Measuring Financial Development in the Middle East and North Africa: A New Database Author(s): Susan Creane, Rishi Goyal, A. Mushfiq Mobarak and Randa Sab Source: IMF Staff Papers, Vol. 53, No. 3 (2006), pp. 479-511 Published by: Palgrave Macmillan Journals on behalf of the International Monetary Fund Stable URL: http://www.jstor.org/stable/30035923 . Accessed: 25/06/2014 10:18 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Palgrave Macmillan Journals and International Monetary Fund are collaborating with JSTOR to digitize, preserve and extend access to IMF Staff Papers.

http://www.jstor.org

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

IMFStaffPapers

Vol.53, No.3 c 2007 International Fund Monetary

MeasuringFinancialDevelopment inthe MiddleEastand NorthAfrica: A NewDatabase SUSAN CREANE, RISHIGOYAL, A. MUSHFIQ MOBARAK, and RANDA SAB* Thispaper develops a methodologyto constructdetailed indices offinancialsector developmentacross countries and uses it to create a new panel database offinancial sector development in Middle East and North Africa (MENA) countries. It combines existingquantitativedata withinformation fromcomprehensivesurveys undertakenin 2000-01 and 2002-03. The data show that some MENA countries have relativelywell-developed banking sectors and regulatoryand supervisory regimes. However, across the region, the nonbank financial sectors and supporting institutionsare in need of reform.The MENA region ranksfar behind industrializedcountries and East Asia infinancial sector development.[JEL E44, E50, 016, 053]

*SusanCreaneis a DeputyDivisionChiefin Division4 of theIMF's Asia andPacificDepartment. RishiGoyalis an Economistin theCaribbeanI DivisionoftheIMF's Western Hemisphere Department. A. MushfiqMobarakis AssistantProfessorof Economicsat the University of Colorado,Boulder. RandaSab is an Economistin DivisionD oftheIMF's MiddleEast and CentralAsia Department. We to MohsinS. Khan,David Burton, aregrateful andPierreDhonte,andtheMENA country economists' desksat theIMF fortheirgenerouscooperation. We thankThamarKechichianand Saeed Mahyoub fortheirable researchassistance;participants in the24thMiddle East EconomicAssociationannual a departmental seminarat theMiddleEast andCentralAsia Department of theIMF,and the meetings, and Dr. RobertFlood and two anonymous refereesfor April2004 IMF MENA regionalconference; We wouldalso like to thankJudith theirhelpfulsuggestions. Rey and ImelYu fortheiroutstanding secretarial support.

479

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal,A. MushfiqMobarak,and Randa Sab

As

countriesin theMiddleEast and NorthAfrica(MENA) considerwaysto further financialsectorreform promoterapidand lastingeconomicgrowth, on The is that be their should high policies aimed at enhancing agenda.' theory andmonwillresultinlowerinformation, financialsectorperformance transaction, allocativeefficiency and raisingoutput(see Levine, itoringcosts,thusimproving evidenceis typicallybased on 1997; and Khan and Senhadji,2000). Supporting is measuredby wherefinancialdevelopment a broadcross sectionof countries, littleworkhas been a smallset of statisticalindicators.2 However,comparatively done on (1) how to measurethespecificsof financialsectordevelopment, taking andinstitutions thatthefinancialsectoris comintoaccountthevarietyofmarkets in theMENA region posedof; and (2) creatingmeasuresoffinancialdevelopment thatgo beyondsimpleaggregate indicators. indicators,such as theratioof Going beyondsimple"standard"quantitative and prioritizeamongdifferent broadmoney(M2) to GDP, is necessarytoidentify The simpleindicators,thougheasily available areas of financialsectorreform. and amenable to cross-regionaland intertemporal comparisons,do not neceswhat is meant financial sector Financial sarilycapture broadly by development. is a multifaceted not development concept,encompassing onlymonetary aggregatesand interestrates,butalso regulationand supervision,degreeof competiof property tion,financialopenness,institutional capacitysuch as the strength a nation's andfinancialproductsthatconstitute rights,and thevarietyof markets financialstructure. The value added of ourpaperis in developingproceduresto createindicesto capturethe developmentof some individualcomponentsof the financialsector, andin thenewdatasetforMENA thatwe havecompiled.Hence,we willfocuson twosurveyswe howthedataweregathered. Thisincludes,importantly, presenting in of IMF economists' desks forMENA 2000-01 and 2002-03 conducted country countries.The data and thesurveyare organizedin six themes,each of whichis meantto capturea distinctcomponentof financialdevelopment:the developmentof themonetarysectorand monetary policy,bankingsectordevelopment, nonbankfinancial sector development,regulationand supervision,financial environment. Based on the data set compiled,we openness,and institutional developedindicesof financialsectordevelopmentfortheregionand an alternativeindexwitha smallersetof variablesfortheworld,to allow forcomparabilityacross regionsand overtime. We findthatwithinthe MENA regionthereis substantialvariationin the Some countries,such as Bahrainand Lebanon, degreeof financialdevelopment. are fairlywell advanced,whereasa fewothers(forexample,Libya,the Syrian Arab Republic, and the Islamic Republic of Iran) have significant room for 'The MENA regioncoverstheIslamicStateofAfghanistan, Algeria,Bahrain,Djibouti,Egypt,the IslamicRepublicof Iran,Iraq,Jordan, Kuwait,Lebanon,Libya,Mauritania, Morocco,Oman,Pakistan, West Qatar,SaudiArabia,Somalia,Sudan,theSyrianArabRepublic,Tunisia,theUnitedArabEmirates, BankandGaza, andtheRepublicofYemen. 2Theseindicators sector usuallyincludetheratiosofbroadmoneytoGDP andofcredittotheprivate toGDP.

480

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLE EASTAND NORTHAFRICA

As a group,MENA countriesappearto perform relativelywell on improvement. as wellas on financialopenness.However,theyneedto and supervision regulation do significantly moreto reinforce theinstitutional environment and promotenonbankfinancialsectordevelopment. Comparedwithmostotherdevelopingcountry well based on thealternative financialdevelregions,theMENA regionperforms countriesand East Asia. opmentindex,butit ranksfarbehindindustrialized

I. Data CollectionProcess We createda 48-questionsurveyaimedat establishing measuresforfinancialsectordevelopment. This questionnaire was partlybased on surveysand tablesused byGelbardand Leite (1999) andMehranandothers(1998). We organizedthedata a different facetoffinancialdevelaccordingto six themes,each ofwhichreflects ofnonbank (2) development opment:(1) bankingsectorsize,structure, efficiency; financialsector;(3) qualityof bankingregulationand supervision;(4) developmentof the monetarysectorand monetary policy;(5) financialsectoropenness; and (6) institutional environment. We describetherationalebehindorganizingthe datain thesesix themesin SectionII. We answeredeach of the48 questionson thesurveyforeach of 20 countries in theMENA regionby drawinginformation froma numberof IMF staffreports, RecentEconomicDevelopments, andEconomistIntelligence We augUnitreports. mentedthisinformation withmacroeconomic, andinstitutional financial, paneldata fromtheIMF's International FinancialStatistics, IMF's WorldEconomicOutlook, World Bank's WorldDevelopmentIndicators,the PRS Group's International RiskGuide (ICRG), theHeritageFoundation(HF), and theIMF's Annual Country on We thencirculated and ExchangeRestrictions.3 Report ExchangeArrangements thecompletedsurveyto IMF country economists'desksin chargeof each of the MENA countriesin our sampleto have themverifythe information we culled fromthesesources.We thenincorporated anycorrections theymade. This entire processwas completedtwice,once in 2000-01 and thenagainin 2002-03, to create two sets of financialdevelopment indicesforeach countryfortwo pointsin time.Table 1 describesthe surveyquestionsthatwereused to ultimatelycreate scoresand rankingsforeach country, and liststheprimarysourcesused foreach response. We consolidatedthe surveyresponsesforeach countryin one table to performsome qualitativeanalyses:assess commontrends,strengths, and weaknesses areas wherereformsare mostneeded.We amongMENA countries,and identify thenreorganizedthe surveyresponses-in some cases combininginformation gatheredfromtheanswersto different questionsand fromadditionalquantitative data collected-in a way thatultimately allowedus to scoreeach countryalong variousdimensionsoffinancialdevelopment. (The scoringprocessoftenforcedus to relyon qualitativejudgmentsofthesurveyresponses.)These scoreswereused 3Seealso Reinhart andRogoff's(2002)dataonexchange rateclassifications availablevia theInternet: http://www.wam.umd.edu/-creinhar/Links.html.

