!on!international!stock!market!comovements!was!first!formalized! by!Longin!&!Solnik!(1995).!In!the!paper!the!correlation!of!monthly!excess!returns!for!seven! major!countries!over!the!period!1960!to!1990.!They!find!that!international!covariance!and! correlation!matrices!are!unstable!over!time.!Specifically,!the!results!indicate!an!explicit! increase!in!the!stock!market!correlations!over!those!30!years.!It!is!also!found!that! correlations!rise!in!periods!when!conditional!volatilities!of!markets!are!large.!As!bear! markets!are!associated!with!increasingly!volatile!markets,!their!research!creates!a!formal! relationship!between!stock!market!correlations!and!bear!markets.!Furthermore,!their! results!show!some!preliminary!evidence!that!economic!variables!such!as!dividend!yield!or! interest!rates!can!contain!information!about!future!volatility!and!correlation!that!is!not! contained!in!past!returns!alone.!Here!effects,!however,!are!often!small!and!not!very! statistically!significant.!!!! ! This!research!would!mean!that!the!benefits!from!portfolio!diversification!are!lost! exactly!when!they!are!needed!most.!For!example,!a!study!by!Roll!(1988)!shows!that!in!a!23! country!indices!studied!over!stock!market!crash!between!12th!and!26th!October!with!the! beta!of!a!country!index!on!the!world!market!index,!calculated!before!the!crisis,!was!the!best! predictor!of!that!index’s!response!to!the!crash!of!the!US!stock!market.!This!suggests!! common!factors,!underlying!the!movement!of!stocks!around!the!world.!The!model!thus! predicts!that!a!macroeconomic!shock!would!affect!all!countries!and!that!diversification!can! only!mitigate!countryespecific!events.!Research!done!by!Fratzscher!(2012)!shows!this!effect! explicitly!for!the!early!2008!global!financial!crisis.!Using!a!dataset!of!highefrequency!capital! flows!of!50!economies,!the!paper!finds!that!common!shocks,!key!crisis!events!as!well!as! changes!to!global!liquidity!and!risk,!have!exerted!a!large!effect!on!capital!flows!during!the! crisis.!By!comparing!and!quantifying!the!data!he!concludes!that!common!factors!were! overall!the!main!drivers!of!capital!flows!during!the!crisis.!It!is!now!interesting!to!go!back!to! the!research!done!by!Roll!(1988)!and!consider!what!determines!the!difference!in!betas! across!countries.! Longin!&!Solnik,!(1995,!p.!4)!suggest!the!following:!“It+is+often+stated+that+the+progressive+ removal+of+impediments+to+international+investment,+as+well+as+the+growing+political,+ economic+and+financial+integration,+affects+international+market+linkages.+This+could+lead+to+ a+progressive+increase+in+the+international+correlation+of+financial+markets+reflecting+the+ 'global+finance'+phenomenon”.!This!suggests!that!rates!of!economic!and!financial!integration! are!important!determinants!of!international!stock!market!correlations.! ! Research!on!how!significantly!theses!factors!affected!stock!market!correlations! during!the!2007!financial!crisis!has!been!incredibly!care.!Research!on!a!number!of!factors! was!conducted!by!Didier,!Love!and!Perìa!(2012)!on!US!correlations.!The!article!finds!that!the! major!factor!driving!increased!comovement!of!stock!market!returns,!is!financial.!Their! research!does!not!find!a!significant!effect!of!trade!on!the!comovement!of!stock!market! returns.!They!conclude!their!article!by!stating!that!the!significant!effect!of!financial! integration!on!the!correlations!between!stock!market!returns!shows!that!countries!need!be! aware!of!a!dark!side!of!financial!integration!and!liquidity.!However,!they!believe!that!a! country!should!always!weigh!the!pros!and!the!cons!of!financial!integration!before! overreacting!to!this!finding.!! !!
!
C.! Possible!factors!that!lead!to!varying!degrees!of!correlations!in!stock!market! returns!across!countries! ! i.! Monetary!environment! ! The!financial!crisis!in!August!2007!and!the!subsequent!meltdown!of!the!global!economy!has! confronted!monetary!policy!in!the!major!industrialized!countries!with!unprecedented! challenges.!The!strong!effects!of!the!crisis!on!the!banking!sector!and!the!associated! economic!downturn!led!central!banks!to!respond!with!previously!untouched!measures!of! monetary!policy.!The!precise!nature!of!response!of!the!main!three!central!banks!examined! differed!significantly.!Due!to!the!strong!relevance!of!bank!lending!in!euro!area!countries!and! Japan!the!Bank!of!Japan!(BoJ)!and!the!European!central!bank!(ECB)!focused!their!policy! mostly!on!providing!relief!to!and!restructuring!of!the!banking!sector.!Contrary!to!the!policy! of!the!ECB,!the!Bank!of!England!(BoE)!and!the!Fed!undertook!largeescale!asset!purchases! predominantly!from!nonebank!counterparties!to!ensure!a!flow!of!credit!to!their!economies.! Despite!the!slight!differences!in!monetary!policies!all!central!banks!experienced!a!massive! surge!in!the!size!of!their!balance!sheets.!In!the!period!between!August!2007!and!November! 2010!the!main!two!central!banks!that!engaged!in!a!variant!of!quantitative!easing!were!the! Fed!and!the!BoE.!! ! The!QE!policy!of!the!Fed!had!been!given!the!name!‘credit!easing’,!by!the!then! president!of!the!Federal!Reserve,!Ben!Bernanke.!This!policy!was!conceptually!distinct!from! pure!QE,!as!in!QE!the!focus!of!the!policy!is!the!quantity!of!the!central!bank!reserves,!and!the! composition!of!lending!and!securities!on!the!asset!side!is!purely!incidental.!In!contrast!the! aim!of!CE!was!explicitly!to!increase!the!amount!to!risk!the!central!bank!was!taking!on!and! decreasing!the!risk!on!the!balance!sheets!of!the!financial!institutions.!In!addition!to!the! credit!easing!in!March!2009!the!Fed!adopted!a!third!unconventional!tool!of!monetary!of! managing!expectations.!With!this!forward!guidance!the!FOMC!stated!that!it!would!maintain! ‘exceptionally!low’!interest!rates!‘for!an!extended!period!of!time’!(Mishkin,!Matthews,!&! Giuliodori,!2013).! ! In!response!to!the!intensification!of!the!financial!crisis!in!autumn!2008,!the!BoE! loosened!monetary!policy,!with!both!conventional!and!unconventional!policy!measures.!In! the!United!Kingdom,!the!most!important!element!of!the!unconventional!measures!was!QE.! Over!the!period!March!2009!to!January!2010,!£200!billion!of!assets!were!purchased,!mostly! made!up!of!government!securities,!representing!around!14%!of!annual!GDP.!“The+evidence+ suggested+that+the+policy+had+economically+significant+effects+–+equivalent+to+a+150+to+300+ basis+point+cut+in+Bank+Rate+–+but+there+is+considerable+uncertainty+around+the+precise+ magnitude+of+the+impact”!(Joyce,!Tong,!&!Woods,!2011).! ! With!these!changes!in!monetary!policy!the!central!banks!attempted!to!achieve!a!set! of!economic!objectives!that!are!expressed!in!macroeconomic!variables!such!as!inflation,!real! output!and!employment.!The!affectivity!of!monetary!policy!measures!on!these!economic! variables!is!at!best!indirect!and!is!known!to!have!considerable!lags.!An!example!of!that!is! more!directly!influenced!by!changing!monetary!policy!is!the!stock!market,!which!is!faster!at! incorporating!new!information!about!fundamentals.!A!trivial!explanation!of!the!way!that! monetary!policy!affects!stock!prices!can!be!seen!through!the!discounted!cash!flow!model.! According!to!the!model!all!stock!prices!are!the!present!value!of!their!future!expected!cash! flow.!This!means!stock!prices!will!be!high!when!interest!rates!are!low,!which!is!caused!by!
