THE POLITICAL ECONOMY OF FINANCIAL REPRESSION IN TRANSITION ECONOMIES

Cevdet Denizer, Raj M. Desai, and Nikolay Gueorguiev

I. INTRODUCTION

Why do governments distort financial markets and impose impediments to capital mobility despite the inefficiencies that accompany such policies? It is well known that financial systems in developing countries tend to be “restricted”or “repressed”through burdensome reserve requirements, through interest-rate ceilings, through foreign-exchange regulations, through rules governing the composition of bank balance sheets, or through forms of heavy taxation of the financial sector. Less understood is why governments are drawn to regulate financial markets to the

point of financial repression.

Before the 1970’s, financial restrictions were often favored in capital-scarce countries on the grounds that usury could be better prevented, money supply better controlled, and higher investment-savings targets met than if financial markets were liberalized. In later years this claim was frequently supported with evidence from high-growth economies in East Asia, where governments supposedly manipulated financial systems in order to promote targeted industrial expansion. In most of the settings where repressive financial controls have been applied, however, the typical outcome has been economic contraction, not sustained growth. This realization has led to the emergence of something of a consensus in development macroeconomics: that financial repression is adopted in developing countries in order for governments to obtain resources to finance their deficits.

In this paper we explore some preliminary evidence from a region of the world where financial regulations have rarely been examined in any systematic manner, namely, the post-Communist economies of Eastern Europe and the former Soviet Union. We show that the public¬finance framework has limited cache in explaining financial repression in the transition economies, given the peculiar financial lineages of the socialist state. Specifically, the weak distinction between the “public” and “private” spheres of finance in transition economies means that the deficit often conveys little information about the real fiscal activities of governments. We find that a more

1. Contact: Raj M. Desai, at Desair@gunet.georgetown.edu A previous version of this paper was prepared for delivery at the annual meeting of the American Political Science Association, Washington, D.C., 1997. We thank Oya Celasun for invaluable research assistance. The electoral data used in this paper were provided by Joel Hellman and prepared by Joshua Tucker, whom we thank for making this resource available to us. The findings, interpretations, and conclusions presented in this paper are the authors’ own and should not be attributed to The World Bank and IMF, their Executive Boards of Directors, their member countries or affiliated institutions.

fruitful approach lies in examining how political institutions, by shaping the incentives politicians face, can determine the choice of financial policy. Our findings suggest that post-Communist governments inhibit the development of financial institutions to ensure adequate flows of external capital to the enterprise sectors, rather than to finance deficits.

The discussion is organized as follows. Section II defines financial repression, its historical debates and rationales. Section III examines the role of financial repression in the unique case of transition economies. The fourth section presents the empirical results of our estimations of financial repression in transition economies. The final section summarizes and offers some concluding remarks.

II. FINANCIAL REPRESSION AND FINANCIAL DEVELOPMENT Financial Policies and Some Key Issues

For the past 25 years, the main analytical basis for studies of the role of the financial sector in development has been the McKinnon-Shaw framework of the “repressed”economy. In this view financial repression refers to a set of policies, laws, formal regulations, and informal controls, imposed by governments on the financial sector, that distort financial prices— interest rates and foreign exchange rates— and inhibit the operation of financial intermediaries at their full potential. The main instruments of financial repression are high reserve requirements and interest-rate ceilings, that is, a combination of (forced) low rates of return on assets and (forced) high portions of “set-aside”or reserve money. Successful financial repression increases the demand for credit, and at the same time, creates disincentives to save.

These conditions permit governments to intervene in financial markets in three ways. First, the imposition of large reserve or liquidity requirements on commercial banks creates an artificial demand for a government’s own securities (Agénor and Montiel, 1996: 152). Second, because of the excess demand for credit, the government invariably begins to ration credit among competing users. Third, because of the disincentive to save, savers begin to switch from holding claims on the banking sector to holding claims in foreign markets. Thus selective and sectoral credit schemes, as well as capital controls on foreign exchange, are typical components of financial repression.

In the neo-classical perspective, the main justification for financial repression derives from an assumption of perfect substitutability of money and “productive”capital. In Tobin’s monetary¬growth model, if the return on capital rises relative to the return on money, it encourages a shift from money to capital in household portfolios, higher capital-to-labor ratios, and increased labor productivity (Tobin, 1965). The central implication of this reasoning is that reducing the rate of return on money— through interest-rate ceilings, but also through an optimal level of inflation, both of which serve as a tax on real money balances— can increase the rate of economic growth.

McKinnon (1973) and Shaw (1973), however, questioned the applicability of the neo¬classical approach to developing countries, and instead argued that the distortions from financial repression crowd out high-yielding investments, create a preference for capital-intensive projects, discourage future saving, and thereby reduce both the quality and quantity of investment in an

economy. In this framework, money and capital are compliments rather than substitutes: the more attractive it is to hold real money balances, the greater the incentive to invest. Productive investment, and therefore capital accumulation, occurs because a large real money stock makes greater amounts of loanable funds available to borrowers (McKinnon, 1973: 59-61

Financial Repression and Market Failure

In the 1980s, critics of the McKinnon-Shaw framework argued that raising institutional interest rates might have strong negative effects on savings, investment, output, and growth. Using models incorporating informal credit or “curb”markets, critics of financial liberalization argued that the lack of effective institutions in developing countries required some degree of government control to be maintained over the financial sector (Taylor, 1983

Proponents of “optimal”financial repression have argued that financial controls can correct market failures in financial markets, lower the cost of capital for companies, and improve the quality of loan applicants by selecting out high-risk projects. In addition, if used in conjunction with export-promotion schemes, or preferential credit schemes, financial repression could encourage the flow of capital to sectors with beneficial technological spillovers (Stiglitz, 1989, 1994). Of course, there are worlds of difference between the claims that financial repression can be efficient and that it will be. Although research on credit controls in Japan, South Korea, and Taiwan indicated that financial repression contributed to the high performance of those economies, this evidence remains somewhat controversial. Moreover, the balance of more systematic, cross-national empirical evidence suggests that there are negative correlations between low real interest rates, high reserve requirements, and low degrees of financial intermediation on one hand, and investment and growth

on the other.2

The Public-Finance Approach

Most scholars of finance and development have rejected the claim that financial repression is adopted on welfare-maximizing grounds alone. Rather, development macro economists have,