481

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 1. Middle East and NorthAfrica(MENA) Countries:Questions Included in the SurveyforFinancial Liberalizationin MENACountries ThemesandComponents

Explanation

1. BankingSectorSize,Structure, Efficiency oflicensed Totalnumber ofbanks, Number/type banks including public,foreign, offshore, development, Islamic,andspecialized banks Totalassetsofbanking Assetsizeofall banksas sector(percent ofGDP) ofGDP percent Publicsectorbanks

Number ofpublicbanks

Islamicbanks

Number ofIslamicbanks

Offshore banks

Number ofoffshore banks

Concentration ofbanking sector

Assetsizeofthelargest three banksas a shareoftotal assetsinthebanking sector ofnewbanks Do banking Entry regulations allow foreasyentry ofnewbanks, orhavetherebeennew banksoverthepastthreeto fiveyearsj Direction ofcredit To whichsectorsdoesmost commercial bankcredit go oftotal)j (as percent Is it Depositinsurance Is theredepositinsurancej orexplicitj implicit Noncashtransaction cards,checks, activity Arecredit/debit automated tellermachines widelyusedj

Sources Staff RecentEconomic report, (RED), Developments EconomicIntelligence Unit IMF country (EIU) reports, desks economists' Staffreport, RED, EIU reports, IMF country economists' desks Staff RED, EIU reports, report, IMF country economists' desks Staffreport, RED, EIU reports, IMF country economists' desks Staff RED, EIU reports, report, IMF country economists' desks Staffreport, RED, EIU reports, IMF country economists' desks Staff RED, IMF country report, economists' desks

Staffreport, RED, EIU reports, IMF country economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks

2. Development ofNonbankFinancialSector Nonbank financial sector

Stockmarket

Arethere Staffreport, markets, RED, EIU reports, mortgage stockmarkets, IMF country economists' funds, pension mutualfunds, insurance desks companies, leasing socialsecurity companies, agencies, money changersj ofstockmarket Staffreport, RED, EIU reports, Gaugeextent IMF country economists' activity usingmeasuresof themarket desks size,including market capitalization, number offirms, turnover, andtrading

482

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 1 (continued) ThemesandComponents

Explanation

financed Housingfinance/mortgage Is housing through Ifnot, market a mortgage marketj howis housingfinancedj Is theinterbank market activej Interbank market activity

Sources Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks

3. QualityofBankingRegulationand Supervision and Bankingregulation supervision

Is banking and regulation supervision adequate;thatis, do regulations complywith BaselCorePrinciplesj What aretheweaknessesj Whatis thecapitaladequacy Bankingsystem capital assetrequirement ratioj Shareofnonperforming Whatis theshareofnonloans(defined loansintotalloans performing as 90 daysinarrears, where available)j of Limitsonexposure tosingle Whatis theconcentration or borrowers orrelated loansjIs thereconnected borrowers family lendingjWhatare thelimitsonexposurej andauditing Describeandassesscurrent Inspection practices Is thepayment system Payments system orby processed manually Is thecentral computerj banktheclearinghousej ofcentral bank Does thecentral bankappear Independence fromother tobe independent branches ofgovernmentj bankborrower Do commercial bankshave Central database accesstoa borrower databasej Central bankwebsite/ Aremonetary dataeasily information dissemination availabletothegeneral publicjIs therea website withcurrent dataj HaveIMF documents been IMF Boarddocuments postedon IMF website postedon theIMF websitej If yes,whichonesj

Staff RED, IMF country report, desks economists'

Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks central bank's Country's IMF country website, economists' desks IMF website

4. Development oftheMonetarySectorand MonetaryPolicy DirectMonetary PolicyInstruments rateliberalization Areinterest ratesfullyor Interest Whenj liberalizedj partially Whatremains underofficial controlj

Staff RED, IMF country report, desks economists'

(continued)

483

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 1 (continued) ThemesandComponents Creditcontrols

Explanation

Haveall creditcontrols been removed (e.g.,ceilings, directed credits forcertain sectors)j Indirect Monetary PolicyInstruments Use ofreserve requirements Arechangesinreserve requirements frequently usedbythemonetary authorityj reserve ratio Whatis therequired reserve Required ratiojWhatis theforeign reserve currency required ratio(ifdifferent from a domestic currency required reserve ratio)j Rediscount window Is therediscount window usedj facility actively Areopenmarket Openmarket operations operations usedj actively Government securities market Howaregovernment securities sold(e.g.,typeofauction)j Whatis therangeof maturitiesj market for Is thegovernment securities Secondary securities market activej government secondary Government securities held Whatshareis heldbynonsector financial sectorj bynonfinancial private Financialinstitutions (e.g.,pensionfunds)j

Sources Staff RED, IMF country report, economists' desks

Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks

Staff RED, IMF country report, economists' desks Staff RED, IMF country report, desks economists' Staff RED, IMF country report, desks economists' Staff RED, IMF country report, economists' desks Staff RED, IMF country report, economists' desks

5. FinancialSectorOpenness Exchangerateregime Article VIII/XIV

Whatis theexchange rate AnnualReportonExchange andin andExchange regime, officially Arrangements Restrictions practicej (AREAER) to AREAER Classify country according Article VIII orXIV.IMF members the accepting refrain from obligations restrictions on imposing themaking ofpayments andtransfers forcurrent international transactions andfrom in engaging discriminatory currency ormultiple arrangements without currency practices UnderArticle IMF approval. continues XIV,thecountry toavailitselfofthe

484

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 1 (concluded) ThemesandComponents

Sources

Explanation

transitional arrangements ofArticle XIV.These allowan IMF arrangements member tomaintain oradapt thoserestrictions tocurrent transactions thatwerein effect atthetimeitbecame a member. Thecountry is expected toeliminate thoserestrictions as soonas andtheintroduction possible, is not ofnewrestrictions allowed. rates Is country AREAER Multiple exchange fleefrommultiple ratesj exchange Parallelexchange market Is country freefroma parallel AREAER marketj exchange Forward market Is therea forward AREAER exchange exchange marketj Restrictions onpurchase/ Areforeigners freeto AREAER sale offinancial assets financial purchase/sell assetsj byforeigners Restrictions on purchase of Areresidents freefrom AREAER of restrictions onpurchase foreign currency by residents foreign currencyj freefromthe AREAER Repatriation requirements Areexporters torepatriate obligation export proceedsj 6. Institutional Environment Legaltradition

Loanrecovery

Law and order

Whatis themainlegal tradition, (e.g.,dominated by orIslamic French, British, law)j Can loansbe recovered thejudicialsystem through easilyandreasonably quicklyj

La Portaandothers, 1998

Indexrating

IndexofEconomicFreedom, Foundation Heritage IndexofEconomicFreedom, Foundation Heritage ICGR ICGR

Index rating

Property rights

Government involvement Indexrating inbanking/finance Indexrating Bureaucracy Democratic accountability Indexrating

Staffreport, RED, IMF country economists' desks InternationalCountryRisk Guide (ICGR)

485

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, Rishi Goyal,A. MushfiqMobarak,and Randa Sab to construct newindicesoffinancialdevelopment, whichare describedin detailin 2 SectionIII. Table describeshow thedata werereorganized, thevariables,their the definitions, weights(describedin detailin SectionIII), score methodology, rationaleforincludingthemin a financialdevelopment index,and datasources. II. Rationale Behind the Organization of the Data Gelbardand Leite (1999) used measuresof marketstructure, financialproducts, financialliberalization,institutional financial environment, openness,and moninstruments to construct a index for38 sub-Saharan etarypolicy comprehensive Africancountriesfor 1987 and 1997. Followinga similarapproach,we created a comprehensive indexfor20 MENA countriesfor2000-01 and 2002-03 using six themes:monetarysectordevelopment and monetarypolicy,bankingsector nonbankfinancialsectordevelopment, development, regulationand supervision, financialopenness,and institutional environment. The monetarysectordevelopment and monetary policy themeexaminesthe extentto whichthe government uses indirectmonetarypolicy instruments, as to direct on of and credit interest rates rates allocacontrols, (or return) opposed tion.It also considerstheefficiency of marketsforgovernment securitiesand the provisionof liquidityservicesby thefinancialsystem. The bankingsectordevelopment themeexaminesthesize, structure, and effitheprofitability ciencyof thebankingsector.Amongotherthings,it investigates of banks,bankcompetition and concentration, paymentssystems,ease of private sectoraccess to bankcredit,and frequency of noncashtransactions. Drawingon recentempiricalresearch,thepresumption is thatbanksoperatingin competitive environments-with less government low marketconcentration, and intervention, and conduciveto growth. The foreignbankentry-arelikelyto be moreefficient financialrepression literature has convincingly shownthatgovernment restrictions on thebankingsystem-suchas highreserverequirements, interest rateceilings, anddirectedcredit-repressdevelopment. In addition,recentworkhas shownthat concentrated bankingsystemsand largergovernment ownershipof bankshave a andrestrictions on foreignbankentry hinder depressing impacton overallgrowth, allocativeefficiency.4 The nonbankfinancialsectordevelopment themeexploresthedevelopment of alternative sources of capitalas well as marketsfor financialproductsand services.These includestockmarkets, mortgageor housingfinanceinstitutions, corporatebond markets,insurancecompanies,mutualfunds,and pensionfunds. Theyreflectthevarietyof productsand marketsthatallow a financialsystemto fulfillits functions:namely,enablingfirmsand householdsto raise financein 4Cetorelli andGambera(2001) findthathighbanking concentration can facilitate ofindustrial growth sectorsthataremorein needofexternal butfinda generalnegativeassociationofconcentration finance, ongrowth acrossall sectorsandfirms. La Porta,Lopez-de-Silanes, andShleifer (2002) showthatcountries withhigher ofbanksareassociatedwithlowersubsequent Levine(2003) government ownership growth. findsthat,controlling forotherfactors, on foreign restrictions bankentryresultin higherbankinterest margins.

