Saving in Transition

Public Disclosure Authorized Public Disclosure Authorized

lwves POLICY RESEARCH

WORKING

1509

Savingsin transition

Saving in Transition Economies

Economies

economieshavedeclined since independence, but

somethingelsehas happened, too: Financial

The Summary Report Public Disclosure Authorized

PAPER

69

assetshaveshiftedfrom bank deposits to alternative financial instruments,

Patrick Conway

including foreign currency, "trust company" shares,and private loans. Governiment are not typically prepared to borrow savingsfrom these new instruments since they are denominated in foreign currency or are offered only

Public Disclosure Authorized

at positive real interest rates That attitude must change if governments are to make needed investments in infrastructure and to avoid creating inflationary credit.

The World Bank Europe and CentralAsia Country DepartmentIV Country Operations Division 2 September 1995

POLICYRESEARCHWORKINGPAPER1509

Summary findings The stimulation of private saving is essential to both stabilization and structural adjustment in the transition economies. Private saving in these countries has declined sharply since independence, and this decline has been a factor in the onset of extreme inflation because governments have resorted to an inflation tax to finance deficit spending. Conway examines evidence on spending in Belarus, Georgia, Kazakhstan, and Ukraine. He examines decisions about whether to save, and in which specific financial or real instruments. He summarizes the evolution of financial sectors in these countries to provide a history of the success or failure of financial institutions to intermediate between private savers and the government as borrower. He concludes that private saving has indeed declined since independence, but less than is indicated by bankingsystem statistics. Concurrent with this downturn has been a shifting of financial assets from bank deposits to

alternative financial instruments, including foreign currency, "trust company" shares, and private loans. The financial sector has reacted slowly to this change, but the most successful commercial banks have recognized the change in demand for financial instruments and have accommodated the savers. The state commercial banks - especially the successor to the Soviet Saving Bank have been slow to adjust to the new environment. As a result, the near-monopoly that bank once held on deposits has been rapidly eroded. Government methods for mobilizing funds must change, contends Conway. Governments are not typically prepared to borrow savings from these new instruments, since they are denominated in foreign currency or are offered only at positive real interest rates. That attitude must change if governments are to make needed investments in infrastructure and to avoid creating inflationary credit.

Thispaper-a product of the Country OperationsDivision,Europe and Central Asia,Country DepartmentIV-is part of a larger effort in the region to analyze credit markets and savingsin transition economies.The study was funded by the Bank's Research Support Budgetunder the research project "Credit Markets and SavingsMobilization in Transition Economics"(RPO 679-07). Copies of this paper are availablefree from the World Bank, 1818 H Street NW, Washington,DC 20433. Pleasecontact Carole Bondarev, room H2-096, telephone 202-473-3974, fax 202-477-3378, Internet address [email protected] 1995. (43 pages plus 16 pages of annexes)

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues.An objective of the series is to get the findingsout quickly, even if the presentations are lessthan fully polished. The paperscarry the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executive Board of Directors, or any of its member countries.

Produced by the Policy Research Dissemination Center

Saving in Transition Economies The Summary Report

Patrick Conway Department of Economics University of North Carolina Chapel Hill, NC 27599-3305

Savingin TransitionEconomies The SummaryReport

ExecutiveSummary. I. Definition of saving. A. B. C. D.

The theory of saving. A theoreticalspecificationof saving in the transition. Private saving and fiscal deficits: intermediatingbetweenthe two. Implicationsof the theory for observed saving.

11. Does private saving still occur? A. B. C. D.

Results from the Georgiansurvey. Results from the Kazakhstansurvey. Conclusionson householdsaving in Georgiaand Kazakhstan. Saving and the distributionof incomein Ukraine.

Ill. The collapseof formal-sectorsaving: evidencefrom four countries. A. The deposit-inflationnexus. B. The attractionof foreignexchangein an inflationaryenvironment. C. Financial-sectorresponse. IV. Other depositoriesfor private saving. A. Financialmarket bifurcation. B. The proliferationof commercialbanks and trust companies. C. Specializationand sluggishresponseby state commercialbanks. V. Conclusionsand suggestionsfor further research. Annex A: Technical presentationof the theory of private saving in transition economies. AnnexB: Survey instrument. Annex C: Profiles of formal financial institutionsin Belarus, Georgia, Kazakhstanand Ukraine.

Savingin TrasidonEconomucs- 2

Acknowledgements This research was fundedby World Bank ResearchGrant 679-07. I have benefitedfrom the assistanceof four groups in preparingthis manuscript. The first is the group of specialistsin the transition economieswho gave of their time and expertiseto further my research: these include in Belarus: Tamara Aksenova,AlexanderBarkun, Yurii Dyurbeiko, SegismundoFassler, Vladimir Grigoriev,Anna Grinkevich,Anatoli Gromovich,SvetlanaIliukevich,AlexanderKudlach, NikolaiLevenkov,Alla Lyubushina,Sumaya Minhko, Nicolai Omelyianovich,Ludmila Rykova, NadezhdaSiburskaya,VapenshinaVarvasheniya,Vladimir Volodko, Alexander Zastavnyuk,GennadyZinovkin, and Messrs. Kernazhitskyand Viaskin. in Georgia: Peter Bakradze,Vano Chkhartishvili,MichaelChkuaseli, MichaelDjibouti, Victor Gonashvili,Merab Kakulia, MamukaKhazaradze,Levan Kistauri,David Kodua, Irakli Koplataidze,Gocha Laphauri, VakhtangMagradze, Irakli Mgaloblishvili,Murad Narsia, Larisa Nikolaishvili,Ghia Nodia, George Tarkhan-Mouravi,Otari Tohadze and Valeri Vibliani. in Kazakhstan: Nurlan Akimov, GulbairanAymanbetova,Marat Bisenov, Liazat Buranbayeva, Andrei Butukhanov,Ergali Dosmagambetov,Oraz Dzhandosov,Zhannat Ertlesova, Erkin Etekbaeva,Nina Krivko, Grigori Martienko, NataliyaMichailova,Serik Sansibaev,Elnar Segizbayev,GulfairuzShaikakova,Juri Shokamanov,Galina Starostenko,Nurjan Subkhanberdinand Erjan Tatishev. in Ukraine: Nina Dorofeyeva,Natalya Grebenyk, SergeiKhodevich, Olga Kruglak, Yelena Martiniuk,Ludmila Mikitenko,AlexanderMruk, Sergei Rabenok,VolodomirSmolenko, VladimirTimchenko, Irina Ukrainetsamd Alexi Volchkov. Some of these representedthe financialinstitutionsreported upon in the text and associatedannexes. Those institutionswere quite open in their cooperationwith this project. The householdsurveysreported upon in this manuscript were conductedin concert with a group of collaborators. In Georgia, these included AmiranTsakadze, losef Gogodze,Rezo Chitashviliand Dato Adeishvili,while in KazakhstanI collaboratedwith Zulfiya Sultanovaand Bakhyt Abdildina. The interpretationin this documentremains my own, but I look forward to further analysisjointly with these colleagues. I owe thanks as well to those who facilitatedmy travel in the transitioneconomiesto gather informationfor this report. These includeWorld Bank representativesand staff Dani Kaufmnann, David Pearce, BonniePacheco,Zhanar Abdildina, Lina Rubinovich,Anna Skliarenkoand Galina Voitsekhovskataya.Also importantwas the assistancefrom Kate Dumbadze,Lali Kikalishviliand Lili Mdzeluri in Georgia,Irina Bazarovain Ukraine, Bakhyt Abdildinain Kazakhstan,and Anna Suvorovaand YevgenyBogomasovin Belarus. Finally, I have had the commentsand criticismsof a number of World Bank economistsat various stages in this report. These include Daniela Gressani, FrancoiseLeGall, MichelleRiboud, Alan Gelb, Wafik Grais, Ricardo Martin and ChandrashekarPant.

Saving in TransitionEconomies- 3

ExecutiveSummary Encouragementof private saving is a central featureof economicsuccess for the transition economiesin both the short run and the long run. The transitioneconomiesare those former republics of the Soviet Union that are introducingthe infrastructurefor market activity into the economicenvironment. In the short run these economieshave been characterizedby extreme inflationand are implementingstabilizationprograms: the current burden of these programs will be less onerous on the citizensif the governmentcan financepart of its expendituresthrough private saving. In the long run, private saving will be the preconditionfor the investmentnecessaryto introduce technologicalimprovementsinto the productionprocessesof these countries. Stimulating private saving is thus a sine qua non of both stabilizationand structuraladjustment. Private saving has unfortunatelydeclinedsharply in the period since independencefor these countries. This declinehas itself been a factor in the onset of extremeinflation, since as private saving was reducedthe governmentsresortedto the inflationtax to financedeficitaryspending. In this research report I summarizethe availableevidenceon saving from four of these countries: Belarus, Georgia, Kazakhstanand Ukraine. I examinethe saving decision as well as the related decisionto place the saving in specificfinancialor real instruments. I also summnarize the evolution of the financialsectors of these countries to provide a history of the success(or failure) of financial institutionsto intermediatebetweenprivate savers and the governmentas a dominant borrower. I concludethat private saving has indeeddeclinedin the period since independence,but to a lesser extent than is indicatedby examinationof banking-systemstatistics. There has concurrently with this downturnbeen a shiftingof financialassets from banking depositsto alternativefinancial instruments, includingforeign currency, "trust company"shares and individuallending. The financial sector has reactedslowly to this change, but the most successfulcommercialbanks have been those that recognizedthis change in demandfor financialinstrumentsand have accommodatedthe savers with appropriateinstruments. The state commercialbanks, and especiallythe successorbank to the Soviet Saving Bank, have been quite slow in adjustingto this new environment. As a result, the near-monopolythat this bank onceheld on deposits has been eroded rapidly. This shift in financialinstrumentsfor saving implies a shift in governmentmethods for mobilizingthese funds. The governmentis typicallynot prepared to borrow the saving in these new instruments,since it is denominatedin foreigncurrency or is only offered at positive real interest rate. This attitudemust change if the governmentsare to avoid inflationarycredit creationand to make the necessaryinfrastructuralinvestmentsfor the future. This report has three novel components. There is a dynamictheory of saving for transition economiesthat is not found elsewhere. There is a sumnmaryof original surveysconductedin Georgia and Kazakhstan(with greater detail in Conway (1994b))to determinethe degree of householdsaving. There is also a summaryof macroeconomicstatisticson saving in the financialsystem that has not been pulled together elsewhere. The text of the report provides a summaryof results, while greater detail can be found in the Annexesand in the earlier report "SavingMobilizationin Transition Economies". An annex of detailedfinancialinformationfrom specificfinancialinstitutionshas been excludedfrom this publishedversion, but can be obtainedfrom the author.

Saving in Transi*zonEconomids - 4

I. Definitionof saving. Private saving is at the center of both present and future concernsfor the economiesof the former Soviet Union. Stabilizationprograms currently in place are contingenton sufficientprivate saving to offset continuedpublic sector borrowingrequirements. Projectionsfor future economic growth are contingentupon a strong private saving and investmentresponseto the planned cutbacks in public investment. Despite its centrality, there has been no attentionpaid to the motivationsand incentivesto private saving in these economiesin current economicresearch. I provide an analytical and empiricalexaminationof the macroeconomiccharacteristicsof private saving behavior in transition economiesin this researchproject. The analysisis new, in that it addressesthe specific features of the transition process; the empirical evidenceon the pattern of saving observed in Belarus, Georgia, Kazakhstanand Ukraine is new as well. The theory of saving. Saving behavior has receivedextensive attentionin the theoreticaland empiricaleconomic literature. However, this literature has paid little attentionto the dynamic evolutionof saving, especiallywhen beginningfrom disequilibrium. There has also been little attentionto the specific incentivesto private saving in transitionaleconomiesdue to the inheritedSoviet financialinstitutions and markets. These two features are crucial to the proper understandingof present saving behavior in the economiesof the former Soviet Union. In the followingsectionsI provide an overviewof the theory of private saving. For those parts that are original to this report, I provide mathematicalrestatementin Annex A. Previousliterature. The determinantsof saving have been examinedextensivelyin the literatureboth for developedand developingcountries.' Most recent analysesrecognizethe intertemporalnature of the saving decisionand begin with the life-cycleor permanent-incomehypothesis(Giovannini(1985), Fry (1988), Campbell(1987), Ostry and Reinhart (1992)). Desired saving (or its "residual",desired consumption)is derived from an Euler equation definingthe evolutionof saving over time. In theory such saving is positively related to the real interest rate through both wealth and substitutioneffects, and positively relatedto increases in real wealth. These analysesare missing, however, the evolution 2 of private saving out of the steady-stateequilibrium. The importanceof the real interest rate in the determinationof saving was also debated in the literature on financialdeepening. McKinnon(1973) introducedthe notionthat economieswith "deep" financialmarkets -- those providingfinancialinstrumentsthat attractedsubstantialsaving flows into the formal financial sector -- would have more impressiveeconomicgrowth than those that did not. 1 Most macroeconomictextbooks devote a chapterto the various theoriesof the consumption function; for given levels of income and tax liabilities,this becomesa theoryof saving as well. Blanchardand Fischer (1989)provides an overviewof life-cycletheories of saving. Smith (1990)is a recent reviewwith empirical evidencefor advancedcountries.

Campbell(1987)does considersuch adjustmentsin an error-correctionframework,but it takes the real interest rate as exogenousand unchangingin its analysis. 2

Saung in TransitionEonomies - 5

McKinnonsaw positive real interest rates as the most direct methodto encouragefinancialdeepening. The concurrentincrease in saving and reductionin the relative price of present consumptionderived in the followingmodels illustratethis point, althoughthe real interest rate is properly endogenous. Van Wijnbergen(1983) and Buffie(1984)demonstrateone channel for that endogeneity: government policies that artificiallylimit interest rate in the formal sector will encouragethe growth of informal sector intermediaries. In the formerSoviet Union. These analysesare suggestive,but do not providea satisfactoryframeworkfor research in the former Soviet Union. Despitethe liberalizationof prices and partial decontrol of financial-sector institutions, substantialinstitutionalrigidities remain. These can either block directly the market outcomeor provide undesirableincentivesto savers and intermediaries. Further, an appropriate theoreticalstructure for this study must be one which allows the tracking of stocks and flows of domestic financialassets so that the "ruble overhang"phenomenoncan be explainedsimulateously with the continuationof saving. Empirical investigationof private saving behavior in these formal institutionsmust acknowledgethat the observedquantitiesdo not correspondto equilibriumat market-clearinginterest rate. Cottarelli and Blejer (1992)provideestimatesof the ruble overhangusing an appropriate methodologyfor the period 1986-1990,and concludethat it was a substantialphenomenonduring the period. This analysiswas not carried forward into the period of independence. The Soviet financialsystem was historicallyone of private saving and governmentdissaving. The governmentdid not cover all investmentexpenditurefrom taxation, and the financialsector intermediatedto ensure that private saving was made availableto finance the excess. Nominal interest rates on this were fixed.3 Private saving includesboth householdsaving and private enterprise saving. Private enterpriseshave historicallybeen a minor source of saving, althoughthe evolutionof this mix of sources of private saving is an importantpart of this study. A key feature of the Sovietsystem in its final years was the existenceof "forced saving". The unavailabilityof goods at the prices fixed by the state led to currency holdings and financial deposits in excess of what the private sectorwould have chosen at those prices. This forced saving accumulatedin the financialinstitutionsas deposits and under mattressesas currency, and grew to such proportions that it was tagged the "ruble overhang" (e.g., Nordhaus(1990)). It representedan accumulationof purchasingpower by the populationthat was to be redeemedfor future production. Financialdeepeningas commonlymeasuredwas quite advancedin the republicsof the Soviet Union in the years precedingits dissolution;for example,householdsaving deposits in 1989 were 36.2 percent of GNP while a comparablefigure in the US was 33.7 percent.4 Interest rates were low 3 Especiallyin later years governmentexcess expenditureexceededprivate saving. In the absenceof a flexiblenominal interest rate this led to central-bankacceptanceof governmentdebt. The consequentinflationarypressure was suppressedby the fixed-pricemarket mechanism.

' The Soviet figure is drawn from McKinnon(1991, p. 121). The US figure is taken from the EconomicReport to the President 1992. Householdsaving is the total of demanddeposits, other checkabledeposits, moneymarket funds and saving deposits (Table B-66)and is dividedby GDP.