expansionary!monetary!policy.!In!contrast,!monetary!policy!tightening!should!coincide!with! lowering!stock!prices!(Ioannidis!&!Kontonikas,!2008).!! ! Ioannidis!and!Kontonikas!(2008)!investigated!the!impact!of!monetary!policy!on!stock! returns!between!1972!and!2002!in!13!OECD!countries.!Their!results!indicate!that!monetary! policy!shifts!significantly!affecting!stock!returns,!thereby!supporting!the!notion!of!monetary! policy!transmission!via!the!stock!market,!even!when!correcting!for!nonenormal!stock!returns! and!various!measures!of!stock!market!returns.!These!results!imply!that!stock!market! investors!should!be!aware!of!international!portfolio!diversification!opportunities!across! countries!with!different!monetary!environments.!Rigobon!&!Sack!(2004)!also!estimate!the! effect!of!monetary!policy!on!asset!prices!using!instrumental!variables!(IV)!regression.!Using! this!test!their!research!requires!less!assumptions!than!the!tests!used!in!previous!research.! The!results!of!the!IV!regression!show!that!on!the!days!that!the!FOMC!announces!changes!in! the!policy!rate!there!is!a!strong!correlation!with!asset!prices.!Moreover,!and!perhaps!more! importantly,!the!results!of!their!model!show!that,!the!eventestudy!method!of!Ioannidis!and! Kontonikas!(2008)!cannot!be!rejected.! ! ! ! ii.! Level!of!economic!integration! ! ! For!the!year!2007,!world!trade!was!measured!to!account!for!around!58%!of!world!GDP! (World!Trade!Organisation,!2011).!This!in!and!of!itself!makes!trade!per!GDP!ratios!an! important!fundamental!for!macroeconomic!analysis.!Eaton!et!al.!(2010)!found!that!during! the!Great!Recession!of!2008e2009,!global!trade!fell!30%!relative!to!GDP.!In!the!analysis!they! show!that!the!decrease!in!world!trade!was!largely!caused!by!a!shift!in!spending!away!from! manufactures!and!particularly!durables.!This!accounted!for!more!than!80!percent!of!the! drop!in!trade!relative!to!GDP.!They!further!show!that!increased!trade!impediments!only! decreased!trade!in!some!countries,!and!globally!the!effect!largely!cancels!out.!Additional! research!by!Bems,!Johnson!and!Yi!(2010)!finds!that!70%!of!the!trade!decline!can!be! accounted!for!by!demand!forces!alone.!Their!research!also!presents!an!elasticity!of!world! trade!to!GDP!of!2.8.!! ! Pretorius!(2002)!states!that!stock!markets!are!expected!to!be!highly!interdependant! if!two!countries!have!strong!bilateral!trade!realtionship.!He!shows!this!using!an!example!of! two!imaginary!countries,!country!A!and!country!B.!If!a!high!proportion!of!country!A’s!total! exports!are!to!country!B,!then!a!downswing!in!country!B!will!cause!a!decline!in!its!imports! from!country!A.!The!total!effect!of!the!downswing!of!the!economy!in!country!B!will!be!a! downswing!of!country!B’s!stock!market!and!economy,!but!also!a!downswing!of!country!A’s! stock!market!and!economy.!Thus!with!a!higher!rate!of!bilateral!trade,!the!stock!markets!of! country!A!and!country!B!will!be!more!interdependent.! ! Johnson!and!Soenen!(2003)!examine!the!degree!of!crossecountry!return!correlations! for!eight!American!equity!markets!with!the!stock!market!of!the!United!States.!Their!results! for!sameeday!relationships!indicate!that!a!higher!share!of!international!trade!in!terms!of! imports!and!exports!with!the!United!States!are!strongly!positively!correlated!with!stock! market!correlations!between!the!two!countries.!Chen!and!Zhang!(1997)!obtain!very!similar! results!for!several!PacificeBasin!markets.!They!find!that!the!countries!in!their!sample!that! have!strong!economic!ties!tend!to!have!financial!market!that!move!together.!They!draw! these!conclusions!from!finding!that!these!correlations!are!important!in!explaining!crosse country!average!returns!beyond!what!is!reflected!in!a!three!factor!model!global!market,!size!
and!booketoemarket.!Specifically,!the!extent!of!trade!explains!between!5!to!40%!of!the! variation!in!the!correlation,!depending!on!the!measure!of!correlation!used.! ! Chor!and!Manova!(2012,!p.!117)!also!find!that!“countries+with+higher+interbank+rates+ and+thus+tighter+credit+markets+exported+less+to+the+US+during+the+peak+of+the+crisis”.+This! implies!that!the!interaction!of!high!interbank!rates!and!lower!exports!would!influence!the! rate!stock!market!returns!during!the!crisis.!This!is!corroborated!by!Chor!and!Manova!(2012)! why!find!that!the!previously!stated!effect!rose!during!the!financial!crisis.! ! ! ! iii.! Rate!of!financial!integration! ! “After+the+collapse+of+Lehman+Brothers+in+September+2008,+the+crisis+spread+rapidly+across+ institutions,+markets,+and+borders.+There+were+massive+failures+of+financial+institutions+and+a+ staggering+collapse+in+asset+values+in+developed+and+developing+countries+alike”!(Didier,! Love!and!Perìa!2012,!p.!183).!This!underlines!that!the!global!financial!crisis!was!largely! caused!by!excessive!bank!lending!in!the!United!States,!and!the!significant!contagion!that!it! had!on!other!countries.!Due!to!the!global!downturn!of!the!stock!markets!following!the! collapse!of!Lehman!Brothers!it!is!interesting!to!study!the!effect!of!the!rate!of!financial! integration!on!international!stock!market!correlations.!Dellas!and!Hess!(2005)!examined!this! effect!for!the!period!between!1980!and!1999.!Their!results!tend!to!differ!somewhat! depending!on!the!indicator!of!financial!development!used!and!the!currency!of! denomination!of!returns.!However,!irrespective!of!the!currency!denomination!of!the! returns,!they!find!that!financial!development!is!significantly!related!to!the!behavior!of!the! distribution!of!stock!returns.!“In+general,+“deeper”+and+more+efficient+banking+systems+have+ been+associated+with+significantly+lower+stock+return+volatility+as+well+as+a+closer+ comovement+with+world+returns”!(Dellas!&!Hess,!2005,!p.!891).!In!their!paper!they!do! suggest!that!further!research!should!produce!more!appropriate!financial!development! indicators—namely,!indicators!that!can!be!uniquely!and!precisely!associated!with!specific! functions!of!the!financial!system—and!then!relate!these!indicators!to!the!properties!of!asset! prices.! ! !! iv.! Exchange!rate!risk! ! For!international!portfolio!diversification!the!implications!of!exchange!rate!risk!are!twofold.! The!degree!of!exchange!rate!volatility!is!an!important!factor!in!the!investment!decision!of! an!international!investor.!First!a!highly!volatile!exchange!rate!between!the!domestic! currency!and!a!foreign!currency!can!lead!to!very!uncertain!returns.!For!example,!the!foreign! currency!could!depreciate!sharply!leading!to!the!domestic!currency!return!on!investment! decreasing,!assuming!all!other!factors!are!held!constant.!Through!perfect!hedging,!this!risk! could!be!eliminated.!Unfortunately,!as!can!be!seen!with!the!interest!rate!parity!formula:! ! ! 8 = 9 ∗(,:;(,:;)∗)!! ! ! ! ! ! ! ! (3.1)! ! !