486

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 2. Data Definitions, Weights,Scores, and Sources forComputing Financial Development Indices ThemesandComponents

Methodology Definition/Score

Sources

1. BankingSectorSize,Structure, (Weight:25 percent) Efficiency RecentEconomic and We includedbanksin our Staffreport, Development ofthe measurebecausetheyare (RED), Developments profitability EconomicIntelligence Unit sector(5 percent) centralto thefinancial banking andpayments of (EIU) reports, system IMF country desks mosteconomies,often economists' playinga criticalrolein theprocessofmobilizing investment savings,funding monitoring opportunities, anddiversifying managers, risk.Thismeasure examineswhether there is largepublicownership, government financing need,or weaksupervision; whether therewerebanking crisesin thepast15 years; whether bankmanagement capacityis adequate; whether banksaresolvent; whether bankshavebeen capitalized. Thescoreis 0 ifbanking sector as a wholeis inefficient; 1 if somebanksareprofitable, butsignificant portion ofbanking sectoris still inefficient orsuffers losses; 2 ifvastmajority ofbanks areprofitable/efficient. Staffreport, Privatization ofbanking Private banksareassociated RED, EIU reports, IMF country sector(3 percent) withhigher financial economists' desks development, stronger andless supervision, intervention. government The scoreis 0 ifthereis substantial presenceof inthe publicinstitutions bankingsectorwithno efforts at privatization; I ifthereis substantial ofpublicinstitutions presence in bankingsector,butsome has occurred; privatization 2 ifbanksarelargely private.

487

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 2 (continued) ThemesandComponents

Definition/Score Methodology

Ratioofcredittoprivate A proxyfortheextent sectorbydepositmoney ofactivity offinancial bankstoGDP (3 percent) intermediaries. Private creditcaptures thefinancial intermediation withthe nonfinancial sector. private sizeindicator Depositmoneybankassets/ Thisis a relative sectorassets thatmeasures theimportance banking ofdepositmoney banks (3 percent) relative tothebanking sector. Reserveratio(3 percent) Bankreserves overmoney andquasi-money (M2),less heldoutsideof currency banks.Thehighrequired ratioofreserves andthelow interest ratesbanksearn onthosereserves reflect desire governments' to maintain a taxdevicecapable ofgenerating substantial revenue. Whenthis implicit taxbecomeslarge,ithasa seriousnegative effect onthe financial system. Interest ratespreads Thedifference between loan anddepositrates.Usedas an (3 percent) indicator ofcompetition in thebanking sector. The scoreis 0 ifthereare highspreads(above6 or interest rates percent) aresetadministratively or 1 ifthereare collusively; moderate spreads(between 2 ifthere 4 and6 percent); arelowspreads(less than 4 percent). inthe A highly Concentration concentrated sector(3 percent) commercial sector banking banking resultinlackof might to competitive pressure attract savingsandchannel themefficiently toinvestors. Thescoreis 0 ifthebanking sectoris highly concentrated (threebanksaccountfor 70 percent ofassets,loans, ordeposits;ortwobanks accountfor60 percent;

Sources International Financial Statistics (IFS), World EconomicOutlook(WEO)

IFS

IFS

Staffreport, RED, EIU reports, IMF country economists' desks

Staff RED, EIU reports, report, IMF country economists' desks

488

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 2 (continued) ThemesandComponents

Presence offoreign banks (2 percent)

Definition/Score Methodology

Sources

orone bankaccountsfor 40 percent);1 ifthereis in moderate concentration thebankingsector(five banksaccountfor70 percent ofassets,loans,ordeposits; orfourbanksfor60 percent; orthree banksfor50 percent; or twobanksfor40 percent; oronebankfor25 percent); 2 ifbankshavelowindustry concentration (theconditions abovedo nothold). A proxyforcompetition and Staff RED, EIU reports, report, inthebanking IMF country economists' efficiency that desks Countries system. their domestic repress also system banking restrict accessto typically thefinancial system. Thescoreis 0 ifthereareno banks;2 ifthereare. foreign

2. Development ofNonbankFinancialSector(Weight:15 percent) Stockmarket Weusethestockmarket Staffreport, RED, EIU reports, (4 percent) turnover ratioas anefficiency IMF country economists' indicator ofstockmarkets. It desks is defined as theratioofthe valueoftotalsharestraded andmarket capitalization. Thescoreis 0 ifthereis no stockmarket, ortrading is verylimited (e.g.,turnover ratio< 20 percent); I ifa stockmarket exists,but is somewhat limited trading ratiobetween 20 (turnover and40 percent); 2 ifthe stockmarket is activewith substantial (turnover trading ratio> 40 percent). to Staff RED, IMF country (2 percent) Examinestheextent Housingfinance report, whichhousingis financed economists' desks markets. through mortgage Thepresence ofhousing financial institutions is inthe also an ingredient ofthenonbank development financial sector.

489

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 2 (continued) ThemesandComponents

Definition/Score Methodology

Sources

Thescoreis 0 ifitis difficult to obtainhousing finance; 1 ifit is possibletoobtainhousing loans(somespecialized finance institutions housing arelargeand exist);2 ifthere activemortgage markets (size > 30 percent ofGDP) and itis easytoobtainhousing finance. Othernonbank financial Examineswhether there Staffreport, RED, IMF country markets andinstruments is substantial in economists' desks activity mutual (4 percent) funds, pensionfunds, bonds,insurance corporate Measuresthe companies. size andactivity ofnonbank financial intermediaries. The scoreis 0 ifat mostone ofthenonbank financial institutions exists,butis notwelldevelopedand is limited;1 if activity at mostthreenonbank institutions exist,but is limited;2 if activity nonbank institutions exist andarewelldevelopedwith substantial activity. Interbank transactions Examinesthedegreeof Staff RED, IMF country report, ininterbank economists' desks (5 percent) trading activity transactions. Thescoreis 0 ifinterbank markets exist,butare inactive;1 ifinterbank markets exist,butneed further and/or development havelimited trading activity; 2 ifinterbank markets exist withsubstantial trading activity. 3. QualityofBankingRegulationand Supervision (Weight:15 percent) Baselcapitaladequacyratio (3 percent) requirements

Measurestheextent to which bankscomplywithBasel capitaladequacyratio (CAR) requirements. in Financialdevelopment, tendstobe higher general, wherebankscomplywith Basel CAR requirements.

Staff RED, IMF country report, desks economists'

490

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 2 (continued) ThemesandComponents

Prudential of monitoring banks(3 percent)

loans Nonperforming (2 percent)

Definition/Score Methodology

Sources

Thescoreis 0 ifmorethanhalf thebanksdo notmeetBasel CAR; I ifmanybanksmeet CAR (between50 and 75 percent), buta significant do not;2 ifthe proportion sectoras a whole banking is largely orfullycompliant of (morethan75 percent banks). Considers thelevelofprudential Staff RED,IMF country report, in banks, economists' desks monitoring auditand including adequate ofdatacollection availability formonitoring. Countries withmoredeveloped financial markets tendto followstricter prudential ofbanks. monitoring Thescoreis 0 ifprudential of banksis monitoring weakandneedssignificant (thatis, strengthening information is not prudential collected andbanks regularly arenotadequately monitored/ audited);I ifprudential ofbanksis monitoring moderate butstillneeds 2 ifprudential strengthening; is adequate. monitoring Ratioofnonperforming Staff RED, IMF country report, loans(NPLs) tototal economists' desks loans.Countries withmore markets developedfinancial tendtohavelowerNPLs. Thescoreis 0 ifNPLs arelarge relative tothesizeofbanks' loanportfolio than (greater 15percent whendefined as 90 daysinarrears); I if NPLsarenotyetlow,but areeither (b) (a) declining, or adequately provisioned, (c) highonlyforsome banks butnotothers;2 if NPLs are small relativeto thesize of banks' loan portfolio(less than6 percentwhendefined as 90 days in arrears).