Saving In Transtion Economfis - 6

in nominalterms, but inflationwas repressedto even lower levels. Further, measuredfinancial deepeningdid not necessarilyreflect desired holdingsof financialassets becauseof the forced nature of saving. Althoughthe Sovietgovernmentachievedone benefit of financialdeepening- channeling of saving - it did not achievethe efficiencyof investmentbecauseof the commandnature of the investmentprocess.5 A second feature of the former Sovieteconomy is importantin the comprehensionof the quantity of depositholdings - the evolutionof prices and inflation. The usual solutionfrom the researchon "fix-price"equilibriais to specify a separate dynamicthat translates excessdemand into price increases. As Conwayand Gelb (1988) illustrate, the evolutionof the nominal commodityprice can be modeled as a sluggishtatonnementprocess in the goodsmarket. Immediateprice adjustment to eliminateexcess demandsis a limiting case of this more general specification. A theoreticalspecificationof saving in the transition. The theory of private saving is best viewed as an intertemporalchoice for the allocationof real resources. In this sectionI begin with a two-periodmodel to illuminatethe basic economic forces determiningprivate saving. I then turn to an extensionof the frameworkin which the dynamic of transition is explicitlyderived. The two-periodmodel. The private sector will be modeled as equatingthe present value of expenditureover the two periods to the presentvalue of after-tax income and the wealth carried forward from the past. The presentvalue of after-taxincome includesincome in both periods, with period-oneincome subjectto a productionshock (representedby ca). The material balancein the period-oneconsumptiongood is then c' + g -y'(a)=(p)

(1)

The governmenthas its own expenditures(g) in period one that are not necessarilyfinanced completelyout of taxationreceipts and contributenothingto utility. A is the volumeof international borrowing(i.e., the negativeof the current account surplus)undertakenby the households. The relativeprice of period-onegoods is denotedp; it can also be thoughtof as the real interest rate relevantto saving decisions.6 Internationallenders are assumedto have a positivep-elasticityof internationalborrowing, with the volume of availablefinancingrising with p.7

This understatesfinancialdeepening in the US in that equityholdings are ignored. 5 Conwayand Gelb (1988)concludedsimilarly for Algeria -- saving rates were astronomically high, but the subsequentinvestmenthad very low returns in terms of output.

The relativeprice p is defined as P'(1+r)/P2 , with P' the price of goods in period i and (l+r) the relevantnominal discountor interest rate to intertemporalsubstitution. 6

' Rising p indicatesthat the real interest rate that the home countrypays on internationaldebt is rising. This specificationwas chosen to provide a parameterizationof two extremecases: an economywith unlimitedaccessto world capital markets at given price p' (infinitep-elasticity)and an economywith no accessto world capital markets (zero p-elasticity). Those special cases are derived

Saving in Transidon Economis

- 7

Saving of the households in period one is defined as S = y'(CI) - c' - T

(2)

and as combination of (1) and (2) indicates is identical in equilibrium to the difference of the government budget deficit (g-r, with T denoting tax revenues) and foreign borrowing ,6. There are thus three sets of determinants to real saving. The first is the government fiscal policy, as summarized by the fiscal deficit. The second comprises shocks to earnings capacity, as indicated by the shock a to period-one product. The third is the relative price p. While the first two are exogenous, the third is jointly determined with real saving; I examine this joint determination in the following sections. Consider the illustration of two-period equilibrium in Figure 1. I illustrate three equilibria corresponding to three p-elasticities of foreign borrowing. The indifference curves u., u' and u(p) correspond to three different cases. At equilibrium A the volume of foreign borrowing is fixed at B. Given y' and g, then, cl is fixed as well. The domestic real interest rate p adjusts to equilibrate period-one and period-two consumption demands, and is given by the (negative of the) slope of the indifference curve u. passing through A.8 It is higher than the international real interest rate pe, indicating that the economy will borrow more if given the opportunity at the real interest rate p-. At equilibrium B the economy is given the opportunity to borrow unlimited quantities at the rate p-, and chooses to obtain fl' in this way. At that quantity of borrowing, the real interest rate p is just equal to the (negative of the) slope of the indifference curve u-. In the intermediate case the economy can borrow along the schedule s(p) illustrated in Figure 1. It will thus choose a level of foreign borrowing intermediate between ,6. and ,l at a real interest rate falling between p0 and p'. This is illustrated at the point E on indifference curve u(p). In this example the greatest amount of saving (equivalently, the lowest period-one consumption) is associated with the highest p, while the smallest amnountof private saving is associated with the lowest p. The rise in p is coincident with the attraction of greater private saving, as two products of the increase in net demand for such saving. The source of this demand is the government budget deficit, while the availability of foreign borrowing provides a source of funds substitutable for private saving. Real saving is rising with an increase in g, although not in a one-for-one manner if foreign borrowing is possible. It is falling with an increase in the accumulated stock of wealth and rising with an increase in real taxation. It is falling with negative output shocks in period one.9 Those

directly in Annex C. a In this case, if there is (as illustrated in Figure 1) a differential between the domestic and international real interest rates, then the definition of n in equations (1) must be expanded to include the term (p - p) 0 . This does not change the analysis, but incorporates the windfall gain from foreign borrowing at the relatively low international interest rate.

9 This last effects are not ambiguous, and provide an interesting contrast with the impact of increased government expenditure. All increase the excess demand in the goods market in period one, but the government's marginal propensity to consume in period one is unity. Given the private marginal propensity of 6, the impact of the output shock or taxation effect is reduced and can be

Samng in Tranitdon Economies- 8

exogenousfactors that causean increasein real saving do so through crowding out private consumption;the mechanismis the rise in real interest rate p. The flexibilityof accessto internationalfinancialmarkets is importantin this result. Consider the extremecase of f,p = 0, or accessto internationalfinance totally insensitiveto the domestic interest rate. Then only shifts in govermnentexpenditure,taxationor period-oneoutput matter to saving, and these enter in a one-for-onemanner. The resource constraint is completelybinding, as in equilibriumA in Figure 1. At the oppositeextreme, with ,B(and A) approachinginfinity, the relative price domesticallyis equated to the internationalrelative price (p = p'), and ,Bis the endogenous variable. Under that alternativespecificationthe results are quite different. The results depend criticallyin this case upon I, the marginalpropensityto consume in period one. Saving is falling with an incomeshock, with an increase in carried-overwealth and with an increase in tax. This reduction is not one-for-onewith the shock, however, since a portion of the shock is carried over into period two through foreign borrowing. A governmentexpenditureincreasehas no impacton saving, as the trade deficitis increasedone-for-onewith this increase. As indicatedin equation(2), private saving in equilibriumis equal to the sum of the fiscal deficit and the trade surplus. The channelby which this equalityis achieved,however, can differ accordingto the structure of the economy. If the fiscal deficit is taken as exogenous,then the two (not mutuallyexclusive)adjustmentchannelswere outlined above: externalborrowingor an increase in private saving. There are further a numberof methodsfor mobilizingprivate saving. One method is through rationing of consumptionat fixed price and nominal interest rate; given the lack of other uses for private income, "forced" saving will occur. Another is through adjustmentof p, the relative price of goods in the two periods. For given period-twoprice, there are two channelsfor adjustment of p to reflect excessdemandsfor goods in period one. The first is through an increase in PI, causinga rise in p and a shiftingof expendituresto c2. The second is through an increase in the nominal interest rate (for given P'), and will have the same effect. Interest rate controls alonedo not avert this adjustment;such efforts will simply occasiona price increase in P', as the commodityprice in period one rises in responseto excess demand. Modellingsaving in the dynamic transition. Life-cyclemodelsof consumption,investmentand saving over an infinitehorizon provide a convenientstarting point for the theoreticalmodel. The optimal behavior in Campbell(1987)or Blanchardand Fischer (1989) can be summarizedin the notation of the previous section, with conditionsintroducedto govern the adjustmentof nominalprice (P') and the stock of financialassets (D') over time: D=

(l+r)

D' + PI st

St= yt - C- tt Pt = P"' +

(3) (4)

[

(gt - tt ) - st - f(p t)]

reversed for extremelyhigh valuationsof period-onerelativeprice p.

(5)

Saving,in TransidonEconomies- 9

The stock of financialassets rises with interest paid to holders and with the value of private saving in the period. The net material balancecondition (1) may not hold in flow equilibrium,allowingfor the possibilityof rationing in periods of excess demand. This excess demandis a functionthe government deficit and the volume of foreign borrowing f(pt).lO As equation (5) indicates, such

excess demand will cause price inflation;the limitingcase, as 0 approachesinfinity, is that of net material balance. The stock equilibriumfor this dynamic economyhas a saddlepointnature. It is illustratedin Figure 2. AA representsthe combinationsof D and P that assure asset market stock equilibrium (i.e., Dt = Dtl ) and GG the combinationsthat ensure zero excess demandin the goods market. The AA curve as drawn is steeper than the GG curve; this is a propertyof the parametersof the system.' The system has a convergentdynamiconly along the saddlepathSS. Consider the adjustmentalong the saddlepathfrom point A in Figure 2. Excess demand is positive (and thus private saving negative)throughoutthe adjustment,as is indicatedby the gradual 2 In the absenceof real shocks to output, the governmentbudget or liquidationof the stock Dt.Y foreign borrowing, real consumption(and real saving)will be constantthrough time but the nominal excess demand in each period is lessenedand the rate of price inflationslowed. Nominalgoods market equilibriumand asset market stock equilibriumare only achievedconcurrently,however, at the equilibriumQ. The analysisof point A does not fully capturethe transition for the Sovietsystem becausethe householdsat independencewere not on their preferred trajectory. The constraintson both price levels and supply of goods led to rationing and forced saving off the SS curve. This forced saving resulted in the "ruble overhang" that could more properly in this context be characterizedas a "deposit overhang" and illustratedat point B. The dynamicfollowedfrom B with price decontrol replicatesa numberof featuresof private adjustmentin the former Soviet republics. First, the quantity of depositswas certainlymore than necessaryto generatedemandfor the quantity of goods available: this is evident in B's distance aboveGG. Second, householdsneverthelessheld these deposits (rather than cash, for example) becauseof their claims on future consumption: this is evident in the distanceby which B is below SS. When prices were liberalizedthere was an immediatejump (to point C), followedby ongoing inflationas householdstemperedthe upward pressure placed on demandfor goods becauseof the effect of the ensuingprice increaseson purchasingpower in the future. The price jump is representedby the movementfrom B to C on the saddlepath;the ongoing inflationby the movement along SS from C toward Q.

10 The equationof motion for foreign debt is subsumedin the asset accumulationequation, and the nominal interest rate is the same on foreignand domestic borrowing. Theseassumptionscouldbe relaxed, but at the cost of modelingthe evolutionof internationaldebt separately.

" The conditionfor this to hold [ (actIaDt-)(gt - tt - ,') > r (8ct/aPt-)J is always true for Dt > 0. 12 This is not a requirementof the model; the saddlepathcould be upward-slopingand still satisfy the saddlepathproperties.

SaWng in Trwsifion &owomif - IO

Real private saving in this adjustmentneed not take a constantvalue if there is access to internationallending. In that case the intertemporaladjustmentin p' can induce foreignfinancial flows that compensatefor variations in private saving. The relative price p' must be interpretedhere as the vector of relativeprices of goods betweentime t and all future time periods. The predictedpath of saving is made more complexby collapseof commodityproduction, as for examplewas observed in the former Soviet Union with independence. This here enters exogenouslyas an increase in the public-sectordeficitto be financedthrough the inflationtax."3 An increase in the steady-statereal fiscal deficitwill have the effect of shiftingdown the asset-market curve to AA' and shifting in the goods-marketequilibriumcurve to GG' as in Figure 3. Beginning from equilibriumQ, pricesjump immediatelyto C' alongthe new saddlepathSS', and then decline somewhatas householdsaccumulatedeposits. The new steady state at E' has a larger nominal stock of deposits but a reducedreal value. This period would be characterizedby excess supply of goods, as prices have overshot the long-runvalue. Alternatively,the increasedfiscal deficit could be recognizedwhile still on the original saddlepatharound C -- if so, there would be an additionaljump in P followedby still further inflationwhile moving down the other arm of SS' to Q'. Similar comparativedynamicsdemonstratethe real saving- and deposit-increasingeffects of exogenous increasesin nominal interest rates. Channels for retention of saving This analysispresumedthat the financialinstrumentsin which saving is placed can be representedby the homogeneousasset Dt with given rate of return. In fact, there is great variationin the characteristicsof assets serving as stores of saving. The traditional instrumentwas the homecurrency deposit in the formal financialsector - during the Sovietperiod, in the Saving Bank. Since independence,however, the potential saving instrumentshave proliferatedto include deposits denominatedin foreign currencies,holdingsof foreign currencies, depositswith informal financial institutionsand trust companies,and stocks of commodities. The former Sovietrepublicshave inheritedthe Sovietfinancialsystem. The formal financial sector in each republic includesinstitutionslicensed by the state anidis dominatedby a few large organizationscontinuingto reprise their roles in the Sovietsystem. The inherited Soviet structurehas been explainedin detail elsewhere- see, for example,the text by Gregory and Stuart (1990), Hardy and Lahiri (1992), Ickes and Ryterman(1992)or Conway(1994a). Three features are of importance to the present study: monopoly,specialization,and the dichotomybetweencash and non-cash transactions. The informalsector includesa collectionof unlicensedintermediarieswhose activities 13 Explanationof the collapse of productionin the former Soviet republicswill be a fascinating task, but one largely outside the scope of the presentstudy. I will examinethe incentivesto enterprise saving, becauseit is anomalousto observe increasedreal time deposits during a period of free-fall in enterprise production. One possible explanationis the growth of an informal productive sector unmeasuredin official statistics;this will reinforcethe conclusionson informal activity discussedbelow. The fall in productionleads directly into large public-sectordeficits. State enterpriseshave reacted slowlyto the downturn, and have consequentlymaintainedpersonnelcosts in excess of revenues. These must be paid from the state budget. Further, the downturnin economicactivity leads to a downturn in taxationrevenuefor the governmentand to an upturn in demandfor pension and unemploymentbenefits.

Sawng in TransidonEconomnia- 11

are poorly monitoredor understood. The monopolisticand specializedfeatures of formal financial intermediarieshave been carried forward into the financialsectors of the former Soviet republics. They have two implicationsof importanceto savers: low (oftennegative)real interest rates in formal-sectorfinancialinstitutions, and an explosionin activityby informal intermediaries. Formal-sectorfinancialinstitutionshave in practicebeen quite sluggish in raising interest rates on householddepositsto rates in excess of inflation. The followingtable, for example, indicatesthe nominalrates on deposits in domestic currenciesat the Saving Banks of various former Soviet economiesin mid-1993.

Kazakhstan Ukraine Georgia Belarus

Saving Banks: Nominal Interest Rates on Deposits, mid-1993 Sight deposit 1-3 year deposit Inflation (2/92 to 2/93) 15 30 800 120 -1400 5 10 800 20 40 800

Source: data collectionnetwork. Conway(1994a) identifiesthree possible reasons for these strongly negativereal interest rates. First, it may be simple inertia in respondingto market forces: these-werethe rates used under the Sovietsystem, and these are the rates to use today. Second, the SavingBank may be respondingto governmentinitiativesto hold depositrates down so that lendingrates can also be depressed- a clear exampleof the financialrepressionphenomenonMcKinnon(1973)discussedin developingcountries. Third, the Saving Bank may have little incentiveto attract more deposits. As Conway(1993) illustrates,the monopolyposition of the bank vis-a-vis borrowersand the problemsof imperfect informationand enforcementall work to reduce the desiredlevel of lendingof the Saving Bank at any real interestrate. It will then set a depositreal interest rate at the lowest level consistentwith attractingfunds to finance the profit-maximizinglevel of lending. This level drops even lower since its deposits are presentlyguaranteed(in nominalterms) by the governmentin all these economies. The financialpositionof the republicsof the former Soviet Union has becomemuch "shallower"in the period since independence. After a controlledrise in prices in mid-1991,a substantialliberalizationof prices in former members of the SovietUnion occurred in January 1992. With price increasesthat reached 1000 percent in the first three months of 1992, the "ruble overhang" was eliminated. Sincethat time, price inflationhas remainedsubstantial. Private saving and fiscal deficits: intermediating between the two. The previous theoreticalexamplesdo not model the processof intermediationbetweenprivate savers and the fiscal deficit. In the Soviet Union this process was nearly automatic: private saving was depositedwith the public uni-bankknown as Gosbank,and Gosbankthen extendedcredit to public investmentand consumptionactivities. The interest rate was positive in real terms, but did not fluctuateto clear the credit market. Rather, the public sector by plan had first claimon goods, and the private sector was able to consumeonly what remained. In a market economythis would have led to increasedcommodityprices (and a fall in p), but in the commandeconomythe result was rationing of consumptionand consequent"forced saving". Saving in Gosbankwas in excessof that desired by the households,but in the absenceof goods to purchasethe depositsaccumulatedinto the

