ONWARD PROSPERITY

MEDICIFIRMASOVEREIGNINVESTMENTSTRATEGIES

TheInvestmentStrategiesofSovereignWealthFunds

Sovereign wealth funds have complex objective functions and governancestructures where returnmaximization and strategic political considerationsmay conflict. SWFs with greater involvement of political leaders in fundmanagement are more likely to support domestic firms and invest insegments and markets with higher P/E levels, especially in their domesticinvestments. But these investments see a subsequent reversal in P/E levelssuggesting that the funds engage in poor market timing. The oppositepatterns hold for funds that rely on external managers. Funds that havestated domestic development goals are more likely to invest at home,especiallyifpoliticiansareinvolved.

1.Introduction

Sovereignwealthfunds(SWFs),inadditiontobeingmajorinvestorsincorporateandrealresourcesworld-wide (Fernandez and Eschweiler [2008]), are particularly interesting tofinancialeconomistsbecauseof theirownershipstructureandmission.Thequasi-publicnatureof these fundsmayhaveunique implications for their investmentobjectivesandthegovernancearrangementstheychoose.TheinvestmentchartersofmostSWFsstatethattheyseektomaximizefinancialreturnstoensure long-termpublicpolicies,suchaspension or economic development needs. But the more closely SWFs are exposed topolitical influences, the more they might show major distortions from long-run returnmaximization. As Shleifer and Vishny [1997] argue, such agency problems within largeinstitutionalinvestorscanhavewide-rangingimplicationsforthebroadereconomy.

Theseagencyproblemsmayhavetwoprimarymanifestations.First,thepoliticalprocesscan introduce short-run pressures on SWFs to accommodate public demands for jobcreationandeconomicstabilizationwithinthecountry.Thesedemandsshouldtranslateintofinancialsupportforlocalfirmsorsubsidiesforindustrialpolicieswithinthecountry.There are two opposing views of the consequences of these investment pressures.Advocates for government-directed investments usually rely on a view that financialmarketsareeithernotdevelopedenoughortoomyopic,andthusleavemanyprofitableinvestmentopportunitieson the table (AtkinsonandStiglitz [1980]; Stiglitz [1993]). Theopposing, less sanguine view of politically directed investments suggests that politicalinvolvementcaneither leadtomisguidedpolicyattemptstopropupinefficientfirmsorindustriesorengageininvestmentactivitiesinindustries,sectorsorgeographiesthatare“hot” (Shleifer and Vishny [1994]; Banerjee [1997]; Hart, et al. [1997]). If the former,benevolent view is accurate, wewould expect to find that government investments inlocal firms are directed at industries that face financial constraints and subsequentlyperform very well. If the latter view is true, we would predict the opposite: SWFinvestments would be disproportionately directed to local firms, follow a pro-cyclicaltrend,andsubsequentlyperformpoorly.

TheseconddistortionasaresultofpoliticalinvolvementinSWFs’investmentsmightstemfrom the appointment of politically connected but financially inexperienced managers.Thishypothesiswouldsuggestthatpoliticallyinfluencedfundsnotonlyshowadistortionin thecapitalallocationbetweenhomeand foreign investments,butwouldalsodisplaypoorerstockpickingabilityevenintheinternationalportfolioofthefund.

Sincewe are interested in understandingwhether the investment behavior of SWFs isshaped by (short-term) political considerations, we focus on the funds’ long-terminvestments—acquisitions,purchasesofprivateequity,andstructuredequitypositionsinpublic firms—on the grounds that thesedistortions shouldbemost evident here.Aftermerging three publicly available investment databases—Dealogic’s M&A Analytics,SecurityDataCompany’s(SDC)PlatinumM&A,andBureauvanDijk’sZephyr—weidentify2,662investmentsbetween1984and2007by29SWFs.

OurresultsshowthatSWFswherepoliticiansareinvolvedinthemanagementofthefundaremore likely to invest indomestic firms,while thoseSWFswhereexternalmanagersplayanimportantrolearelesslikelytoshowapropensitytoinvestathome.WeseethatSWFsthathavepoliticalleadersplayinganimportantroletendtoinvestinsegmentsandmarketswithhigherP/Elevels.Thisvaluationeffectisparticularlystrongforthedomesticinvestments of SWFswith politically connectedmanagers. In their foreign investments,thesefundsdonotshowastrongerlikelihoodofinvestinginindustriesandmarketswithespecially high P/Es. In contrast, SWFs with external (professional) managers invest inindustriesandmarketswithlowerP/Es.Whenlookingatthepost-investmentreturns,wefind that the investments of SWFs with politically connected managers tend to see areduction in P/E levels after the investment, while investments of external manager-influenced funds on average experience an increase in the P/E levels. This trend isparticularlypronouncedforinvestmentsathome,whichcouldsuggestthatthepressurestoinvestinhotmarketsathomeareespeciallystrongforpoliticallyconnectedSWFs.

Interestingly,wealso find thatpoliticallyconnected fundsaremore likely to take largerstakes in the firms they invest in, while external managers take smaller equity stakes.Insteadofinvestinginsmallliquidstakes,asthosewithexternalmanagersdo,politicallymanaged SWFs take larger and potentially controlling equity stakes in domestic firms.Againthissuggeststhattheformerinvestmentsaretargetedatsupportingandpotentiallyproppinguplocalfirms,ratherthanoptimizingtheinvestmentreturnsoftheSWF.

It is difficult to reconcile the benevolent view of government-directed investmentwithsomeof our results. In particular, it is hard to understandwhy economic developmentneedswouldcompelfundstoinvestdomesticallywhenequitypricesarerelativelyhigher,whichpresumablyshouldbeatimewhencapitalconstraintsarelesslimiting.Similarly,itishard toexplainwhysocialwelfareconcernswould leadpolitician-influenced funds toinvest in the highest P/E industries, especially in light of the negative returns thatsubsequentlycharacterizethesesectors.Whiletheseresultsareonlysuggestivegiventhe

lack of more micro-level data, they raise a number of important questions about theinvestmentstrategiesandmanagementstructuresofSWFs.

Finally,we investigatewhether the stated objectives of the SWFsmake a difference intheirinvestmentbehavior.Wefindthatfundswhichhavestatedstrategicagendas—e.g.,economic development of the country, macroeconomic smoothing, and the like—aremorelikelytoinvestathome,butonlyifpoliticiansareinvolvedinthefund.Otherwise,the stated objective per se does not seem to lead to a change in behavior. Butinterestingly,wefindthattheexpostchangesinindustryP/Eratiosarenegativeforfundswithstatedstrategicobjectives independentofwhether theyhavepoliticallyconnectedmanagers. These results strengthen the interpretation that thedual pressureson SWFswithpoliticalandfinancialobjectivesareassociatedwithlowerfinancialreturns.