it!is!very!difficult!to!predict!exactly!the!amount!of!foreign!currency!to!sell!forward!as!the! return!on!investment!is!also!uncertain!unless!one!invests!in!a!riskefree!security,!or!said! differently!&∗!should!be!a!known!number!for!covered!interest!parity!to!hold.!! ! The!degree!of!exchange!rate!volatility!can!also!affect!the!international!investment! decision!of!an!investor!through!an!increase!or!decrease!in!the!correlations!of!international! stock!returns.!On!this!issue,!theory!offers!only!ambiguous!conclusions,!along!two!tangents.! A!first!line!of!reasoning!is!based!on!the!fundamental!approach!of!asset!prices!and!suggests! that!a!low!volatility!in!the!exchange!rate!increases!international!correlations.!When!a!group! of!countries!has!a!credible!peg,!common!variables!like!strongly!synchronized!business! cycles,!absence!of!shocks!in!the!tradables!balance!due!to!absence!of!exchange!rate!shocks! and!the!simultaneous!conduct!of!monetary!policy!lead!to!strongly!correlated!stock!market! returns!(King!&!Wadhwani,!1990).!The!second!line!of!reasoning!is!based!on!the!contagion! explanation!of!change!in!asset!prices.!It!suggests!that!correlation!of!international!returns! increase!instead!of!decrease!when!exchange!rate!markets!are!more!volatile.!Because!of! heard!behavior!and!possible!other!factors,!contagion!effects!are!highest!in!volatile!markets! (King!&!Wadhwani,!1990).!When!investors!have!different!expectations!about!fundamentals! they!look!at!asset!prices!abroad!to!seek!likely!trends!in!the!domestic!market.!International! correlations!are!then!higher!in!the!case!of!increased!exchange!market!volatility!a!lower!in! the!case!of!countries!with!a!credible!peg!(Bodart!&!Reding,!1999)! ! Bodart!and!Reding!(1999)!conduct!empirical!research!to!discover!which!of!these! tangents!has!been!more!applicable!during!the!period!in!which!numerous!European! countries!decided!to!peg!their!currencies!to!the!Deutsche!Mark,!more!formally!known!as! the!ERM.!They!use!the!daily!returns!on!the!bond!and!stock!markets!of!six!European! countries!over!the!period!from!2!January!1989!to!19!December!1994.!Their!main!conclusion! relevant!to!this!research!is!that!during!the!period!the!analysis!makes!clear!that!the!degree! of!exchange!rate!variability!exerts!an!influence!on!international!bond!and!stock!market! correlations.!In!particular,!it!is!shown!that!an!increase!in!exchange!rate!variability!during!the! ERM!causes!a!drop!in!the!correlation!between!German!and!other!European!countries’!stock! markets.!The!research!also!shows!that!this!effect!is!slightly!less!significant!than!on!bond! markets,!which!the!authors!attribute!this!result!to!“the!larger!effect!of!idiosyncratic!shocks! on!the!behavior!of!domestic!stock!prices”.!Their!research!thus!supports!the!fundamental! approach!of!asset!prices,!that!a!low!volatility!in!the!exchange!rate!increases!international! correlations.!Building!on!the!research!of!Bodart!and!Reding!(1999),!McPherson!(2006)! conducted!similar!research!for!Canada.!In!his!sample!he!finds!that!there!is!a!connection! between!foreign!exchange,!stock!market!volatilities!and!correlations.!Specifically,!he!finds! that!they!are!significantly!different!for!various!time!periods.! ! ! ! ! ! ! !
v.! Dominant!position!of!the!United!States,!EMU!effects!and!other!monetary! union!effects! ! To!conclude;!in!research,!one!should!want!to!include!factors!that!increase!stock!market! comovement!between!countries!that!cannot!be!captured!by!other!economic!fundamentals.! Along!with!most!of!the!research!cited!in!this!paper,!Eun!and!Shim!(1989)!find!that!there!is!a! substantial!amount!of!comovement!of!between!national!stock!markets.!They!analyze!nine! time!series!of!daily!stock!market!returns,!which!are!then!interpreted!via!vector!auto! regression!in!order!to!obtain!insights!into!interdependence!structure!of!major!international! stock!indices.!Their!research!also!corroborates!their!hypothesis!that!the!US!stock!market! turns!out!to!be,!by!far,!the!most!influential!in!the!world.!Innovations!in!the!US!stock!market! are!rapidly!transmitted!to!other!markets!in!a!clearly!recognizable!pattern,!whereas!no!single! foreign!market!can!significantly!explain!the!US!market!movements.!They!suggest!that!this! might!reflect!the!dominant!position!of!the!US!in!the!world!economy,!which!makes!the! country!the!most!important!producer!of!information!affecting!the!world!stock!market.!! Kim,!Moshirian!and!Wu!(2005)!find!another!effect!for!stock!market!correlations!that! might!not!be!captured!by!the!other!factors.!They!examine!the!influence!of!the!European! Monetary!Union!(EMU)!on!the!dynamic!process!of!stock!market!integration!over!the!period! 2!January!1989!to!29!May!2003.!They!find!that!there!has!been!a!clear!regime!shift!in! European!stock!market!integration!with!the!introduction!of!the!EMU.!For!this!effect!they! find!a!unidirectional!causality,!which!means!that!it!does!not!suffer!from!simultaneous! causality!bias.!! !Newer!research!by!(Wälti,!2011)!also!focuses!on!the!relationship!between!stock! market!correlations!and!monetary!integration.!In!his!sample!he!calculates!the!bilateral!stock! market!return!correlations!of!fifteen!developed!economies!over!the!period!1975!to!2006.! His!findings!vindicate!much!of!the!above!research.!Specifically,!he!finds!that!trade!and! financial!integration!contribute!significantly!to!higher!stock!market!return!correlations.! Furthermore,!he!finds!that!monetary!integration!leads!to!stronger!stock!market! synchronization,!through!the!elimination!of!exchange!rate!volatility,!through!common! monetary!policy!and!convergence!of!inflation!expectations.!The!earlier!research!done!by! Kim,!Moshirian,!&!Wu!(2005)!for!the!group!of!EMU!members!also!reveals!that!increasing! stock!market!correlations!can!be!explained!with!the!overall!macroeconomic!convergence! process!associated!with!the!introduction!of!the!euro!rather!than!the!specific!effects!of!the! elimination!of!foreign!exchange!rate!risk!due!to!a!common!currency.! ! Although!further!research!on!this!topic!is!scarce,!it!seems!important!to!correct!for! other!currency!unions!in!the!sample.!As!research!by!Kim,!Moshirian!and!Wu!(2005)!has! shown!this!factor!might!contain!relevant!information!about!stock!market!comovements!that! cannot!be!captured!by!other!factors!used!in!the!model.!Due!to!the!ambiguity!of!these! findings,!both!options!will!be!considered!in!this!paper.!! ! !
III.! Methodology!
! A.! Model! ! To!examine!what!affected!the!difference!in!stock!market!correlations!before!and!during!the! financial!crisis,!two!models!have!been!estimated.!In!these!two!models!the!difference!in! correlations!are!a!function!of!the!factors!described!in!the!literature!review.!In!the!research! two!time!series!are!distinguished.!One!for!precrisis!and!one!for!the!crisis!period.!From!here! on!out,!the!precrisis!period!will!be!defined!as!the!period!between!July!1st!2004!and!August! 1st!2007.!The!crisis!period!shall!be!defined!as!the!period!between!August!1st!2007!and! October!1st!2010.!This!research!uses!the!start!of!the!financial!crisis!as!defined!by!Taylor! (2009)!to!be!August!1st!2007.!The!two!empirical!models!can!be!summarized!by!equations! (5.1)!and!(5.2)!respectively:! ! 6<=><?,)1 = @A+ @,8C<=><?,)1+ @0%C<=><?,)1+ @D/EFGHI*JE;I<E,<=><?,)1+ @K/LM*E<I;N;I<E,<=><?,)1 + @OP9 + @Q%RP + S)1! ! ! ! !!!!!!!!(5.1)! ! ! 6<=><?,)1 = @A+ @,8C<=><?,)1+ @0%C<=><?,)1+ @D/EFGHI*JE;I<E,<=><?,)1+ @K/LM*E<I;N;I<E,<=><?,)1 + @OP9 + @Q%RP + @T/LM*E<I;N;I<E,<=,)1∗ %C<?,)1+ S)1!!(5.2)! !In!these!models!subscripts!U,!and!U0!denote!the!precrisis!and!crisis!period!respectively.!The!
subscript!VW!denotes!the!country!pair!that!is!being!considered.!@,!to!@Q!denote!the!
coefficients!of!the!variables!used!in!the!model.!In!both!models!@A!is!a!constant!and!S)1!is!the!
error!term!of!the!regression.! ! B.! Variable!description! ! The!definitions!of!the!variables!are!presented!in!Table!1.!For!the!absolute!difference!in! correlations,!daily!stock!returns!have!been!obtained!from!respective!indices!20!countries.1!! ! !