Saving in Trwnsidon Economici - 12

"ruble overhang". Since householdsin the Soviet Union were not integratedinto internationalcapital markets, imports did not serve as an alternativefor the scarce domesticproduct. With independence,fiscal deficits rose and private income fell. Together, these led to an imbalancein the capital market. There was further a looseningof restrictionson commoditymarkets; the result was an initial burst of inflation. The capital-marketimbalancewas worsenedby the maintenanceof artificiallylow nominal interest rates by the state banks. This combinedwith the inflationaryburst to turn sharply negativethe real interestrate on deposits and discourageprivate saving in the banking system. There was thus an ex ante shortageof saving. The short-termsolutionto this shortage relied upon a variant of the "inflationtax". The private sector receivedincome in nominal assets (cash or bank deposits). The fiscal deficit was financedthrough the bankingsystem extensionof directed credits refinancedby the centralbank. The fiscal purchaseswere made with these credits; when the private sector spent its income it found the supply of goodslimited. Excess demand then bid up the price of goods. In the notation of the previous sectionthis registers as saving, since the real value of income exceedsthe real value of consumption. From the householdperspective,however, it is a tax becausethe real value of purchasingpower in the present has been decreasedby the inflation. The key to the effectivenessof the tax was the private sector's continuedholding of nominal assets denominatedin the home currency. The private sector quicklyrecognizedthis and switched its holdings into assets denominatedin other currencies. I denote these other assets as foreign-exchange holdings, either of cash or of deposits in the banking system. The tax becameless and less effective as the tax base becameless and less, and in the spirit of Cagan(1956) inflationentered an upward spiral. Implications of the theory for observedsaving. Private saving is the deferral of consumptionfrom the presentto future periods. As noted above, it respondsto price incentives. However, it is also subject in aggregateto the fundamental material balanceconstraints,especiallywhen foreignlending is inelasticallysupplied. Saving reductionswill be triggered by negativetemporaryoutput shocks and increasesin government expenditureand taxation; it will also fall through increasesin the value of the stock of inherited wealth. All of these scenarioswill be accompaniedby an increase in the relative price of goods at present, either by rising nominal interestrate or by increasein period-oneprice level. The dynamicsof adjustmentto these shocks can be complex, as noted in the transition analysisabove. The economiesof the former Soviet Union faced two adjustmentssimultaneously. Negativeoutput shocks reduced the volume of saving, ceteris paribus, while the deposit overhangled to a concurrent "shock" after price liberalizationthrough appropriateadjustmentof the current price for given interest rate to attain the equilibriumreal interestrate consistentwith the availabilityof consumptiongoods. This latter "shock"is analyzedin detail above, and indicatesthe evolution of nominal price through time to ensure material balance. The stock of nominal assets is decliningover time, and the flow of real saving is kept constantthroughthe adjustmentin real wealth. Governmentexpenditureas consideredin this section includesboth current expenditureand investment. Reduction in governmentexpenditurecauses lower real interestrates and increased private consumption,other things equal. This expenditurereductioncouldfollow from separate constraintsthrough the governmentbudget and financing. If the governmentis unable to obtain

Saving in Transidon Economnid-13

resourcesfrom the private sector through taxationor saving mobilization,then private desired consumptionwill rise. A reduction in formal-sectorsaving deposits need not be a reliable indicator of the flow of real saving in the economy. These depositswill fall into disfavor with savers if financialauthorities impose artificiallylow real interest rates through nominal interest rate controls and persistentinflation due to governmentbudget deficits. Other saving instrumentswill then provide a more effectivehedge against interest-raterisk, and will absorb an increasingshare of observed saving activity. These saving instrumentswill not at present provide resourcesfor governmentinvestmentor consumption, but will be tapped by others in the private sector to finance their own consumptionor investment. II. Does private saving still occur? The theoreticalmodelpresented earlier indicatesthat private saving will declinewith worseningof the income of the private sector, and that this declinewill be coincidentwith either high nominalinterest rates or inflationas methodsof achievingthe equilibriumquantityof consumption. In practice, the adjustmentmechanismhas workedthrough inflation, since the economieshave controllednominal interest rates. The income shocksin the former Soviet economieshave further been continuousand negative.Governmentexpenditureprograms have been reduced only little in the period of independence. Therefore it is a real possibilitythat private saving has disappeared,or dwindledto very little, during the post-independenceperiod. Despitethe importanceof informalsources and uses of householdincome, statistical informationis unavailable. The governmentalstatisticalagencies in a number of these countries continueto administersurveys that ask only about the formal sources of income; as a result, much 4 economicactivity goes unmeasured." The mismeasurementof sources and uses of incomehas profoundeffectsfor policy discussions. In the absenceof a completecataloguingof resourceflows, the analysisof consumerdemand and saving will be skewed: there may be an underestimationof the volume of householdsaving that could be attractedto formal financialintermediationthrough appropriateuse of incentives(e.g., positivereal interest rates). In collaborationwith statisticiansin Georgiaand Kazakhstan,I have written and implemented a survey instrumentthat examinesthe importanceof non-formalincome and expendituresin the householdbudget. It is based upon the householdflow of funds, and is designedto elicit information about the inflow of resources from all sources and the usage of this total inflow. The survey instrumentsas prepared for Georgiaand Kazakhstanare provided in AnnexB in English and Russian. The flow of funds is illustratedin Figure 4. The sources of funds will equal the uses of funds. Formal-sectorincome will be only one of many sources of funds, while deposits in the formal banking system will be only one of many non-consumptionuses of funds. Truthful responsesto the This lacuna in the reach of the statistical instrumentswas recently remarked in a Wall Street Journal article on Russianstatistics. For example,Mr. Yuri Yurkovof Goskonistat-Russiais quoted as saying that a reported 29 percent drop in industrialoutput is "just a sign of a more market-oriented economy". Misha Belkindasof the World Bank states that "the old statistical system is breaking down, and the new one has yet to be built". See ClaudiaRosett, "Figures Never Lie, but They SeldomTell the Truth about the Russian Economy",Wall Street Journal, 1 July 1994. 14

*Savng in Transidon Economies - 14

series of questionsof the survey provide a more detailed examinationof householdeconomicactivity and saving than has heretoforebeen available. The Flow of Funds survey was administeredin Augustand November1994to 500 householdsin Georgiaand in Kazakhstan. Its scope is quite limitedin depth of questionnaireand in geographicalreach of the survey. It does, however, provide a first look at the private-sectoractivity of householdsin these transition economies. In the followingsectionsI report results from the first round of each of these surveys. The results reported here are not exhaustive,but are focused upon the saving behaviorand foreign exchangeuse of the households. Conway(1994b)provides further detail on the income and expenditurecharacteristicsof the sample. Results of the Georgia survey. The Tbilisi economyin each samplingperiod was one in which most expenditureis undertakenfor subsistence. Of the aggregatesample, 81 percent of respondentsindicatedthat more than half of availableincome in July was spent upon food alone; 33 percent indicatedthat over 80 percent of incomegoes for this use. In October, 92 percent of those sampledindicatedthat more than half of availableincome was spent upon food, and 64 percent indicatedthat over 80 percent of availableincome went to food purchase. Durablepurchaseswere only reported in 7 instancesin July and 10 instancesin October. Roughly one-thirdof householdsreported clothingpurchases in July, and only 22 percent did so in October. The heads of householdswere asked whethertheir householdshad depositedfunds in the last month in two types of accounts. The first was a commercialbank deposit in coupons (the national currency), whilethe second was a deposit in foreignexchange(typicallyUS dollar) holdings. In July, only 19 householdsindicatedmaking commercial-bankdepositsin coupon;these were on average for 30 percent of total monthly income. 22 householdsindicatedsome depositinto hardcurrencyholdings during the previous month, with an averagevalue of nearly 40 percent of total monthlyincome amongthose depositing. These activitieswere primarilyundertakenby those with incomeprimarily from the private sector: only 10 percent of public-sectorhouseholdsindicateduse of hard-currencydeposits, while the correspondingrate for private-sectorhouseholdswas 32 5 By October, 7 householdsindicatedmaking coupondeposits, for an averageof around 10 percent." percent of total monthly income. The heads of householdswere asked to indicatewhetheranyonein their householdhad receivedfunds (through income, borrowingor exchange)in foreigncurrencies. These currencies includeRussianrubles and US dollars. Of the total, 63 percent indicatedyes and 37 percent indicated no in each sample. Those with substantialforeign-currencyincome were predominantlyfrom the private sectors. Of those indicatingin that a majorityof household incomein Octobercame from the govermmentsector, 92 percent indicatedreceivingless than half of total income in foreigncurrency. The correspondingpercentagesfor "private sector" and "other", respectively,were 22 percent and 29

Those utilizingthe coupon deposits were predominantlyfrom the "other-income"sector, perhaps becausetheir rental activitywas conductedprimarily in coupons. 15

Sanng in Tranaidon Economiewr- 15

percent."6 Of total households,over 49 percent (for July) and 56 percent (in October)indicatedthat they relied upon these foreign currencyreceipts for more than half of total incometo the household. The uses of foreign currencyran the gamut of possible activities. Over 60 percent of those receivingforeign currency in July and over 80 percent in Octoberindicatedthat they used it for purchaseof goods and services." Another 14 percent (in July) and 10 percent (in October)indicated holding it at home as saving, while 5 percent (in July, but zero in October)placed some in bank deposits. One respondentin Augustand 2 in Novemberindicatedthat a portion of foreign currency had been placed with a trust company-- despitethe wholesalefailure of those institutions during the preceding months. Other uses includedthe paymentof debts, lending to individualsand gifts to friends and family. Results of the Kazakhstan survey. The dominant source of income in the sampleswas the formal-sectorsalary, with 48 percent of the July householdsand 50 percent of the Octoberhouseholdsindicatingit as the sole source. Another one-third in each sampleindicatedthat more than half of total income came from that source. Only 11 percent of householdsin July and 6 percent in October indicatedthat no income came from formal salaries. Other sourcesof income were importantto much smaller groups of respondents. In July, only 8 percent indicatedthat all or more than half of income came from government support (pensions, stipends), and 5 percent indicatedsale of agriculturalproducts as a source of more than half of income. Private businesswas responsiblefor a majorityof income in 13 percent of the households, and interest from depositsor rentals was the majoritysource of income in only one household. In October, the percentagesfor non-formalincomesources were the same, except that no one indicatedsale of agriculturalproduceas a source of a majorityof income and 2 households indicatedinterest or rental income as responsiblefor more than half of total income. Householdsalso obtained funds by selling possessions. Nearly20 percent of the households surveyed in each sample indicatedselling possessions. Of those, 42 percent sold possessionsworth less than $22 in July, while another 21 percent sold items valuedbetween $22 and $44. In October, 26 percent of those selling possessionssold goods valuedat $20 or less, while an additional22 percent sold items valued at $20-40. In each sample, the remaindersold possessionsof even higher value. In each sample these sales were disproportionatelydone by those in the non-formalsector, with 28 or 29 percent of those in the non-formalsector doing so and 16 percent from the formal sector. Borrowingand dissaving by householdswas a commonmethod for obtaining funds. Over 50 percent of the sample indicatedthat they had borrowedor dissavedduring July, with the value of this dissavingexceeding$40 during the month for one-thirdof these individuals. Others respondingfor July reported receivingback loans previouslymade to others; 22 percent of the sampled households indicatedreceivingsuch repaymentof principal. In the Novembersample, 39 percent of those 16 Even this low percentagemay be too high. It is illegal, though rarely enforced, to accept payment in foreign currencieswithout a licenseto do so. Those indicatingno may be misrepresenting their position for fear of legal reprisal.

" The percentagesquoted in this sectionwill not sum to 100 percent, for some respondents indicatedtwo or more uses for foreigncurrency.

SaWng in Transition Economidf -16

surveyed indicatedborrowingor dissaving in October, with 29 percent of these householdindicating dissavingexceeding$40 during that month. In October, 14 percent of the populationreceived principalback from prior loans. The responseson expenditurequestions in Kazakhstanindicatedless stringencyof the budgetary situationwhen comparedwith Georgia. The households' share of funds devotedto food purchases, for example,were high but not extreme: only 7 percent of the population in both July and Octoberindicatedspending more than 80 percent of funds on food, with 50 percent (in July) and 33 percent (in October)reporting expenditureon food of 40-80 percent of income. Clothing and communalservices (utilities)were significantexpendituresfor the majorityof households;in July and October, for example,32 and 26 percent, respectively,indicatedexpendituresof more than 20 percent of resources on clothing. Deposits in tenge accountsat commercialbanks were reportedby only 10householdsin July, and for relativelysmall sums. In Octoberthere were no reported deposits amongthe 500 respondents. Holdingsof foreign currencywere indicatedby 22 households in July and 31 householdsin October. Durable goods purchaseswere reported by 22 and 35 householdsin the two samples. The responseson uses of funds were for the most part statisticallyindistinguishableby source of majorityof income. Those categorizedabove as "formalsector" incomeearned spent a significantlylarger (at the 95 percent level of confidence)share of resources on food, and a significantlysmall share on private businessand trading activity. Of the respondentsfor July, 18 percent (89 households)indicatedreceiving incomeor transfers denominatedin foreign currency. This number was drawn disproportionatelyfrom those householdsin the non-formalsector: 38 percent of non-formalhouseholdsreported foreign-currency income, while only 12 percent of formal-sectorhouseholdsdid so. The majorityof those receiving incomereported receivingless than half their funds in foreign currency, while only 4 households indicatedthat all funds receivedwere denominatedin foreigncurrency. Five percent of households reported using foreigncurrency frequentlyor exclusivelyfor food purchases,while 5 percent also indicateduse of foreign currencyfrequentlyor exclusivelyfor saving. Only 2 householdsreported using foreign currencyfrequentlyor exclusivelyin bank deposits. Three householdsreportedusing foreigncurrency in trust companydeposits. In October, 15 percent of householdsindicatedreceiving incomeor transfers denominatedin foreigncurrency. Thirty-threepercent of non-formalhouseholdsreported foreign-currencyincome, whileonly 9 percent of formal-sectorhouseholdsdid so. The majorityof those receivingincome once again reported receivingless than half their funds in foreign currency,while only 4 households indicatedthat all funds receivedwere denominatedin foreign currency. Four percent of households reportedusing foreign currencyfrequentlyor exclusivelyfor food purchases, while 5 percent also indicateduse of foreign currencyfrequentlyor exclusivelyfor saving. Only 3 householdsreported using foreign currencyfrequentlyor exclusivelyin bank deposits, while there were no reported dealingswith trust companies.

Saving h Transtion &onomiei - 17

The use of foreign currency is centeredin Almaty. The percentageof respondentsin each city indicatinga use of foreign currency and a relianceon foreign currencyfor denominatingincome is:

July Use Almaty 37 Aktyubinsk 13 Chimkent 23 Ust Kamenogorsk 9 Kokshetau/Kustanai 24 Zhezkazgan II

Receiptof over half of income 10 4 5 4 9 0

Use 28 10 27 5 14 18

October Receiptof over half of income 11 3 7 0 4 3

The dependenceupon foreign currencyas a store of value also differs greatly across regions. In Chimkent and Kustanaiover 20 percent of respondentsin Octoberindicatedholdings of foreign currency as saving instruments,while in the other cities the comparablepercentageswere below 5 percent. Conclusions on household saving in Georgia and Kazakhstan. The survey results confirm certaincharacteristicsof transactionsin these countriesthat are evident to all observers. First, currencysubstitutionis advancedin both economies(and throughout the former Soviet Union), with US dollars the currencyof choice for saving and durable purchases and Russian rubles popular in Georgiafor non-durablepurchases. Second, expendituresare concentratedon necessities, especiallyfood, in both countries. In Georgiathis has reached an extremeconcentration. Third, the incomedistribution in both countriesis unequal, with evidence both of greater inequalityand greater absolutepoverty in Georgia. There are other featuresof the two economiesthat observershave suspectedbut have been unable to affirm. The true usefulnessof this survey has been in the quantificationof these. The sources of income for householdsin the two economieshave diverged sharply since independence. At that time, nearly all householdsreceiveda majorityof income from formal-sector (mostlygovernment)sources. In Kazakhstanthis share of householdsremainshigh at 80 percent, while in Georgia the transition has led to a minority of households(29 percent in October 1994)with predominantlygovernment incomes. Private businessactivitieswere predominantin 39 percent of Georgianhouseholdsbut only 12 percent of Kazakhstanhouseholdsin October. * The sale of possessionshas become a significantsource of incomefor householdsin both countries. In July, 25 percent of Georgianhouseholdsand 20 percent of Kazakhstanhouseholds indicatedsuch sales. In both countries, the percentageof householdsinvolvedin such activities fell from July to October, althoughmore strikingly in Georgia. There is an interestingdivergencein these activities as well -- in Georgiasuch sales were more commonamongthose dependentupon governmentsalaries, while in Kazakhstanthey were morecommon amongthe familiesdependent upon non-formalincome. ' Borrowing and dissavingare also endemicamongthe householdsin both countries. Of

- 18 Savingin TransitionEcononmies

Georgianrespondents34 percent indicatedthat they had undertakenthese activitiesin July, while in Kazakhstanover 50 percent of respondentshad done so. The amountsborrowed or dissavedwere also larger in Kazakhstan. Once again, there was a declinein this percentagefrom July to October in both countries, althoughin Kazakhstanthe percentageremainedrelatively high at 39 percent of households. Expenditureson food take up a large share of householdpurchasingpower, but the Georgiancase is quite extreme. In Georgiaand Kazakhstan33 and 3 percent of households, respectively,reported spendingover 80 percent of incomeon food in July. Those households reporting expendituresof over 60 percent of income on food were 63 percent of the Georgiansample and 17 percent of the Kazakhstansample. In Octoberthe Georgiansituation had worsened: 44 percent reported expendituresin excessof 80 percent of income on food, and 74 percent expenditures in excess of 60 percent. In Kazakhstan,by contrast, the percentageof householdswith dominant food expendituresshrank: 3 percent in the over-80 percent category and 11 percent in the over-60 percent category. * Saving activitiesstill continue in both countries, but they rarely take advantageof the formal bankingsystem. National-currencyaccountswere reported by a handfulof householdsin each country. Foreign-currencydepositswere equally likely in the two countriesin July with 4 percent of householdsin Kazakhstan. By Octobersuch deposits were more likely in Kazakhstan(6 percent of households)but not observed in the Georgiansample. Holding of foreign currency at home was reported by 10 percent of both Georgianand Kazakhstanhouseholdsin July; by Octoberthe percentagein Kazakhstanremainedconstantwhile that in Georgia fell to 7 percent of households. Despitethe prominenceof "trust companies"in recent news in both countries, only a handfulof respondentsin each country indicatedparticipationin these financialintermediaries. Theseresults contrastwith the evidencefrom Family BudgetSurveys in Belarus: Conway (1995). These surveys indicatethat 30 percent of the population continuedin July 1994to rely upon the formal-sectorfinancialinstitutions(primarilythe Saving Bank) for intermediationand deposit services. Saving and the distribution of income in Ukraine. I was unableto organize a survey of householdactivity for Ukraine, and thus have not comparablesurvey results on saving or the use of saving instruments. I do have some evidenceon saving activityfrom the householdbudget survey conductedin Ukraine for the fourth quarter of 1993; these results are reported in Table 1. Saving as a percentageof income was apparentlywell below its performancebefore independence. For the period 1985-1990,nationalsaving in the Soviet Union averaged25 percent of national income, while in the last quarter of 1993the Ukrainianhousehold saving rate was only 5 percent of householdincome. As Table 1 makes clear, this small positive rate masks a wide dispersion in the degree of savingor dissaving amonghouseholds. Those at the low end of the incomedistributionwere dissavingsubstantially,either through liquidationof bank balances(although these had depreciatedgreatly by this time) or through sale of possessions. Those at the upper end were saving on average about 15 percent of householdincome -- a substantial amount,but below the averagefor the entire economyof 5 years before.