2. Studies establishing this relationship are: Lanyi and Saraco_lu, 1983

generally speaking, reached a strong consensus regarding the reasons for financial repression: fluctuations in government revenue. A financial sector under administratively-imposed restrictions is a potential source of “easy money”for the public budget. In the classic cases of financial repression, the proliferation of financial instruments from which governments can extract seignorage is encouraged, mainly a relatively oligopolistic banking system, since obligatory holdings of government bonds can be imposed on commercial banks. Private securities markets are suppressed through a variety of taxes and duties, since seignorage cannot be so easily extracted from these markets (Fry, 1995: 20-22). In short, financial repression has the overall effect of transferring funds from the financial system to public borrowers.

The “fiscal”choice that governments make in choosing the degree of financial repression, in the public-finance approach, depends on the revenue losses due to the failure of more direct tax instruments. It has been argued, for example, that economies subject to large amounts of income¬tax evasion are likely to turn to implicit taxes in the form of the inflation tax, indirect taxes on the financial sector through interest-rate ceilings or high reserve requirements, or some combination of both (Roubini and Sala-i-Martin, 1995). In developing countries facing sustained deficits, it is argued, porous or weak systems of revenue collection force governments to rely on inflation taxes. But when governments allow financial systems to develop fully, the need for people to carry money is reduced, eroding the inflation tax base, along with the opportunities for seignorage. By enforcing restrictions on the activities, services, and products of financial institutions, on the other hand, governments can maintain a ready base for the inflation tax.

The implications of the public-finance approach are simple. First, efficient use of the inflation tax requires certain repressive measures to increase the demand for money (Giovannini and de Melo, 1993

III. FINANCIAL REGULATION IN THEORY AND PRACTICE

Our main conceptual objection to the “easy-money”thesis is that it presents an incomplete picture of the financial-policy choice at the heart of the matter. In assuming that governments respond to potential revenue shortfalls in uniform ways, this approach falls short in explaining why different governments facing similar budgetary constraints might choose to regulate their financial systems differently. This, more or less, axiomatic depiction of policy-making processes as homogenous is sharply disputed by research on both the domestic and international aspects of financial policy-making. Institutional differences among political systems have economic effects

A more serious objection is that the public-finance explanation for financial repression relegates preferential credit policies to a secondary role. In the public-finance view, preferential credit schemes are an unintended result of interest rate controls, not their cause. In the standard, hypothetical sequence of events: governments facing a large public sector deficit introduce a variety of financial restrictions, including interest-rate ceilings, which allow deficit financing at a lower interest rate. Since interest-rate controls cause the demand for credit to surge, the government is unwittingly drawn into the process of allocating credit among competing users. A significant body of case-study evidence, however, suggests that selective credit schemes are themselves the primary reasons governments repress their financial systems in the first place (Haggard, Lee, and Maxfield, 1993

Fiscal and Monetary Policy in Post-Communist Economies

The implication that inter-temporal revenue losses may predict the level of financial repression assumes that the fiscal responsibilities of the state are well-defined and enforceable, and that any financial flows from the public to the private sector are controlled and appear in consolidated governmental balance sheets. In other words, “fiscal” property rights are

assumed to be exclusive.3 For this reason, the public-finance framework carries little explanatory power with respect to the formerly state-socialist economies of Central and Eastern Europe and of the Former Soviet Union.

In practice, of course, all governments can engage in a variety of activities that render “deficits” meaningless numbers. Unbudgeted expenditures, outside the supervision of budget offices, are typical culprits in developing countries (Pradhan, 1996

Certain legacies of the socialist financial system suggest this possibility. First, banking sectors remain among the more state-controlled parts of these economies, with few governments having taken steps towards their full commercialization. Thus the line between “public” and “private” finance often remains blurry, with governments prompting banks to act as quasi-fiscal agents of the state through interest-rate controls or, more directly, through directed credit programs. Second, even in the few cases where banks have been fully or partially commercialized, the lending portfolios that formerly state-owned commercial banks have inherited are heavily concentrated among a few firms or industrial sectors. In addition, all formerly state-owned banks inherited bad loans from their public enterprise clients— loans that continue to list on the asset side of their balance sheets. Under such circumstances, foreclosing on the biggest borrowers often threatens the banks themselves (Desai and Pistor, 1997). Finally, the enforcement of debt

3. On the concept of “fiscal” property rights, see Tanzi, 1993a, 1993b.

contracts, in most cases, is impaired due to the incapacity of courts, ambiguities in bankruptcy legislation, and ad hoc government interventions on behalf of certain companies.4

While it is difficult to estimate the size of directed credits from the banking systems in transition economies in a reliable way, the experience of Russia is indicative as to how large they can be. Easterly and Vieira da Cunha Cunha (1994), for example, show rapid inflation in Russia eroded the financial savings of households and it was this group that paid the inflation tax. Their estimates suggest that the losses of households becasue of highly negative interest rates, about -78 percent, in 1992 were about 12 percent of Russian GDP. Enterprises, like anybody who had bank deposits, also lost and their losses were about 18 percent of GDP in 1992. However, as the Russian Government issued credit to enterprises this sector was a beneficiary of inflationary financing. Easterly and Vieira da Cunha’s estimates suggest that the amount they received in 1992 was about 16 percent of GDP with their net inflaton tax being only 2 percent as opposed to 12 percent by the households. Hence, the costs of financial repression has been large, at least in Russia and there were large income transfers from one group to another.

4. Over the long run, of course, the inability of insolvent enterprises to make payments on loans should lead commercial banks to cut them off as clients. But in the short term, these factors peculiar to the transition economies may actually encourage banks to lend more, rather than less, to loss-making firms.