Variable+ Description+ Time+ Source+ Considered++in+
model+ 6<=><?,)1! The!absolute! difference!in!the! correlations! between! country!the!two! sample! countries! From!precrisis!to! crisis!(absolute! difference)! Yahoo!!Finance! 1!and!2! 8C<=><?,)1! The!percentage! difference!in! rates!of!financial! integration! between!two! sample! countries!! From!precrisis!to! crisis!(absolute! difference)! World!bank! 1!and!2! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 1!The!national!stock!market!indices!that!have!been!used!are!summarized!in!section!VII.!
%C<=,)1! The!percentage! difference!in! rates!of! economic! integration! between!two! sample! countries!
Crisis! World!bank! 1!and!2!
/EFGHI*JE;I<E,<=><?,)1! The!volatility!of! the!exchange! rate!between!to! sample! countries! From!precrisis!to! crisis!(absolute! difference)! OANDA! 1!and!2! /LM*E<I;N;I<E,<=,)1! The!percentage! difference!in! volatility!of!the! monetary!rate! between!two! sample! countries! From!crisis! FRED!and!Central! bank!statistical! bureaus! 1!and!2! P9! Equals!1!if!on!of! the!sample! countries!is!the! US! ! ! 1!and!2! %RP! Equals!1!if!both! of!the!countries! are!part!of!the! EMU! ! ! 1!and!2! /LM*E<I;N;I<E,<=,)1∗ %C<=,)1! An!interaction! term!between! the!volatility!of! the!monetary! rates!and!the! level!of! economic! integration! For!the!monetary! rate:!crisis! ! For!the!rate!of! economic! integration:!crisis! Same!as!highlighted! previously! 2! Table+1:+Description+of+the+variables+ For!this!research!the!G20!have!been!chosen,!because!they!accounted!for!about!85%!of! world!GDP!in!2012!Eurostat,!(2015).!It!is!assumed,!because!of!this!big!share!of!world!GDP,! these!countries!will!also!have!been!the!focus!of!investors!interested!in!international! diversification,!both!through!the!amount!of!available!information!and!the!relative!limited! amount!of!barriers.!Two!G20!countries!have!been!omitted!from!this!sample.!Firstly,!Saudi! Arabia!has!been!omitted!due!to!the!use!of!Islamic!banking!in!their!stock!exchange,!which! could!potentially!cause!noise!in!the!variables!that!are!beyond!the!scope!of!this!research.! Secondly,!Europe!has!been!omitted,!both!due!to!the!limited!availability!of!a!European!stock! index!and!the!similar!results!that!it!could!have!with!other!European!countries.!The!two! countries!have!been!replaced!by!the!Netherlands!and!Greece.!The!reason!for!adding!the! Netherlands!has!come!from!its!relative!importance!in!the!EU!and!high!GDP!per!capita!(after! the!countries!that!have!already!been!included!in!the!sample!like!Germany,!UK!and!France).! Greece!has!been!added!to!the!sample!because!of!its!precarious!situation!in!the!EU.!Hong! Kong!has!been!taken!as!a!proxy!for!China!due!to!the!capital!restrictions!China!imposes!on! the!rest!of!the!economy.!From!the!daily!stock!return!data,!190!correlations!have!been! calculated!and!are!defined!as!6),1!where!i!and!j!are!two!different!countries!considered.!To! analyze!the!increase!or!decrease!of!the!correlations!in!stock!returns,!several!factors!are!
considered!that!are!deduced!from!the!review!of!the!literature!conducted!in!section!IV2.!For! the!monetary!environment,!the!policy!rate!of!the!central!banks!has!been!used.!This!is!the! interest!rate!that!central!banks!use!to!coordinate!their!monetary!policy.!For!most!of!the! countries,!this!is!the!30eday!interbank!rate!in!their!country.!! The!metric!used!for!economic!integration!has!been!the!share!of!the!tradebles!sector! with!respect!to!the!GDP.!The!metric!that!has!been!chosen!for!financial!integration!is!the! amount!of!credit!provided!in!the!country,!considered!as!a!share!of!GDP.!For!the!two! previous!metrics,!averages!of!the!yearly!rates!have!been!used!over!the!precrisis!and!crisis! periods.!For!the!exchange!rate,!the!volatility!has!been!calculated!for!the!two!periods.!For! the!EMU!countries,!this!volatility!has!been!set!to!zero!due!to!the!equivalence!of!the! currencies!considered.!Lastly,!two!dummies!have!been!created.!For!the!dominant!position! of!the!US,!the!dummy!is!1!if!the!correlation!considered!involves!the!US!as!one!of!the! countries!and!zero!otherwise.!Similarly,!for!the!EMU!the!EMU!dummy!is!1!if!the!two! countries!considered!are!both!involved!in!the!EMU!and!0!otherwise.!As!the!Indian!central! bank!held!its!monetary!policy!rate!constant,!those!correlations!are!omitted!from!the! regression.! Lastly,!an!interaction!term!has!been!added!as!suggested!by!Chor!and!Manova!(2012).!! This!variable!considers!the!interaction!between!the!volatility!of!the!monetary!rates!and!the! level!of!economic!integration!between!country!pairs!on!the!difference!in!stock!market! correlations.! !
IV.! Data!description!and!hypotheses!
! A.! Data!Description! ! The!summary!statistics!of!the!variables!are!shown!in!table!2.! !! Observations! Mean! Standard!
deviation! Minimum! Maximum! 6<=><?,)1! 190! 0.0715! 0.2284! ^0.6062! 0.5819! 8C<=><?,)1! 190! ! 0.1339! 0.5603! ^2.2207! 2.8712! %C<=><?,)1! 190! 0.0554! 0.2119! ^0.3224! 1.5627! /EFGHI*JE;I<E,<=><?,)1! 190! 0.0050! 0.0034! ^0.0017! 0.0183! /LM*E<I;N;I<E,<=,)1! 161! 0.9379! 2.9586! ^0.8779! 22.9048! P9! 190! 0.1000! 0.3004! 0! 1! %RP! 190! !0.0789! 0.2700! 0! 1! /LM*E<I;N;I<E,<=,)1∗ %C<=,)1! 161! 0.5411! 2.7937! ^2.3287! 22.5371! Table+2:+Summary+statistics+of+the+variables+used+ The!deviations!from!the!expected!amount!of!observations!for!the!volatility!of!the!monetary! rate!and!the!interaction!term!have!a!specific!reason.!The!missing!observations!of!the! monetary!rate!are!due!to!the!nonexistent!volatility!of!monetary!rates!between!countries! that!under!the!charter!of!the!ECB!and!due!to!the!lack!of!volatility!of!monetary!rates!in!India.! As!we!examine!percentage!difference!in!monetary!rates!between!countries,!some!of!the! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! ! !