SaWngin Transidon Economiei.- 19

m. The collapse of formal-sector deposit of saving: evidencefrom four countries. In the last years of the Soviet Union, the share of nationalproduct representedby saving depositswas steadily rising. This phenomenonbecameknown as the ruble overhang,and represented the forced saving of individualsunable to find goods to purchaseat the controlledprices of the economyat that time. As Figure 5 indicates,the share of currency in circulationto GNP rose from under 10 percent in 1985 to nearly 15 percent in 1990, while the share of householddeposits in the bankingsystem rose from 28 percent in 1985to nearly 40 percent in 1990. This suppressed purchasingpower would fuel the burst of inflationobservedwith independencein the newly independentstates. Availableindicators suggestthat Belarusand Ukrainereactedsimilarly to the Soviet Union as a whole during this period of overhang. Despite the suppressionof purchasingpower, there was no apparentbias away from deposits with the formal banking system. The relativereturns on cash and deposits remainednearly constant in the period through 1990, and this was reflected in the proportionalallocationof the ruble overhang. During the build-upfrom 1985to 1990, the ratio of householddepositsto cash in circulationin the SovietUnion remainednearly constantat 3.2 to unity. The deposit-inflation nexus. With the onset of inflation in 1991, the relative attractivenessof bankingdepositsdenominated in domestic currencydeclinedsharply. Those holding wealth in nominaldomesticassets (including governmententitlementprograms) were made poorer by inflationrelativeto those with the ability to diversify assets into real property and hard-currencyinstruments. This was evident in the evolution of real currencyand depositholdings for the period 1989 - 1993in Belarus illustratedin Figure 6. The data points representthe currency and deposits as of the end of the stated year definedin terms of 1989 rubles."' Inflation eroded the real value of nominalassets, as indicatedby the downward sloping lines. In the year 1992, for example,the real value of the stock of deposits of the population fell from 13.3 billion to 2.2 billion rubles. In each period there was an additionto the real value of the stock of depositsthrough the inflow of saving resources to those instruments: this is indicatedin the figure by the vertical arrows. Althoughthe magnitudesfor 1993 appearsmall in 1989 rubles, the combinedreal increase in holdingsof these two instrumentsrepresentedsaving of over 5 percent of GDP in Belarus. Only in 1990did the increase in holdingscompensatefor the loss in value due to inflation; the real stocks of these instrumentswere falling thereafter. The lesson to private savers of 1992and 1993 in Figure 6 must be not to do it in domesticcurrency-denominatedassets: the value of holdings at the beginningof each year dwindledto almostzero by the end of the year. There was also a shift in the nominal instrumentsheld by the private sector. The share of currency emissionin the inducementof private saving had been growing over the years -- nearly zero in 1990,but playingthe dominantrole in 1993. As noted above, this indicatedthe growingreliance of the governmentupon the "inflationtax" to effect the internal transfer. The loss in private purchasingpower is striking. Conversionof the loss in real purchasing power impliedby the inflation-induceddevaluationof currencyand depositholdings of the population "' The GDP deflator is used to calculatethese: inflationrates for the years 1990through 1993 were 5, 102, 1099 and 1425 percent per annum, respectively.

Savingin TransitionEconomis - 20

into 1993 rubles yields a total devaluationof the stock of nominalwealth equal to 53 percent of GDP in 1993. Inflation was also a problem in Ukraine, as was the flight from the karbovanets-denominated instrumentsof the financialsector. This is illustratedin Figure 7. Bankingdeposits are separated into four componentsin the two panels; the first panel indicatesthe ratio of sight deposits (both by enterprises and by households)to cash in circulation, while the secondpanel reports the ratios of time deposits to cash in circulation. These includeonly deposits in karbovanets(or, before that, in rubles), and does not includethe use of foreign exchangedeposits. A similar pattern is evident in each panel. Householddepositholdings in karbovantsybegan at large multiplesof the quantity of cash in circulationbut rapidly dwindledto nearly zero. From the beginningof 1993it is clear that householdshave ceasedto rely upon the banking sector as a store of wealth. Enterprise depositholdings have followedan inverted-Upattern in both cases. Sight deposits and cash are the traditionalmethod for enterprisesto hold balancesfor day-to-dayexpenditures,and reliance on sight deposits grew with the governmentuse of directedcredits through the banking system to finance enterpriseoperations. Time deposits also becamerelatively attractiveduring the early months of 1993; given the low interest rates on directed credits, maintenanceof time deposits may well have been more attractivethan production. Negative real interest rates on deposits were one causeof the households' flight from karbovanetsbankingdeposits. As Table 2 indicates,ex post monthly real interest rates were negative throughout 1993. Not surprisingly,householdssoughtto reallocatesaving away from the banking sector. However, no matterhow negativethe depositrate, the refinance credits were usuallymore negative-- enterpriseswere thus shieldedfrom these losses as they held borrowed funds on account. There was a movementto positivereal interest rates on deposits in early 1994. This should have causedan uptick in the use of depositsby households. Such an uptick is there, though small, in the holdings of time deposits relativeto cash in the first four months of 1994; see Figure 7. The attractionof foreign exchange in an inflationary environment. The holdingof foreign exchangeprovides a hedge againstthe inflationtax on accumulated wealth. This alternativehas become quite prevalentin the former Soviet economies,as is evident in the very liquid informal markets in foreign currencies. One hedge is of course to hold the banknotes of the foreign currency, while another is to make foreign-exchangedenominateddeposits with a financialintermediary. This latter strategyhas proved quite lucrativefor private savers. Commercial banks have been quite activein offeringthese deposits, as noted in the followingsection of this report, as have "trust companies". The surveysreported above providethe only availableconsistent evidenceon holdings of foreignbanknotesor use of non-bankfinancialintermediaries,but commercial-bankdeposits in foreignexchangeprovidea lower bound on the importanceof these deposits in attractingwealth. In the Ukraine, foreign currencydepositsgrew rapidly relativeto cash in circulation. These deposits were not offered in Ukraineprior to independence,but they becameattractive in the environmentof extreme inflationin 1992 and 1993. The value of these peaked in July 1993and was

Saving in Transiion &onomies-- 21

decliningthereafter.'9 There was notably a decline in foreign-exchangedeposits in early 1994 that was a mirror image of the upturn in householdholdings of karbovantsytime deposits. Foreign-exchangedeposits also becamemore popular in Georgia. The monetarysurvey for Georgiais presentedin Tables 3a and 3b for the period since December1991. As the memorandum item indicates,domesticcredit has grown strongly during this period, with averagemonthlygrowth rate of 38.9 percent over the period January 1993- August 1994. This has not been uniform growth; domesticcredit creationhas peaked in the spring and the fall as planting- and harvest-related activitieshave been financed.' A relatively small fraction of the domestic credit creationwas allocatedto financingof general government(i.e., the governmentbudget);the overwhelming majoritywas provided to state enterprisesto finance their activities. However, the nominal interest rate on these credits was in general so low relative to inflationthat repaymentdid not replenishthe credit fund; the cost was borne through extreme inflation. Broad moneyrepresentsthe liabilitiesof the financialsector. These liabilitieshave evolved over time to include an increasingshare of foreign-currencydeposits. Bankingdeposits denominatedin couponshave declinedas a share of domestic credit over this period, while currency outsidebanks has followeda cycle from rather low percentagesof domesticcredit in early 1993 (the period of currency shortages)through elevatedpercentagesin late 1993, to decliningshares in 1994 concludingat a quite low percentagethat reflects the little real use for the coupon in the current economy. Bankingdeposits in foreign currency, by contrast, rose from zero at independenceto a peak of 51 percent of domestic credit in January 1994. From that time until August 1994the share of foreign-exchangedeposits in domestic credit declined,reflectingin part the movementof such funds to the trust companies. Foreign exchangehere refers to non-menaticurrencies,althoughRussianrubles and US dollars were used for the majorityof transactionsin Georgia. There were liquid markets for foreign exchangein rubles or dollars -- the bazaar and exchangeshops provide exchangefor individual transactions,while the Tbilisi InterbankExchangeprovidedauctionsfor the larger transactionsof commercialbanks for their clients (and on their own account). Financial-sectorresponse. The banking system in these countries was slow to adjust to the inflationaryenvironment,but private banks led the way in introducinginnovationsto attract deposits. One such innovationwas the foreign-exchangedeposit. Another was the adjustmentof nominal interest rates and maturitieson domestic currency-denominatedinstrumentsto reflect the more inflationaryand uncertain environment. The state commercialbanks, and in particular the successorsto the SovietSaving '9 This decline after July 1993may be due to the exchangerate used to convert the foreigncurrency values into karbovantsy. There was a substantialovervaluationof the official exchangerate during the last half of 1993 and early 1994, and this overvaluationmay have placed a downwardbias on the share of foreign-currencydeposits in the monetary survey. ' The indicationof 88 percent growth in domestic credit in August 1994 is misleading,as this reflects the creation of temporary credit facilities to clear arrears from state enterprisesto state commercialbanks. This operation allowedthe state commercialbanks to retire overdraftswith the NBG, and left little net effect in credit creation.

Saving in Transidon Economnic - 22

Bank, were much slower in adjustingto these new financialconditions;as a result, the share of saving depositedin those traditionalrepositoriesfell sharply during this period. The situation in Ukrainereflects this evolution. By May 1994the bankingsystem in Ukraine had adjusted to the preferencesof its depositorsand was offeringa menu of deposits variegatedby maturity and return. As Figure 8 indicates,the depositors (and the banks) in Ukraine favored shorter maturitiesfor deposits as a meansfor hedgingrisks.2 ' The annualizedinterest rates on household deposits exhibiteda positivelysloped yield curve, with interest rates well in excess of the annualized monthly inflationrate observedat that time. The enterprise yield curve turned down in the maturitiespast 6 months, indicatingthat those deposits may not be as responsiveto real interest rates as they are to other factors (e.g., the contingencyof future lendingon present deposits). This evolutionhas importantconsequencesfor governmentdeficits and inflation. The instrumentsrepresents an everholdings of foreign exchangeor foreign exchange-denominated increasing share of private saving that the governmentcannotat presenttap. Commercialbanks with foreign-exchangedeposits will not willingly intermediatefor the government'sborrowing needs in domestic currency at present low nominal interest rates -- this representsa subsidyto the government that the commercialbanks cannot afford. The governmentis forced in this instance either to issue its debt at positivereal interest rates or to raise its return on own foreign exchange-denominated domesticcurrency-denominateddebt to positivereal levels. In either case, the real cost of the borrowing rises sharply. Kazakhstanis a good exampleof this process. It was characterizedin 1994 by a small volume of private saving. Private investorsand commercialbanks attractthe funds of those who did save, and this saving was made availableto private investors and traders. The governmentbudget was not competitivein its efforts to attract saving, and relied upon currencyemissionfor financing. The government is able to "attract"saving in this case only through exploitationof the private sector's need for currency; it has as cost the reduction in the real value of original nominalfinancial assets known as the inflationtax. The bankingsystem was slow to attract saving from the population or enterprises in domesticcurrency, as it was able to rely upon National Bank of Kazakhstan(NBK) refinancingcredits. In the end, this circuit of financingonly enriched the bankers without attracting any real resources -- the NBK was in the end been dependentupon inflationto achieveits financial goals. The change in the pattern of saving also has implicationsfor resource allocation. Saving in foreign exchangewould be beneficialif intermediariesexistedto ensure that the funds were allocated to high-return investmentactivities. In fact, however, the intermediariesin foreign exchangefocus upon the financingof trade expeditionsto bring back consumergoodsto the countries. These goods are imperfectsubstitutesfor those availablefor domesticcurrency. Consumptionthus rises, inflation hits the domestic economyand the domesticcurrencydepreciatesagainst foreign exchange. The populationalso made "deposits"with non-bankingintermediariessuch as Smagulovand Company; these investmentswere denominatedin tenge for one-monthmaturity and were yielding 600 percent annual interest (50 percent per month). This saving was however financingthe purchasesof the The bars in this figure indicatethe shares of total deposits(householdplus enterprise) that fall into each category. The total shares sum to 500 so that the scalesof the two sets of informationcan be placed in the figure. 21

Saving In Transition EconomIes- 23

company. This financial-marketexperiencein Kazakhstanprovides two lessonsto the government's economicstrategists. First, the real cost of financingthe ongoing budget deficit was quite high. The credit auctionscoordinatedby the NBK and the interest rates offeredon dollar deposits at the commercialbanks indicatedthat saving was quite scarce. This scarce saving was being attracted by private actors -- commercialbanks and "trust companies"like Smagulovand Company-- through offeringhigher real interest rates than were available in the Narodnyi Bank or other state commercial banks. Second, there is an opportunityfor the governmentto attract this saving if it is willing to raise the interest rate that it pays on borrowed funds to competewith these private actors. IV. Other financialdepositories for private saving. As noted above,saving in the SovietUnion has historicallytaken place through deposit in long-termfinancialinstrumentsin the formal bankingsector, primarily in the SavingBank (or Sberbank). These depositsbecameunattractiveas stores of wealth during the extremeinflationafter independence,and savers were attractedto alternativeinstruments. Some of these alternative instrumentswere providedby private commercialbanks, while others were made availableby nonbank financial institutions. Although there has been a general tendencywithin the former Soviet economiestoward disintermediation,certain individualbanks have been quite successful at operatingand attracting deposits in this environment. I have surveyed commercialbanks in Belarus, Georgia, Kazakhstanand Ukraine. The successfulbankshave offeredhigh nominal interest rates, often on hard-currency deposits, and have introducedother innovationsdesignedto enticedepositors and establishtheir confidence. Their major competitionhas come from non-bank financialinstitutionsalso promising high returns on depositsor from individualhoarding of foreign exchangebanknotes. The successor banks to the Soviet Saving Bank have as a rule been quite slow to innovatein the post-independence inflationaryenvironment;as a result, their historical monopolyin the provision of depositshas been diminishedin manyof these countries. In this sectionof the report I summarizeconclusionson other depositoriesfor private saving. Greater detail on individualintermediariesand the financialsystems of individualcountriesis provided in Annex C. Althoughthere has been substantialvariation in the developmentof financialsectors in transition economies,there are two stylized facts that characterizethis evolution. First, the financial sectorhas split into two markets with one denominatedin the domestic currencyand the other denominatedin foreign exchange. Second, there has been a profusion of commercialbanks chartered in the transition economies,but very few of these offer opportunitiesfor saving. Third, these banks have chosen one or the other market to work in -- most specializein either domestic-currencyor foreign-exchangefinancialinstruments. Fourth, the legal frameworkhas been conduciveas well to the growth of non-bankfinancialintermediariesoffering financialtransactions, oftenpromisinghigher returns than the charteredinstitutions. Financialmarket bifurcation. A bifurcationof financialmarkets emerged in these countries soon after independence. The traditional financialmarket used the ruble, or the successorcurrencyof the state. The dominant

Saving in Transidon Economice - 24

financialintermediarieswere the successorsto the state banks -- Promstroibank,Agroprombank, Kredsotsbank,Vnesheconombankand Sberbank. The governmentor state enterpriseswere net users of funds in this market, whilethe householdswere on net the source of funds. An alternative financialmarket grew up in hard currency. The private commercialbanks and trust companieswere the intermediariesin this market, and the sourcesand uses of funds were largely householdsand private enterprises. The real interest rate on deposits divergedsharply in the two markets; in the traditional market the real interest rate was strongly negative,while in the alternativemarket the real interest rate was strongly positive. The credits in local currency were not allocatedby an economiccompetition,but were directedby the governmentto enterprises, collectivesor individualsas it suited the advancementof governmentpolicy. These credits were in the pre-independencedays offeredat positive real interest rate, but since independenceand the onset of inflationthe nominal interest rate on these loans has lagged significantlybehindthe inflationrate. The government'scredit policy has thus had a substantialsubsidy component,favoringthose able to receivethese credits. The maturityon these loans has varied, but in many cases (e.g., the loans of Agriprombankfor the planting season)these are for 6 months to one year. These directed credits were not offeredby the centralbank, but by the state commercialbanks with whom the targeted enterpriseshad accounts. These state-ownedbanks were then provided refinancecredit by the centralbank at a slightly lower nominal interest rate. Althoughthe central bank in principleoffset these credits with a claim againstthe resourcesof the Ministryof Finance, in practicethe government resources were insufficientto cover the volumeof refinance credit issued. There were similarlyonly small amountsof private saving depositedwith the state commercialbanks (e.g., Sberbank). The public-sectorborrowing requirementwas thus financedon the whole through the "inflationtax" described earlier. Competitiondid not drive up the real interest rate and thus encouragegreater voluntarysaving becausethose enterprisesreceivingcredits were not profitableon the whole -- the governmentsupported them at least in part throughthe credit subsidy inherent in the low nominal interest rate. These enterpriseshad no incentiveto bid up the cost of funds. Most credits extendedwithout central-bankrefinancingare now denominatedin foreign currencies,mainly US dollars or Russianrubles. These loans are extendedalmost always as working capital for trade transactions.' These are short-termloans with maturitiesusually in the 1- to 3month range. The interest rates chargedon these loans are quite high by internationalstandards, with the most conservativebanks charging 15-25percent annual rate of interest in US dollars, and with more aggressivebanks charging that amountof interestper month. There is in additiona requirement of collateralin preciousmetals or real estate on these loans, with the value of the collateralbetween ' There is a paradoxinvolved in the rapid growth of these foreign-exchangefinancialmarkets. They rely upon the existencewithinthe economyof a stock of foreign currencythat can be borrowed and lent, but this stock of foreign currency can only be accumulatedthrough a successionof current account surplusesor capital outflows. The officialstatistics for these countriessince independence indicatea series of current account deficits and capital inflowsgreater than outflows; if true, the stock of foreigncurrency in circulationshould be reducedover time rather than increased. Transactionsnot captured in officialstatistics are seeminglycharacterizedby large trade surpluses (as for examplein internationalsales of resourcestocks) -- or the countriesare home to an abilityto counterfeitthe foreign currencies.