One might expect deficits and revenue losses to be an overriding concern of post-Communist governments, most of which have witnessed tax-evasion epidemics following the dismantling of the socialist revenue-collection apparatus, the multiplication of taxpayers following privatization, and the fall in information available to tax authorities (Barbone and Marchetti, 1995

IMF, 1995).5 Meanwhile, just as public finances were being restructured, governments in this region faced their most severe budgetary pressures in decades, as they were forced to assume spending responsibilities that previously were fulfilled by enterprises— pensions, medical care, and welfare. If these revenue concerns were the primary reason for adopting restrictive controls on the activities of financial intermediaries, then the inflation tax should have been used in conjunction with these restrictions in the transition countries, as the public-finance approach hypothesizes. A graphic illustration of the relationship between reserves in the banking system (an indicator of financial repression) and deficits between 1989 and 1996 reveals no clear-cut answer. The scatter¬plot (Figure 1) shows two “clusters” of countries, the first in the lower left-hand portion of the chart, the second farther to the upper-right. But within each of these clusters, one might also see a negative relationship between size of the deficit and reserve ratios. The lack of data impair a more systematic evaluation of this kind.

We examine some hypothesized relationships between financial restrictions and certain fiscal indicators as a second test of the general public-finance framework. The resulting correlations, based on annual data from 25 transition economies, are presented in table 1. Two results seem to confirm the public finance thesis: (i) that tax losses are associated with higher real reserve ratios in deposit-taking banks

5. Revenue collection in the classic socialist system was facilitated to a great extent by the unitary organization of the Party-state, and the relatively small number of tax instruments— namely, turnover, profit, and payroll taxes secured from a small number of large state-owned enterprises under the supervision of branch directorates. For a description, see Kornai, 1992

Finally, for the same thirteen countries, we list maximum and minimum annual real central bank discount rates for the period 1990-1996 in table 3. This table also lists the rank correlations between real discount rates and net government credit on the basis of quarterly IMF figures. In comparison to table 2 above, the correlations generated here are weaker

We propose an alternative approach to understanding financial regulation. Our aim here is to provide, as much as possible, a systematic analysis of the political economy of financial liberalization in the post-Communist region, explaining how different policy decisions are reached in different political settings.

Modeling the Politics of Financial Policy

Recent studies of how political institutions shape economic policy are of two views. In the first view, two features of political systems matter: stability and polarization. Unstable governments are claimed to behave more “myopically”, that is, discount the future at a greater rate than more cohesive systems, while polarized governments exacerbate the coordination problems inherent in adopting economic reforms and lead to protracted stalemates (Rodrik, 1993).

Governments that are more likely to be thrown out in future elections, and governments that are characterized by divisive and sharp disagreements between alternative policy makers, therefore, are more likely to delay the adoption of stabilization programs (Alesina and Drazen, 1991), and especially more likely to run higher deficits (Roubini and Sachs, 1989

A more recent, second view has challenged the argument that strong, centralized governments are needed for liberalization. This alternative view suggests that economic reforms are a by-product of struggles for political authority, and thus the major obstacle to economic liberalization is not the stalemate among groups fighting over the distribution of costs and benefits of reforms, but the internal opposition to such reforms within governments. Thus reforms are more likely to occur when political outsiders challenge the authority of incumbents (Bates and Kreuger, 1993

This second approach has been used to explain stabilization delays in the case of the transition economies. Hellman (forthcoming) examines the effects of party fragmentation, coalition structure, and executive power on the “survival” rates of inflationary periods, and finds

that those countries whose political institutions are characterized by fragmented parliaments, multi¬party coalition governments, and weak executives, stabilize faster than those countries with unified party structures, majority or single-party governments, and authoritarian executives. In the section that follows, we analyze the impact of different political structures on financial policy and financial liberalization, focusing on three main institutional features of legislatures in twenty-five transition

countries: the share of seats held by the Communist Party or its direct descendants, the degree of party fragmentation, and the degree of parliamentary support for the government in power.

The first measure we consider to be an approximation of anti-reform incumbent power, the second to be a measure of polarization, the third a measure of expected stability. Following Hellman, we consider the effect of each institutional feature on financial repression separately, in order to test specific, broader hypotheses about the relationship between the characteristics of political institutions and economic policies. The financial policies of transition countries have displayed a great deal of variation between 1989 and 1996 (World Bank, 1996

IV. DATA, MODELS, AND RESULTS Measuring Financial Repression

Financial repression is a reference to a specific set of policies involving a variety of controls on the activities of the main financial institutions in an economy. The analysis of such policies, however, does not easily lend itself to systematic study. First, financial restrictions may be implicit, imbedded in intricate tax codes or financial regulations. Second, there are also problems of coding such policies across countries in a way that is standard and comparable, especially when the countries in question lack precedents or conventions needed for “benchmarks”against which variation in such policies could be measured. Finally, a financially “repressed”economy may exhibit certain effects— low interest rates for financial transactions, distortions in savings and investment, and low levels of financial intermediation— that may or may not be attributable to those policies. In table 1 above, for example, there are several countries that maintain extremely high reserves in the banking system even if the legal reserve requirements are modest. In Hungary, Estonia, and the Czech Republic, real reserve ratios are 48%, 27%, and 20% respectively despite a common 10%

required-reserve ratio in each these countries.6 It is therefore desirable to consider a range of different indicators of financial repression.

Our solution is to examine three different “proxies”for financial repression and financial liberalization, two of which are policies, one of which measures a hypothesized effect of these policies. The first measure is directed credit, which we code as a binary variable. Following earlier studies, countries in which directed credits constitute more that 25% of the total credit volume in the economy are assigned a value of one

6. In several countries, banks often kept reserves in excess of what they were required to hold by law as hedge against poorly-enforced creditor rights— a consequence of, among other things, deficiencies in bankruptcy laws, lengthiness of court-adjudicated proceedings, and the limited variety of financial instruments that could yield reasonable risk-adjusted returns.