observations!would!have!been!undefined.!For!this!reason,!these!observations!are!not! considered.!For!this!reason,!the!number!of!observations!of!the!interaction!term!are!also! reduced.!! ! The!mean!of!the!difference!in!stock!return!correlations,!means!correlations!on! average!increased!by!0.0715!in!absolute!terms.!When!expressed!in!terms!of!percentages! the!stock!market!correlations!increased!on!average!by!42.86%!during!the!crisis.!The!stock! market!correlations!between!Argentina!and!India!increased!to!most!during!the!crisis.!The! correlation!increased!by!more!than!3000%!from!0.018!to!0.6.!The!largest!decrease!in!stock! market!correlations!was!between!Argentina!and!Russia,!with!a!decrease!in!correlation!of! nearly!130%!from!0.309!to!e0.088.!A!graph!of!the!absolute!differences!between!precrisis!and! crisis!correlations!are!presented!in!figure!2.! ! Figure+2:+Absolute+differences+in+correlations+for+the+190+countries+pairs+in+the+sample,+between+the+period+before+the+crisis+ and+during+the+crisis.+ Here!we!can!see!that!most!of!the!stock!market!return!correlations!increased.3!From!the! summary!statistics,!we!can!see!that!on!average!that!the!percentages!of!financial!integration! and!economic!integration!increased!during!the!crisis.!Similarly,!the!volatilities!of!exchange! rates!and!monetary!rates!also!increased!during!the!crisis.! ! ! ! ! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 3!A!test!for!the!noneequality!of!correlations!between!precrisis!and!crisis!can!be!found!in!
B.! Hypotheses! ! In!the!following!section,!a!number!of!hypotheses!will!be!tested!using!OLS!regression.!The! models!used!for!this!OLS!regression!have!been!described!in!section!V.!Using!the!information! in!section!II,!the!following!hypotheses!are!presented!in!the!table!3.! ! Hypotheses! During!the!crisis,!stock!market!correlations!will!increase!significantly.! If!the!two!sample!countries!have!a!decreasing!difference!in!in!the!rate!of!financial! integration,!it!is!expected!that!their!stock!market!correlations!will!have!increased.! If!the!two!sample!countries!have!a!big!difference!in!the!rate!of!economic!integration! during!the!crisis,!it!is!expected!that!their!stock!market!correlations!will!have!increased.! If!the!exchange!rate!volatility!between!two!countries!increased!during!the!crisis,!it!is! expected!that,!stock!market!correlations!will!have!decreased.! If!two!sample!countries!have!high!differences!in!the!volatility!of!their!monetary!rates,!it!is! expected!that!their!stock!market!correlations!will!have!decreased.! If!one!of!the!two!sample!countries!is!the!US,!it!is!expected!that!stock!market!correlations! will!have!increased!by!more!than,!if!a!country!pair!does!not!contain!the!US.! If!both!of!the!sample!countries!considered!are!part!of!the!EMU,!it!is!expected!that!stock! market!correlations!increased!by!more!than!country!pairs!that!are!not!part!of!the!EMU.! Monetary!policy!rates!and!levels!of!economic!integration!will!interact!with!stock!market! correlations.! Table+3:+Hypotheses+that+are+tested+in+section+V+ Using!the!literature!in!section!II!the!intuition!behind!these!hypotheses!can!be!explained. ! Firstly,!as!found!by!Longin!and!Solnik!(1995),!Roll!(1988)!and!Didier,!Love!and!Perìa! (2012)!it!is!presumed!that!the!results!will!show!an!increase!in!the!in!stock!market!return! correlations!during!the!crisis.!! Secondly,!it!is!expected!that!if!the!two!sample!countries!have!a!decreasing! difference!in!in!the!rate!of!financial!integration,!their!stock!market!correlations!will!have! increased,!as!was!found!by!Didier!Love!and!Perìa!(2012).!! Thirdly,!ambiguity!exists!about!the!way!the!difference!in!economic!integration! precrisis!to!crisis!would!affect!stock!market!correlations.!On!the!one!hand!Didier,!Love!and! Perìa!(2012)!did!not!find!a!significant!effect!for!the!rates!either!in!economic!integrations!or! for!bilateral!trade!linkages.!However,!research!by!Johnson!and!Soenen!(2003),!Chen!and! Zhang!(1997)!and!Pretorius!(2002)!would!suggest!a!significant!effect!of!economic!integration! on!stock!market!correlations.!It!is!therefore!important!to!test!the!effect!of!economic! integration!on!stock!market!correlations!again!using!a!different!metric.!With!this!metric!it!is! presumed!that!there!will!be!an!effect!for!the!difference!in!the!rate!of!economic!integration! on!stock!returns.!As!the!research!by!Didier,!Love!and!Perìa!(2012)!only!considered! correlations!with!the!US,!this!metric!might!have!more!effect!on!correlations!that!don’t! contain!the!US!in!the!country!pair.!! Fourthly,!it!is!assumed!that!an!increase!in!precrisis!to!crisis!exchange!rate!volatilities! negatively!affects!stock!market!correlations!as!shown!by!(Bodart!&!Reding,!1999).!! Fifthly,!in!accordance!with!the!literature!the!difference!in!the!volatility!of!monetary! policy!rates!during!the!crisis!between!the!two!countries!should!decrease!the!stock!market! return!correlations.!This!is!asserted!from!the!studies!of!Ioannidis!and!Kontonikas!(2008)!and!
Rigobon!and!Sack!(2004).!It!is!also!assumed!that!most!central!banks!would!have!decreased! their!monetary!policy!rates!in!order!to!dampen!the!effects!of!the!crisis.!If!for!example! Central!bank!A!decided!to!cut!their!monetary!policy!rates!to!a!higher!degree!than!that!of! Central!bank!B!the!difference!in!volatilities!would!be!high,!it!is!assumed!from!Ioannidis!&! Kontonikas!(2008)!that!the!stock!returns!would!differ!after!the!announcements.!This!would! decrease!the!correlations!of!stock!market!returns!during!the!crisis.!! Sixthly,!it!is!expected!that!a!correlation!between!two!EMU!countries!or!one!of!the! countries!being!the!US!will!show!a!positive!effect!on!stock!market!correlations!that!are!not! captured!by!the!other!factors.!Lastly,!as!research!of!Chor!and!Manova!(2012)!has!found!that! monetary!policy!rates!and!levels!of!economic!integration!interact!on!stock!market! correlations!it!is!expected!that!the!results!off!the!research!will!show!this!effect.! !
!
V.! Results!
! A.! Increase!in!correlations! ! First,!a!Wald!test!is!conducted!to!compute!if!the!correlation!matrices!are!significantly! different.!For!the!computation,!the!reader!is!referred!to!Jennrich!(1970).!The!test!statistic!is! distributed!X0~LY>Y, Z ZY>Y,0 !.!The!results!are!presented!in!Table!1.! ! Wald!test! X0~190! p9value! Value! 4388.70! 0.0000! Table+4:+Wald+test+for+equal+correlations+ This!implies!that!at!a!significance!level!of!1%!the!hypothesis!of!equal!correlations!before!and! after!the!crisis!can!be!rejected4.!This!test!does!not!say!anything!about!a!positive!difference! in!correlations!or!a!negative!difference.!To!evaluate!the!sign!of!the!difference!in!correlations! one!can!use!the!data!given!in!table!2!and!figure!2.!The!mean!of!the!difference!in!correlations! is!positive!which!implies!that!most!of!the!observations!will!have!been!positive!if!the! negative!data!are!mostly!outliers.!Figure!3!presents!a!boxplot!of!the!difference!in! correlations.!Using!the!data!in!figure!2!and!3!we!see!that!most!of!the!negative!data!have! been!outliers.! Figure+3:+Boxplot+of+the+difference+in+correlations+ Furthermore,!figure!2!shows!that!for!more!than!77%!of!the!observations!the!difference!in! correlations!increased.!This!leads!to!the!tentative!conclusion!that!during!the!financial!crisis,! stock!market!return!correlations!for!the!countries!used!in!this!sample!have!significantly! increased.!Thus,!the!first!hypothesis!seems!to!be!confirmed!by!the!data.! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 4!Strictly!speaking,!this!test!assumes!that!the!data!are!multivariate!normally!distributed.!A! test!of!multivariate!normality!on!this!dataset!is!rejected!using!a!DoornikeHansen!test.! However,!the!data!from!figure!2,!we!can!safely!assume!that!the!result!of!the!Wald!test!is! valid!as!nearly!none!of!the!data!is!observed!around!a!zero!difference!in!correlations.!