Theplanofthispaperisasfollows.Inthesecondsection,wereviewrelevanttheoreticalperspectives and the earlier studies on SWFs. Our data sources and construction aredescribed in Section 3. Section 4 presents the analysis. Section 5 presents robustnesschecks.Thefinalsectionconcludesthepaper.

2.TheoreticalPerspectivesandRelatedWork

Numerousaccountsbybothobserversandpractitionerssuggestthatthereissubstantialvariation in the investment criteria and sophistication of institutional investors. Inparticular,practitioneraccounts(e.g.,Swensen[2009])suggestthatsomeinstitutionsrelyonoverly rigiddecision criteria and lacka sufficientunderstandingof keyasset classes.Observersattributethesefailurestounderlyingfactorssuchasinappropriateincentives—for example, the limited compensation and autonomy that investment officers enjoy,whichleadstofrequentturnoverandapredilectiontoselect“safe”investments,eveniftheexpectedreturnsaremodest—andconflictingobjectives,particularlythepressuresbyfund overseers to invest in projects sponsored by local entrepreneurs, even if theexpectedinvestmentreturns(andinsomecases,socialbenefits)aremodest.

RecentpapersbyGompersandMetrick[2001]andLerner,etal.[2007]havehighlightedtheheterogeneityininvestmentstrategiesandultimatelyreturnsacrossdifferenttypesofinstitutionalinvestors.However,theevidenceonSWFshasbeenlimiteduntilrecentlydueto many data restrictions. Several papers conduct international stock market reactionanalyses to SWF investment announcements. Kotter and Lel [2008] collect 163 SWF

investment announcements and find a positive market reaction in the two dayssurroundingtheannouncement.Dewenteretal.[2010]findapositivemarketreactionto196stockpurchasesandnegativereactionsto47SWFstocksales.Inthelongterm,theyfindslightlynegativeabnormalreturns.Bortolotti,etal.[2010]reachsimilarconclusions:a positive abnormal return on the day of the announcement and deterioration in firmperformanceoverthefollowingtwoyears.Knill,etal.[2010]analyzeasampleof232SWFinvestments in publicly traded companies.While they find positivemarket reactions toannouncements, the one year abnormal returns varies by SWF characteristics such aswhethertheSWFisfromanoil-producingcountry,thedegreeofopacity,andinvestmentsinnon-financialtargets.

ChhaochhariaandLaeven [2009] takeadifferentperspectiveregardingSWF investmentchoices.TheyconsiderthegeographicaldecisionsmadebySWFswheninvestingabroadinpubliccompanies.Theyfindthatfundslargelyinvesttodiversifyawayfromindustriesat home, and do so mainly in countries that share the same ethnicity, language, andreligion. Fernandes [2009], rather than exploring transactions, focus on SWF holdings.Using data on 8,000 firms between 2002 and 2007, he finds that the stakes in SWFs’public investmentsare small:at the95thpercentileof thesample,SWFshold less than1.5%ofthecompany.Moreover,hefindsapositivecorrelationbetweenSWFownershipandfirmperformanceandvaluation.

WhilethepapersoutlinedaboveexploreSWFsthroughthelensesofinternationalequitymarkets,ouranalysis’sfocusistheirmostsubstantialtransactions(themedianacquisitionstake is50% inoursample).Thepapermostcomplementarytoours isDyckandMorse[2011], who construct a sample of SWF investments in public equity, real estate, andprivate equity between the years 1999 and 2008. They find that relative to variouscapitalization benchmarks, there is a significant bias towards private equity. Thecombined private equity and real estate holdings account for almost half of SWFs’portfolios.DyckandMorsefindthatsubstantialexplanatorypowercanbeattributedtoeitherfinancialreturnmaximizationorstateplanningmotives,demonstratingthetensionbetweenthetwoobjectives.OurpaperdiffersfromDyckandMorseintwodimensions.First, our focus is on SWFs’ investmentperformance, rather thanonportfolio holdings.Second,whileDyckandMorseinferfunds’objectivesfromportfolioholdings,wetakeadifferentapproach,byusingSWFgovernancestructureandstatedobjectivestoexplorehowtheseinteractwithinvestmentdecisions.

3.DataSourcesandConstruction

To analyze the direct investment strategies of SWFs, we combine three sets of data:informationontheSWFsthemselves,thedirectinvestmentsthatthefundsmade,andtheinvestment climate around the time of the transaction. The data for all the threecomponentsaredrawnfrompubliclyavailablesources.

SWFsampleconstruction:WestartwithapreliminarysampleofSWFsbycombiningtheprofiles of the funds published by J.P.Morgan (Fernandez and Eschweiler [2008]) andPreqin(Friedman[2008]). Inthecaseswherethetwodatabasesusedifferentnamesforthe same SWF, we employ the fund address and related information to eliminateduplicates. We add five funds to the sample that were not included in these twocompilations but are frequently described as SWFs in at least one of the investmentdatasetsnotedbelow.This initial searchyieldsapopulationof69 institutions, includingsomeSWFsthathavebeenannouncedbutarenotyetactive.

WethenmergethisinitialsampleoffundswiththeavailabledataondirectinvestmentsandcharacteristicsofSWFs.WearecarefultoextractinvestmentdataforboththeSWFsand their “subsidiaries,” which we define as entities in which SWF has at least a 50%ownershipstake.ThetwoSWFdirectoriesandtheinvestmentdatasetsnotedbelowdidnot always explicitly note the links between SWFs and their subsidiaries. To extracttransactions involving SWF subsidiaries, we supplement our list of SWF subsidiaries byemployingownershipdataintheDirectoryofCorporateAffiliationsandBureauvanDijk’sOrbis.

SWF Characteristics: The fund profiles in the J.P.Morgan and Preqin databases containinformationonthesizeandoperationsofthefunds.Ifthereisadiscrepancybetweenthetwodatabases,wereconfirmtheaccuracyoftheinformationthroughwebsearchesandnewspaperarticles.Thekeyvariablescollectedare:

· Assets under Management—J.P.Morgan and Preqin profiles containestimatesof fundsizes. In caseofdiscrepancies, J.P.Morgan’sestimateofassetsunder management is given preference. Preqin’s estimate of assets undermanagementisusedonlywhennoJ.P.Morganestimateexisted.