Denizer, 1997). Directed credit programs, used by former socialist governments to maintain employment levels in certain industries, represent a fairly severe restriction on the portfolio composition of banks’ balance sheets. Under the typical directed-credit scheme, banks are instructed to lend to certain sectors or enterprises a specified portion of their total assets or total lending in a certain time period. These quantitative restrictions, coupled with the interest-rate controls often attached to these loans, tend to segment financial markets in these countries, and constitutes a significant obstacle to their development.

The second dependent variable is the real discount rate used by central banks for their refinance operations. Since the discount or refinance rate is a policy measure used at the discretion of the central banks, however, and since it becomes the base rate for other rates, this is likely to be a better indicator of policy stance than deposit or lending rates. We calculate the real central bank discount rate as the nominal discount rate deflated by the price level (annualized).

Finally, we choose a third variable to measure the effect of financial repression on the level of financial intermediation. The most accurate way to gauge these effects would be to estimate a money¬demand curve at different periods in time in order to measure shifts in the curve, along with changes in cost and income elasticity (Fry, 1995: 21). Such an estimation is difficult in the context of the transition economies due to lack of observations and shortness of time-horizon. We use, instead, the ratio of money supply (defined to include currency in circulation, sight and time deposits, or M2) to GDP— a standard measure of financial “depth”that is consistently found to be higher in market economies (indicating that most transactions are intermediated within a formal financial system) and lower in financially repressed economies (Barro, 1991

Explanatory Variables and Estimation

We estimate these four variables using cross-sectional, time-series analysis with country¬years as units of observation. Such a pooling of data, in addition to increasing the degrees of freedom, is also more sensitive to the inter-temporal properties of the sample than would be period averages for each country. Ours is not, however, a standard panel as only observations for the year following parliamentary elections are selected, and our electoral variables are lagged once with respect to the dependent variables. Such matching is preferred to a standard panel given that the data exhibit limited variation between elections. In a standard panel this would create problems with standard errors, as one would be trying to approximate a continuous line with a crude stepwise function involving fairly long steps. The units of the time series in our panels, therefore, represent electoral periods rather than calendar years. Although we lose data in between elections for the dependent variables (and some independent variables), we maintain all the inter-temporal variation in the sample and eliminate most of the flat segments. (For an explanation of certain estimation problems and how they were dealt with, see appendix.)

Using the electoral database compiled in Hellman (forthcoming), we calculate three

indicators of parliamentary structure.7 Our first variable is the share of seats in parliament held by

7. A complete set of political-institutional indicators should also, naturally, include some measure of executive powers. In fact, Hellman (forthcoming) codes the transition countries according to degree of presidential and prime

members of the Communist Party, or its direct descendant or set of descendants (COMMP). A high percentage for COMMP can be interpreted, ceteris paribus, as an indicator of several things, including “insider”control of the legislature, low development of a party system, lack of political succession in parliament, and so on. In general, then, we consider COMMP to measure the degree of “persistence”of pre-reform elites in legislatures. To measure the degree of polarization in parliaments, we use the Rae fractionalization index (FRACTION), which measures the likelihood

that two legislators chosen at random belong to different political parties.8 Inevitably, the FRACTION panel suffers from two defects: it does not control for the sometimes large number of independent members of parliament, nor does it include “extra-electoral”changes in party affiliation or party unity. Where parties splintered between elections, the increased Rae index is

reflected in the next electoral period.9 Finally, we calculate level of parliamentary support for the government by simply adding up all seats that the government coalition can claim, divided by the total number of seats in parliament (GOVSTR). This measure may be interpreted as a proxy for government stability, as low values indicate a higher probability, ceteris paribus, that the government in power may be removed at some future date.

We hypothesize that the influence of electoral changes should be the strongest within a finite period following an election. Given that elections take part in different months of the year, and that there is a necessary “transition”period of a few months, during which changes in the composition of parliaments and government produced by elections occurs, the year of election would not be the best choice for the independent variables. After a first full year, moreover, future

ministerial strength, using the methodology in Shugart and Carey (1992). We included these variables in our estimations, but in no specifications did they turn out to be significant, either independently, or in conjunction with the parliamentary¬structural variables.

Install Equation Editor and double

click here to view equation.

8. The Rae index is defined as:

where pi is the fractional share of seats of the i-th party (Rae, 1967).

9. As Hellman (forthcoming) notes, this is most commonly the case in countries where broad anti-Communist reform movements won certain critical elections (held in 1990 in Czechoslovakia, Estonia, Latvia, Lithuania Macedonia, Moldova, Romania, Slovenia, 1991 in Russia), but splintered shortly thereafter. In most of these countries, the Rae index rises following the next electoral period, which occurs within two years in all cases except Latvia (three years) and Macedonia (four years). Only in Lithuania (between 1990 and 1992) and Macedonia (between 1990 and 1994) does the fractionalization number fall. Certainly this may hide— or even worse, reverse— some of the inter-temporal variation that occurs, but there is simply no reliable way of accounting for these changes. Part of this problem, however, is solved through our use of “matched”observation panels.

electoral considerations may change politicians’ behavior. For these reasons, all of our relevant political variables are lagged once with respect to the dependent variable.

Several conditioning variables have been included to control for certain factors in the regressions (see appendix for data sources): the logarithm of per-capita GNP (based on the World Bank Atlas method), on the premise that richer nations repress their financial systems less

structural similarities that former Soviet republics may share.10 In the regressions for financial depth and the real discount rate, we include the logarithm of the growth in base money to control for currency substitution and portfolio switching that is expected during periods of rapid money growth. and attendant inflation. In the directed credit regression, we include Freedom House’s Freedom Index, which ranks countries according to their protection of civil and political liberties, in

order to test the effect of general regime “openness”on state-enterprise relations.11 With the exception of per-capital income (which is lagged once), all conditioning variables are contemporaneous to the dependent variable.

In sum, our regressions take the following basic formats:

Install Equation Editor and double¬click here to view equation.

