Financial Reforms in Eastern Europe

Kanhaya Gupta, Robert Lensink

• Publisher: Routledge
• Year: 2006
• Print ISBN:
  9780415166683
  9780415166683
• Edition: 1
• eBook ISBN:
  9781134706617
• Topic: Business
• Subtopic: Business General
• Index: Business

This book presents a model which simulates the effects of financial reforms in transitional economies, which is then applied to Poland for a variety of policy simulations. The authors develop models for households, commerical banks and firms, expanding their enquiry into the government sector, the central banking sector, the external sector and finally the supply side. These sub-sector models explicitly incorporate institutional features specific to the Polish economy. The estimated model is used to simulate the effects of a wide array of financial policies introduced in Poland, and these results are then used to assess the effectiveness of the policies analyzed.
This timely and authoritative study sheds new light on how a country's overall economic system responds when it pursues a 'big-bang' approach to marketization.

1
INTRODUCTION

Ever since the demise of the former Soviet Union, the economies of Central and East European countries have drawn considerable attention. This is because of their attempts to move away from the highly centralized systems to a free market regime. While many other countries have also been trying to move in that direction, these economies constitute a special case because of their centralized structures. In a way they offer almost a controlled experiment going from one extreme to another. And no country exemplifies such an attempt more than Poland. This is because of its adoption of the so called “big bang” approach when Solidarity first came to power in 1989. While it would be interesting to analyze not only all of the countries of the former Soviet empire and also all policies designed to achieve the desired market orientation in these countries, that would be a task beyond any single volume. Our aim is more modest and restrictive. What we hope to achieve in this short monograph is to confine ourselves to three countries: namely, the Czech Republic, Hungary and Poland as a background and then concentrate on Poland. However, while the exercise is specifically carried out with reference to Poland, our hope is that the model proposed and the estimates and the data series constructed are prototypes of the transition economies and thus have broad applications in this region.

Even for Poland, our interest is narrowly defined. We are entirely concerned with the implications of financial reforms and their effects on the Polish economy. This particular aspect of economic policies, both in Poland and other transition economies, has attracted considerable attention, as will become evident in the chapters to follow. In this introduction, we briefly explain what we want to do and provide the outlines of the chapters to follow.

As pointed out above, Poland is a fascinating case study because of the fundamental break from the past in 1989. The break was so radical that one could not really appeal to its past to examine its current economic policies. This poses both analytical and practical challenges. Analytically, while the temptation to model the Polish economy along the lines of standard neoclassical lines is enormous, that would be a mistake because of its special institutional features and its very rapidly changing environment, not only economic, but also institutional in the widest sense of the word. The data problems are formidable. Most of the new series that are relevant to our study start in 1990 or 1991, so that on an annual basis there is virtually nothing one can do. Even using quarterly data we would not have enough observations for any meaningful estimates. Therefore, the only option is to use monthly data. Here, while monthly data on some of the financial variables are generally available, such is not the case on virtually any of the real variables. The preparation of the appropriate data series was thus a major task. Therefore, we consider the data Appendix to be a major contribution of this study. We also provide a relatively comprehensive bibliography in the area.

It is not a ground-breaking observation to say that for a proper understanding of the role of financial reforms in the Polish—or for that matter in any—economy we should construct some sort of a general equilibrium model, which can at least clearly identify some of the mechanisms through which financial reforms of various types affect different aspects of the economy concerned. It is also widely accepted by now that single equation reduced form approaches are not very enlightening because they reveal virtually nothing about the mechanisms via which any given variable affects the dependent variable concerned. But, as we shall see later on, precisely such models have been widely used in the analysis of various types of policies in the transition economies. This is not the approach used in this study. Instead, we follow a general equilibrium approach. However, the term “general equilibrium” is used more in the nature of macro models rather than the detailed CGE type models.

The basic approach of this study then is as follows. We distinguish five economic agents: households, commercial banks, firms, the central bank and the central government. We then use the balance sheet of each of these agents and model its components. The modelling of the components is consistent with the requirements imposed by the balance sheet. Since the assets and the liabilities in each balance sheet also appear in one or more of the other balance sheets, the cross balance sheet requirements are also imposed. This allows us to design an internally consistent model, which allows for interactions between the behaviour of the five agents identified above. Once the behaviour of these five agents has been modelled, we move on to the supply side as well as the external side of the economy. The model introduces nominal wage and price rigidity. A rationale for these features is offered. In the specification of the models for the three private agents, namely, the households, the commercial banks and the firms, we follow a uniform approach. That approach consists of the Tobin type portfolio selection approach. It has not only the virtue of providing us with a unified framework, but also provides a very rich menu of mechanisms through which financial variables may affect the real and the nominal variables. However, within this framework special consideration is given to features specific to the Polish economy. In the assets and the liabilities modelled for the three agents identified above, we have carefully followed the ones identified in the official Polish data rather than impose any prior categorization. This has been very instructive because the composition of the various portfolios in Poland has undergone radical changes over the six years studied and our disaggregation allows us to capture some of these changes. Because of data constraints and the need to keep the model both understandable and useful, we have confined ourselves to modelling the effects of those aspects of financial policies which are being used most often by the central bank and the commercial banks in Poland, although the model we propose is flexible enough to allow for further extensions.