· The Presence of Politicians in the Managing Bodies—The J.P.Morganreportemphasizesgovernancestructuresoffunds.Wereadcarefullytheprofiles,andformadummyvariablethat indicatesifafund’sprofilecontainsevidenceof

presence of politicians in the governance of the fund. For example, KhazanahNasional’s profile indicates that the fund’s board of directors “has an eight-member Board comprising representatives from the public and private sectors.AbdullahAhmadBadawi, theRightHonorablePrimeMinisterofMalaysia, is theChairmanoftheBoardofDirectors.”TheAlaskaPermanentReserveFund’sprofileindicatesthatthefund’sBoardofTrustees“iscomprisedoffourpublicmembers,the Commissioner of Revenue and one additional cabinet member of thegovernor'schoosing.”Similarly,theJ.P.Morganreportindicatesthattheboardofdirectorsof theGovernmentofSingapore InvestmentCorporation (GIC) includesLee Kuan Yew as the chairman, and Lee Hsien Loong as deputy chairman. TheformerisSingapore’sministermentor,andthelatterisSingapore’sprimeminister.

·RelianceonExternalManagers/Advisors—TheJ.P.Morganvolumealso indicateswhether the governance of the fund is in the hands of a board consisting ofinvestment professionals and/or outside business leaders. We create a dummyvariable that is one if the report contains evidence that the institution reliesheavilyonexternalmanagementoradvisors.1Forexample,theJ.P.MorganprofileindicatesthattheHongKongExchangeFund“employsexternalfundmanagerstomanageaboutonethirdoftheFund’sassets, includingallof itsequityportfoliosand other specialized assets.” Similarly, the profile of Abu Dhabi InvestmentAuthority (ADIA) indicates that “approximately70% to80%of theorganization’sassetsaremanagedbyexternalfundmanagers”.

· Stated InvestmentGoal—The Preqin tabulation reports the stated goals of theSWF. Some funds have multiple goals. To simplify the analysis, we combine insomeanalyses theobjectives into twogroups: strategicobjectives (Managementof Government Assets, Acquisition of Strategic Assets, and DomesticDevelopment) and non-strategic objectives (Investment of Oil/CommodityRevenues, Currency Reserve Management, and Pension Funding). When SWFs’descriptions include objectives from both groups, we included all thesetransactions inthenon-strategicgroupandverifiedthatresultsaresimilarwhenincludedinthestrategicgroup.

Thesemeasures, itmust be acknowledged, have important limitations. First, these arereportedasof2008:wedonothaveatimeseriesonthegovernanceoforadvisorusagebythefunds.Second,thesemeasuresareextremelycrudecharacterizationsoftheSWFs’

organizationalstructures.

1 We classify a fund as relying on external managers if the reliance is sufficiently important that it isdocumentedinthereport.

InvestmentData:InformationregardingSWFtargetinvestmentsisidentifiedinDealogic’sM&A Analytics, SDC’s PlatinumM&A, and Bureau van Dijk’s Zephyr. All three of thesedatabases compile information on direct investments by institutional and corporateinvestors.Transactionsincludedinthedatabaseencompassoutrightacquisitions,venturecapital and private equity investments, and structured minority purchases in publicentities(frequentlycalledPIPEs,orprivateinvestmentsinpublicentities).Thedatabasesdonot include investments intohedge,mutualorprivateequity funds,oropenmarketpurchasesofminoritystakesinpubliclytradedfirms.

In each of the three datasets,we runmultiple keyword searches for every fund in thesample. We also search for investments carried out by their subsidiaries. Finally, textfieldsofacquirerdescriptionsaresearchedforphrasessuchas“SWF,”“sovereignfund,”or“sovereignwealthfund.”Theseadditionaltransactionsareexamined,andifthereisamatch intheSWF’s identity (e.g., if there isaslightmisspellingof theSWF’sname)andlocation,theentriesareaddedtothedatabase.Thevariablesweobtainabouteachdealare theannouncementdate, transactionsize, shareof theequityacquired,andcountryandindustryofthetarget.Inthecaseofdiscrepanciesacrossthedatabases,weusepressaccounts and web searches to resolve the differences. Some of the databases includeproposed deals that were not consummated. If the transactions are described in thedatabasesas“withdrawn”or“rejected,”wedropthemfromtheanalysis.

Aftermergingthethreedatabases,weare leftwith2,662transactionsbetweenJanuary1984andDecember2007by29SWFs.Weconfirmthatthebulkofthefundsthatarenotincludedareeither verynew (indeed, somehadnot yet commencedoperationsby theendof2007)orverysmall.Ofthe29institutionswithtransactionsinoursample,24areprofiledineithertheJ.P.MorganorPreqinvolumes,orinbothpublications.Thereexist23J.P.Morganand16Preqinprofilesforthefundsinoursample.WedescribearobustnesscheckthatseekstoassesswhetherselectionbiasesaffectourresultsinSection5.

Inthebulkoftheanalysesbelow,wealsoexclude36transactionswherethetargetswereinCentralAmerica,SouthAmerica,orAfrica.Thisdecisionreflectsourdesiretofocusoninvestments in the major markets—i.e., Asian, Middle Eastern, andWestern countries(NorthAmerica,Europe,andAustralia)—wherethevastmajorityofthe investmentsareconcentrated.

Environment Data: We also characterize the pricing and subsequent returns in theindustryand thenationof the transaction. Ideally,wewouldhave liked toanalyzedealpricingusing the actual target firm’s P/E ratio.However, sincemost SWFs’ investmentsareinprivatefirms,thesedataarenotavailable.

Theperformancedatawhichweuseare:

· Industry P/E ratios— To obtain a measure of deal valuations, we use theweighted

average of the P/E ratios of firms in the target company’s industry and nation(determinedbycompanyheadquarters).TocalculatetheP/Eratiosforthetargetcountries, we use the P/E ratios of public companies in the same industry andcountry from the Datastream database, dropping companies with negative P/Eratios.2WeightedaverageP/E ratioswere formed foreach target investmentatthecountry-industry-year level(usingmarketvaluesofthefirmsasweights).Weused industry classifications based on the Standard Industrial Classificationscheme.3ThedistributionofP/Evalueswaswinsorized4atthe5%and95%levelin order to reduce the impact of extreme observations. We also construct anapproximate performance measure for each deal: the change in the weightedmeanindustry-countryP/Eratiointheyearfollowingthetransaction.

·HomeP/EandversusOutsideP/E—TomeasuretheP/Elevelsinthehomenationversusoutsidethenation,weconstructHomeP/EandOutsideP/EvariablesusingtheMSCIdatabase(downloadedfromDatastream).Theseratiosareweightedbymarket capitalization and measured at the country-year level. We completemissing country-level P/E ratios using the Zawyadatabase andDatastream’s P/Eindexesforemergingmarkets.Forinvestmentsmadeabroad,thevariablesHomeP/EandOutsideP/EcorrespondtotheP/ElevelofthehomecountryoftheSWFandthetargetcountry,respectively.Ifinvestmentsaremadeathome,theOutsideP/Evariableequalstheweightedaverage(bythetotalamount investedbySWFs

over the sample period) P/E ratios of all countries in which investments weremadebySWFs,excludingthehomecountry.