Install Equation Editor and double¬click here to view equation.

where _i,t is any relevant measure of, alternatively, financial repression or liberalization, Ci,t is a vector of contemporaneous conditioning variables, POLi,t-1 is any relevant lagged measure of parliamentary structure, ß0 is a constant, ß K for K>0 is a coefficient (or vector of coefficients) on each variable (or vector), ei,t is a random disturbance term, and where i=1,…,N, t=1,…,T for N countries and T time periods. To gauge differences in the effects of the political variables between those countries of the former Soviet Union and those countries in Eastern Europe, we test a second model which disaggregates POLi,t-1 into two interactive terms, one with an East-Europe dummy variable (EE) another with the Post-Soviet dummy (FSU). All regressions— with the exception of the directed credit estimation— were performed using ordinary least squares (OLS).

Results

Tables 4, 6, and 7 present our regression results for each of the dependent variables. In table 4 we report the results for directed credit— generated using the probit method, by which we

10. Following convention, we did not code the Baltic states as “Post Soviet,”on the grounds that this might be an artificial way of inducing intra-Soviet variation into the results. Coding them 1, however, does not alter our results significantly.

11. In this sense the Freedom Index is a proxy for “access”to decision making authority, and its inclusion controls for differences across countries.

estimate the probability a given country will implement a directed credit program. In our first

model, both the Freedom Index12 and Communist share of seats are significant. Note that, by itself, COMMP is significant and has the expected (positive) sign. When we split this variable between former Soviet and East European countries we see that only the Post-Soviet, interactive coefficient is significant. This is so even with the inclusion of the Post-Soviet dummy, suggesting a robust result. In the next step we analyze fractionalization, which carries a significant, negative sign, indicating that fractionalized parliaments are less likely to implement directed credit programs. Fractionalization also has a differential impact between Soviet and non-Soviet states, as it appears that party fragmentation does not increase the probability of having a directed program in East European countries while it does in the former Soviet Union. When we performed similar steps with the government strength variable we found that this variable did not have explanatory power.

For the remaining two regressions, therefore, we examine “real”measures— the ratio of money supply to GDP and the central bank real discount rate. We begin with some averages in table 5. For each parliamentary measure we took averages for each country over the entire available time period. Next we split the sample into “high”or “low”explanatory categories depending upon whether the country fell above or below the median. Finally we calculated the corresponding M2/GDP ratio and real discount rate averages for the appropriate sample of countries over the same time period. These values are listed in table 5, along with the individual countries in their respective groups. From this rough picture, we see that a high percentage of seats held by the Communist Party, a high level of party unity (low fractionalization) in parliament, and a high degree of government support, all correspond to lower real discount rates, but moderately higher levels of financial depth

To test these relationships more systematically, we focus first on the M2/GDP ratio, and perform the same series of tests as we did for directed credit and liberalization. We first specify, however, an equation that may be considered a test of an “economic”model, and thus include only the conditioning variables and the Post-Soviet dummy. As shown in table 5, only Soviet membership turns out to be significant (with the expected negative sign), while per-capita income, fiscal balance, and base money growth are not. Following the same steps as in the first two regressions, we individually add variables for Communist Party share of seats, fractionalization, and government strength, decomposing each measure between East-European and Post-Soviet regions.

With the exception of government strength, the coefficient on each measure is significant individually. Note that the statistically significant coefficients for FRACTION and COMMP carry the signs we expect from table 5. Indeed, according to our results, countries with legislatures having a greater percentage of Communists tend to have higher M2/GDP ratios than countries that do not. Upon splitting this variable, it remains significant for the FSU only, although the positive result is still surprising, especially when the Post Soviet dummy remains significantly negative. This would suggest that former Soviet countries as a whole tend to have lower M2/GDP ratios, but that

12. Note that in the Freedom Index, freedom is measured on an inverted scale, with lower values implying greater political freedom.

within this region, having a larger group of Communists in parliament has a positive effect on the ratio. The same is not true of the East European countries, where Communist share of seats has no significant effect. Hence, using a dummy variable for the former Soviet Union, as is conventionally done, may not be a useful device

Finally, we turn to the central bank’s real discount rate, presented in table 6. Our baseline “economic”model includes the identical conditioning variables used in the previous set of regressions: logarithms of per capita income and base money growth, fiscal balance as a proportion of GDP, and a dummy for the former Soviet countries. In this specification, only base money growth is significant. Base money growth, moreover remains strongly negative for all our specifications, indicating that an increase in this term— that is, an increase in the rate of base money growth— is correlated with a lower real discount rate, as should be expected. Adding, individually, COMMP, FRACTION, and GOVSTR reveals a common pattern. By themselves, each variable carries significant coefficients with the signs we expect from table 4. That is, higher shares of seats held by Communists and former Communists, greater unity of parties, and greater government strength, all lead to lower real discount rates.

When split between East Europe and the former Soviet Union, these effects are significant for the latter (the exception is government strength, which is significant in interaction with both regional dummies). Note again two surprising findings about the former Soviet Union. First, the Post-Soviet dummy, in each specification where it is significant, carries a positive sign. Surprisingly, therefore, this suggests that former Soviet countries have managed to raise their real discount rates above that which has been set in non-Soviet countries. Second, in two of the three equations where the Post-Soviet dummy is significant (equations 3 and 7) the interactive coefficient carries the opposite sign to the Post-Soviet coefficient. While former Soviet countries may set higher discount rates, within these countries party greater party unity and government strength lead to lower rates. Again, this result suggests that there is a substantial amount of variation in monetary policy within the former Soviet Union that is not captured by a single grouping. Ironically, this last set of regressions leaves the fewest degrees of freedom, but produces the best overall fit.

These results do not lend much support to the thesis that financial liberalization will be less in politically unstable or polarized systems. The regressions shown here, in fact, suggest that overall financial repression is greater in countries with higher percentages of Communists in parliament, less party fractionalization, and greater government support. We see that COMMP and FRACTION are significant for all three dependent variables, GOVSTR for two of them. Moreover, with the exception of our estimation of financial depth, higher COMMP values, lower FRACTION values, and higher GOVSTR values are correlated with a higher degree of financial repression.