B.! OLS!factors!regression! ! ! Using!the!models!described!in!section!III,!an!OLS!regression!is!performed.!Using!the!results! of!the!pevalues!of!the!coefficients!the!hypotheses!that!were!stated!in!section!VI!will!be! evaluated.!Using!the!signs!on!the!coefficients!that!are!that!are!obtained!in!the!OLS! regression!the!results!obtained!in!the!literature!will!be!compared!to!the!results!of!this! thesis.!Table!5!presents!the!results!of!the!two!OLS!regressions.! ! Dependent!variable!:!Change!in!correlation!between!country!i!and!j!between!crisis!and!precrisis;!161!observations!
Regressor! Model!1! Model!2!
8C<=><?,)1! 0.1055***! (0.0282)! 0.1018***! (0.0274)! %C<=,)1! 0.0149! (0.0082)! 0.0170**! (0.0079)! /EFGHI*JE;I<E,<=><?,)1! 922.0005***! (5.1824)! 917.8110***!(5.2006)! /LM*E<I;N;I<E,<=,)1! 90.0102*! (0.0060)! (0.0090)!0.0100! P9! 0.0451! (0.04143)! (0.0529)!90.0123! %RP! 0.0378! (0.0967)! (0.0940)!0.0339! /LM*E<I;N;I<E,<=,)1∗ %C<=,)1! ! 90.0297***! (0.0093)! C^U_&`_aU! 0.1608***! (0.0327)! 0.1354***!(0.0328)! F9!Statistic!and!Summary!Statistics! ! ! ! F9statistic!test!on!coefficients!of!the! interaction!term! ! ! ! ! ! ! ! ! ! ! ! 19.67***! SER! 0.2119! 0.2058! bc! 0.2112! 0.2559! Table+5:+The+F[statistic+tests+the+hypothesis+that+the+coefficient+on+the+interaction+term+is+zero.+Robust+standard+errors+are+ given+in+parentheses+under+coefficients,+and+p[values+are+given+in+parentheses+under+F[statistics.+Individual+coefficients+are+ statistically+significant+at+the+*10%,+**5%+or+***1%+level.+Both+test+robust+for+heteroskedastic+standard+errors+ When!considering!the!first!model!the!following!is!shown.!At!a!significance!level!of!1%!the! null!hypotheses!that!the!coefficient!of!financial!integration!is!zero!is!rejected.!This!confirms! the!hypothesis!and!the!research!done!by!Didier!Love!and!Perìa!(2012).!The!result!shows!that! for!every!percentage!increase!in!the!difference!of!financial!integration,!during!the!crisis!the! stock!return!correlations!would!have!increased!by!0.0011,!ceteris!paribus.! The!coefficient!on!the!difference!in!level!of!economic!integration!is!significant!at! neither!significance!level.!With!this!model!that!leads!to!the!conclusion!that!the!results!found! by!Didier,!Love!and!Perìa!(2012)!can!be!transferred!to!noneUS!correlations.!Using!this!model,! the!results!also!further!disprove!the!results!found!by!Johnson!and!Soenen!(2003),!Chen!and! Zhang!(1997)!and!Pretorius!(2002).!A!more!mild!conclusion,!would!be!that!their!results! simply!did!not!hold!during!the!2007!financial!crisis.!! The!difference!in!exchange!rate!volatility!over!the!two!period!has!been!shown!to!be! significant!at!the!1%!level.!The!coefficient!on!the!difference!in!exchange!rate!volatility!shows! that!every!0.001!increase!in!exchange!rate!volatility!will!have!decreased!stock!market!return!
correlations!by!0.022.!This!result!confirms!the!findings!of!Bodart!and!Reding!(1999)!and!the! fourth!hypothesis!from!the!previous!section.!An!interesting!finding!is!that!this!factor!is! rejected!even!at!the!1%!significance!level.!For!their!sample,!Bodart!and!Reding!(1999)!could! only!reject!the!null!hypothesis!at!an!inferior!significance!than!found!in!this!research.!This! leads!to!a!tentative!conclusion!that!exchange!rate!volatility!has!a!more!significant!effect!on! stock!market!correlations!during!crisis!periods.! ! The!coefficient!for!the!difference!in!volatility!of!monetary!rates!has!shown!not!to!be! significant!5%!significance!level.!Using!this!model,!the!results!found!by!Ioannidis!and! Kontonikas!(2008)!and!Rigobon!and!Sack!(2004)!are!not!economically!significant!for!the! precrisis!to!crisis!period.!At!the!5%!significance!level!the!hypothesis!that!a!high!difference!in! monetary!rates!between!a!country!pair!leads!to!a!lower!correlation!in!their!stock!market! returns!is!rejected.!It!is!interesting!to!note!however!that,!the!sign!on!the!coefficient!is!in!line! with!what!was!expected!in!the!literature.!In!addition,!the!results!show!that!the!hypothesis!is! rejected!at!the!10%!significance!level,!there!is!some!suggestion!that!the!difference!in! volatility!of!monetary!rates!might!be!significant!using!a!different!metric.! ! Lastly,!the!explanatory!power!of!the!two!dummy!variables!has!been!show!to!been! insignificant!at!the!10%!significance!level.!The!coefficients!on!both!dummy!variables!are! shown!not!to!be!statistically!different!from!zero!at!both!significant!levels.!This!means,!that! this!model!does!not!the!support!the!evidence!presented!by!Kim,!Moshirian!and!Wu!(2005)! for!the!EMU!and!the!number!of!findings!on!the!dominant!position!of!the!US.!Using!this! model!evidence!is!found!to!reject!the!hypothesis!that!if!one!of!the!countries!in!a!country! pair!is!the!US,!stock!market!correlations!will!have!increased!more!during!the!crisis!than!the! other!factors!in!the!model!could!have!explained.!Using!this!model!evidence!is!also!found,!to! reject!the!hypothesis!that!if!the!two!countries!in!a!country!pair!are!both!members!of!the! EMU,!stock!market!correlations!during!the!crisis!will!have!increased!by!more!than!other! factors!in!the!model!could!have!explained.!It!is!interesting!to!note!that!the!signs!on!both!the! dummy!variables!are!in!line!with!what!was!predicted!in!the!literature.! ! As!explained!previously,!in!the!second!model!an!interaction!term!between!the! difference!volatility!in!modetary!rate!and!the!difference!in!economic!integration!during!the! crisis!has!been!added.!The!addition!of!this!variable!leads!to!two!interesting!conclusion.!! ! Firstly,!the!coefficient!on!this!variable!is!shown!to!be!statistically!significant!from! zero!at!the!1%!significance!level.!The!effect!of!the!interaction!term!can!shown!best!when! taking!the!first!derivative!of!the!dependant!variable!with!respect!to!the!difference!in!rate!of! economic!integration:!dρt2−t1,ij
dEIt2,ij = 0.0170 − 0.0297/LM*E<I;N;I<E,<=,)1!.!This!expression!shows! that!if!a!country!pair!has!a!large!difference!in!the!volatility!of!their!monetary!rate!the!effect! that!the!difference!in!economic!integration!during!the!second!period!will!be!decreased.!A! possible!economic!interpretation!of!this!effect!is!as!follows.!In!this!research!it!is!assumed! that!if!a!country!pair!has!a!large!difference!in!volatility!of!monetary!rates,!one!of!the! countries!has!a!central!bank!that!is!commited!to!fighting!the!effect!of!the!financial!crisis!and! the!central!bank!of!the!other!counry!is!indiffirent!to!the!effect!of!the!financial!crisis.!If!this! variable!is!high,!it!could!be!that!investors!cease!to!invest!in!the!country!with!the!indiffirent! central!bank!and!use!that!capital!to!invest!the!country!with!the!commited!central!bank.!This! would!mean!that!firms!in!the!country!with!the!indiffirent!central!bank!would!decrease!in! value!relative!to!the!firms!in!the!country!with!the!commited!central!bank.!This!would!imply! that!the!firms!in!the!country!with!the!indiffirent!central!bank!would!now!have!less!potential! capital!available!to!use!on!imports.!Ceteris!paribus,!this!would!decrease!the!effect!of!