2ThemainchallengewastogetP/EratiosforMiddleEasterntargets,particularlyinthePersianGulfregion.In73cases,wecouldnotcompute a P/E ratio using the Datastream information.3 We use a broader definition than the 2-digit SIC level, since under thisclassificationthenumberofcompaniesperindustryisverysmallinsometargetcountries.

4Most of the affected transactions are concentrated in the Asian region. The estimators aremore robust when we winsorize theextremeobservations rather than trimming the sample (and thusdisproportionately removingAsian regiondeals).Nevertheless, theresultsremainunchangedwhenwetakethelatterstrategyandexcludeoutliers.

4.Analysis

4.1.Samplestatistics

Table 1 presents the descriptive statistics of the 2,662 transactions made by the 29sovereign wealth funds in our sample. Panel A of Table 1 sorts the funds into threeregions:Asia,MiddleEast,5andWesterngroups.TheWesterngroupincludesfundsfromNorth America, Australia, and Europe.Our sample consists of seven funds in the Asiangroup,15fundsintheMiddleEasterngroup,andsevenfundsintheWesterngroup.ThenumberoftransactionsofAsianfunds(2,045observations)issubstantiallylargerthantheMiddleEasterngroup(533observations)andthe84observationsoftheWesterngroup.

Onepossibleexplanationforthesedifferencesinsamplesizeisthatwehaveonlypartialcoverage of the deals. We believe, however, that this can only explain part of thedifferences. More important, we believe, are the differences in fund sizes and thewillingness toengage indirect investments.Forexample, theaverageAsianandMiddleEastern funds have $132B and $124B under management, respectively, and aresubstantially larger than the averageWestern fund ($40B). Moreover, to estimate thecoverageofoursample,wecomparetheaggregatetransactionvalueofoursampletotheestimate in a J.P.Morgan publication (Fernandez and Eschweiler [2008]). They estimateoutstanding SWF investments in alternatives (encompassing hedge funds and privateequity)attheendof2007as$316billion.Inoursample,theaggregatetransactionvaluein the years 2003-2007 (excluding the public investments) is $198 billion (expressed in2008 U.S. dollars). Given that we exclude private equity partnerships and hedge fundinvestments,thecomparisonsuggestswehavereasonablesamplecoverage.Ourestimate

5 We add the single investment by the Venezuelan SWF to the totals for the Middle East, given thepetroleum-drivennatureofthateconomy.

(as all analyses in this paper) relies on winsorized transaction values.Were we not towinsorizetransactionvalues,oursamplecoverageisevenbetter.

While the sample consists of transactions between the years 1984 and 2007, Panel Bdemonstrates that more than 97% of the transactions are after 1991. While both theAsian and Middle Eastern funds’ investments go back to the mid 1980s, the Westernfunds’ investments are more recent, beginning around 2003. Panel C shows that theaverage transaction size is $158 million (in 2008 U.S. dollars) and the average stakeacquiredbytheSWFsissubstantial(56.6%).PanelCalsodemonstratesthattheaverageP/E level inthe industry-country-yearof thetargetofaSWFtransaction is25.6andthetypicalinvestmentsegmentexperiencesadropof-1.2%inthemeanP/Eratiointheyearafter the investment. For the approximately 20% of the investments in publicly tradedfirms, we also examine the market-adjusted returns in the six months after thetransaction(seethedetaileddescriptionbelow).

PanelDreportsvariablesthatcapturethegovernancestructureofthefunds.Recallthatforeachfund,wedevelopindicatorvariablesforwhetherpoliticalleadersareinvolvedinthe board andwhether the fund relies on externalmanagers. About 24% of the fundshavepoliticiansinvolvedinthefundand28%ofthefundsrelyonoutsidemanagers.Wesee that both funds with political leaders and external managers tend to make largerinvestments. Interestingly,whenpoliticiansare involved,funds investmore inthehomecountry(44%ofthedealsinthesample),relativetofundswithouttheirinvolvement(only31%ofthetransactions).Fundswithexternalmanagersinvolvedinvestlessinthehomecountry(8%)relativeto36%forfundsthatdonotrelyonexternalmanagers.

The final panel of Table 1 reports the stated fund objectives. Currency reservemanagement is the objective associatedwithmost funds andmost transactions. Fundswhose stated goal is themanagement of government assets have the largest share ofdomestic investments; thosewhose goal is the investment of oil/commodity revenues,thefewest.

WenowanalyzewhetherthecharacteristicsoftheSWFsareassociatedwithdifferencesin their investment strategies, focusing on the likelihood of political interference. In

particular,wefocusontheimpactofthegovernancestructureofthefunds—i.e.,whetherthe SWFs rely on externalmanagers for investment advice andwhether politicians areinvolvedinthefund—andtheSWFs’statedobjectives.WeanalyzeinvestmentstrategiesofSWFs,lookingattheirpropensitytoinvestathome,theindustry-countryP/Elevelsatthetimeoftheinvestments,thesubsequentchangesintheP/Eratios,andthesizeoftheacquisitionstakesoftheirinvestments.

Theunitofobservationinouranalysisisatthetransactionlevel(thatis,foraspecificSWFandtarget),withstandarderrorsatclusteredatthelevelofthenationinwhichthefundisbased.Inmanyregressions,wecontrolfortheyearthattheinvestmentismadeandthesovereignwealth fundmaking the investment. Inmost specifications,we useweightedregressions,witheachobservationweightedbythetransactionsize(transactionsizesareallexpressedin2000U.S.dollars).Sinceweonlyhavesizesfor67%ofourtransactions,we imputemissingweights by constructing the fitted values from a regression of dealsizes on fixed effects for the investment year, target industry, target region, and fund.Afteraddingimputedobservations,wewinsorizethedealsizevariableatthe5%and95%level,inordertoreducetheimpactofextremeobservations.

4.2.Propensitytoinvestathome

Inordertoanalyzehowfundsvaryintheirallocationofinvestmentsbetweenthehomenation and outside, we estimate a probit model. The dependent variable is a homeinvestmentdummy,whichequalsoneifthetarget investment ismadewithinthehomenation of the SWF and zero otherwise. In Table 2, columns (1) and (2),we regress thehome dummy on indicator variables for the presence of political leaders in themanagementofthefundandtherelianceonexternalmanagers.WeclusterthestandarderrorsatthelevelofthecountrywheretheSWFisbased.Thedisplayedcoefficientsaremarginal effects. In the specificationswhere year dummies are added, the sampleonlyincludes transactions from 1991 onward. All regressions are weighted by winsorizedtransactionsizes(expressedin2000U.S.dollars),asdescribedabove.