The role of deficit financing is less conclusive. Certainly, in some specifications, fiscal balances are significant

liberalization index, fiscal balances were correlated with the dependent variable. But in all four equations, the sign of the coefficient is the opposite of what is expected: larger positive fiscal balances are correlated with lower levels of financial depth, and with lower degrees of overall liberalization. Moreover it is interesting to note that, in these four equations, only after the GOVSTR variable is added does the government’s fiscal balance generate a significant negative correlation, suggesting that the effect of deficits on liberalization and financial liberalization is, to a certain extent, conditioned by the strength of the government. One must allow for the possibility that, when one controls for the strength of the government (measured by its parliamentary support), deficit-spending governments are more likely to liberalize, and more likely to develop their financial systems. When we control for the other political variables, deficits have no effect. Indeed, we find only two estimations which support the public-finance thesis. In two regression equations for the real discount rate, positive fiscal balances were correlated with higher real discount rates. But again, when we control for the strength of the government, the significance of the fiscal balance coefficient disappears.

V. CONCLUSION: GOVERNMENTS AND FINANCIAL MARKETS

This paper has presented an exploratory analysis of the political correlates of financial repression in the transition economies. We examined some preliminary evidence for the thesis that, in the face of excessive costs for certain forms of taxation, governments will levy an implicit tax on domestic financial markets. Such claims were not borne out fully in the data. We suggested that the unique lineages of the socialist financial system leave this public-finance framework with limited applicability to post-Communist economies.

As an alternative, we outline a rudimentary approach that examines the effects of the dispersion and concentration of political power in governments on financial policy. We find that, for four separate proxies of financial repression, legislatures with larger proportional numbers of Communists, with less party competition, and with less governmental opposition, tend to extract rents from financial markets in the form of repressive controls. The most reliable predictors of whether governments in transition economies will liberalize financial markets are degree of Communist party control and degree of parliamentary polarization, both of which we found to be significant in most of our specifications. As expected, Communists and their descendants are the most opposed to financial liberalization. But the surprising finding is that polarized, fragmented parliaments are more likely to liberalize and develop financial systems than those that are more unified. Finally, government stability, where it was significant, was found to inhibit the liberalization of financial markets. Our evidence seems to support the view of policy reform predicting that reform periods will occur as politicians struggle to enhance their authority and snatch the levers of fiscal and monetary control away from pre-reform elites.

These findings suggest that repressive financial controls may be adopted not to finance deficits more cheaply than would be the case under financial liberalization, but to maintain the authority and ensure the survival of those in power. In those countries where pre-reform elites are plentiful in legislative bodies, where inter-party competition is low, and where governing parties are well-represented in parliaments, elites have been able to perpetuate a system of implicit subsidies by “softening up”the financial sector— particularly the commercial banks— in order to assure the

continued flow of cheap credit to specific borrowers. The main beneficiaries of these policies— large formerly state-owned industries with tight financial links to the largest commercial banks— are thus able to convert their well-established claims on public resources into preferential access to credit lines..

If this is the case, then financial repression serves a special role in transition economies, namely, a mechanism for solidifying main-bank, main-firm relations. Through a combination of partial state ownership of financial institutions and interest-rate controls, governments assure that commercial banks will maintain the largest enterprises as their chief clients, even once the cash flows of those enterprises have been privatized. In general our results lend some support to the common claim of smaller, cash-starved Eastern European entrepreneurs that the commercial banks have “taken over the role of the old planning ministries.”

Our results, finally, revealed a wide degree of variation in the financial policies of the post-Soviet republics, and cast some doubt on the usefulness of generalizations about the former Soviet Union. In addition, one of our more surprising findings was that, in the post-Soviet republics, governments in which Communists are well-represented tend to have “deeper”, more developed financial systems than the former Soviet countries in which Communists are scarce

If this is true, it suggests a possible explanation for the cross-regional differences between the formerly socialist countries and, for example, Latin America or other developing countries in which financial reforms took place after the consolidation of political power, not before— if those accounts are true. It is likely that the conflicts reformers face in setting financial policy— and economic policy in general— are conditioned by the peculiar political settings inside which they operate, in particular, the relationships between anti-reform elites and existing financial institutions. In the transition economies, the heavy concentration of pre-reform elites in the Communist Party¬state apparatus, including the planning bureaucracy and the public-enterprise sector, consisted of extremely tight relationships between these institutions and a few commercial banks. In such a setting, those forces leading to a general dispersion of political authority— away from a single party, away from the central government, and so on— would more likely lead to successfully implemented financial reforms.

By contrast, in countries were the power of anti-reform elites was more decentralized— fragmented along regional, ethnic, or linguistic lines, or where the sources of elite authority were more closely connected to property ownership— and were financial institutions were regionally¬based or specially linked to local industries or landholdings, there it may have been necessary for the opposite to occur, that is, for political authority to be centralized before financial reforms could be implemented. We can only suggest this possibility here, to mention it briefly, leaving a more systematic exploration to future research.

APPENDIX

Variables, Definitions, Measurement, and Sources (Dependent variables in bold)

Variable

Mean (Std. Dev.) Definition

Directed Credit Dummy variable for

0.479 (0.503) directed credit programs

Real Discount Rate Base rate charged to Discount rate n deflated by Same as above

-0.350 (0.412) financial institution, set annualized change in price level

at the discretion of the p, in percent:

central bank (1+n/1+p)-1

COMMP

0.383 (0.311)

FRACTION 0.601 (0.271)

GOVSTR 0.649 (0.192)

Communist Party control in legislature

Index of party fractionalization

Government strength

Share of seats in parliament held by members of the Communist Party or its descendant(s), in percent. (lagged)

Herfindahl index of shares of seats held by all parties in parliament, subtracted from unity.

(lagged)

Share of parliamentary seats held by party or parties represented in governing coalition, in percent.

(lagged)

Calculated on the basis of data and in Hellman (forthcoming), updated by the authors

Same as above

Same as above

1 if country was a constituent Authors’ coding republic of the USSR (with the

exception of the Baltic states),

otherwise 0 (EE coded in the

reverse).