491

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 2 (continued) ThemesandComponents

Definition/Score Methodology

Sources

ofthecentral Independence bank(3 percent)

Examineswhether a central Staff RED, IMF country report, bankneedstoextendcredit economists' desks tothegovernment and/or is required toconsult withothergovernment offices. central Independent bankstendtohavea more market developedfinancial becausetheycanconduct theirmonetary policy without interference. The scoreis 0 ifthecentral bankis notindependent; I ifthecentralbankis somewhat (but independent is required toconsultwith othergovernment offices); 2 ifthecentral bankis largely independent. andavailability Examineswhether financial Website forthecountry's Transparency offinancial andmonetary andmonetary dataaremade central bank,IMF country data(4 percent) availabletothepublicand economists' desks whether in theyareprovided a timely A country manner. withmoredeveloped financial markets tendsto also makepublictheirdata. Thescoreis 0 iffinancial andmonetary dataarenot availabletothepublic,or limited dataareavailable withlonglags(fourmonths ormore);1 ifbasicfinancial andmonetary dataare availabletothepublicina 2 ifa range manner; timely ofdetailedfinancial and data,including monetary lawsandprocedures, are easilyavailabletothepublic ina timely manner. 4. Development oftheMonetarySectorand MonetaryPolicy(Weight:20 percent) RatioofM2 toGDP (5 percent)

Thisisa commonly available IFS,WEO indicator offinancial M2 isa typical intermediation. measure offinancial "depth" andthusoftheoverall sizeof thefinancial sector.

492

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 2 (continued) ThemesandComponents

Definition/Score Methodology

Sources

of Indirect instruments monetary policy (4 percent)

Examinesthedegreetowhich Staff RED, IMF country report, countries usechanges economists' desks inreserve requirements, rediscount and window, openmarket operations Theuseofindirect actively. instruments ofmonetary associated policyis typically withhigher financial andless development financial repression. Thescoreis 0 ifmostly direct monetary policyinstruments areused;I ifsomeindirect areused, policyinstruments butarenotregularly and used;2 ifa range flexibly ofindirect monetary policy instruments areactively used(e.g., andflexibly through regular openmarket operations). Creditcontrols anddirected Considersthedegreeto Staff RED, IMF country report, credit(3 percent) whichallocationofcredit economists' desks is closelycontrolled and directed ormoralsuasion is heavilyreliedupon. and Highercreditcontrols directed creditcharacterize a financially repressed economy. Thescoreis 0 ifcredit is closely allocation anddirected, or controlled moralsuasionis heavily reliedupon;I ifcredit allocation is notmandated butceilings byauthorities tocertain sectors exist,or moralsuasioninallocating creditis used;2 ifthereis nogovernment involvement increditallocation. Interest rateliberalization Market-determined interest Staff RED, IMF country report, ratesareassociated witha economists' desks (5 percent) moredevelopedfinancial system. The scoreis 0 ifinterest rates aresetbytheauthorities;

493

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 2 (continued) ThemesandComponents

Government securities (3 percent)

Definition/Score Methodology

Sources

I ifinterest ratesare liberalized partially (e.g., setminimum authorities or range);2 or maximum ifinterest ratesarefully liberalized. ofsecurities Staff The availability RED, IMF country report, ofa economists' desks is also an indication moredevelopedfinancial system. Thescoreis0 ifgovernment securities do notexistorare notauctioned ordistributed viamarket mechanisms; securities 1 ifgovernment or existandareauctioned distributed usingmarket butthere is mechanisms, noactivesecondary market; 2 ifgovernment securities or existandareauctioned distributed some through market andthere mechanisms, markets. areactivesecondary

5. FinancialSectorOpenness(Weight:10 percent) AnnualReportonExchange forcesallowed marketAremarket Appropriate rate todetermine theexchange andExchange determined Arrangements exchange Restrictions ratejHighintervention in (AREAER) (2 percent) theforeign market exchange rate tomaintain an exchange ata certain "desired" level couldcreateimbalances and, difficulties inthe eventually, financial system. Thescoreis 0 ifnotappropriate; 1 ifsomewhat appropriate; 2 ifappropriate. AREAER rates Thepresence ofmultiple Multiple exchange orparallelmarkets ratesorparallel exchange markets couldsignal (1 percent) intheforeign imbalances market, hindering exchange investment andcausing arbitrage. speculative has The scoreis 0 ifcountry multiple exchangerates orparallelmarkets; 2 ifit doesnot.

494

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 2 (continued) ThemesandComponents

Definition/Score Methodology

Restrictions on foreign currency purchases by residents (2 percent)

A measure ofcapitaltransaction AREAER controls. Thescoreis 0 ifthereare restrictions on foreign currency purchases by 2 ifnot. residents; A measure ofcapitaltransaction AREAER controls. Thescoreis 0 ifthereare restrictions on thefinancial activities ofnonresidents; 2 ifnot. AREAER Thepresenceofa forward exchangemarket signalsa developedforeign exchange market. Thescoreis 0 ifthereis no forward market; exchange 2 ifthereis. AREAER Repatriation requirements coulddiscourage exports andinvestment. Thescoreis 0 ifthereare repatriation requirements; 2 iftherearenot. AREAER Has thecountry acceptedthe ofArticle obligations VIIIj IMF members accepting theobligations refrain from on restrictions imposing themakingofpayments andtransfers forcurrent international transactions or fromengagingin discriminatory currency ormultiple arrangements currency practices without IMF approval. Ifcountries havenot under acceptedobligations Article VIII,theymaintain on current restrictions transactions. Thescoreis 0 ifa country has notacceptedArticle VIII 2 ifithas. obligations;

Restrictions on thefinancial activities ofnonresidents (2 percent)

Forward market exchange (1 percent)

Repatriation requirements (1 percent)

Article VIII status (1 percent)

Sources

495

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab

Table 2 (concluded) ThemesandComponents

Definition/Score Methodology

Sources

6. Institutional Environment (Weight:15 percent) Ease ofrecovering loans thejudicial through (4 percent) system

Countries withstronger institutions tendtohavea moredeveloped andefficient financial system. Thescoreis 0 ifitis difficult torecover loansthrough thejudicialsystem; I ifit 2 if is moderately difficult; thejudicialsystem helpsthe processofloanrecovery. Law andordertradition The"law"subcomponent ofthe is an assessment (1 percent) andimpartiality strength ofthelegalsystem; the "order"subcomponent is an assessment ofpopular ofthelaw. observance index Theextent towhichthe Property rights (4 percent) government protects private property byenforcing thelawsandhowsafe is from private property expropriation. Bureaucratic quality Highpointsaregivento countries wherethe (2 percent) hasthestrength bureaucracy andexpertise togovern without drastic changesin in policyorinterruptions services. government Government involvement Thedegreeofgovernment inbanking/finance involvement inbanking/ finance. The morebanks (2 percent) arecontrolled bythe theless government, freetheyaretoengage infinancial servicesthat facilitate economicgrowth. ofhow Democratic accountability Thisis a measure thegovernment is (2 percent) responsive toitspeople.