Inthebasespecification,column(1)ofTable2,wecontrolforthehomecountry’sgrossdomesticproduct (expressedusingthe logarithmofGDP, in2000U.S.dollars),andGDPgrowth in the calendar year prior to the year of the investment. In subsequentregressions,wecontrolfortheyearoftheinvestmentandthegeographiclocationoftheSWF(Asian,MiddleEastern,andWestern).Theresults incolumn(1)showthat incaseswhere political leaders are involved with the management of the funds, domestic

investments are more common, while involvement of external managers is associatedwithfewerdomesticinvestments.Themagnitudeoftheeffectsislarge:thecoefficientonthepoliticiandummyreflects36% increase in the likelihoodof investingathomewhenpoliticiansareinvolved.Incomparison,thecoefficientontheexternalmanagerdummyisequivalenttoa26.3%lowershareofdomesticinvestmentswhenexternalmanagersareemployed. Incolumn(2),weaddtransactionyear fixedeffects.Theresultschangeonlyslightly,from+36.0%to+32.9%forthepoliticiandummyandfrom-26.3%to-24.7%fortheexternalmanagerone.

To get a better understanding of the decision where to invest, we include controlvariablesforthepricinglevelsoftargetmarketsathomeandabroad,asdiscussedinthedataconstructionsection.Incolumn(3)ofTable2,werepeattheregressionfromcolumn(1),butaddmeasuresoftheHomeP/EleveloftheSWF’snationandOutsideP/E,theP/Elevelof thecountry that the fund invests in.Theresults showthat there isasignificantcorrelation between the Home P/E level and the likelihood of investing at home orabroad.SWFsaremore likely to investathomewhenprices therearerelativelyhigher.Themagnitudeofthiseffectissubstantial:anincreaseofonestandarddeviationofHomeP/Eincreasesthelikelihoodofinvestingathomeby4.5%.

Similarly,higherP/Elevelsintheothercountriesarecorrelatedwithalowerpropensitytoinvest at home. An increase in one standard deviation of Outside P/E decreases thelikelihoodof investing athomeby3.1%. Ifweadd year fixedeffects in column (4), thecoefficienton

7 Inunreported regressions,weexplore the impactof theAsian financial crisison the likelihood to investathomeby

adding a Post-Crisis dummy variable. We add interaction terms with Asian and Middle Eastern groups to capture

differentialreactiontotheAsianfinancialcrisis.Whilewefindpositiveeffectonlikelihoodtoinvestathome(withthe

effectlargestatMiddleEasternSWFs),theresultsaresensitivetothenumberofpost-1998yearsweusetodefinethe

Post-Crisisvariable.One importantreasonwhytheresultsarenotveryrobust is thatpriorto1995oursamplesize is

relativelylimitedcomparedtolateryears.

Home P/E is still positive, but much smaller and insignificant. The coefficient on theOutside P/Ebecomes significant at 5% confidence level and themagnitude is larger. Incolumn (5),we find that the results hold evenwhenwe use both the governance and

valuationmeasures.

In column (6), we add a corruption index8 for the country, taken from InternationalCountry Risk Guide (2008). This index is scaled from zero to ten, with higher numbersrepresentinglesscorruptnations.Morecorruptionisassociatedwithagreateremphasisondomesticinvesting:adecreaseofoneinthecorruptionscaleisassociatedwitha10.8%increase in the likelihood of domestic investment. In an unreported regression, theinteractionsbetweenthegovernancedummiesandthecorruptionscoreareinsignificant:e.g.,politicalinvolvementwithfundmanagementisassociatedwithastronghomebiastoinvestments,regardlessofthelevelofcorruptioninthenation.

The cross-sectional results suggest that SWFs invest less at home if their local equitymarketshaverelatively lowP/Elevels.9OnepossibleexplanationforthispatternisthatSWFs shun low-valued local markets because these financial markets are not as welldeveloped.Butthishypothesishasdifficultyexplainingawaythefactthatthepropensitytoinvestabroad

8Anothervariableofinterestisthehomecountrylegalorigin,anditsimpactonSWFinvestmentbehavior.However,this

variable has very little variation in our sample, asmost countries have common law legal origin.9We also explore

whether different determinants affect domestic investments in public versus private firms. To do so, we re-estimate

Table2separatelyforprivateandpublicfirms.Wefindthatthedecisiontoinvestathomeinpubliclylistedcompaniesis

sensitivetomarketconditionsathomeandabroad,whileinvestmentsatprivatecompaniesaresensitiveonlytoforeign

stockmarketconditions.However,thedifferentialsensitivityofprivateandpublicinvestmentstomarketvaluationlevels

cannotbeunambiguouslyinterpreted,sincemarketvaluationmightbeaworsemeasureofinvestmentopportunitiesfor

privatefirmscomparedtopublicfirms.

Increasesasthepricinglevelinforeignmarketsrises.Rather,itappearsmoreconsistentwiththeinterpretationthatSWFsengagein“trendchasing,”thatis,gravitatetomarketswhereequityvaluesarealreadyhigh.

4.3.Valuationlevels

In a second step, we examine whether there are significant differences in the markettimingofthetransactionsundertakenbySWFs.InthefirstthreecolumnsofTable3,thedependentvariable is theweightedaverage (by firmvalue)of theP/E ratiosofpubliclytradedfirmsintheindustry,country,andyearofthetransaction.Theestimationmethod

isaweightedordinaryleastsquaresmodel,usingasweightswinsorizedtransactionsizes(convertedtoU.S.dollarsasoftheyear2000).Robuststandarderrors,clusteredbythecountriesinwhichtheSWFsarebased,areshowninparenthesis.

Asbefore,ineachspecification,wecontrolforthelevelandgrowthofGDPintheSWF’snation, the location of the SWF, and the year of the investment.We also add controlsacrossallspecificationsforwhethertheinvestmentisadomesticoneandthegeographicregionofthetargetcompany(usingthebroadregionalclassificationasbefore).InthefirsttwocolumnsofTable3,weincludeadummyfortheinvolvementofpoliticalleadersandofexternalmanagers.Wefindthathavingpoliticiansinvolvedisstronglyassociatedwithinvestments inhigher-priced sectors (apremiumof three to four timesearnings),whileexternalmanagersareassociatedwithinvestmentsinlower-valuedsectors.Incolumn(2),we include a control for the country corruption index. SWFs based in more corruptnationsareconsistentlyassociatedwithinvestmentsinhigher-valuedindustries(anaddedmultipleoftwotothreetimesearnings).