Saving in Transtion Economies. - 25

150-300percent of the loan. Not surprisingly, giventhe high monthly interest rate, the more 3 aggressivebanks report a large share of loans delinquentat maturity.2 These foreign-currencycredits are offeredto both enterprisesand individuals. The prototypicalcredit will finance an internationaltrading transaction. Goods from abroad are purchased in foreign markets -- Turkey, India, Germany -- and are imported into Georgia. They are then sold in the bazaar at a substantialmark-up, with the proceedsbeing used to retire the loan. There is clearly substantialrisk involvedin these transactions,given the fickle taste and small effectivedemand of the consumers,but enough successstories have surfacedto encouragea populationdesperate to retain past living standardsto pledgejewelry and apartmentsagainst such a loan. The result of this intermediationis the use of saving for the financingof consumption(usually,luxury consumption) expenditures. The high cost of financingworking capitalfor a productive transactionis evident,'and leaders of the bankingsector indicatethat very few loans are provided for productive or investment purposes. The high positive real interest rate on US dollar credits appearsto be the product of a number of factors. First is the probabilityof non-repayment(as noted above). Second, the variable costs of banking activityare quite high. Third, as theory demonstrates,an economywith excess demandfor present consumption(as above)withoutaccess to internationalcredits will drive up the real interest rate to induce those with loanablefunds to part with them. Fourth, and most fundamentally,there is little saving at present. Althoughsuch saving is held predominantlyin foreign exchangeor in real assets, this flow of saving is small relativeto the demand by traders. This excess demandfor loanablefunds causes the high real interest rates in the hard-currencyfinancial market; these more accuratelyreflectthe premium put upon consumptiontoday by the residentsof these countries. The bifurcationin credit markets extendsto the variety of deposits observed in the banking system. There is little depositactivity in domesticcurrency by individuals;what depositactivity there is representsenterprisesholdingthe balanceof directedcredits until its disbursement. Evidencefrom Georgia's monetary survey indicatesthat bankingdeposits in couponswere roughly stable at 20 percent of domesticcredit in the months January-August1994, while bankingdeposits in foreign currenciesaveraged in excess of 40 percent of domesticcredit. This ascendancyof foreign-exchange deposits will be accentuatedin later periods, as notedbelow. The large differentialin interest rates betweenthe domestic-currencyand foreign-exchange financialmarketsprovides a lucrativespeculativeopportunity. Those actors with an abilityto borrow in domestic currencyand to lend in foreign currencyhave gained large windfall profits. The interest rates are not bid together by speculationbecausethe supply of domestic-currencycredits at low 24 However, the continuingsupply of interest rate is limited to the directedcredits of the government. 23 For example, Iberiabankreported 30 percent delinquencyat maturity -- although they were rolling over the loans and anticipatedbeing paid in full (with interest)eventually.

4 Althoughthere are no systematicdata, there were numerousreports that recipientsof directed credits found it profitablesimplyto convertthe principalto foreign exchangeand to hold that currency to the term of the loan. Then a small fraction of the principalwas re-convertedto domestic currencyto provideboth principalrepaymentand interest. In this case, the directed credit clearly did not go as directed. This use of directedcredits, however, provideda persistent demandfor foreign

Sgirng in Trasiton Ec&onaml-26

these credits provided constantdepreciatingpressure on the country's exchangerate. The proliferationof commercialbanks and trustcompanies. There has been tremendousgrowth in the number of commercialbanks in operation in the former Soviet economiessince independence. The growth began in 1987 with the restructuringof the operationsof Gosbankand the charteringof non-governmentalcooperativebanks. However, the numberhas increasedexponentiallysince independencein these countries. From less than 10 banks in each country at independencethe number rose by mid-1994to over 200 in Kazakhstan,230 in Georgia,and similar explosionsin Belarusand Ukraine. This growth did not in general indicateincreasedopportunityfor or competitionto attract depositors. Commercialbanks found "origination"of directed credits refinancedby the nationalbank to be quite profitable. Attractionof deposits was more expensive,giventhe high costs of establishing branch offices, and required more detailed record-keeping. Finally, that activity involvedsome risk to the bank if assets and liabilitieswere not matchedas to maturity and interest rate. This risk did not occur in the originationmarket. Finally, the deposits denominatedin the nationalcurrencywere an inherentlypoor saving instrumentduring a periodof extreme inflation,but the banks were not in a strong enough financialpositionto offer inflation-hedgingimprovementsto these instruments(e.g., indexedinterest rates). Commercialbanks were charteredin great number. A large subset of these also qualifiedfor "generallicenses"that permitted the bank to denominateits depositand lending activity either in nationalcurrencyor foreign exchange. The competitionfor depositors in fact surfaced in the foreignexchangetransactions, since the banks had an excessdemand for foreign-exchangecredits. Interest rates in these markets were bid up to extremelyhigh levels. This reflected in part the profitabilityof internationaltrading transactions, but also in part the emergenceof Ponzi schemesand ill-advised financialtransactionsin the underregulatedmarkets of these economies. The "trustcompany"phenomenon. Commercialbanks were not the only financial institutionsto thrive in the volatile financial environmentafter independence;non-bankfinancial institutionsalso proliferated. Thesetook a numberof forms, from foreign exchangebureaus to Lombard companies(or pawnshops),but the most famous (or infamous)were the "trust companies". These were non-bankinstitutionsthat offered depositsor shares with interest rates or capitalgains promisedto exceedreturns availablein the bankingsystem. The MMM corporationof Moscowwas the best-knownof these, but similar institutionsgrew up in all the countriesobserved. These trust companiesoperated in a standard manner. Each sold monthlycertificates. The firm set a buying and sellingprice each month, and these prices rose from month to month to guaranteea positive return to investors. For example,certificatesin the Golden Bowl Corporationof Georgiahad risen in value from $.98-$l in June 1994 to $2.30-$2.50in October 1994 for a compoundedmonthlyreturn of about 25 percent. These trust companieswere not charteredby the governmentand were not typicallyregulated by the nationalbank or any other government organization. exchangethat causedthe nominal exchangerates at the bazaar in these countriesto depreciatemore rapidly than the inflationdifferential, thus causingsubstantialundervaluationof the currency.

Sawing in Transidon Economies - 27

The first wave of trust companiesattractedclients with the promise of -- and initial success in returning - 20 percent return per monthon deposits. Some of these may have been pure "Ponzi" schemes, but others were financialintermediarieswith ties to corporationsable to turn high short-term profits through trading operations. These organizationswere able to offer such high returns because of the profits from trading activitiesand becausedeposits were loaned out completely- negligible reserves were held for repurchaseof depositcertificates. The success of these corporationsattracted competitors,both from newly createdtrust companiesand from more aggressivecommercialbanks. The cash flow from depositreceipts and repaymentsof lending was sufficientto cover withdrawals until the failure of the Russianfirm MMM. During the time of the MMM crisis in Moscow, depositors in other countriesapproachedtheir trust companiesto redeem their certificates. The trust companywas typicallyunableto meet these demands,and was forced to close. In Georgia there were 15 to 20 such companiesduring their heyday in the first half of 1994, but few remainedafter the MMM crisis. These failures were liquidity-driven. Subsequentfailures were relatedto insolvency,and occurred in one of two ways. The first was through theft: for exarnple,the Golden Bowl Corporationfailed when the chairmanfled the country (reportedlywith $60 million). Smagulovand Companyin Kazakhstansimilarlyclosed in late 1994after a year of operation as a "trust company"; Mr. Smagulovwas reported to be in Australia. The second was through ineptitude. Here, the case of InnovationBank in Georgiais instructive. This bank was backed by the Ministry of Internal Affairs and offered high interest rates on deposits. The leadersof the bank made a number of improper loans, includinga 400,000 USD loan to a woman friend, whilepaid-in capital was valued at less than a tenth of that. Most loans were made to the foundersof the Bank. Each borrower proposed a productiveuse for the loan, but these were not well thought out and not well screenedby the bank. The bank was warned about its behaviora number of times by the National Bank of Georgiabefore it finally failed. Much "trust company"activitywas fraudulent. There were also reports, however, of companiesresorting to formation of trust companiesbecauseof the shortageof available credit in the bankingsystem. This is comparableto, though less dramatictharn, the formationof a commercial bank by a trading company. In each case the shortageof credit from the banking system encouraged the formation of a "captive"financial intermediary. Dynamic growth of foreign-exchangedeposits with private commercial banks. The market for foreign-exchangedepositsand credits was a much freer and more dynamic environmentfor financialactivity. The fastest-growingbanks were private commercialbanks that specializedin this market. Althoughthese private banks had internationalcorrespondents,these relationshipsdid not extendto credit lines; as a result, the banks were constrainedin their ability to offer credits by the volume of deposits they could attract. The interest rates offered on these deposits were thus well in excess of the internationalrate. Foreign-exchangelicenseswere necessaryin each countrybefore banks couldparticipatein these transactions,and the qualificationsfor obtainingthese licenseswere fairly strict. As a result, private commercialbanks were separated into two groups. Those that had obtained the proper licenseswere able to grow quickly, while those without such licensesstagnated. Banksthat made the strategicdecisionto specializein domestic-currencytransactionswere typicallyhamstrungby the unattractivenessof nominal domestic-currencyinstrumentsin a periodof extremeinflation. In Kazakhstanand Georgia,respectively,Alataubankand Kerdsobankwere two

Savwng in Transition6conomime- 28

examples: private banks, but without the dynamismof those specializingin foreign-exchange transactions. This tendencyexisted, but was less pronounced,in Ukraine; there, Inkobankbecame the largest private bank through a concentrationon foreign-exchangeinstrumentswhile Agiobankwas able to expandits operationsthrough reliance upon innovativeand high-yieldinstrumentsdenominated in karbovanetsfor attractingprivate depositors. Specializationand sluggish adjustment of state commercial banks. With the bifurcationof the financialmarkets, the former state commercialbanks (otherthan the Saving Bank) found it profitableto act as the originator of directedcredits in domestic currency that the nationalbanks refinanced. The return to the bank was in effect indexedto inflation,since the "fee" receivedwas a fixed percent of the value of the directed credit. Giventhis secure and relatively stable source of funds, these banks did not find it profitableto attract depositsin domestic currency. This proved to be true across countriesfor the successor banks to Agroprombank(Agro-Industry Bank) and Promstroibank(IndustryBank). The sole exceptionproved to be during periods of cash shortage; in those circumstancesthese banks (for example, Promstroibankof Belarus)solicited deposits from householdsbecausethe deposits would be made in domesticbanknotesthat could be redistributedto other clients. The successorsto the Saving Bank in each country were rather slow to adjust to the new economicenvironment. The banks remainedstate banks for some time in all countries, and their managementthus remainedunder the direction of the state bureaucracy. As a result, the Saving Bank was not allowedin most instancesto capitalizeupon its existing monopolyin the collectionof domestic-currencydeposits. Interest rates paid on deposits were adjustedinfrequently,and by too little to be attractiveto depositors;the maturity and other characteristicsof deposits were not adjusted to the inflationarypost-independenceeconomicclimate. As a result, the Saving Bank became progressivelyless dominantas an intermediaryfor private saving, althoughits extensivenetworkof branchesmade it the sole provider of financial servicesfor much of the population. V. Conclusionsand suggested extensions. The major conclusionsof this researchproject can be summarizedas follows. Private saving has assuredlydeclinedin the period since independencefor the four countriesconsideredhere. It is necessaryto correct for the "forced saving" of the last years of the Sovietregime and to recognizethat there has been substantialsaving outsidethe formal financial system since independence. However, after those correctionsthe evidenceof the surveysand official statisticssuggeststhat private saving is down. * Private saving is greater than that observed as deposits in the bankingsystem. With independenceand the sluggishnessof the state commercialbanks in adjustinginterest rates to respond to the inflationaryepisodes, alternativefinancial instrumentsbecamemore attractivethan formal deposits. These includedholdingof foreign exchange,self-financeand shares in "trust companies". deposits to attract some of Private commercialbanks eventuallyoffered foreign exchange-denominated these funds, but other forms of saving still abound. I am unable to provide a specificestimateof that alternativesaving, but I do provide evidenceof its existencein survey results and reports from the countries examined. ' Private saving does respondto the incentivesof positive real interest rates. Here it is

Sawngin Transzion Fxonomie - 29

impossibleto disaggregatetwo effects: the pure saving effect and that of portfolioallocationof that saving. The evidenceprovided here suggeststhat portfolioshave been skewed strikinglytoward those instrumentswith positive real interest rates. It is impossiblewith present data to separate convincinglythe real interest rate effect from the "incomeeffect" that has led to lower saving overall, so that I have not yet identifiedthe contributionof real interest rates to the saving process. This is an interestingdirection for future research. The governmentsof these economieshave not yet tapped effectivelythe continuingsaving. Rather, they have relied upon the use of credit creationto finance expenditures;this is effectiveonly through stimulating extremeinflation. Proper use of existingprivate saving will require a reconsiderationof governmentpolicies to recognizethe true opportunitycost of borrowedfunds, and the offeringof either (a) debt instrumentsindexedto the exchangerate or (b) debt instrumentswith interest rates as high as those on deposits at reputablecommercialbanks. This step will no doubt be coincidentwith a decision to cut governmentexpenditures,since use of these debt instrumentsimplies a positive real cost to borrowing. It promises, however, a non-inflationarysource of deficit finance. This research report representsthe beginningof study of an importantquestion. As noted above, there are issues not resolvedin this report; these can only be addressed with confidenceby the amassingand consolidatingof additionaldata. The surveys of private saving behaviorreported here are an importantfirst step, and I hope that such informationwill be collectedin future survey efforts. In addition, more work on the transition in the financialsector will be fruitful. There is a thin line betweenthe dynamicbanks and the "Ponzi"banks, and regulationmust be sophisticatedto cull the latter while allowing the former to thrive. The informationof this report representsa first step in examiningthat question.

Saving in Transition &onomnie - 30

Bibliography Aslund, A.: "Key Dilemmasin the RussianEconomicReform", processed, 1983. Blanchard, 0. and S. Fischer: Lectureson Macroeconomics. Cainbridge, MA: MIT Press, 1989. Branson, W. and D. Henderson: "The Specificationand Influenceof Asset Markets", chapter 15 in Jones, R. and P. Kenen, eds.: Handbookof InternationalEconomics. Amsterdam: North-Holland,1985. Bruno, M., G. di Tella, R. Dornbuschand S. Fischer, eds.: Inflation Stabilization: The Experienceof Israel. Argentina.Brazil. Bolivia and Mexico. Cambridge, MA: MIT Press, 1988. Buffie, E.: "FinancialRepression,the New Structuralists,and StabilizationPolicy in SemiIndustrializedCountries",Journal of DevelopmentEconomics 13, 1983. Cagan, P.: "The MonetaryDynamicsof Hyperinflation",in M. Friedman, ed.: Studies in the OuantityTheory of Money. Chicago: Universityof Chicago Press, 1956. Campbell,J.: "Does Saving AnticipateDecliningLabor Income? An AlternativeTest of the PermanentIncomeHypothesis",Econometrica55, 1987, pp. 1249-1274. Conway, P.: "The Problemof Intermediationin a Country in Transition: The Exampleof the Sberbanks",processed, 1993. Conway, P.: "Rubles, Rubles, Everywhere...: Cash Shortagesand FinancialDisintermediationin the Former SovietUnion", processed, 1994. Conway, P.: "HouseholdIncomeand Saving in Transition Economies",processed, 1995. Conway,P. and A. Gelb: "Oil Windfalls in a ControlledEconomy: A 'Fix-Price' EquilibriumAnalysisof Algeria", Journal of DevelopmentEconomics28, 1988, pp. 63-81. Cottarelli, C. and M. Blejer: "Forced Saving and RepressedInflation in the Soviet Union, 1986-1990",IMF Staff Papers 39, 1992, pp. 256-286. Dornbusch,R.: "Lessonsfrom Experienceswith High Inflation", World Bank Economic Review6, 1992, pp. 13-32. Fry, M.: Money. Interest and Bankingin EconomicDevelopment. Baltimore, MD: Johns HopkinsPress, 1988. Giovannini,A.: "Savingand the Real InterestRate in LDCs", Journal of Development Economics 18, 1985, pp. 197-217.