Staff RED, IMF country report, economists' desks

Staff RED, IMF country report, economists' desks

IndexofEconomicFreedom, Foundation Heritage

Risk Country International Guide(ICRG)

IndexofEconomicFreedom, Foundation Heritage

ICRG

496

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLE EASTAND NORTHAFRICA

cost-effective managers,and diversifying ways, mobilizingfinance,monitoring risk.Researchon stockmarketshas shownthathighlyliquid stockmarketsare in promoting an important complementto bankingsectordevelopment growth.5 as opposed to the size of stock markets, Liquidityor the ease of transacting, and assets,thus is important because it facilitatesthe exchangeof information and As Levine resource allocation notes, (2002) "Simplylistgrowth. improving allocaon does not foster resource the national stock exchange necessarily ing tion."Therefore, in additionto theexistenceof nonbankfinancialintermediaries and markets, we pay particular attention to liquidity. The regulationand supervisionthemeassesses banks' performancewith Amongotheritems,it respectto minimum(Basel) capitaladequacyrequirements. of banksand thetransparency and openness evaluatestheprudentialmonitoring of theregulatory environment. of the exchange The financialopennessthemeassesses the appropriateness on and examines whether there are restrictions the tradingof significant regime and residents. Restrictions on current financialassets or currencyby foreigners hindertradeingoodsandservices.Similarly, couldsubstantially accounttransactions multipleexchangepracticesand misalignedexchangeratescould hindertradeand on capitalaccounttransactions, resourceallocation.Restrictions however, mightbe neededunless appropriate institutional includingprudential regulaarrangements, tionsand supervision, are in place.As is beingdebatedin thecontextof currency an opencapitalaccount andfinancialcrisesand theoptimalorderof liberalization, disclosure could increasetheriskof and information without oversight appropriate fromthe With an financial institutions,openeconomybenefits collapse. appropriate domesticinvestment worldwide projectsandfrom pool offundstofinance promising theallocationof local savingstopromising investment alternatives globally. environment themetriestojudge thequalityof instiFinally,theinstitutional bureaucratic suchas law and order,property tutions, quality,accountability rights, and the ease of loan recoverythroughthejudicial system of the government, of thefinancialsystem.Some componentsof this thatinfluencetheperformance are admittedly less directly related theme(forexample,democratic accountability) thanothers(forexample,judicial loan recovery).This is to financialdevelopment accountedforin therelativeweightsplaced on thevariouscomponents. III. Index Creation To capturea country's"development" along each of the six themesmentioned subindices,whichwe above,we developedand scoredcountrieson six different a comprehensive index(thatis, a compositemeasure thencombinedto construct Each of thesesix subindiceswas of "financialdevelopment")foreach country. thatallowedus to measurethe indicators basedon betweenfourandeightdifferent and Levine(2001), Levine(2002), and Beck and 5See Levineand Zervos(1998), Demirgig-Kunt andeconomicgrowth. Levine(2004). Notethatthisresearch development mainlylooksat stockmarket has notbeendoneon the ofcross-country data,research Owingto thelimitedpresenceand availability on growth ofotherfinancial effect markets andinstruments suchas bondsandcommercial paper.

497

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab various subfacetsof each area. The comprehensiveindex,therefore, was a combinationof 35 different indicatorsdrawnfromthe48 surveyquestions,as described in Tables 1 and 2. For instance,underthe bankingsectordevelopmenttheme,the relevantsubindex attemptedto create indicatorsrelatingto banking sector size, and efficiency.Box 1 shows the specificaspects of bankingsectorsthat structure, we triedto measure in buildingthe subindexforbankingdevelopment. Much of the data appears in a mix of quantitativeand qualitativeforms.For example, in coding concentrationin thebankingsector,we were able to determine thatin Bahrain,two of 21 commercialbanks accountfor57 percentof bank assets, whereas in Jordan,the threelargestbanks hold 90 percentof bank assets, with theArab Bank alone accountingfor60 percentof assets. Because the two sets of numbersare notdirectlycomparable,we had to make some qualitativejudgments in ratingthe bankingsectorconcentrationon a 0-1-2 scale. For each countryand indicator,we coded thescore 0-2 based on theresponsesin thesurvey,withhigher values indicating a higher degree of development.We firstdeveloped "rules" for the coding to limiterrorsintroducedby qualitativejudgments (Table 2). For instance, for the indicatoron interestrate spread (differencebetween loan and deposit interestrates),we used the followingrule forscoring: High spreads (above 6 percent)or interestrates set administratively or collusively Moderate spreads (between 4 and 6 percent) Low spreads (less than4 percent)

0 1 2

For thesequalitativeassessments,the0-1-2 scoringscale was chosen insteadof a binaryscale, because thismore-detailedscale allows us to differentiate between countriesin termsof various dimensionsof financialdevelopmentmore appropriately.For example, withrespectto the stock market(underthe nonbankfinancial sectorindex),a binaryquestionallows us to testonlyforstockmarketpresence;the differentiation of countrieswhere stock 0-1-2 scale, however,allows fora further marketsare merelypresentfromcountrieswheretheyare active and characterized is not by substantialtradingactivity.In some cases, where such differentiation

Box 1. SubindexforBankingSector Development:SpecificIndicators * Development andprofitability ofthebankingsector * Privatization of commercial banking * Creditto privatesectorbydepositmoneybanksas a shareof GDP * Deposit moneybank assetsas a shareof depositmoneybankplus centralbank assets * Reserveratio * Bankingsectorcompetition ratespreads,ease of newentry) (interest * Geographicandmarket in thebankingsector concentration * Presenceofforeign banks

498

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA sensibleor possible (forexample,is thecountryfreefrommultipleexchangeratesj), a binaryscale is used. Because scores fromthe variousquestionsare combinedto constructtheindices,all scoresare rescaled so thattheyfallin therange0 to 2, with to higherlevels of financialdevelopment.For example, higherscorescorresponding all the strictlyquantitativedata withmore continuousvariationthatwe had at our disposal (forexample,creditto privatesectoras a shareof GDP) were normalized to yield numericalresultsin the 0-2 range. Finally,the combined subindicesand financialdevelopmentindex were normalizedto generatecomposite scores on a 0-10 pointscale. The resultsare presentedin Tables 3 and 4 and Figure 1.

WeightingofVariables Thecomprehensive financial indexis a weighted development averageofthe35 differentindicators, whichimpliesthatwe hadto assignweights toeachindicator. Weights wereselectedon thebasisoftheauthors' oftherelative of interpretation importance

Table3. MENA Countries: FinancialDevelopment Comprehensive Indexi data, scale: O-10)2 (Based on qualitativeand quantitative 2000-01

2002-03

Bahrain

7.5

7.7

Lebanon Jordan Kuwait

7.0 6.8 6.7

7.0 6.9 6.8

Oman Qatar Tunisia Morocco Egypt Sudan

5.9 5.6 4.8 4.8 5.5

5.9 5.7 5.6 5.5 5.4

ArabEmirates United SaudiArabia Pakistan

6.6 6.2 4.8

3.3

6.6 6.4 6.0

4.7

3.3

4.1

Algeria

3.5

3.2

Iran,IslamicRep.of SyrianArabRepublic Libya Average

1.6 1.2 1.2 4.7

Djibouti

of Yemen, Republic Mauritania

3.8 3.2

3.9 3.5

2.5 1.1 1.0 5.0

Source:Authors' calculations. index. 'Original"subjective" weighted 2Scale:Verylow = 2.5 and below,Low = 2.6-4.9, Medium= 5.0-5.9,High = 6.0-7.5, Veryhigh= above7.5.

499

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

SUSAN CREANE

RISHI GOYAL, A. MUSHFIQ MOBARAK

AND RANDA

SAB

8.95.25.45.95.94.23.94.86.35.03.83.24.52.02.24.52.32.42.41.04.2 Institutional Environment

8.07.08.08.08.08.04.08.08.05.04.06.07.07.09.05.04.04.00.00.05.9 Financial Openness

Sector Policy 7.88.36.56.65.86.47.44.25.74.56.85.66.26.05.03.94.40.50.90.55.1 2002-031 and 0-10)2Monetary

7.5.

Index,

above =

scale:

high

data,

9.37.78.78.06.78.07.78.36.75.37.35.33.75.03.33.03.54.70.02.05.7 Supervision Regulation and Development

Very 6.0-7.5, =

quantitative

High

Sector 5.03.36.35.05.03.36.35.00.74.74.76.30.71.30.70.73.03.30.70.73.3 Nonbank Financial

and Financial

5.0-5.9, =

qualitative on

Countries:

Medium

7.38.77.17.47.97.85.86.16.87.75.66.05.73.84.13.82.51.91.91.35.5 Sector Banking

MENA 4. (Based Table

2.6-4.9, = Low index.