Incolumn(3),weincludeaninteractiontermbetweenpoliticiansandhomeinvestments,as well as an interaction between external managers and home investments. Whenincludingtheinteractions,weseethatthecoefficientonpoliticiansinteractedwithhomeinvestment is statistically significant and economically large, reflecting a 6.9 higher P/Eratio when politicians invest at home.10 The direct effect of the dummy for politicianinvolvement becomes insignificant and small. We do not find a similar significantcoefficient on the interaction term for the external managers and home investment.TheseresultssuggestthatpoliticianstendtoparticularlyinvestinhighP/Edealsathome,while the effect is much less pronounced abroad. For external managers, there is nodifferenceintheinvestmenttimingbetweendomesticandforeignmarkets.

4.4.Investmentperformance

Bythemselves,theseresultsabovecouldbeconsistentwithverydifferentinterpretations.First,thefactthatSWFswithpoliticians’ involvementare investing inhigherP/Esectorsmightnotbetroubling.Iftheseratiosaretruereflectionsoftheinvestmentopportunitiesof the firms, these institutions may be rationally choosing sectors that are likely toperformwell. This view (Gordon [1959]) was recently used by Bekaert et al. [2007] tomeasure countries’ growth opportunities. A second interpretation relies on theassumption thatmarkets can be inefficient and that higher P/E ratios reflect investors’overreaction(Lakonishok,Shleifer,andVishny[1994]).ThisviewsuggeststhathighP/Es

mayreflectoverpaymentonthepartoftheSWFs.

10Intheequallyweightedregressions,theinteractiontermisinsignificantwithsmallermagnitude.

Toshedsomelightonthesetwocompetinginterpretations,welookattheperformanceof investments in the industry and country in the year after the deal. If the firstinterpretation is true,weshouldsee thatSWFsoutperform inhome investments,whiletheoppositewouldholdunderthesecondexplanation.

TheregressionsinColumn(4)-(6)inTable3arestructuredtobeparalleltothefirstthreecolumns,butnowthedependentvariableisthepercentagechangeinthemeanP/Eratioof firms inthatcountryand industry intheyear followingthe investment.By lookingatthe subsequent performance of the sector, we can address some of the interpretativechallenges highlighted above. As in the previous section, we use a transaction size-weightedordinaryleastsquaresspecification.

In column (4),we see that SWFswhere political leaders play a role select sectorswithsignificantdropsinP/Egoingforward(-3.9%).Thisisincontrasttothecasewhenexternalmanagers are involved,whereP/E values increase in the year following the investment(+2.0%).We repeat this estimation in column (5), nowcontrolling for corruption levels.Whilecorruptionby itselfhasnosignificanteffectonreturns, thegovernancemeasuresbecomestronger.Thecoefficientonpoliticiansbecomes-4.6%,whilethepositiveeffectof external managers grows to +2.7%. When interactions with home investments areadded in column (6), the interaction term between politician influence and homeinvestments is negative and substantial, reflecting a decline of 7.5% in returns wheninvestingathome.Thecoefficientissignificantatthe10%threshold.11

TheanalysissuggeststhatSWFswithpoliticianinvolvementdonotselecthighP/Esectorsbecausetheyhavebetterprivate informationabout investmentopportunities.Rather, itseems to reflectawillingness to trendchaseandoverpay for investments. Theanalysissuggests, at least weakly, that these effects are stronger when it comes to domesticinvestments.

In Table 4, we undertake a robustness check of the performance regressions. A keydrawbackoftheanalysesinTable3isthatweexaminedtheperformanceatthelevelofthechosenindustryandnationasawhole,andnotthatoftheindividualcompaniesthat

theSWFsinvestedin.Theperformanceoftheseselectedfirmscouldbeunrepresentativeofthereturnsmoregenerally,ortheexperienceofprivatefirmscoulddifferfromthoseofthe public concerns whose P/Es we can trace. To partially address the first of theseconcerns,weexaminethesubsetoffirmsthatwerepubliclytradedatthetimeoftheSWFinvestment. We search the Datastream database for all target companies that werepubliclytradedandextracttheirmonthlyreturns.Wedeterminethebenchmarkreturnsfor the stock exchanges in which the target companies were traded and extract thosereturnsaswell.Wecomputecumulativeabnormal returns relative to thebenchmark inthesixmonthsafterthetransaction,which leadstoaconsiderably largercoveragethanone-yearreturns.

11Theequallyweightedregressionsdemonstratesimilarresults,however,thecoefficientsarenoisier,andinsomecases

insignificant.

WeestimateinTable4transactionsize-weightedOLSregressionssimilartothoseinthefinal three columnsof Table 3, but nowwith the differencebetween the return of thetarget in the six months after the transaction and the return of the correspondingbenchmarkoverthesameperiodasthedependentvariable.Weuse533observationsintheseestimations.

We find once again that in the basic specifications, politician-influenced SWFs areassociatedwith lowerreturns.Thesetransactionssignificantlyunderperform,generating16%lowerreturnsinthesixmonthsaftertheinvestments.Thehomeinvestmentdummynow has a significantly negative coefficient, suggesting underperformance amongdomestic investments.While the sample of publicly traded transactions is considerablysmaller,thesimilaritytotheresultsinTable3isreassuring.

Inanunreportedregression,werepeattheanalysisusingameasurewhichaddressesthesecond concern. We use as the dependent variable the percentage change in theweighted(byfirmvalue)averageEBITDA/Assetsratioofallpubliclytradedfirmsiftargetispublic,orifthetargetisprivate,allprivatelyheldfirmsinthecorrespondingthree-digitSICindustry,country,andyearofthetargetinthetransaction.

Wedeterminetheratiosforthecorrespondingfirmsfromthe2009editionofOrbisfromBureau van Dyck, which includes financial information about private firms for manynations. The important advantage of Orbis is that it includes data on both public and

private firms (in fact,most of the firms in this database are private). Unfortunately, inmany cases, the information is quite scanty, so we can only obtain a ratio for thecorresponding industry, country, and year for 796 firms: far worse than the P/E ratio,where we have a benchmark for 2,553 firms. The results are quite weak. In the basicregressions, the politicians variable retains a negative coefficient and the externalmanagersoneapositiveone,butneitherarestatisticallysignificant.