Savingin TraJidon Econaoml.-31

Giovannini,A. and M. deMelo: "GovernmentRevenuefrom FinancialRepression", AmericanEconomicReview83, 1993, pp. 953-963. GOSKOMSTAT-USSR:NarodnoeKhoziaistvoSSSR v 1999 g. Moscow: GOSKOMSTAT,1991. Gregory, P. and R. Stuart: SovietEconomicStructure and Performance. New York: Harper-Collins, 1990. Hardy, D. and A. Lahiri: "BankInsolvencyand Stabilizationin EasternEurope", IMF Staff Papers 39, 1992, pp. 778-800. Lipton, D. and J. Sachs: "Prospectsfor Russia's EconomicReform", BrookingsPapers on EconomicActivity 2, 1992, pp. 213-265. McKinnon,R.: Money and Capital in EconomicDevelopment. Washington,DC: Brookings Institution, 1973. McKinnon,R.: The Order of EconomicLiberalization: FinancialControlin the Transitionto a MarketEconomy. Baltimore, MD: Johns HopkinsUniversityPress, 1991. Nordhaus,W.: "Soviet EconomicReform: The LongestRoad", BrookingsPapers in EconomicActivity 1, 1990, pp. 287-308. Ostry, J. and C. Reinhart: "PrivateSaving and Terms of Trade Shocks", IMF Staff Papers 39, 1992, pp. 495-515. Pleskovic,B.: "RegionalDevelopmentand Transition in the Former SovietUnion: A Comment", InternationalRegional ScienceReview 15, 1993, pp. 297-305. Pleskovic,B. and J. Sachs: "PoliticalIndependenceand EconomicReformin Slovenia", in 0. Blanchard,K. Froot and J. Sachs, eds.: EconomicTransition in Eastern Europe. University of Chicagoand NBER, 1993, pp. 191-230. Sachs, J. and D. Lipton: "RemainingSteps to a Market-BasedMonetarySystem in Russia", WorkingPaper 54, StockholmInstitute of Sovietand East European Economics, 1992. Smith, R.: "Factors affectingSaving, Policy Tools, and Tax Reform: A Review", IMF Staff Papers 37, 1990, pp. 1-70. van Wijnbergen,S.: "Credit Policy, Inflationand Growthin a FinanciallyRepressedEconomy", Journal of DevelopmentEconomics 13, 1983, pp. 45-65. World Bank: "RussianEconomicReform: Crossingthe Threshholdof Structural Change", processed, 1992.

Saving in Trwnsidon Econonles- 32

Table 1

Distribution of income and Expenditure in Ukraine for the fourts quarterof 1993, takenfrom householdbudgets

Less than 120

All Total Income Wages(of public and private sectors) Pensions,stipends andgifts Saleof products Incomefrom other sources

Nlonthly Household Income (in thousandsof karbovanets) From 120 From 160 From 200 From 240 From 300 lo 160 to 200 to 240 to 300 to 400

From 400 to 600

More than 600

100

100

100

100

100

100

100

100

100

58.5

60.3

60.3

60.1

62.0

59.9

59.5

57.9

55.8

16.9 19.7

23.8 15.6

21.5 18.0

20.0 18.4

18.4 18.4

18.3 19.0

17.1 19.1

16.1 20.5

11.5 22.6

4.9

0.3

0.2

1.5

1.2

2.8

4.3

5.5

10.1

57.4 20.1 2.5 2.1 3.9 5.5 3.5

87.9 11.9 1.2 2.2 4.0 4.0 3.3

75.4 13.1 1.9 2.3 4.6 4.6 1.9

71.0 16.8 2.1 2.0 4.2 4.9 2.5

66.5 16.1 2.6 2.0 4.5 4.7 2.6

62.2 17.6 2.5 2.0 4.4 5.3 3.1

57.4 19.8 2.5 1.8 4.1 5.6 3.8

51.2 21.6 2.7 1.7 4.3 6.1 4.4

38.0 29.3 2.6 1.5 3.4 5.9 4.1

5.0

-14.5

-3.8

-3.5

1.0

2.9

5.0

8.0

15.2

Expenditures (as a shareof income) Foodstuffs Other goods Alcoholic beverages Commtunalservices Other services Taxes Other expenses

Saving (as residual)

Memorandum: Lessthan US dollar equivalent $4.70 on 15 Novemberin informal market

$4.70 6.27

Source: Ministry of Statistics;data collection network.

$6.27 $7.84

$7.84 $9.40

$9.40 $11.76

$11.76 $15.68

$15.68$23.52

More than $23.52

SaWngIn TrwiAo

&emlea-

Table 2 Real InterestRates in Ukraine Average interest rate minus observed inflationrate

l________ 1993

__________ _________

Refinance

Commercial

Credits

Credits

January

-41.1

-40.8

-41.5

February

-22.4

-18.8

-19.1

March

-13.8

April

-10.8

-8.2

May

-1.2

-7.7

June

-26.0

-20.4

-24.2

July

-19.1

-8.1

-8.0

August

-19.0

-9.4

-11.5

September

-36.8

-28.5

-30.4

October

-37.1

-29.7

-32.7

November

-26.7

-11.6

-14.5

December

-32.9

-24.7

-28.7

-7.3

5.6

-0.4

6.6

17.3

9.9

March

13.3

25.4

20.5

April

9.9

20.1

20.2

May

10.6

19.1

16.9

!__________ 1994

Deposits

January February

_

Source: NationalBank of Ukrane

-10

-11.2 -10.1

-1.9

- 33

S.i*g in Tr'l*W

Ecov-ner - 34

Table 3a

Mometasy Survey for Georgia (As percentof DomesticCredit)

DEC 1991 DEC 1992 JAN 1993 FEB 1993MAR 1993APR 1993MAY 1993JUN 1993 JUL 1993 AUG 1993 SEP 1993 OCT 1993NOV 1993DEC 1993

Net ForeignAssets Convertibleassets Russianrubleassets NetDomesticAssets - -

0.0

-94.4

-66.0

-67.6

-62.7

44.9

-51.6

-36.2

-29.4

-21.7

1.3

-11.9

2.5

13.5

0.0 0.0

1.9 -96.3

2.2 -68.2

2.5 -70.1

2.3 -65.0

1.9 -46.8

5.5 -57.1

6.3 -42.5

6.1 -35.6

5.9 -27.6

25.9 -24.6

16.8 -28.7

19.5 -17.0

23.6 -10.0

105.3

112.8

91.5

87.7

94.5

84.5

71.9

73.1 --

- ,. - --

- -

140.4 ..

.

.

--

--

--

-

-

92.7 -

-.

75.7 -_ -_ --

-.

-_--

96.5 - --

--

94.1 --

--

- --

--

96.6 - --

--

-

--

-

DomesticCredit Claimson Generalgovemment Claimson restof the economy o/w claims on entcrpfises

100.0 7.6 92.4 77.3

100.0 14.9 85.1 64.5

100.0 22.5 77.5 56.8

100.0 9.3 90.7 58.4

100.0 6.2 93.8 69.2

100.0 4.2 95.8 70.5

100.0 6.7 93.3 58.7

100.0 7.0 93.0 57.0

100.0 2.4 97.6 59.5

100.0 5.9 94.1 56.4

100.0 8.2 91.8 61.7

100.0 10.0 90.0 60.8

100.0 12.5 87.5 65.7

100.0 23.3 76.7 61.7

Other Assets(net)

-26.9

40.4

5.3

12.8

-7.2

-24.3

-3.5

-5.9

-3.4

-8.5

-12.3

-5.5

-15.5

-28.1

73.1

46.1

39.3

45.2

30.1

32.1

44.9

57.9

67.2

69.8

89.0

82.6

87.1

85.4

73.1 34.9 38.2 0.0

45.2 16.8 28.5 0.8

38.8 13.4 25.4 0.5

44.5 13.0 31.5 0.8

28.9 7.7 21.1 1.2

31.4 10.9 20.5 0.7

41.7 11.3 30.5 3.2

53.9 23.8 30.1 4.0

64.0 31.6 32.4 3.3

65.8 31.5 34.3 4.0

64.7 33.4 31.3 24.3

66.6 37.3 29.3 16.0

58.3 33.2 25.1 28.8

48.7 22.0 26.7 36.8

17.7

158 795.5

211 33.1

234 11.0

402 71.9

478 18.9

444 -7.0

515 15.9

543 5.4

648 19.4

858 32.4

1305 52.0

1934 48.3

3769 94.9

Broad Money Broad money(coupon-denominated) o/w Currencyoutsidebanks o/w Bankingdeposits Bankingdepositsin foreign currency

Meitorandum items: Domesticcredit (billions of coupons) Domenstic credit (monthlygrowth rate) Source: NationalBankof Georgia *

- Annual growth rate

Sawngin Tasdtion Bconomies - 35

Table 3b Monetary Survey for Georgia (As Percent of Domestic Credit)

JAN 1994

Net Foreign Assets

FEB 1994 MAR 1994 APR 1994 MAY 1994 JUN 1994

JUL 1994 AUG 1994

26.5

17.9

11.2

-16.7

-18.7

-26.2

-24.3

-25.1

29.1 -2.7

20.0 -2.1

12.0 -0.8

-16.6 -0.1

-18.6 -0.1

-26.0 -0.2

-24.2 -0.1

-25.1 -0.0

78.1

90.9

79.5

82.2

91.3

92.4

96.4

89.7

DomesticCredit Claims on General government Claims on rest of the economy o/wclaims on enterprises

100.0 14.0 86.0 71.8

100.0 15.7 84.3 68.4

100.0 13.5 86.5 72.5

100.0 6.2 93.8 82.1

100.0 8.6 91.4 76.9

100.0 6.0 94.0 81.8

100.0 9.0 91.0 79.0

100.0 6.4 93.6 85.9

Other Assets (net)

-21.9

-9.1

-20.5

-17.8

-8.7

-7.6

-3.6

-10.3

104.5

108.8

90.7

65.6

72.7

66.2

72.1

64.5

41.7 17.5 24.2 51.3 11.6

39.6 15.5 24.1 47.8 21.4

30.2 8.9 21.3 49.6 10.9

17.1 3.3 13.7 42.9 5.6

27.4 4.3 23.0 36.6 8.7

23.3 3.7 19.6 34.2 8.7

26.3 3.6 22.7 30.7 15.1

18.3 3.6 14.7 38.8 7.4

4708 24.9

5870 24.7

11438 94.9

38573 237.2

39701 2.9

50138 26.3

60209 20.1

113172 88.0

Convertible assets Russian ruble assets Net Domestic Assets

Broad Money Broad money (coupon-denominated) o/w Currency in circulation olw Banking deposits Banking deposits in foreign currency Balances on banks' correspondentaccounts

Memorandum items: Domestic credit (billions of coupons) Domestic credit (monthly growth rate) Source: National Bank of Georgia

Ech ni4&- 36 SaWngin 7&uLWon

Figure 1 Endogenous Determination of Saving and the Real Interest Rate

C2

y2

2~~~~~~~~~~~~~~

1-

y .. g+j# Y1.g+#(P) 0

yLg+16(p

Savng in TranultJ

ECnMie, - 37

Figure 2 Steady State Equilibrium in Price and Wealth Accumulation

b~~~~~~~~~~~~~~

AA

I S~~~~~~~~~~~~~~~~~~~~~~~~~~~~ IS

Saving in Trududon Eceoined

- 38

Figure 3 Comparative Dynamics of a Negative Production Shock

bI

\~~~~~~~~~~~~s

b

_~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I-

.~~~~~~~

S&nngin TIhmdonEconomdes- 39

Figure 4 HouseholdSourcesand Uses of PurchasingPower (on a monthlybasis) Sources

Uses

Total income Formal wage and/or salary Saleof produceor livestock Private trading or business activity Rental of land or property Interest from loans given Pensionsor other governmentpayments Gifts from family and/or friends

Expenditureson consumergoods

Saleof possessions

Holdingsof currency (domesticor foreign)

Borrowing,withdrawalfrom bank accountsor use of hoardedcurrency Receiptof previouslyloaned funds

Expenditureson communalservices Businessexpenditures Deposits in banks or other financial institutions

Capitalpurchases

SaWngin Transidon&onomdes-40

Figure 5 Asset-holdingand the "Ruble Overhang" in the Soviet Union

Currency and HouseholdDeposits as shares of GNP

45 40

Household Deposits/GNP

35

-

30 25 20 15

Currency/GNP

5 _

US Currency/US GNP

0 1985 1986 1987 Shaded areas in diagram represent estimate of overhang.

1988

1989

1990

Saving in TranaliionEconomies - 41

Figure 6 Currency Holdings and Deposits of the Population in Belarus in 1989rubles

20

|

^

|Deposks of the | ~~~populatbion |t

rn10

Cufrency

0 o000

1000

100 Yer

102

10I3

- 42 Savng in TrmWon Econowmic

Figure 7 Ratio of BankingSystemFinancialInstrumentsto Cash in Ukraine

|Ukraine: Trends in Saving Activity (SightDeposits as a ratio to Cash in Circulation

4

___

O -3

\lou.eholds

1/91

1/92

1/93

4/93

7/93

10/93 11/93 12/93 1/94

2/94

3/94

4/94

Ukraine: Trends in Saving Activity Time Deposits as a Ratio to Cash in Circulation 2.5 0

*0

O.5 ho19 I9 41dokbv. 1/91

1/92 1193 4/93 7/93 10/93 11/93 12/93 1/94 2/94

3/94

4/94

SaN

In Th=ad 7

EcWMi

Figure 8 Yield Curveand Concentrationof Depositsin BankingSystemin Ukraine

[Distribution

Ukraine: Deposits in Banking System and Interest Rates on Deposits, May 1994

....... .......... .

..-

500

U,

0

400

/

400

households

0

0300 enterprises

'200

A0 less than I

6 to 12 3 to 6 I to 3 Maturity of deposit (in months)

[Source: NationalBank of Ukraine

Over 12

- 43

Annex A Theoretical specificationof saving in the transition. The theory of private saving is best viewedas an intertemporalchoice for the allocationof real resources. In this section I beginwith a two-periodmodel to illuminatethe basic economic forces determiningprivate saving. I then turn to an extensionof the frameworkin which the dynamic of transition is explicitlyderived. The two-periodmodel. The private sector will be modeled as equatingthe present value of expenditureover the two periods to the present value of after-taxincome and the wealth carried forward from the past. This budget constraint can be written as

(1)

E(p,1,u) = 0

0 = p(D - t) + Y E = pc' + C2 y = py'(Ca)+ y2 E is the present value of expenditure(here, simply consumption)in the two periods, and has the usual properties of expenditurefunctions.' Consumptionin period two (c2) is taken as the numeraire, while consumptionin period one (c') has a relativeprice p.2 The level of utility achievedthrough this pattern of expenditureis denotedu. D is the inheritedwealth, while Y is the presentvalue of income in both periods and t is a lump-sumtax paid in period one.27 a representsthe possibility of an income shock in period one. The material balance in the period-oneconsumptiongood is then c' + g - Y' =

(2)

The governmenthas its own expenditures(g) in period one that are not necessarilyfinanced completelyout of taxation receipts and contributenothingto utility. , is the volume of international borrowing (i.e., the negativeof the current accountsurplus)undertakenby the households. Internationallenders are assumedto have a positivep-elasticityof internationalborrowing, with the volume of available financing rising with p.2 S

5 The conventionin the followingwill be that derivativesare indicatedby subscript while time periods are indicatedby superscript. I This relative price is the amalgamnof the ratio of price levels in the two periods and the

nominal discount factor R = (1+r); i.e., p = P'R/P2 where superscripts indicateperiods. 1 I assume for simplicitythat no taxes are levied (and no govermnentexpendituresmade) in period two. 2 Rising p indicatesthat the real interest rate that the home country pays on internationaldebt is rising. This specificationwas chosento provide a parameterizationof two extremecases: an economywith unlimited accessto world capital marketsat given price p' (infinitep-elasticity)and an economywith no access to world capitalmarkets (zero p-elasticity). Those special cases are derived directly in Annex C.