7.77.06.96.86.66.46.05.95.75.65.55.44.74.13.93.53.22.51.11.05.0 Index Financial Development

below, and weighted 2.5 calculations. =

of Emirates Arab Arabia

low "subjective" Rep. RepublicAuthors' Very Republic Arab Islamic 'Original Source: 2Scale:

Kuwait Lebanon United Saudi Morocco Sudan Mauritania Iran, Libya Bahrain Pakistan Oman Tunisia Egypt Djibouti Yemen, Algeria Syrian Average Jordan Qatar

500

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA Figure 1. MENA:Comprehensive Index of Financial DevelopmentComparing VeryHigh,High,Medium, Low, and VeryLow Development Countries,2002-03 (Scale 0-10)

environment Institutional

Bankingsector 10 9 8 7 6 5 4 3 2 1 0

Nonbankfinancialsector

andsupervision Regulation

Financialopenness

sectorandpolicy Monetary Averageforveryhighfinancial development Averageforlow financial development

financial Averageformedium-high development development Averageforverylow financial

Source: Authors'calculations.

on each relevantvariable,whichpartlystemsfromourreadingoftheexistingliterature

thesetofweights invarious financial (Table2). Wealtered ways development assigned thefinancial indexwe computed is robust toalternative tocheckwhether development weightingschemes.6We groupedcountriesaccordingto thecompositeindexunderfive categoriesof financialdevelopment:veryhigh (above 7.5), high (6.0-7.5), medium (5.0-5.9), low (2.6-4.9), and verylow (2.5 and below) (see Table 3). The grouping

was robust to thedifferent ofcountries intothefivefinancial development categories within each grouping the relative of countries schemes, ranking weighting although changedslightly.

dimensionsof an issue in a single or in a small set 6Indicesthatattemptto captureseveraldifferent of measuresinvariablyinvolvechoices of variablesto use and weightsto assign.This impartsan element of subjectivityto the analysis,and a biased choice of variablesor weightscould lead to incorrectinferto ences. Our choice of variablesand weightsreflectsour understanding of whatis likelyto be important distinguishmoredevelopedfinancialsystemsfromless developedones, and whatis commonlyfoundin It also reflectsconstraints on whatcould be measuredquite easily.By alteringtheassigned the literature. weights,we confirmthatour qualitativeinferencesare notsensitiveto theparticularchoice of weights.

501

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal, A. MushfiqMobarak, and Randa Sab There are also alternativestatisticalmethods, such as factor analysis and principal components analysis, to assign weights that would not be affectedby any subjective errors thatwe make. Although our primaryapproach was to rely on our qualitative judgment to identifyand then assign relative weights to differentcomponents of financialdevelopment,we also used principal components analysis (PCA) to generatean alternativeset of weights. Roughly speaking,PCA examines the statisticalcorrelationsacross scores on the differentindicators,and assigns the largest weights to those indicators of financial development most correlated with the other indicatorsin the data set. Intuitively,this method tries to uncover the common statisticalcharacteristicsacross the various indicatorsto combine theminto a composite index of financialdevelopment. Because each one of our indicatorsis meant to capturesome aspect of the concept we term"financial development,"the variable most correlatedwith the others was judged to be the most accurate indicatorof financialdevelopment. The PCA-generated weightsserve as a check forthe sensitivityof our results. We recreated each of the six subindices using PCA. The index values generated forthe 40 data points (20 countries,two time periods) are highly correlatedwith the original index values thatused weightsbased on our subjectivejudgment.The correlationranges from0.915 forthe openness index to 0.988 forthe institutional quality index, and the average correlationcoefficientacross the six subindices is greaterthan 0.97. As a result,generatingweights using PCA instead would not change our conclusions significantly. PCA also helps identifya subset of variables that,according to the correlations in the data, are the most crucial indicatorsof financial development.Table 5 lists 18 variables (out of 35) thatwere assigned a weight of 3 percentor greater (of the total weight across all 35 indicators)by PCA. These 18 variables jointly account for approximately 80 percent of the total weight. The last column in Table 5 also reportsthe weightswe chose to assign to those same variables based on our own judgment of what mattersmost in defining financial development. Comparison of the two columns indicates thatthe correlationsin the data do not always correspond perfectlywith our a priorijudgments. However, all the variables thatbear a direct relationshipto financialdevelopment appear in this list of thetop 18 indicatorsof financialdevelopment,whereas variables only tangentially relatedto financial development(forexample, democratic accountability,housing finance,quality of the bureaucracy,law and order) get close to a zero weight.The PCA does yield a sensible set of resultsand allows us to reduce our reliance on qualitativejudgments in developingindicatorsof financial development. We added the weights assigned by PCA to the individual variables to create a set of percentage weights thatmeasure the contributionof each of the six subindices to the summaryindicatorof financialdevelopment (Table 6). Althoughwe had chosen to assign the largestweights to the "banking sector" and "monetary sector/policy"themes in our originalconstructionof the index, the PCA suggests that the variables that make up "banking regulation/supervision"and "banking sector"are jointlythe mosttellingindicatorsof financialdevelopmentin ourMENA data. Comparison across the two columns of Table 6 also indicate thataccording to PCA, our subjective judgmentsoveremphasized the roles of "monetarysector/ 502

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 5. Comparison of Index Weightsby Variable (Inpercent) Principal Components Weights

Variables Ease ofloanrecovery thejudicialsystem through andprofitability ofthebanking sector Development Government inbanking andfinance involvement (Heritage Foundation) offorward market Existence exchange Privatization ofbanking sector sectorassets Depositmoneybankassets/total banking index(Heritage Foundation) Property rights Prudential ofbanks monitoring offinancial andmonetary data andavailability Transparency Basel capitaladequacyrequirements ofthecentral bank Independence Credittotheprivate sector/GDP Restrictions onforeign currency purchase byresidents Interbank transactions markets Interest rateliberalization ofmonetary Indirect instruments policy Government securities loans Nonperforming

Qualitatively Assigned Weights

6.2 6.1

4 5

5.8 5.3 5.3 5.3 4.9 4.8 4.8 4.7 4.2 4.0 3.9 3.7 3.6 3.5 3.0 3.0

2 1 3 3 4 3 4 3 3 3 2 5 5 4 3 2

Source:Authors' calculations.

policy" and the "nonbankfinancialsector" in constructingmeasures of financial development.

AlternativeIndex neededto construct the comprehensive financialdevelopBecause information mentindexis notavailableat therequiredlevelofdetailfortheMENA countries Table 6. Comparison of Index Weightsby Theme (In percent) Principal Components Weights

Themes Bankingsectordevelopment sectorandpolicy Monetary andsupervision Bankingregulation Institutional environment Nonbank financial sector Financialsectoropenness

21.3 12.8 21.4 19.0 7.8 17.7

Qualitatively Assignment Weights 25 20 15 15 15 10

Source:Authors' calculations.

503

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal,A. MushfiqMobarak,and Randa Sab over long periodsof timeor forothercountries,we developed an alternative, simplerindex based solely on a few widelyavailable quantitativedata series, similarto an index developedby Beim and Calomiris(2001). This alternative of financialsystemswithinMENA index allows us to analyzethedevelopment countriesovera longerperiod,and also allows us to comparetheMENA region to otherregions. To constructthe index,we combinedfourvariablescommonlycitedin the literatureas measuresof financialdevelopmentusing PCA. The fourvariables were (1) ratioof broad money(M2) to GDP, (2) ratioof the assets of deposit moneybanksto assetsof thecentralbankplus depositmoneybanks,(3) reserve ratio,and (4) ratioof creditto theprivatesectorby depositmoneybanksto GDP. andrelative These variablesmeasurethesize ofthefinancial sector,theimportance ease withwhichbanksprovidefunds,and theextentto whichfundsare provided across thevariables to the private(as opposedto thepublic) sector.Aggregating in a single to capturedifferent notonlyattempts aspectsof financialdevelopment measure,butalso reducesbiasesor errorsthatmayplagueanyone ofthefourdata in keepingwiththestandard series.Furthermore, practiceofaveragingthevariables in either5-yearpanelsor 10-yearpanelsto smoothoutbusinesscyclefluctuations we averagedthedatain 10-yearpanelsto obtainobservations and focuson trends, forthe 1960s, 1970s, 1980s,and 1990s.All scoresare rescaledso thattheyfallin to higherlevelsof financial therangeof 0 to 10, withhigherscorescorresponding indexcomputedfor Figure2 comparesthevaluesof thisalternative development. financialdevelMENA countriesforthe 1990s to the 2000-01 comprehensive Figure 2. MENA Countries: Comparing the Comprehensive and AlternativeIndices 7 Lebanon

6 SJordan

5

Tunisia

index 4

Morocco4 ,Qaata

Kuwait

Egypt

3 Alternative 2

Algeria Iran,I.R. of

*UnitedArab Emirates

Oman

Pakistan Mauritania

Syrian Arab Republic Yemen, Rep. of

1

Sudan

0 0

1

2

3

4

5

6

index Comprehensive Source: Authors'calculations.

504

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

7

8

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLE EASTAND NORTHAFRICA

in thisregion.Thereis a positive opmentindexwe have developedforcountries correlation in thevaluesforthetwoindicesacrossMENA countries, butFigure2 also pointsout some important differences. For example,Lebanon appearsmore financially developedrelativeto therestof theMENA countriesaccordingto the index.This reflects themuchhigherweightplaced on theM2simpleralternative to-GDPratioin thealternative index,andthefactthatLebanon'svalueforthisratio is relatively because Lebanon high.This highvalue couldbe somewhatartificial, shocks to GDP denominator in this (the ratio)duringits experienced largenegative civilwaryears.