4.5.Acquisitionstake

InTable5,weexplorehowthesizeoftheacquisitionstakesvariesintheinvestmentsthatdifferent SWFs make. Unlike the other tables in the paper, we use equally weightedregressionshere,sinceweightingbasedondealsizeswillbiasourresults.Otherwise,theset-up of the table is parallel to the specifications in Tables 3 and 4, with an OLSspecification.

Incolumn(1),wefindthattheacquisitionstakeofSWFswithexternalmanagerspresentis lower by 10.9%. Investments at home are 12.2% larger, an effect that is statisticallysignificant. Political leaders’ involvement is associatedwith 7.9% increase in acquisitionstake, although the coefficient is not statistically significant. When we control forcorruptionlevelsincolumn(2),wefindthatthecoefficientonthedummyforpoliticians’involvementbecomessignificantandhasalargermagnitude(a10.7%increase).Similarly,the coefficient on externalmanagers remains statistically significant (13.1%). In column(3), we add interaction terms with home investment dummies. Interestingly, theinteraction term between politicians and home investment dummy is insignificant;suggesting that politician-influenced SWFs’ propensity to acquire larger stakes isindependentofthelocationofthetargetcompany.

4.6.Statedinvestmentobjective

These institutions often differ in their stated investment goals. Some SWFs profess adesiretofocusonmorestrategicobjectives,suchastheacquisitionofusefulcompaniesordomesticdevelopment.Othersaimmoreatthelong-termreturngoalsthatareakintothoseofauniversityendowment.

InTable6,werepeattheanalysesofTables2and3,butlookspecificallyattherolethatinvestment objectives play. Using the Preqin data, we define funds’ objectives to be

strategic if stated goals include Management of Government Assets, Acquisition ofStrategicAssets,orDomesticDevelopment.Weconsidertherestoftheobjectivesasnon-strategic (Investment of Oil/Commodity Revenues, Currency Reserve Management, orPensionFunding).Mostfundsincludemultiplegoals,whichtypicallyfallunderthesamebroadcategory.In220investments,fundgoalsincludedbothstrategicandnon-strategicobjectives.Weincludedallthesetransactionsinthenon-strategicgroup,andverifiedthatresultsaresimilarwhentheseareincludedinthestrategicgroupinstead.

The first twocolumnsrepeat thespecifications inTable2,estimatingaweightedprobitmodelwhere the dependent variable is the home country indicator. In column (1),weregressadummyforwhethertheinvestmentisathomeonourstandardvariablesplusaninteraction between a dummy forwhether politicians are involved in the fund and thestrategicobjectiveofthefund.Wealsoincludeadummyvariableforstrategicobjectives.Wefindthatthecoefficientonthestrategicobjectiveindicatorisinsignificantandclosetozero, suggesting no increase in the likelihood to invest domestically. However, whenpolitical leadersare involved, those fundswhichhave strategicobjectives showamuchhigherprobabilityof investingathome: thecoefficientof the interaction termbetweenpoliticiansandstrategicobjectivesissignificantat1%level,andindicatesa74.5%increaseinthelikelihoodtoinvestathome.Werepeattheanalysis incolumn(2),controllingforlevelofcorruption,andfindthattheresultsaresimilar.

ThesecondpairofregressionsinTable6usesasthedependentvariabletheaverageP/Eratio in the target country, industry, and year of the transaction. We repeat thespecification in Table 3, and add the direct effect for strategic objectives and aninteractiontermwithpoliticians’involvement.Incolumn(3),wefindthatthecoefficientof the strategicobjectivedummy is significantat the5% level and reflectsahigherP/Eratioof2.5. This effect is independentofhavingpoliticians involved in theSWF, as theinteraction term is now insignificant. We repeat the analysis in column (4), addinginteraction termswith home investment dummy. The result from the previous columnremains,withtheinteractiontermbecomingbigger(3.8)andstatisticallysignificantatthe1%level.Interestingly,theinteractiontermofpoliticiansandhomeinvestmentdummyisstill significant and positive, similarly to our previous findings. This suggests that theincreased tendencyofpoliticians to investathighP/E ratiosathome is independentofthefund’sstrategicobjectives.

ThelastpairofregressionsinTable6usesasthedependentvariabletheone-yearchangeinP/Eratios. Incolumn(5),wefindthat fundswithstrategicobjectivesperformpoorly,

with a significant drop of 1.7% in sector performance. Similarly, the direct effect ofpoliticians’involvementishighlysignificantatthe1%level,withadropof3.8%.However,the interaction term of these two factors is insignificant and close to zero. Thespecification incolumn(6)adds interaction termsof thehome investmentdummywiththepoliticalinvolvementmeasure.Thedirecteffectofstrategicobjectivesnowbecomesevenmoresignificantanditsmagnitudegetsbiggerinabsoluteterms(-2.7%).TheearlierresultfoundinTable3ofpolitician-influencedfundsperformingpoorlywheninvestingathomeexistshereaswell,evenwhencontrollingforthestrategicobjectivesofthefund.Theinteractiontermbetweenpoliticians’involvementandthehomeinvestmentdummyreveals a large dropof 7.5% in such investments.Overall, this implies that the politicalleader-influenced investments underperform irrespective of their stated strategicobjectives,andthattheunderperformanceisparticularlyconcentratedathome.12

5.RobustnessChecks

AsthedescriptivestatisticsinTable1show,thesampleissomewhatskewedinthesizeofthedeals,andthetransactionsareclusteredtowardstheendoftheperiodunderstudy.Moreover,thereisheterogeneityindealsizeacrossdifferentfunds.Onecouldbeworriedthat our results might be driven either by some of the smaller deals or the valuationtrendsintheyearsimmediatelybeforethefinancialcrisis.Alternatively,onemightworrythat there is selection biases in the deals included in our sample,which is doubtless agreaterproblemamongthesmallertransactions.Toverifythatourresultsarerobusttotheseconcerns,weundertakeanumberofadditionalteststhatexaminedifferentsubsetsofthedata.

Werepeatall the regressionspresented in thispaperusing twosubsamples,onewhichincludesthelargest75%ofthedeals,andtheotherwiththelargest50%ofthedeals.Inthese

12 In equally weighted regressions, the direct effects of politicians and external managers on P/E levels and sector

performanceare similar to the resultspresentedhere,and thecoefficientsexhibiteven largermagnitudesandhigher

statisticalsignificance.However,theeffectofstrategicobjectivesonthesemeasuresbecomessmallerandinsignificant.

samples, the cutoff value is either $11 million or $68 million (in 2008 U.S. dollars),respectively. We re-run the regressions of the paper, both size-weighted and equal-weighted. Even after removing 50% of the transactions, the remaining transactions

maintain the same distribution across the groups. And in both subsamples, the resultsremainedsimilartotheonesreportedinthepaper.Wealsoruntheregressionswithoutwinsorizingthedata.Finally,werepeatouranalysisexcludingeitherthelasttwoyearsorthelastyearofthesampleandfindthattheresultsremainunchanged.