Sawng in Transidton ononmies Annex - ii

Saving of the householdsin period one is defined as (3)

s = Y, - cT- T

and as combinationof (2) and (3) indicatesis identical in equilibriumto the differenceof the governmentbudgetdeficit (g-T)and foreign borrowings. There are thus three sets of determinantsto real saving. The first is the governmentfiscal policy, as summarizedby the fiscal deficit. The second comprisesshocksto earningscapacity, as indicatedby the shock a to period-oneproduct. The third is the relative price (or real interest rate) p. While the first two are exogenous,the third is jointly determinedwith real saving; I examinethis joint determinationin the followingsections. Equilibrium in the two-period model. The determinantsof real saving and the real interest rate p can be derived from total differentiationsof the equationsystem (1) through (3). Through appropriatesubstitution,this derivationyields the followingsolutionin terms of endogenousvariablesp and u. Define y = (cO-y'-fl-D+T), and assume that it is positive in equilibrium(i.e., that the economywill have excess demandfor goods in period one). Define as well a (EP 1 JE,) as the marginalpropensityto consumein period one from discountedwealth 0 and X = G8 - Ep) > 0 the increase in availabilityof period-oneresourcesto the governmentbudgetthrough real interest rate increase. Then

=

(1/a) 1

E..du

_

A = be

+n >

-p

dD _y+pq , Y' da_

j

The relative price of period-onegoods is rising (either through an increase in P' or an increasein R) with an increasein g or D and a decrease in r. A shock a that tends to reduce period-oneincome also has this effect for non-extremevalues of p. The shifts in g, r and a all limit the quantityof goods availablefor consumptionin period one, and thus raise the relativeprice of those goods. The stock D representsa claim by consumerson the existingstock of goods based on previoussaving, and an increasein that stock will increasedemand and thus the relative price of period-onegoods. The multiplierterm A is positive by the assumptionof excess demand for periodone consumption(,yb)and due to the positive effectsof real interest rate rises in causingdomestic saving (the Slutskysubstitutioneffect -Epp)and increasedforeign borrowing(a). Utility is falling with an increase in g or r; governmentexpenditureis not valued in the utility functionbut crowdsout private consumption,while taxes subtract directlyfrom disposableincome. I By definitionE(l,p,u) = D. It is also the case that in equilibriumthe Slutsky and Marshallian demand curves coincide. For period-oneconsumption,that implies Ep(p,l,u) = c'(p,). Differentiationof these two equalitiesfor constantp yields E. = Duand E, = cQ0k.. Substitution and rearrangementyields EpF,E.= cQ,as was to be shown.

Saing in TrondizionEconomdesANer - id

A negative shock to period-oneoutput causes a fall in utility as well. Utility is increasingin wealth carried forward. Real saving behavior as definedin this macroeconomiccontextcan be derived from equations (3) and (4). ds = y' dct - Ep dp =

-

5 E. du - dr

(1-(6SpA))dg - (l/A)pb#p dD - (1 - (pS6,BpI))dT - (1 - (l-pb)1,B/A)y' dot (5)

Real saving is rising with an increase in g, althoughnot in a one-for-onemanner if foreignborrowing is possible. It is falling with an increase in the accumulatedstock of wealth and rising with an increase in real taxation. It is falling with negativeoutputshocks in period one.' Modelling saving in the dynamic transition. Life-cyclemodels of consumption,investmentand saving over an infinitehorizonprovide a convenientstarting point for the theoreticalmodel. The optimalbehavior in Campbell(1987) or Blanchardand Fischer (1989)can be summarizedin the notationof the previous section, with conditionsintroducedto govern the adjustmentof nominalprice and the stock of financial assets over time:

Dt

=

(l+r) D'I + PI s'

St= y - c

-

(6)

t(7)

Pt = P-' + 4)[ (gt tt)

-

st - B(pt )J

(8)

The stock of financial assets rises with interest paid to holders and with the value of private saving in the period. The net material balancecondition(2) may not hold in flow equilibrium, allowingfor the possibilityof rationing in periods of excess demand. This excess demandis a functionthe governmentdeficit and the volume of foreign borrowingf(pt).` As equation (8) indicates,such excess demandwill cause price inflation;the limiting case, as X approachesinfinity, is that of net material balance.

x This last effectsare not ambiguous,and provide an interestingcontrastwith the impactof increasedgovernmentexpenditure. All increasethe excess demandin the goods market in period one, but the government's marginalpropensityto consumein period one is unity. Given the private marginalpropensityof 6, the impact of the output shock or taxationeffect is reduced and can be reversed for extremelyhigh valuationsof period-onerelative price p. 31 The equation of motion for foreign debt is subsumedin the asset accumulationequation, and the nominal interest rate is the same on foreign and domesticborrowing. These assumptionscould be relaxed, but at the cost of modelingthe evolutionof internationaldebt separately.

SaWg i 2wisidon Ectomies Anne. - iv

Expansion of this equation system and evaluation at net material balance (with stable price P) yields the following matrix equation system: rDt- DtI

LP-1*1

r - P' (Oct /OD")

L e(lCl8D)'-)

(gt - t' -

(a-

t)

-

PS(act/aPt ')

D' I

(9)

L-#P(P')) P (

Consumption is assumed to respond to price and financial-asset stocks as in the two-period case, with (act IaDl)) > 0 and (act 8P'-?)< 0. The former condition is derived from the marginal propensity to consume from an increase in purchasing power 0, ceteris paribus, while the second stems from the impact of nominal price increases on the purchasing power of nominal financial assets. The effect of D includes both its direct impact and its indirect impact through the change in the relative intertemporal price of goods p'. The trace and determinant of the matrix indicate the saddlepoint nature of the equilibrium to this dynamic system.3 2 It is illustrated in Figure 2 of the text.

3 The trace of this dynamic system is r - P (ac/8D'l)) + (act /lP" - #), and this will be negative

for lower values of r. The determinant of the system is 44r (8ct /aP'1 - #P) - (g' - t')] C 0 always. These signs will assure a saddlepoint-stable equilibrium.

Annex B: Survey instrument Saving Survey - Georgia

The followingsurvey was prepared by Patrick Conwayand the GeorgianCenterfor Transition EconomicSystems and SustainableDevelopment. It was administeredtwice in the city of Tbilisi, in August and in November 1994, to 500 heads of household. The results will be used in a comparativestudy of saving behavior in transition economiesunder preparationby Patrick Conway, and by the GeorgianCenter and the CaucasianInstitutefor Peace, Democracyand Developmentin their periodic reports. The heads of householdsinterviewedwere selectedaccordingto the followingcriteria. 1.) The proportionof householdsin the sample from each administrativedistrict (raion)of Tbilisi was set equal to the existingpopulationdistribution. 2.) The proportionof householdsof certainsize (e.g., 4 members)in the sample was set equal to the existingpopulation distribution. Given these selectioncriteria, householdswere selectedby randomfrom the population. 20 individuals trained by the CSEI conducted the first survey during the period 15 - 22 August 1994. Demographic questions: 1.How many individualsreside in your household? 1

2

3

4

5

6

7

8

9

10 or more.

2. How many individualsresiding in your householdhave monetaryincome? 1

2

3

4

5

6

7

8

9

10 or more

6

7

8

9

10

3. In which region of Tbilisi (raion) do you live? 1

2

3

4

5

4. Do you (or any membersof your household)own your home? Yes 5. What is your gender?

Male

No

Female

6. What is your age? 18-24

25-34

35-49

50-60

61 or older

7. What is the highest level of educationyou have received? Compulsory

TechnicalTraining

Higher education

Saing In Tra'adon EconomiesAner - W

Economicquestions: 8. Within the past month, considerthe total incomeof your household.From what sourcesdid the income come? More than Less than All half half None Formal salary Saleof agriculturalproduce or livestock Private trading or business activity Rentalof land or property Interest from lending funds Pensionsor other governmentpayments Gifts from familyor friends 9. In which of the followingrangesdoes the total monthlyincome of your householdlie? Check one Less than 1 millioncoupons Between I million and 10 millioncoupons Between10 milion and 20 millioncoupons Between20 million and 30 millioncoupons Between30 million and 50 millioncoupons Between50 miUionand 80 millioncoupons More than 80 millioncoupons 10. In the last month, did you sell any possessionsto obtain money? Yes

No

11. If yes, was the value of those possessions Check one Less than I miUioncoupons Between I million and 10 millioncoupons Between 10 milion and 20 millioncoupons Between20 millionand 30 millioncoupons Between30 millionand 50 millioncoupons Between50 millionand 80 millioncoupons More than 80 million coupons 12. In the last month, did you borrow any money,withdraw moneyfrom bank depositsor use currencysaved at home? No Yes 13. If yes, was the combinedvalue of these activities Checkone Less than I milion coupons Between1 millionand 10 millioncoupons Between10 millionand 20 millioncoupons

Saving in Transidon Economits Annsx-- vii

Between 20 million and 30 million coupons Between 30 million and 50 million coupons Between 50 million and 80 million coupons More than 80 million coupons 14. In the last month, did you receive back any money previously lent to others? Yes

No

15. If yes, was the value of the returned money Check one Less than 1 million coupons Between 1 million and 10 million coupons Between 10 million and 20 million coupons Between 20 million and 30 million coupons Between 30 million and 50 million coupons Between 50 million and 80 million coupons More than 80 million coupons 16. For the last month, add together the total of those sources of funds described above -- income, sale of possessions, borrowing, reduction of bank deposits, use of accumulated currency, and return of previously lent money. Of this total, how much was used on each of the following? Food expenditures Housing rental Clothing expenditures Communal services Entertainment expenses Business expenses Deposits in banks and trust companies Holdings in currency (coupons or other) Large purchases (autos, houses, refrigerators, and others)

None None None None None None

Less Less Less Less Less Less

than half than half than half than half than half than half

More More More More More More

than half than half than half than half than half than half

All All All All All All

None

Less than half

More than half All

None

Less than half

More than half

All

None

Less than half

More than half

All

17. For the last month, add together the total of those sources of funds described above -- income, sale of possessions, borrowing, reduction of bank deposits, use of accumulated currency, and return of previously lent money. Of this total, what percentage was used on each of the following: Food expenditures Housing rental Clothing expenditures Communal services Entertainmentexpenses Business expenses Deposits in banks and trust companies Holdings in currency

None None None None None None

0-20 0-20 0-20 0-20 0-20 0-20

None 0-20

20-40 20-40 20-40 20-40 20-40 20-40

40-60 40-60 40-60 40-60 40-60 40-60

20-40 40-60

60-80 60-80 60-80 60-80 60-80 60-80

80-100 80-100 80-100 80-100 80-100 80-100

All All All All All All

60-80 80-100

All

Savingin Transidon Economies Annex- viii

(couponsor other)

None 0-20

20-40 40-60

60-80 80-100

All

Large purchases (autos, houses, refrigerators and others)

None 0-20

20-40 40-60

60-80 80-100

All

18. Do you expect the past month to be similar to this month economically? Yes

No

19. During the past month, did anyone in your household receive funds (income, borrowing, exchange) in foreign currency(rubles, dollars or other)? Yes

No

20. How much of your household funds came in the form of foreign currency? None

Less than half

More than half

21. Your household used this foreign currency in what way? No Spent on goods and services Held at home as saving Placed in bank deposit Placed with trust company Given to friends Lent to individuals Paid debts

Partial Use

Primary Use

All

All

Annex C Profiles of Formal Financial Institutions in Belarus, Georgia, Kazakhstan and Ukraine In the course of this research project I intervieweda numberof officialswith commercial banks in Belarus, Georgia,Kazakhstanand Ukraine. The text of the report summarizesthe conclusionsI reached based upon these interviews,but the details of the financialsectors and banking operation in these countries may be of independentinterest. For this purpose I include below short summariesof the financialsectors of the various countriesand the operationsof specificcommercial banks. For each countrythe section on banking operationsbegins with a descriptionof the successor bank to the Soviet Saving Bank, given the historicalmonosonyof this institutionin attractingsaving. Belarus.

In Belarusthe bifurcationin financialmarkets was strongly in evidence. The markets in zaichik (the nationalcurrency, de facto) in mid-1994were a channelfor distributingdirected credits 33 These credits did not represent intermediationof to favored industrialand agriculturalborrowers. private saving, but rather a claim on the National Bank of Belarus(NBB). Markets in foreign exchangealso appeared,with the private commercialbankstaking the lead in these activities. Lendingvolumesdependedupon the prior attraction of foreign-exchangedeposits, and banks were thus more competitivein seekingthose depositors. The state commercialbanks, especiallyAgroprombankand Promstroibank,were the prime channelsfor distributionof these directed credits. Agroprombankin mid-1994relied upon NBBor budgetaryrefinancingfor over 75 percent of its resources, while time deposits (includingforeign exchangedeposits)representedonly about 10 percent of resources. Promstroibankwas less reliant on these governrmentrefinancings,but very dependentupon the correspondentbalancesof industrial enterprisesheld at minisculenominal interest rates. Promstroibankdid offer some foreign-exchange credits, but these were typicallyrefinancedthrough foreign lines of credit arrangedby foreign exportersor joint-venturepartners. Private commercialbanks were also able to refinance credits with the NBBin mid-1994. Thus, the most successfulbanks were not necessarilyspecializedin foreign-exchangetransactions. For example,Priorbank was roughly equally split in value betweenzaichikand foreign-exchange transactionsat the beginningof 1994. It obtained a significantshare of zaichik funds through NBB refinancingbut sought the rest through interbank borrowingand enterprisedeposits depositsof zaichik or foreign currency. BelarusBank, by contrast, specializedalmost completelyin foreign-exchange credits and deposits. The Saving Bank of Belaruswas rather slow to respondto the changingeconomicclimate, in part because it remaineda state-controlledinstitution. In mid-1994it continuedto offer only zaichikdenominateddeposits, with interest rates much below the inflationrate. Deposits did increasefive times in nominalterms over 1993, but the consumerprice indexhad increased20 times in the same period. Foreign-exchangedeposits were under preparationat that time, nearly a year after competing '3 The local currencyat that time was de jure a ruble supplement,exchangeableat par for Russianrubles. There was de facto a large premium on Russianrubles in this exchange,especiallyin non-cashtransactions. The ruble supplementwas known in Belarusas the "zaichik", or rabbit, becausethat animalis pictured on the 1-rublesupplement.

Saving in Transiton Economues Annex - x

banks had begun these operations. The Saving Bank activitiesremainedconstrainedby governmental regulation. For example,40 percent of Saving Bank deposits by law were to be relent for housing constructionat an annual interest rate of 140percent. The borrower paid 10 percentagepoints of this, while the governmentwas to reimbursethe Saving Bank for the balance. Not only was this lendingrate well below inflation, but the governmentwas typicallynot prompt in its reimbursements. This greatly constrainedthe interest rate the Saving Bank could offer on deposits. The structure of interest rates on deposits in Belarusunderwentsome adjustmentin 1993. Table Cl Interest Rates on Deposits at the Saving Bank in Belarus (in percent) Maturity of deposit

July 1993

January 1994

Sight deposit

20

60

One month time deposit

40

160

Three month time deposit

40

190

Six month time deposit

40

240

One year time deposit

40

250

Three-fiveyear time deposits

50

More than five year deposits

60

Source: Saving Bank Historically,the Saving Bank lent large volumesof its funds to the National Bank of Belarus (NBB)and the Ministry of Finance. By mid-1994this had almostdisappeared,with no borrowing by the Ministry of Finance, and only small amountsby the NBB. An April 1994 decree of the presidium of SupremeSoviet governedthis. There were no other government-directedloans, but the NBBdid require reserve holdings of between 10 and 12 percent of depositvolumeto be held at the NBB. In mid-1994the Saving Bank was just preparing to introduceforeignexchangedeposits. It expectedto fill a niche in the market, for commercialbanks requireda minimumbalanceof $300 for such deposits; the Saving Bank had no such minimumbalance. The interest rate was projectedto be 6 percent annually. The details remainingincludedthe receipt of a licenseto trade foreign exchange from Belvnesheconombankand the buildingup of foreign exchangereserves. Georgia. In Georgiathe bifurcationof financialmarkets was quite far advanced,and the financial intermediarieshad for the most part chosen to specializein one or the other currency. The market for deposits and credits in the domesticcurrency, or menati, was stagnantin mid-1994except for the credits refinancedby the National Bank of Georgia(NBG). The Agroprombankand the Promstroibank(here known as the Industriabank)were channelsfor distributionof these credits;

SaWngIn Tra,fsion Economisc Amnna- xi

Industriabankhad branched as well into foreign-exchangetransactions. In October 1994 there were 250 commercialbanks licensed by the NBG for operation in Georgia. Most of these were in operation on only a limited scale, and may have been the "in-house" bank for a firm or group of firms. Of the total, only the five state commercialbanks and perhaps 10 others were consideredactive in solicitingbankingdeposits and clients for lending. The NBG was at that time involvedin an effort to tighten restrictionsupon these intermediaries. The New GeorgianBank (the Georgianname for the Saving Bank) remaineda state bank until mid-1994. Its depositsless than doubledin nominal terms during the first half of 1994despite a tenfold increase in consumerprices, due in part to a nominal interest rate of 150 percent per annumon those deposits. With privatizationin July 1994 it began its transformationinto a full-servicebanking system. 50 enterprisesper day becomedepositors. It has correspondentbanks throughoutEurope, and is presently searchingfor a Europeanpartner in a joint venture in internationalbanking. Foreign exchangedepartmentswere put in place in all affiliatesand the bank acceptedforeign-exchange(US dollar, not Russianruble) deposits. The rate on these deposits was seeminglynot very competitive, however, since the bank experienceda drop in deposits in mid-1994as it lowered its return to 3 percent per month. Private commercialbanks have left the menati financialmarkets nearly entirely. For example,Ivertbankin mid-1994had accepteddeposits from individualssplit amongRussianrubles and dollars in a 9:1 ratio. Ruble depositsreceiveda 5 percent/monthinterest rate. US dollar 3month time deposits earned 3 percent per month. There was solid growth (roughly20 percent) in dollar deposits during the first half of 1994. These funds were loanedout at interest rates of roughly 10 percent per month for trading transactions. The TBC Bank, a member of the TBC Group, operated in a similar manner. Of deposits, only 20 percent of the total is found in coupons,with 60 percent in US dollars and 20 percent in Russian rubles and German marks. The bank's lending occurs almost exclusivelyin US dollars. The imbalancebetweencoupon (or ruble) deposits and lendingis rectified by immediateconversionof coupon (or ruble) to US dollars on the auctionor bazaar markets. Both banks have limited resources, as deposit volume is not large, but were demonstratingpositive growth at a time of depressionelsewherein the economy. Other banks, such as Kerdsobank,have chosen a strategyof specializationin menatitransactions,but these are not exhibitingrapid growth. Trust companieswere very visible alternativesto depositorsin Georgia. A large number of these companieswere in evidenceat the end of 1993, but most failed in mid-1994with the failureof the MMMcorporation in Moscow.' Others continuedthrough to the end of 1994, but these failed in turn. Depositorsand shareholdersplaced great pressure on the governmentto provide ex post guaranteesfor their funds, pointing to involvementby governmentagenciesand officialsin the organizationand managementof some of these. These companiessolicitedUS dollar deposits offering24 percent interest per month. They apparentlyhad lined up loans backing trade that could offer 100 percent per month. Unfortunately,these loans often failed, and so also did the trust companies..