Freedom Index Index of political Ranks countries from 1 Freedom House, Freedom in

7.466 (3.211) freedom (most free) to 14 (least free) the World (various issues) (contemporaneous)

Log (GNP/Capita), where GNP is in US$, based on World Bank Atlas method.

(lagged)

Log (1+? M/M),

where M is base money. (contemporaneous)

Annual surplus/deficit as percent of GDP. (contemporaneous)

World Bank, World Development Indicators

IMF, International Financial Statistics (various issues), IMF staff country reports

IMF, Government Financial Statistics (various issues), IMF staff country reports

Pooled Estimation with “Matched” Observations

Our technique of matching of independent and dependent variables limits the time-series, and thus places certain limits on robustness tests. First, we are forced to ignore any possible contemporaneous correlation between the same variables for different countries. This would be a

problem in a standard N¥T panel for these countries, because N (the number of units) is 25, while T (the number of time periods) varies between 4 and 6, making it difficult to estimate contemporaneous correlations. Second, it becomes difficult to test for serial correlation in limited T panels. We experimented estimating a single first-order autocorrelation coefficient by a maximum likelihood method, but it did not turn out to be significant in any specification. Third, it is also difficult to test for so-called ‘panel heteroskedasticity’, where observations for the same unit are assumed to be homoskedastic, but different units may have different variances (Beck and Katz, 1995). As an admittedly imperfect substitute, we applied the White and the Breusch-Pagan tests for heteroskedasticity. Note that these tests are more general, because their alternative hypothesis is that every observation will have different variances

REFERENCES

Agénor, Pierre-Richard and Peter J. Montiel (1996). Development Macroeconomics. Princeton, N .J.: Princeton University Press.

Alesina, Alberto and Allan Drazen (1991). “Why are Stabilizations Delated?” American Economic Review 81(5): 1170-1188.

Alesina, Alberto and Howard Rosenthal (1995). Partisan Politics, Divided Government, and the Economy. Cambridge: Cambridge University Press.

Alt, James E. and Robert C. Lowry (1994). “Divided Government, Fiscal Institutions, and Budget Deficits: Evidence from the States.” American Political Science Review 88(4): 811-828.

Amsden, Alice H. (1989). Asia’s Next Giant: South Korea and Late Industrialization. New York: Oxford University Press.

Barbone, Luca and Domenico Marchetti, Jr. (1995). “Transition and the Fiscal Crisis in Central Europe.” Economics of Transition 3(1): 59-74.

Barro, Robert J. (1991). “Economic Growth in a Cross-Section of Countries.” Quarterly Journal of Economics 106: 407-43

Bates, Robert H. and Anne O. Kreuger (1993). “Generalizations Arising from the Country Studies,”in Robert H. Bates and Anne O. Kreuger, eds., Political and Economic Interactions in Economic Policy Reform. Cambridge, Mass.: Blackwell.

Beck, Nathaniel and Jonathan N. Katz (1995). “What to Do (and Not to Do) with Time-Series-Cross-Section Data in Comparative Politics.” American Political Science Review 89(3): 634-47

Begg, David K. H. (1996) “Monetary Policy in Central and Eastern Europe: Lessons After Half a Decade of Transition.” IMF Working Paper 96/108, International Monetary Fund, Washington, D.C.

Bencivenga, Valerie and Bruce D. Smith (1992). “Deficits, Inflation, and the Banking System in Developing Countries: The Optimal Degree of Financial Repression.” Oxford Economic Papers 44(4): 767-790.

Brock, Philip L. (1989). “Reserve Requirements and the Inflation Tax.” Journal of Money, Credit, and Banking 21(1): 106-121.

Buffie, Edward F. (1984). “Financial Repression, the New Structuralists, and Stabilization Policy in Semi-Industrialized Economies.” Journal of Development Economics 14(3): 305-322.

Campos, Ed and Sanjay Pradhan (1996). “Budgetary Institutions and Expenditure Outcomes.” Policy Research Working Paper 1646, The World Bank, Washington, D.C.

de Melo, Martha, Cevdet Denizer, and Alan Gelb (1996). “From Plan to Market: Patterns of Transition.” World Bank Economic Review 10(3): 397-424.

de Melo, Martha and Alan Gelb (1997). “A Comparative Analysis of Twenty-Eight Transition Economies in Europe and Asia.” Post-Soviet Geography and Economics 37(5): 265-285.

de Melo, Martha and Cevdet Denizer (1997). “Monetary Policy during Transition: An Overview.” Policy Research Working Paper 1706, The World Bank, Washington, D.C.

Desai, Raj M. and Katharina Pistor (1997). “Financial Institutions and Corporate Governance,”in Ira W. Lieberman, Stilpon S. Nestor, and Raj M. Desai, eds., Between State and Market: Mass Privatization in Transition Economies. Washington, D.C.: The World Bank.

Díaz-Alejandro, Carlos (1985). “Goodbye Financial Repression, Hello Financial Crash.” Journal of Development Economics 19(1-2): 1-24.

Easterly, William (1993). “How Much do Distortions Affect Growth?” Journal of Monetary Economics 32(4): 187-212.

Edwards, Sebastián (1984). “The Order of Liberalization of the External Sector in Developing Countries.” Princeton Essays in International Finance 156, Princeton University, Princeton, N .J.

Fry, Maxwell J.(1995). Money, Interest, and Banking in Economic Development, 2nd Ed. Baltimore: Johns Hopkins University Press.

— — — (1982). “Models of Financially Repressed Developing Economies.” World Development 10(9): 731-750.

Garvy, George (1977). Money, Financial Flows, and Credit in the Soviet Union. Cambridge, Mass.: Ballinger.

Geddes, Barbara (1994). “Challenging the Conventional Wisdom.” Journal of Democracy 5(October): 104-118.

Giovannini, Alberto and Martha de Melo (1993). “Government Revenue from Financial Repression.” American Economic Review 83, (4): 953-963.