economic!integration!on!the!difference!in!stock!market!correlation.!The!rejection!at!the!1%! significance!level!leads!to!a!strong!conformation!of!the!results!found!by!Chor!and!Manova! (2012),!and!a!conformation!of!the!hypothesis!that!monetary!policy!rates!and!levels!of! economic!integration!will!interact!with!stock!market!correlations.! ! Secondly,!using!the!second!model,!the!coefficient!on!difference!in!levels!of!economic! integration!during!the!crisis!is!now!statistically!different!from!zero!at!the!5%!level.!This! shows!that!if!one!corrects!for!the!interaction!between!the!monetary!rate!and!the!level!of! economic!integration!the!coefficient!on!the!difference!in!the!rate!of!economic!integration! during!the!crisis!itself!will!be!statistically!significant.!!This!model!has!shows!a!possible! justification!for!the!reason!that!Didier,!Love!and!Perìa!(2012)!failed!to!find!a!statistically! significant!effect!for!difference!in!the!rate!of!economic!integraton!on!stock!market! correlations!during!the!financial!crisis.!Using!the!interaction!term!the!model!thus!finds!a! conformation!of!the!research!done!by!Johnson!and!Soenen!(2003),!Chen!and!Zhang!(1997)! and!Pretorius!(2002)!who!did!find!a!statistically!significant!effect!for!the!impact!of!the! difference!in!rates!of!economic!integration!on!stock!market!correlations.!Surprisingly,!this! same!effect!is!not!found!for!the!other!variable!in!the!interaction!term.!The!test!on!the! coefficient!of!the!difference!in!volatilities!of!the!monetary!rate,!still!fails!to!be!rejected!at! either!significance!level.!Finally,!it!is!interesting!to!note!that!the!coefficients!on!the!variables! are!almost!the!same!in!model!2!as!opposed!to!model!1.!For!the!second!model,!two!tests!on! the!validity!of!the!regression!is!presented!in!part!B!of!section!VII.! ! C.! Robustness!analysis!
!
The!data!on!the!correlations!shows!that!about!24%!of!correlations!decreased!during!the! crisis!and!about!76%!of!the!correlations!increased!during!the!crisis.!This!can!also!be!deduced! roughly!from!the!data!in!figure!2.!Using!the!factors!that!have!been!described!in!the!second! model,!two!more!regressions!can!be!run.!Specifically,!increasing!correlations!and!decreasing! correlations!can!be!regression!separately!on!the!factors!described!previously.!The!results! from!the!regressions!with!increasing!and!decreasing!correlations!respectively,!are!presented! in!table!6.!When!comparing!these!results!to!the!results!in!the!second!model!in!table!5!a! number!of!interesting!conclusions!can!be!drawn.!! The!results!of!the!regression!show!that!the!significance!of!the!difference!in!financial! integration!and!difference!in!economic!integration!has!increased!significantly!if!only! increasing!correlations!are!considered.!For!the!decreasing!correlations,!the!difference!in! rate!of!financial!integration!is!no!longer!significant,!while!the!difference!in!rate!of!economic! integrations!has!increased!in!significance.!The!coefficients!have!also!changed!remarkably.! The!results!show!that!a!large!increase!in!the!difference!in!financial!integration!affected!stock! market!correlations!more!for!the!country!pairs!that!experienced!an!increase!in!stock!market! correlations.!The!difference!in!rate!of!economic!integration!during!the!financial!crisis!is! shown!to!have!affected!decreasing!correlations!more!than!when!all!correlations!are! considered.!The!difference!in!exchange!rate!volatility!is!now!insignificant!for!country!pairs! with!increasing!correlations!and!the!coefficient!has!even!become!positive.!For!country!pairs! with!decreasing!correlations!the!difference!in!exchange!rate!volatility!has!decreased.!This! implies!that,!an!increase!in!the!exchange!rate!volatility!affected!country!pairs!with! decreasing!volatilities!more!than!if!all!correlations!are!considered.! !Dependent! variable! :! Change! in! correlation! between! country! i! and! j! between! crisis! and! precrisis;! 122! and! 39! observations! respectively!
Regressor! Increasing!correlations!(Model!3)! Decreasing!correlations!(Model!4)!
8C<=><?,)1! 0.0585***! (0.0130)! (0.0168)!0.0137! %C<=,)1! 0.0101***! (0.0017)! 0.0965***!(0.0332)! /EFGHI*JE;I<E,<=><?,)1! 4.0225! (2.7058)! 926.5851***!(3.6375)! /LM*E<I;N;I<E,<=,)1! 90.0053! (0.0090)! 0.0323***!(0.0045)! P9! 90.0703***! (0.0165)! (0.1300)!0.0048! %RP! 90.0995***! (0.0209)! !! /LM*E<I;N;I<E,<=,)1∗ %C<=,)1! 0.0054! (0.0130)! 90.0348***!(0.0063)! C^U_&`_aU! 0.1552***! (0.0164)! 90.1493***!(0.0296)! F9!Statistic!and!Summary!Statistics! ! ! ! ! ! SER! 0.0761! 0.1040! bc! 0.3465! 0.7309! Table+6:+The+F[statistic+tests+the+hypothesis+that+the+coefficient+on+the+interaction+term+is+zero.+Robust+standard+errors+are+ given+in+parentheses+under+coefficients,+and+p[values+are+given+in+parentheses+under+F[statistics.+Individual+coefficients+are+ statistically+significant+at+the+*10%,+**5%+or+***1%+level.+Both+test+robust+for+heteroskedastic+standard+errors.+ ! The!significance!of!this!effect!also!increased!from!the!5%!level!to!the!1%!if!only!country! pairs!with!decreasing!correlations!are!considered.!For!country!pairs!with!increasing! correlations!the!effect!of!the!volatility!in!the!monetary!rate!remains!nearly!the!same!as! opposed!to!model!2,!both!in!statistical!significance!and!coefficient.!For!country!pairs!with! decreasing!stock!market!correlations!the!significance!of!the!difference!in!volatilities!of!the! monetary!rate!during!the!crisis!is!now!statistically!significant!at!the!1%!level.!The!coefficient! for!the!volatility!in!monetary!rate!suggests!that!an!increasing!difference!in!volatility!of! monetary!rate!correlations!for!country!pairs!with!decreasing!correlations.!The!ability!of!the! EMU!and!US!dummy!variable!to!explain!increasing!stock!market!correlations!beyond!other! factors!has!increased!significantly!for!country!pairs!with!increasing!correlations.!In!this! model!the!dummy!variables!suggest!that!if!one!of!the!countries!considered!was!the!US!or!a! country!pair!was!part!of!the!EMU,!stock!market!correlations!would!have!gone!down,!ceteris! paribus.!For!country!pairs!that!experienced!decreasing!correlations!the!US!dummy!variable! remains!insignificant.!For!country!pairs!with!decreasing!correlations!the!coefficient!and! significance!of!the!interaction!term!remains!largely!the!same.!For!country!pairs!with! increasing!correlations!this!interactions!term!loses!its!significance,!even!at!the!10%!level.!! ! Lastly,!it!is!interesting!to!note!that!for!country!pairs!that!have!decreasing! correlations!the!factors!now!explain!73.09%!of!the!variance!as!opposed!to!the!same!factors! explaining!25.59%!of!the!variance!when!all!country!pairs!are!considered.!For!country!pairs! with!increasing!correlations!the!factors!now!explain!34.65%!of!the!variance!as!opposed!to! explaining!25.59%!of!the!variance!when!all!country!pairs!are!considered.