IV.DescriptiveAnalysisofthe Data SurveyAnalysis Some basic descriptiveanalysis of the data set we compiled is suggestiveof commonstrengths, trends,and weaknessesin financialsectorsacross MENA countries,and pointsto areas in greaterneed of reform.MENA countriesin generalperformreasonablywell in regulationand supervision.But theyneed to do more to strengthen the institutional environment and promotenonbank financialsectordevelopment.Withinthe region,progresson financialsector reformshas been uneven. Some countrieshave well-developedfinancialsectors,particularly bankingsectors,such as theGulfCooperationCouncil (GCC) countries,Lebanon, and Jordan.Others,such as Egypt,Morocco, and Tunisia, have made important advancesin recentyears.Overall,however,moreremains to be done. The main findingsforthe MENA regionare summarizedbelow by our six broadthemes.7 Monetarypolicy. For the most part,interestrates (or rates of return)are indirectmonetary freelydetermined, policytoolsare employed,and government securitiesexist.However,in some cases, despitede jure liberalizationof interest rates(or ratesof return)and removalof creditceilings,continuing public sector in practiceprevents involvement determination ofratesand allocompletemarket cationof credit. In nearlyall cases,government securities billsorcentralbank (whether treasury ofcountries, some openmarketoperapaper)existto somedegree.In themajority tionstakeplace. However,in mostcountries, theincompletedevelopment or nonexistenceofsecondarymarkets forgovernment securities hindersthebroaderuse of a fewcountries do notfollow openmarketoperations by centralbanks.In addition, a comprehensive framework fordesigning andconducting monetary policy. and Lebanon,thebankingsecJordan, Bankingsector.In theGCC countries, toris well developed,profitable, and efficient. However,in abouthalftheregion, thisis notthecase. In sevenof the20 countries, thebankingsectoris dominated holds significant by public sectorbanks,and in anothereight,the government 7See Creaneandothers(2004,AppendixII) forthesurveyofthe20 MENA countries, availablevia theInternet at http://www.imf.org/external/pubs/ft/wp/2004/wp04201.pdf.

505

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, Rishi Goyal,A. MushfiqMobarak,and Randa Sab stakesin financialinstitutions. Thesecountries aregenerallycharacterized bygovin creditallocation,losses and liquidityproblems,and wide ernment intervention interest ratemargins(or spreadsin ratesof returns).In manypartsof theregion, thereis an urgentneed fordevelopingmodernbankingand financialskills. In sevenof thecountries, noncashtransactions, suchas creditcarduse or automated tellermachineaccess, werelimitedor nonexistent. The bankingsectoris highlyconcentrated in eightMENA countries.For example,assets of the threelargestbanksin these countriesexceed 70 percent of totalbank assets;thesameholdstrueforloans and deposits.In anotherseven thereis moderateconcentration countries, with,forexample,fourbanksaccountfor more of than60 percent totalbankassets,loans,and deposits.In halfthe ing countries,theentryof newbanksis difficult. thereis somecorrelation attributes ofthebankGenerally, amongthedifferent ing sector.For instance,countrieswitha highlyconcentrated bankingsector,in addition,are generallyalso dominatedby public sectorbanksand have limited noncashtransactions. Nonbankfinancialsector.In mostoftheregion,thenonbankfinancial sectorof the stock bond insurance market, market, consisting corporate companies,pension funds,and mutualfunds-needsfurther Wheresuch markets development. exist,tradingis usuallyquitelimited.For instance,stockmarketsin theregion tendto be characterized fewlistingsresulting in byhighconcentration, relatively andno separateregulatory low levels of liquidity, Moreover,stateownauthority. in somecountriesdeprivesthemarketofan ershipofutilitiesand otherenterprises sourceof newissues.The development of thesemarketsis complicated important on limitations the need for a clear and stablelegislativeframebylegal ownership, weak investor and inactive or nonexistent for work, confidence, secondarymarkets financialinstruments. have been developedin mostMENA countries, Housing financeinstitutions state-owned tendto primarily through specializedhousingbanks.Theseinstitutions subsidizecreditto low-andmiddle-income households.However,thesequasi-fiscal in government operationsare oftennotreflected transparently budgets.Banks are involvedin mortgagefinancein countries wherespecializedmortgageinstitutions arenotpresent. Regulation and supervision.Many MENA countries-such as the GCC countries,Jordan,Lebanon,Morocco,and Tunisia-have strengthened banking establishedup-to-dateproceduresto collectprudensupervisionand regulation, and regularly tial information, inspectand auditbanks.Theyhave takenstepsto Basel standards conformto international byincreasingcapitaladequacyratiosand loans. However,success in the latterhas been limited, reducingnonperforming and formostcountriesnonperforming loans remainin therangeof 10 percentto 20 percentof totalloans. The independenceoftheregulatory and supervisory authority, usuallythecentralbank,could be enhanced,and supervisors'skillscould be improved.In six of thecountriessurveyed,thecentralbankis notconsideredto be independent, and an additionalsix have onlylimitedindependence.Moreover,thedegreeof transin surveyed parencycould be improved.Abouthalfof themonetaryauthorities 506

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLE EASTAND NORTHAFRICA

countrieshad createdwebsitesto disseminatetimelymacroeconomicdata and relevantfinancialsectorlaws and decrees,althoughcoveragecould be increased. Reflectinglimitedoveralltransparency, onlyhalfof the countriespost country staffreportson theIMF website. Financial openness.MENA countries havegraduallyopenedup theircurrent as well as capitalaccounts.However,nearlyhalfthecountriescontinueto mainon repatriation ofearningsas well as on thedomesticpurchaseof tainrestrictions currency. foreign Mostof thecountriesin theregionmaintainsome formof a peggedexchange ratearrangement, withmorethanhalfof thecountriessurveyedpeggingto the U.S. dollar. Half of the countrieseitherhave or can easily access a forward exchangemarket. Threeofthe20 countries continue tomaintain and/or parallelexchangemarkets multiplecurrencyrates.At thesame time,thesethreecountries,and two others, continueto maintainrestrictions on current international and have not transactions, acceptedtheobligationsofArticleVIII (Sections2, 3, and4) oftheIMF's Articles ofAgreement. Institutionalenvironment. In muchoftheMENA region,thequalityof instiand property tutions,includingthejudicial system,bureaucracy, rights,is poor. This hindersbankingand commercialactivityas well as investment, and hence growth. In severalcountries,thejudicial systemis susceptibleto politicalpressure and long delays,resultingin poorlegal enforcement of contractsand loan recovin onlytwowas it consideredeasy to recover ery.Of the20 countriessurveyed, loans throughthejudicial system.The ICRG assigns a low rankto countriesin theregionforthequalityof thebureaucracy, at a level significantly below thatof more industrializedcountries,includingthe fast-growing, industrialized newly Asian economies. tendsto be weak in theregion.On theHF index Property rightsenforcement of privatepropertyprotection, only one countryin the region(Bahrain) has a and two (theUnitedArab Emiratesand Kuwait) ratingof veryhighprotection, have a ratingof highprotection. Similarto theresultspresentedabove, the HF notessignificant involvement in bankingand financein the region. government Its index(whichweighsgovernment influenceovercredit restrictions, ownership, and freedom to offer services in the financial allocation,regulations, sector)rates one as low restrictiveness in the only country(Bahrain) havingvery government financialsectorfor2002, and two (Jordanand Lebanon) as havinglow governmentrestrictiveness.

ComprehensiveIndexAnalysisofFinancialDevelopment We foundthat,reflecting reform efforts in theregion,Tunisia,Pakistan, continuing and Morocco moved intoa higherlevel between2000-01 and 2002-03. Within groups,therelativerankingof somecountries changed;forexample,theincrease in Sudan's rankingreflectedreforms carriedout duringthattimeacross mostof thesix categories(Tables 3 and 4). 507

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal,A. MushfiqMobarak,and Randa Sab On average,countriesat higherlevelsoffinancialdevelopment outperformed countriesat lowerlevels in each of thesix aspectsof financialdevelopment. The countriesscoredrelativelywell on regulationand supervisionand on financial of a stronginstitutional environopenness,butfaredpoorlyon thedevelopment mentand thenonbankfinancialsector(Figure1).8 In comparisonto othercountriesin theregion,MENA countrieswithhigher levels of financialdevelopment tendedto have (1) a greateruse of indirectmona smallerdegreeof public ownershipof financial instruments; (2) etarypolicy smaller or no of thefiscaldeficit;(4) stronger institutions; (3) monetary financing and humanresources,includ(5) higher-quality prudential regulation supervision; and financialskills;and (6) a stronger ingmanagement legal environment.