Finally,weconductsimpleweightedmeanteststoexploretherobustnessoftheresults.The results are provided in the appendix and exhibit similar patterns to the onesdescribedinthemultivariateanalysisexceptwhennotedabove.

6.Conclusions

Sovereignwealth fundsaredistinguishedby their complex setofobjectives.Takenasawhole,ourresultslendsupporttothehypothesisthatthemoreintenselythesefundsareexposed to political influences, the more they show major distortions from long-runreturnmaximization. In particular, SWFs with politician involvement aremore likely toinvest domestically,while those SWFswhere externalmanagers play an important rolearemorelikelytoinvestinternationally.PoliticallyinfluencedSWFsalsoconcentratetheirfundsinsectorsthatbothhavehighP/Elevelsandthenexperienceadropintheselevels,especially intheirdomestic investments,patternsthatdonothold in fundsthatrelyonexternal managers. Funds that have stated strategic goals are more likely to invest athome,butonlyifpoliticiansareinvolved.

While these results are only suggestive, given the preliminary nature of the data, theyraiseanumberofimportantquestionsabouttheinvestmentdecisionsandmanagementstructure of SWFs. A logical extension of this analysis would be to investigate thestrategiesofSWFsacrossawidersetofassetclasses.Suchananalysis,however,wouldbechallenginggiventheopacityofmanyofthesefunds.

More generally, we believe that much interesting work remains to be done inunderstanding the underlying objectives of SWFs, their investment strategies, andorganizational differences, as well as the constraints they face due to internal andexternalpressures.Forexample,manyreportssuggestthatSWFsareoftenemployedtofurther the geopolitical and strategic economic interests of their governments (WorldEconomic Forum, [2011]). A recent example is the emphasis of Singaporean SWFs oninvestingintoIndiaandChina,whichhasbeeninterpretedasbeingmotivatedbyadesiretoforgestrategictieswiththecity-state’s largerandmorepowerfulneighbors. Inothercases, political considerations have led to the abandonment of prescient investment

strategies,aswhentheNorway’sGovernmentPensionFundcausedanuproarin2006byshorting the shares of Icelandic banks. Thus, SWFs present an ideal setting inwhich tounderstandtheinteractionbetweenfinanceandpoliticaleconomy.

References

Atkinson,AnthonyB.,andJosephE.Stiglitz,1980,LecturesonPublicEconomics,London,McGrawHill.

Bekaert,Geert,CampbellR.Harvey,ChristianLundblad,andStephanSiegel,2007,Globalgrowthopportunitiesandmarketintegration,JournalofFinance62,1081-1137.

Banerjee, Abhijit V., 1997, A theory ofmisgovernance.Quarterly Journal of Economics,112,1289–1332.

Bortolotti, Bernardo, Veljko Fotak, William L. Megginson, and William Miracky, 2010,Sovereign wealth fund investment patterns and performance, Unpublished workingpaper,availableatSSRN:http://ssrn.com.ezp-prod1.hul.harvard.edu/abstract=1364926

Chhaochharia,Vidhi,andLucA.Laeven,2009,Sovereignwealthfunds:Their investmentstrategies and performance, Unpublished working paper, available at SSRN,http://ssrn.com/abstract=1262383.

Dewenter,KatherineL.,XiHan,andPaulH.Malatesta,2010,Firmvaluesandsovereignwealthfundinvestments,JournalofFinancialEconomics98,256-278.

Dyck, I. J. Alexander, andAdairMorse, 2011, Sovereignwealth fundportfolios, ChicagoBoothResearchPaperNo.11-15,availableatSSRN,http://ssrn.com/abstract=1792850

Fernandez,DavidG.andBernhardEschweiler,2008,SovereignWealthFunds:ABottom-UpPrimer,NewYork,J.P.MorganResearch.2008.

Fernandes,NunoG.,2009,Sovereignwealthfunds: Investmentchoicesand implicationsaround the world, Unpublished working paper, available at SSRN,http://ssrn.com/abstract=1341692.

Friedman, Tim, 2008, Preqin Sovereign Wealth Fund Review, London, Preqin Limited.Gompers,PaulA.,andAndrewMetrick,2001,Institutionalinvestorsandequityprices,

Quarterly Journal of Economics, 116, 229-259.Gordon, Myron J., 1959, Dividends,

earnings,andstockprices,ReviewofEconomicsand

Statistics41, 99-105.Hart,Oliver, Andrei Shleifer, and Robert Vishny, 1997, The properscopeofgovernment:

Theoryandanapplicationtoprisons,QuarterlyJournalofEconomics,112,1127–1162.

Knill, April M., Bong-Soo Lee, and Nathan Mauck, 2010, Is sovereign wealth fundinvestment destabilizing?, Unpublished working paper, available at SSRN,http://ssrn.com/abstract=1328045.

Kotter, Jason, andUgur Lel, 2008, Friends or foes? The stock price impact of sovereignwealth fund investments and the price of keeping secrets, International FinanceDiscussionPaperNo.940,FederalReserveBoard.

Lakonishok, Josef, Andrei Shleifer, and RobertW. Vishny, 1994, Contrarian investment,extrapolation,andrisk,JournalofFinance49,1541-1578.

Lerner, Josh,AntoinetteSchoar,andWanWongsunwai,2007,Smart institutions, foolishchoices?Thelimitedpartnerperformancepuzzle,JournalofFinance,62,731-764.

Shleifer, Andrei, and Robert Vishny, 1994, Politicians and firms, Quarterly Journal ofEconomics,109,995–1025.

Shleifer,Andrei, andRobertVishny,1997,A surveyof corporategovernance, JournalofFinance52,737-789.

Stiglitz, JosephE.,1993.Theroleof thestate in financialmarkets, inProceedingsof theWorld Bank Annual Conference on Economic Development, Washington, InternationalBankforReconstructionandDevelopment/WorldBank,pp.19–56.

Swensen,David,2009,PioneeringPortfolioManagement:AnUnconventionalApproachtoInstitutionalInvestment,2ndedition,NewYork,SimonandSchuster.

Tirole,Jean,1994,Theinternalorganizationofgovernments,OxfordEconomicPapers,46,1–29.

World Economic Forum, 2011, The Future of Long-Term Investing, New York, WorldEconomicForumUSA.

MEDICI FIRMA

GOVERNMENT

AFFAIRS

www.medicifirma.com

MEDICIFIRMASOVEREIGNFUNDSTRATEGIES