VI.! Conclusion!and!discussion!
!This!paper!examined!the!significance!and!possible!determinants!of!the!difference!on!stock! market!return!correlations!between!a!precrisis!period!defined!as!July!1st!2004!to!August!1st! 2007!and!a!crisis!period!defined!as!August!1st!2007!to!October!1st!2010.!To!calculate!190! stock!market!correlations,!national!stock!market!indices!of!20!countries!have!been!used.! Using!a!Wald!test,!a!significant!difference!between!precrisis!and!crisis!stock!market! correlations!found.!Specifically,!it!is!found!that!over!the!two!time!periods!the!correlations!in! the!country!pairs!in!the!sample!have!on!average!increased,!in!line!with!the!findings!of! Didier,!Love!and!Perìa!(2012)!for!US!correlations.!! Using!two!different!models!a!number!of!conclusions!can!be!drawn!with!respect!to! previous!research!done!on!the!deteriminants!of!stock!market!correlations.!Specifically,!for! several!factors!the!first!model!has!confirmed!previous!research!that!has!been!conducted.! Conformations!have!been!found!for!the!rate!of!financial!integration!and!exchange!rate! volatility.!Conformations!have!not!been!found!for!other!determinants.!Specifically,!the! volatility!of!monetary!rates!and!levels!of!economic!integration!have!been!shown!to!be! insignificant.!! In!the!second!model,!an!interaction!term!between!volatility!in!the!monentary!rate!and! the!level!of!economic!integration!is!added!as!suggested!by!Chor!and!Manova!(2012).!This! variable!has!been!shown!to!have!a!significant!effect!on!stock!market!correlations.!Using!this! model,!a!significant!effect!is!also!found!for!the!level!of!economic!integration,!contrary!to!the! first!model.!This!model!has!thus!shown!a!possible!justification!for!the!reason!that!Didier,! Love!and!Perìa!(2012)!failed!to!find!an!effect!for!difference!in!the!rate!of!economic! integraton!on!stock!market!correlations!during!the!financial!crisis.!A!similar!effect!for!the! volatility!in!monetary!rates!has!not!been!found.!In!both!models!two!dummy!variables!were! added!to!correct!for!possible!other!factors!that!could!have!increased!or!decreased!stock! market!correlations!beyond!what!is!explained!by!the!other!factor!included!in!the!model.! This!effect!was!not!found!for!either!dummy!variable.! When!only!increasing!correlations!are!considered!as!the!dependent!variable,!the! factors!of!economic!integration!do!a!much!better!job!of!explaining!the!variance!that!the! interaction!term!and!differences!in!exchange!rate!volatility.!Furthermore,!for!increasing! correlations,!the!dummy!variables!now!significantly!some!of!the!variance!in!stock!market! correlations.!Specifically,!the!dummy!variables!suggest!that!if!one!of!the!countries! considered!was!the!US!or!a!country!pair!was!part!of!the!EMU,!stock!market!correlations! would!have!gone!down,!ceteris!paribus.!The!amount!of!variance!that!the!factors!explain!has! increased!as!opposed!to!the!model!when!all!correlations!are!considered.! When!only!decreasing!correlations!are!considered,!the!increase!or!decrease!in!the! difference!of!financial!integration!is!no!longer!significant.!The!significance!of!the!difference! in!rate!of!economic!integration!has!increased.!The!results!show!that!the!rate!of!economic! integration!has!affected!country!pairs!with!decreasing!correlations!by!nearly!a!factor!10! more!that!when!all!correlations!are!considered.!Finally,!as!opposed!to!the!model!where!all! correlations!are!considered,!country!pairs!that!experienced!decreasing!correlations!are! strongly!affected!by!high!differences!in!monetary!rate!volatilities!during!the!crisis.!! It!is!suggested!that!further!research!be!conducted!to!discover!possible!other! determinants!of!differences!of!stock!market!correlations.!Furthermore,!is!this!suggested!to!
and!with!varying!definitions!for!precrisis!and!crisis!time!periods.!Following!the!robustness! analysis,!it!is!suggested!that!further!research!considered!that!effects!of!interpreting! increasing!and!decreasing!correlations!separately.!It!is!also!acknowledged!that!some! assumptions!of!the!tests!in!this!paper!are!not!fulfilled.!Namely,!multivariate!normality!of!the! Wald!test!and!the!absence!of!simultaneous!causality!bias,!in!the!OLS!regression.!The!issue!of! simultaneous!causality!bias,!is!due!to!the!increase!in!correlations!and!the!metric!used!for! financial!integration.!This!issue!is!somewhat!negated,!as!Love!and!Perìa!(2012)!also!found!a! significant!effect!for!financial!integration!using!a!different!metric.!Ongoing!research!on!this! metric!is!important!and!a!no!consensus!has!been!reached!in!the!literature!which!metric!is! most!approprate.!Similar!to!Dellas!and!Hess!(2005)!it!is!suggested!that!further!reseach!is! conducted!to!find!the!most!appropiate!measure!of!financial!integration.! !
!
VII.! Appendix!!
! A.! Table! ! Country! Stock!Index!Used! United!States! S&P! Argentina! MERVAL! Australia! ASX!200! Brazil! IBOVESPA! Canada! TSX! Hong!Kong! HSI! France! CAC!40! Germany! DAX! India! BSE! Indonesia! JSKE! Italy! FTSE!MIB! Japan! NIKKEI! South!Korea! KOSPI! Mexico! MXX! Russia! RTS! South!Africa! JSE! Turkey! XU100! United!Kingdom! FTSE! Netherlands! AEX! Greece! ASE! Table+7:+Countries+and+indices!
! B.! Test!for!normality!of!error!terms!and!serial!autocorrelation!for!model!2! ! Two!important!criteria!outlined!by!Stock!and!Watson!(2011)!to!fulfill!the!validy!of!a! regression!model!is!the!absence!of!ommited!variables!and!normallity!of!error!terms.!Two! other!tests!of!model!validity!will!be!left!out.!Firstly,!a!test!of!heteroskedasticity!is!not! considered!as!the!standard!errors!in!the!regression!of!section!V!are!corrected!for! heteroskedasticity.!Secondly,!a!test!for!simultaneous!causality!bias!is!not!considered!as! possible!sources!of!this!effect!have!been!discussed!in!section!VI.!! ! A!test!for!ommited!variable!bias!is!presented!in!table!8,!as!suggested!by!Ramsey! (1969).! ! Ramsey!RESET!test! F~(3,150)! p9value! Value! 6.34! 0.0004! Table+8:+Ramsey+RESET+test+for+omitted+variables+ At!the!1%!confidence!level!this!test!implies!that!the!model!used!in!this!paper!contains! omitted!variables.!This!finding!is!in!line!with!what!has!been!suggested!in!the!discussion!in! section!VI.!! !! Using!the!data!on!the!residuals!of!the!OLS!regression!of!the!second!model,!one!can! test!for!normality!in!two!different!manners.!Specifically,!normality!can!be!verified!both! visually!and!analytically.!For!the!visual!verification,!in!figure!4!and!5!two!plots!of!the! residuals!is!presented!alongside!a!graph!of!the!normal!distribution.! ! ! Figure+4:+Histogram+of+residuals+and+graph+of+normal++ ++Figure+5:+Graph+of+residuals+and+normal+distribution+ distribution+