MENAand the RestoftheWorld:Alternative IndexAnalysis The rankingsofcountrieswithintheMENA regionunderboththecomprehensive and thealternative indicescloselytrackeach other(Figure2). This providessome confidencein usingthealternative indexto makeintertemporal and interregional index producesrankingsof financial comparisons.In addition,the alternative similarto thosedevelopedin otherresearch. development Accordingto the alternativeindex, we find that most MENA countries fromthe 1960s throughthe 1980s (Table 7 experiencedfinancialdevelopment and Figure3). In the 1990s,manycontinuedto experiencefinancialdeepening, or conflictresultedin a deterioraalthoughin a fewcountriespoliticalinstability tionof the index.The MENA regionrankswell below the industrialized countriesin financialdevelopment, butabove mostotherdevelopingcountry regions. However,it is interesting that,althoughtheMENA regionrankedwell above the industrialized economies of East and SoutheastAsia in the 1960s,it fell newly behind them in the 1980s and the 1990s, as theseAsian countries considerably steppedup financialdeepening.Withtheexceptionof sub-SaharanAfrica,financial developmentin all otherregionshas progressedconsiderablymorerapidly thanin mostcountriesin theMENA region.The countriesin theMENA region in whichtherehave been important advancesin financialdevelopment sincethe 1960s are Egypt,Jordan,Morocco,and Tunisia.In the remainingcountries, the level of financialdevelopment overthesefourdecades has improvedonlyslightly or,in a couple of cases, declined. V. Conclusions MENA countrieshavereformed theirfinancialsectorsoverthepastthreedecades. Althoughtheyhave made progress,theireffortshave been eclipsed by faster reform and growthin otherpartsoftheworld.Againstthebackdropofan increasin moving inglyglobalizedworld,one of thechallengesforMENA policymakers acrossthesix subindices shouldbe treated withsomecaution.The limited coverageof 8Comparisons variablesin each subindeximpliesthatsomepotentially which variablesmaynotbe included, important couldresultinincorrect inferences.

508

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLEEASTAND NORTHAFRICA

Table 7. AlternativeFinancial Development Index, 1960s-90s (Averages,scale: 0- 10) Countries andRegions MENA Algeria Egypt Iran,IslamicRepublicof Jordan Kuwait

1960s

1970s

1980s

1990s

2.9 2.4 1.7 3.8 3.1

3.3 4.2 1.9 3.3 3.7

3.9 5.0 3.5 2.4 5.3

3.5 2.7 3.8 2.4 5.4

5.7

5.0

...

...

5.1 2.4 2.6

6.7 2.7 2.8

9.6 3.0 3.0

6.4 2.6 4.0

Pakistan

2.7

2.9

2.6

2.8

3.3

Qatar

.........

Lebanon Mauritania Morocco

...

Oman

Sudan SyrianArabRepublic Tunisia

2.5 2.2 3.3

UnitedArab Emirates Yemen,Republicof

...

...

MENA(without Lebanon) Industrial countries "Asiantigers" LatinAmerican andCaribbean SouthAsia Sub-Saharan Africa

2.6 3.9 1.8 2.4 1.6 2.3

3.1

3.0

4.2

2.1 1.9 3.8

1.5 1.8 4.5

0.9 2.3 4.8

2.9 ...

3.8 ...

4.3 1.3

2.9 4.6 2.9 2.9 1.7 2.5

3.3 5.1 4.1 3.0 2.4 2.3

3.1 5.9 5.7 3.4 2.7 2.1

Sources:IMF,International FinancialStatistics; andauthors' calculations.

Financial Figure3. MENACountriesand Global Comparators:Alternative DevelopmentIndex,1960s-1990s1 7

7 6

Industrial countries

5 4 3

MiddleEastandNorth Africa_

6 Newly 5 industrialized economies 4 LatinAmerica andCaribbean

2

Africa Sub-Saharan

3 2

0 1960

1970

1980

1990

Sources: IMF, InternationalFinancial Statistics;and authors'calculations.

IDecadeaverages;scale0-10.

509

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

Susan Creane, RishiGoyal,A. MushfiqMobarak,and Randa Sab repressive away fromfinancially policieswill be to createan enablingstructural environment forfinancialdevelopment, intervenincludingreducedgovernment tionin creditallocationand strengthened institutional of the quality,particularly shouldbe concentrated wherefinancialdevelopment legal system.Efforts appears to have been theweakest.For some countries, thismeansless involvement of the in thefinancialsystem,includingcuttingbackon publicownership of government financialinstitutions, ofbudgetdeficits, minimizing monetary financing enhancing in humanresources,promoting nonbankfinancialdevelopinvesting competition, thelegal environment. ment,and strengthening This paperpresentsa methodology to createdetailedindicatorsof financial developmentfordevelopingand emergingeconomiesthatallows fornuanced and weaknessesof particularfinancial analysisof the strengths policy-relevant of specificareas wherereformeffortsshouldbe prianalysisand identification oritized.Althoughourspecificapplicationconcentrated on financialdevelopment in theMiddle East and NorthAfrica,themethodology we outlinecan be easily appliedto a broadercrosssectionof countriesand to longerperiodsof time.In are alreadyunderwayby economistsat theWorldBank and theIMF fact,efforts to replicateouranalysisto otherregionsofparticular interest to them. REFERENCES andRossLevine,2004,"StockMarkets, PanelEvidence," Beck,Thorsten, Banks,andGrowth:

Journalof Bankingand Finance,Vol. 28 (March), pp. 423-42. Beim, David, and CharlesCalomiris,2001, EmergingFinancialMarkets(Boston: McGraw-Hill). Cetorelli, Nicola, and Michele Gambera, 2001, "Banking Market Structure,Financial Evidence fromIndustryData," Journalof Finance, Dependence and Growth:International Vol. 56 (April), pp. 617-48.

Creane, Susan, Rishi Goyal, A. Mushfiq Mobarak, and Randa Sab, 2004, "Financial Sector Development in the Middle East and NorthAfrica," IMF WorkingPaper No. 04/201 (Washington:International MonetaryFund), pp. 26-48. Demirgtii-Kunt,Asli, and Ross Levine, eds., 2001, Financial Structuresand Economic Growth:A Cross-CountryComparisonof Banks, Markets,and Development(Cambridge, Massachusetts:MIT Press). Gelbard,EnriqueA., and S6rgioPereiraLeite, 1999, "MeasuringFinancialDevelopmentin SubSaharanAfrica,"IMF WorkingPaper 99/105(Washington:International MonetaryFund). Khan, Mohsin S., and Abdelhak S. Senhadji, 2000, "Financial Development and Economic Growth-An Overview,"IMF WorkingPaper 00/209(Washington:InternationalMonetary Fund). La Porta,Rafael, FlorencioLopez-de-Silanes, and AndreiShleifer,1998, "Law and Finance," Journalof Political Economy,Vol. 106 (December), pp. 1113-55. 2002, "GovernmentOwnershipof Banks," Journalof Finance, Vol. 57 (February), pp. 265-301. Levine, Ross, 1997, "Financial Development and Economic Growth:Views and Agenda," Journalof Economic Literature,Vol. 35 (June),pp. 688-726.

---,

"Bank-Based or Market-BasedFinancial Systems:Which Is Betterj" Journalof ---,2002, Financial Intermediation, Vol. 11 (October),pp. 398-428.

510

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions

MEASURINGFINANCIALDEVELOPMENTINTHEMIDDLE EASTAND NORTHAFRICA ---, 2003, "DenyingForeignBank Entry:Implicationsfor Bank InterestMargins," Working PaperNo. 222 (Santiago:CentralBankofChile). ---, and Sara Zervos,1998,"StockMarkets, American Banks,and EconomicGrowth," Economic Review,Vol. 88 (June),pp. 537-58.

Mehran,Hassanali,PieroUgolini,Jean-Pierre Briffaut, GeorgeIden,TonnyLybek,Stephen Swaray,and PeterC. Hayward,1998,FinancialSectorDevelopmentin Sub-SaharanAfrican

IMF OccasionalPaperNo. 169(Washington: International Countries, Fund). Monetary CarmenM., and KennethS. Rogoff,2002, "The ModernHistoryof Exchange Reinhart, Rate Arrangements: A Reinterpretation," NBER WorkingPaper No. 8963 (Cambridge, Massachusetts: NationalBureauofEconomic Research).

511

This content downloaded from 195.78.108.60 on Wed, 25 Jun 2014 10:18:56 AM All use subject to JSTOR Terms and Conditions