Grilli, Vittorio, Donato Masciandro, and Guido Tabellini (1995). “Political and Monetary Institutions and Public Financial Policies in the Industrial Countries,”in Torsten Persson and Guido Tabellini, eds., Monetary and Fiscal Policy, Vol. 2. Cambridge, Mass.: MIT Press.

Haggard, Stephan, Chung H. Lee, and Sylvia Maxfield, eds. (1993). The Politics of Finance in Developing Countries. Ithaca, N.Y.: Cornell University Press.

Haggard, Stephan and Steven B. Webb (1993). “What Do We Know About the Political Economy of Economic Policy Reform?” World Bank Research Observer 8(2): 143-168.

Haggard, Stephan and Robert R. Kaufman (1995). The Political Economy of Democratic Transitions. Princeton, N .J.: Princeton University Press.

Hall, Peter A. and Robert J. Franzese, Jr. (1996). “Mixed Signals: Central Bank Independence, Coordinated Wage-Bargaining, and European Monetary Union.” Unpublished.

Hastings, Laura (1993). “Regulatory Revenge: The Politics of Free-Market Financial Reforms in Chile,”in Stephan Haggard, Chung H. Lee, and Sylvia Maxfield, eds., The Politics of Finance in Developing Countries. Ithaca, N.Y.: Cornell University Press.

Hellman, Joel S. (forthcoming). “Competitive Advantage: Political Competition and Economic Reform in Post-Communist Transitions.” British Journal of Political Science.

International Monetary Fund (1995). “Weakening Performance of Tax Revenues.” Economic Systems 19(2): 101-124

Iversen, Torbin (1996). “The Real Effects of Money.” Paper presented to the Political Economy of Integration Study Group, University of California, Berkeley.

King, Robert G. and Ross Levine (1993a). “Finance and Growth: Schumpeter Might Be Right.” Quarterly Journal of Economics 108(3): 717-738.

— — — (1993b). “Finance, Entrepreneurship, and Growth: Theory and Evidence,”Journal of Monetary Economics 32(3): 513-542.

Kornai, János (1992). The Socialist System: The Political Economy of Communism. Princeton, N .J.: Princeton University Press.

Lanyi, Anthony and Rü_dü Saraco_lu (1983). “Interest Rate Policies in Developing Countries.” Occasional Paper 22, International Monetary Fund, Washington, D.C.

Levine, Ross (1996). “Financial Development and Economic Growth.” Policy Research Working Paper 1678, The World Bank, Washington, D.C.

— — — (1993). “Financial Structures and Economic Development.” Revista de Análisis Economico 8(1): 113-129.

Lukauskas, Arvid J. (1994). “The Political Economy of Financial Restriction: The Case of Spain.” Comparative Politics 27(1): 67-89.

McKinnon, Ronald I.(1973). Money and Capital in Economic Development. Washington, D.C.: Brookings Institution.

Nelson, Joan M. (1990). “The Politics of Adjustment in Developing Nations,”in Joan M. Nelson, ed., Economic Crisis and Policy Choice. Princeton, N .J.: Princeton University Press.

Pradhan, Sanjay (1996). “Evaluating Public Spending: A Framework for Public Expenditure Reviews.” Discussion Paper 323, The World Bank, Washington, D.C.

Quinn, Dennis P. and Carla Inclán (1997). “The Origins of Financial Openness: A Study of Current and Capital Account Liberalization.” American Journal of Political Science 41(3): 771-813.

Rae, Douglas (1967). The Political Consequences of Electoral Laws. New Haven, Conn.: Yale University Press.

Rodrik, Dani (1993). “The Positive Economics of Policy Reform.” American Economic Review 83(2): 356-361.

Roubini, Nouriel and Jeffrey D. Sachs (1989). “Political and Economic Determinants of Budget Deficits in the Industrial Democracies.” European Economic Review 33(x): 903-933.

Roubini, Nouriel and Xavier Sala-i-Martin (1995). “A Growth Model of Inflation, Tax Evasion, and Financial Repression.” Journal of Monetary Economics 35(2): 275-301.

— — — (1992). “Financial Repression and Economic Growth.” Journal of Development Economics 39(1): 5-30.

Shaw, Edward S. (1973). Financial Deepening in Economic Development. New York: Oxford University Press.

Shugart, Matthew S. and John M. Carey (1992). Presidents and Assemblies: Constitutional Design and Electoral Dynamics. New York: Cambridge University Press.

Stiglitz, Joseph (1994). “The Role of the State in Financial Markets.” Proceedings of the Annual World Bank Conference on Development Economics. Washington, D.C.: The World Bank.

(1989). “Financial Markets and Development.” Oxford Review of Economic Policy 5 (October): 55-68.

Tanzi, Vito (1993a). “Fiscal Policy and the Economic Restructuring of Economies in Transition.” IMF Working Paper 93/22, International Monetary Fund, Washington, D.C.

— — — (1993b). “Financial Markets and Public Finance in the Transformation Process,”in Vito Tanzi, ed., Transition to Market: Studies in Fiscal Reform. Washington, D.C.: International Monetary Fund.

Taylor, Lance (1983). Structuralist Macroeconomics: Applicable Models for the Third World. New York: Basic Books.

Tobin, James (1965). “Money and Economic Growth,”Econometrica 33(4): 671-684.

Van Wijnbergen, Sweder (1985). “Macroeconomic Effects of Changes in Bank Rates: Simulation Results for South Korea.” Journal of Development Economics 18(2-3): 541-554.

— — — (1983a). “Credit Policy, Inflation, and Growth in a Financially Repressed Economy.” Journal of Development Economics 13(1-2): 45-65.

(1983b). “Interest Rate Management in LDCs.” Journal of Monetary Economics 12(3): 433-452.

Wade, Robert (1990). Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization. Princeton, N .J.: Princeton University Press.

Williamson, John, ed. (1994). The Political Economy of Policy Reform. Washington, D.C.: Institute for International Economics.

World Bank (1996). World Development Report: From Plan to Market. Washington, D.C.

(1989). World Development Report: Financial Systems in Developing Countries. Washington, D.C.