3' These were not linked financially. However, Georgiandepositors in such companieslost faith after observing the MMM difficultiesand demandedtheir deposits returned. This "run on the trust company"led to closures in Georgiaas well.

Saving in Transition Economies Annex- - xii

The first wave of trust companies(includingAchi, Golden Fish, Geico and others) attracted clients with the promise of

-

and initial success in returning -- 20 percent return per month on

deposits. Some of these may have been pure "Ponzi" schemes,but others were financial intermediarieswith ties to corporationsable to turn high short-termprofits through trading operations. These organizationswere able to offer such high returns becauseof the profits from trading activities and becausethe deposits were loaned out completely-- negligiblereserves were held for repurchase of deposit certificates. The successof these corporationsattracted competitors,both from newly createdtrust companiesand from more aggressivecommercialbanks. The cash flow from receipts and repaymentsof lending was sufficientto cover withdrawalsuntil the failure of the Russianfirm MMM. During the time of the MMM crisis in Moscow, depositorsin Tbilisi approachedthese trust companiesto redeem their certificates. The trust companywas typicallyunable to meet these demands,and was forced to close. These failures were liquidity-driven. Subsequentfailures were relatedto insolvency,and occurred in one of two ways. The first was through theft: for example,the Golden Bowl Corporationfailed in October 1994 when the chairmanfled the country(reportedlywith $60 million). The second was through ineptitude. Here, the case of InnovationBank is instructive. This bank was backedby the Ministry of Internal Affairs and offeredhigh interest rates on deposits. The leaders of the bank made a numberof improperloans, includinga US $400,000loan to a woman friend, while paid-in capitalwas valued at less than a tenth of that. Most loans were made to the foundersof the Bank. Each borrower proposed a productiveuse for the loan, but these were not well thoughtout and not well screenedby the bank. The bank was warned about its behaviora numberof times by NBG before it finally failed. Depositorshave been quite vocal in demandingthat the governmentmake restitutionfor the losses in these trust companies. In part, this demand is due to the understandabledesire of investors to recoup an investmentgone sour. It is also the case, however, that many of these failed companies had closeties with governmentagencies;in some cases (like InnovationBank)the governmentagency was a part owner. Kankhstan. The banking system of Kazakhstanhad its beginningsin 1987with the "break-up"of the Gosbankand the passage of a new law on corporations. The AlmatyCooperativeBank in 1988 receivedthe fourth banking license issued in the SovietUnion. By mid-1994there were over 200 35 banks in Kazakhstan. 35 of these were private, while the remainderwere joint-stock corporations. Four that had begun as specializedstate banks -- Turanbank(previouslyPromstroibank), Agroprombank,Narodnyibank(previouslySaving Bank)and Credsotsbank-- were joint stock companieswith the government retainingpart of the stock. There were also 9 banks with partial foreign ownership. The bifurcationof financialmarkets has occurred in Kazakhstanas well. However, as in Belarus, the most successfulbanks are participatingin both markets. In the tenge market (the national currency),funds were obtainedpredominantlythrough sale of capital, founders' deposits and either credit auctions or the refinancingof directedcredits with the NationalBank of Kazakhstan 35 Cooperativeshave fallen from favor, so the cooperativeshave been convertedto joint-stock corporations.

Savingin Transidon&conomles Annc:

..

xiii

(NBK). Tenge depositswere in general an unprofitableway to obtain funds. For small banks (mainly those createdsince independence)there was an insufficientnetwork of offices to attract deposits. For most there was only one office

-

in Almaty -- and for some there were no offices at

all. In that situation attractionof deposits from the populationbecamealmost impossible. For large banks (mainlythe specializedstate banks Agroprom,Turan and Kredsots)the deposits of the population were unprofitableand undesirable. Despitetheir network of offices, the cost of funding through NBK refinancingfell below that of attractingmany small depositors. Further, these banks were also awarethat the directedcredits were extendedto many essentiallybankrupt producers. If the banks attracted their own deposits, they would be pressed to lend those as well to these corporations,and thus couldnot protect the funds of the depositors. The NBK was not unhappy with this outcome. Keepingall banks dependentupon them for funds allowedthem to supervise and organize the bankingsystem. Only at Narodnyibankwere deposits a primary source of tenge funds. Non-bankfinancial institutionssteppedinto the breach and soliciteddeposits from the populationfor trading activities. They promiseda large return on investment. These activitieswere popular with investorsbut poorly monitored. These includedtrust companies, Lombardcompanies, pension funds and investmentfunds.6 The governmentwas unsure how to handle these new businesses,since legislationdid not at that time provide the governmentwith the ability either to regulate the new businessesor to obtain data from those businesses. For example, Smagulovand Companywas a private firm offering600 percent annual interest on tenge "deposits"throughout 1994. The firm deliveredas promised in the earlier months, and by August 1994 a "large share of the population" wishesto place moneywith it. In one case, people waited in line 3 days to make their deposits. The people liked the return, and also liked the fact that the agreementswere in effect 1-monthdeposits. Unfortunately,Smagulovand Companyclosedits doors in late 1994, without repaying principalor interest on currentbalances. Smagulovwas reportedly in Australiaat the end of 1994. Narodnyi Bank becamea joint-stockcompany in May 1994 after a history as a state-run bank. Changesin bankingpolicy were not evident in August 1994, however. Interest rates had not been adjusted since January 1994 (see Table C2). The minimum-maturitytime deposit was one year. The bank did not have a licenseto trade foreign exchange. There were many plans for improvements-short-term deposits, foreign-currencydeposits, and higher interest rates on deposits to name a few but none was in place. Deposits of individualswith Narodnyi Bank and the commercialbanks rose 58 percent in nominal terms from the beginningof 1994until I June 1994, from 202.5 million tenge to 320.8 million tenge. However, the inflationrate during that same period was 266 percent, indicatinga large fall in the real value of accumulatedtenge deposits. As of I July 1994, the volume of tenge deposits at Narodnyi Bank was 1.15 billion tenge, while there were only .3 billion tenge at all other conunercialbanks combined. However, from the beginningof the year the commercialbanks have been more successfulat attracting new deposits: there was a 58 percent increase in all depositsand a 19 percent increase in deposits at Narodnyi Bank. In early 1994 the NBK introduceda less expansionarycredit policy. Less credit was directed to state enterprisesand refinancedby the NBK. A greater share of credit was distributed through 3

A Lombard companyruns on pawnshopprinciples.

Saving in Transiuon Economies Annev - - xiv

NBK auction, often with consequentlyhigher interest rate. This made reliance upon NBK funds less attractiveto commercialbanks, and they beganto solicit householdand enterprisedeposits. In mid1994, commercialbanks accepteddeposits in both tenge and fbreignexchange. The latter were more popular with commercialbanks. The dynamicbanks in Kazakhstancould be characterizedin mid-1994in two ways. First, they transactednot only in tenge but in foreign exchange. The latter attracted depositorswishing to minimizethe risk from unanticipatedinflation. The second major differencewas the interest rate on deposits. On 1 January 1994, Narodnyi Bank offered deposit rates of 60 percent (for sight deposit) to 250 percent (for children's deposits, maturity 10 years) per annum. Kazcommertsbankoffered rates of 230 - 550 percent per annum on deposits. Alataubankoffered 188 percent per annum. Only Kazcommertsbankof the three offeredforeign-exchangedeposits in July 1994, and the 35 percent per annum return on those depositsrepresented an extremelypositive real interest rate. Ukraine. The financialsystem in mid-1994in Ukraine illustratedthe strains of transition from extreme inflationto a more moderateprice environment. There were two markets in existence. The majority of transactionsremaineddenominatedin karbovanets(nationalcurrency), althoughforeign-exchange transactionswere ubiquitous. In the market for karbovanetsinstrumentsthe interest rates on deposits and credits reflectedthe more-inflationaryenvironmentof 1993; in real terms these interest rates were extremelyhigh, given the rapid declinein inflationfrom hyperinflationaryrates. The successfulcommercialbanks were transactingin both currencies. Inko Bank, for example,was the largest private commercialbank in Ukraine when measuredby number of branches, assets, liabilitiesor profits. It accepteddeposits in both karbovanetsand US dollars. On its karbovanetsdeposits, it offered in mid-1994very short maturities (as little as 2 weeks)with annualizedinterest rates of 150 - 200 percent. With US dollar deposits, the interest rates were 6 percent per annum for small amountsup to 24 percent per annum for amountsgreater than $10 million. In mid-1994enterprisesmade 78 percent of deposits in the bank, with private persons makingthe other 22 percent. The share of enterprise deposits in karbovanetsis greater than the share in US dollars, with private depositsjust the reverse. Agiobank, a medium-sizedbank, thrived by specializingin the offeringof karbovanets deposits to individuals. It was rankedsecond in depositsto Otshadbank(SavingBank) in mid-1994, with its success attributableto the variety of deposit instrumentsit created. For example,it was the first bank to introducehigh-yieldcertificatesof deposits in the form of bearer bonds (in April 1993). It also introduceddoor-to-doorsalespeopleto sell Agiobankinstrumentsto enterprises. Finally, it beganpaying creditinginterest monthly to accounts while Otshadbankwas only doing this onceper year. The karbovanetsdepositsof the bank remained70 percent from the populationwhile 30 percent are from enterprises. Lendingin karbovanetswas 90 percent to enterprises. Agiobank'sannual interest rates on hard currencydeposits ranged from 18 percent to 25 percent. Roughly80 percent of total deposits are made by private individuals. The lendingis about 50 percent to enterprises,50 percent to individuals. Of total deposits, roughly90 percent were in karbovanetswith the remaining10 percent in hard currency.

SaWng in Trwwidon Econondies

Anne

- xv

Table C2 Narodnyl Bank Annual Interest Rates on Deposits Type of Deposit

Soviet Era

Demand deposits Deposit of uncertain maturity Time deposit: one to three year Time deposit: three to five year Time deposit: greater than five year "Childrendeposit" - ten year Annuity for auto or housing purchase Special accounts

1 March 1992

1 August 1992

1 March 1993

1 July 1993

1 January 1994

2

3

6

15

20

60

3

7

14

30

40

100

5

7

14

30

40

100

7

10

20

40

50

120

9

15

30

60

80

180

9

15

30

70

100

250

2 7

5 7

14 7

30 20

40 20

100 60

Source: Narodnyi Bank, 19 July 1994 Deposit interest rates have changedonly on these days. The types of instrumentsavailablehad not changedsince the Soviet period.

SaWngin Transition Economies Annm - x

The Otshadbankin Ukrainewas the most innovativesuccessorto the Saving Bank of those in the countriesstudied in adaptingto the post-independenceeconomicclimate. It modernizedits operations,and could guaranteenext-daypayment to any location in Ukraine. Its technicalfacilities were the most advancedof all the banks. For example,all lending and depositingoperationsin the capitalcould be governedby magneticcard access. However, its deposit-takingabilities were limited by the interest rates enforcedby the NationalBank of Ukraine (NBU). On 1 July 1994the Otshadbankcould offer only 30 percent per annum on time depositsof less than 6 month maturity. Its traditionalbusinesshad been cut way back since independence. In 1988through 1992,there were roughly 250 customersper day at each affiliate. In 1994, that average had fallen to 5-7. Large deposits were no longer the rule. In the Soviet days, 8 percent of the customersheld 72 percent of the value of Saving Bank deposits. Most accountsin mid-1994were very small in value. Its strategicplan for 1995 was to focus upon providingan effectivemedium of payments, while leaving depositand lending operationsto others. The Otshadbankaccepteddeposits in foreignexchangeat a 3 percent annual interest rate. It lent these funds carefully, and insistedupon land collateral. These 37 deposits (at the market exchangerate) representedroughly 50 percent of deposits in mid-1994.

The chief of the Kiev operationsof Saving Bank insistedthat the foreign-exchangedeposits Il were "mainlyheld for foreign studentsin Kiev".

Policy Research Working Paper Series

Title

Author

Date

Contact for paper

WPS1491 Equilibrium Incentives for Adopting PeterW. Kennedy Cleaner Technology Under Emissions Benoit Laplante Pricing

August 1995

E. Schaper 33457

WPS1492 Trade Policies, Macroeconomic Sarath Rajapatirana Adjustment. and Manufactured Exports: The Latin American Experience

August 1995

J. Troncoso 37826

WPS1493 Migration and the Skill Composition of the Labor Force: The Impact of Trade Liberalization in Developing Countries

Ram6n L6pez Maurice Schiff

August 1995

J. Ngaine 37947

WPS1494 Adjustment and Poverty in Mexican Agriculture: How Farmers' Wealth Affects Supply Response

Ram6n Lopez John Nash Julie Stanton

August 1995

J. Ngaine 37947

WPS1495 Raising Household Energy Prices in Poland: Who Gains? Who Loses?

Caroline L. Freund Christine I. Wallich

August 1995

G. Langton 38392

WPS1496 Reviving Project Appraisal at the World Bank

Shantayanan Devarajan August 1995 Lyn Squire Sethaput Suthiwart-Narueput

C. Bernardo 37699

WPS1497 Public Choices between Lifesaving Programs: How Inportant are Lives Saved?

Maureen L. Cropper Uma Subramanian

August 1995

A. Maranon 39074

WPS1498 Decentralized Rural Development and Enhanced Community Participation: A Case Study from Northeast Brazil

Johan van Zyl Tulio Barbosa Andrew N. Parker Loretta Sonn

August 1995

M. Williams 87297

WPS1499 The Dynamics of Poverty: Why Some Christiaan Grootaert People Escape from Poverty and Ravi Kanbur Others Don't-An African Case Gi-Taik Oh Study

August 1995

A. Sachdeva 82717

WPS1500 Agricultural Trade Liberalization in Merlinda D. lngco the Uruguay Round: One Step Forward, One Step Back?

August 1995

J. Ngaine 37947

WPS1501 Are Partner-Country Statistics Useful Alexander J. Yeats for Estimating "Missing" Trade Data?

August 1995

J. Ngaine 37947

WPS1502 Active Labor Market Policies in the OECD and in Selected Transition Economies

August 1995

WDR 31393

Hartmut Lehmann

Policy Research Working Paper Series

Title

Contact for pape-

Author

Date

WPS1503 Africa's Growth Tragedy: A Retrospective, 1960-89

William Easterly Ross Levine

August 1995

R. Martin 39120

WPS1504 Savings and Education: A Life-Cycle Model Applied to a Panel of 74 Countries

Jacques Morisset C6sar Revoredo

August 1995

N. Cuellar 37892

WPS1505 The Cross-Section of Stock Returns: Stijn Claessens Evidence from Emerging Markets Susmita Dasgupta Jack Glen

September 1995

M. DaJis 39620

WPS1506 Restructuring Regulation of the Rail Industry for the Public Interest

September 1995

J. Dytang 37161

WPS1507 Coping with Too Much of a Good Morris Goldstein Thing: Policy Responses for Large Capital Inflows in Developing Countries

September 1995

R. Vo 31047

WPS1508 Small and Medium-Size Enterprises Sidney G. Winter in Economic Development: Possibilities for Research and Policy

September 1995

D. Evans 38526

WPS1509 Saving in Transition Economies: The Summary Report

September 1995

C. Bondarev 33974

loannis N. Kessides Robert D. Willig

Patrick Conway