It is important to point out that we are not interested in building a model for its own sake, but rather so that it can be estimated and then used for policy analysis. In short, the parameterization is not based on guestimates or borrowed values, but on actual estimates of the model using Polish data. In estimation and specification of the complete model, we follow a sequential approach. That is, we first specify and estimate models for each chapter separately, then we bring all of the components together and test for consistency as defined above. The robustness of both the sectoral as well as the entire model is checked by ex post forecasts of the dependent variables. For illustrative purposes, we also report policy simulations for each sector. This may be construed as a partial equilibrium approach. These experiments have some merit in themselves in that they are derived, not from some reduced forms or single equation models, but from the portfolio approach, so that they provide some check on the suitability of the approach to the policy issues being examined. A detailed outline of the chapters now follows.

Chapter 2 sets the stage with a brief review of the major economic changes in the Czech Republic, Hungary and Poland since 1990, as well as summaries of some of the work done to explain those changes. This chapter tries to highlight the important changes that need explanation.

Chapter 3 models the behaviour of the Polish commercial banks. Following the data given in the various issues of the bulletin of the central bank (henceforth called NBP), it first identifies the main assets and liabilities. Their portfolio then consists of the following assets: loans to firms; loans to households; net foreign assets; excess reserves; and government securities, separately in the form of T-bills and treasury bonds. The liabilities consist of borrowing from the NBP.

The model in this chapter tries to explain the observed changes in the portfolio composition. This is done by specifying a Tobin like portfolio model. Using the budget constraint for the commercial bank, the adding-up and the symmetry restrictions are derived. These restrictions are used in the estimation procedure. An important feature of the estimates reported is that the interest rates, wherever relevant, are the real rates and net of taxes. Furthermore, an important feature of the commercial banks in Poland is the very high default rates of loans to firms. These high rates imply a lower loan rate. We pay special attention to the modelling of this feature. Effectively, this means specifying loan rates as being net of the default rate. The “fit” of the estimated model is judged by comparing the within-sample dynamic forecasts with the actual values. Finally, the chapter reports some simulations regarding the effects of changes in the various rates of return.

Chapter 4 models the household sector. This chapter tries to build on the model presented in Gupta and Lensink (1996). There are two distinctive features of this model. The first is that it treats the decision to consume and the portfolio decision as being simultaneously determined rather than sequential, which is the common practice. As Gupta and Lensink (1996) explain, this simultaneity has far-reaching implications for examining the effects of financial policies. The second feature is the incorporation of the financial balance sheet of the Polish households into the model, just as we do in the case of the commercial banks. A very special feature of the household portfolio in Poland is the role that foreign currency-denominated deposits, both demand and time, play. To ignore this feature would greatly distort any analysis of the Polish economy as far as modelling the effects of financial policies is concerned. The main assets for this sector then are: currency; zloty demand and time deposits; foreign currency demand and time deposits; and net foreign assets. On the liabilities side, there is borrowing from the commercial banks.

Since the portfolio behaviour and consumption behaviour of this sector are modelled as being simultaneous, particular attention is paid to defining the income of the households. This, as can be seen from the next chapter on firms, has implications for their behaviour. The basic model in this chapter is the same as in Chapter 3 for the commercial banks. In the specification of the equations for the various assets and consumption, particular attention is paid to the role of credit constraints on the households and to the role of taxes and inflation. Quite apart from the fact that it is reasonable to assume that households react to after-tax real rates of return and not to gross nominal rates, this kind of modelling allows financial policies to affect real behaviour in interesting ways. The issue of adding-up and symmetry restrictions is carefully addressed.

The estimated model is once again used for two purposes. First, to examine whether it fits the data adequately. This is done, once again, by performing within-sample dynamic forecasts. Second, a number of illustrative partial equilibrium policy simulations are carried out to check the effects of changes in the various rates of return on the portfolio composition of the households.

Chapter 5 is devoted to the behaviour of the Polish firms. The basic approach here is the same as in Chapters 3 and 4. Therefore, we proceed by first identifying the components of their balance sheet and then by using a similar portfolio framework, specifying a model explaining the behaviour of various assets and liabilities. As explained in Chapter 2, one of the more popular but contentious issues in Poland, when it comes to firms, is the role of credit and its cost. We pay particular attention to this issue, by specifying the role of commercial bank loans and their cost in the various assets and liability equations, including the investment equation. This allows for a much more comprehensive and integrated role for credit. Not only does it allow for the direct effects of credit constraints on investment, but it also allows for indirect effects via changes in the entire portfolio composition of the firms. These changes may well be important if firms hold significant other assets: say, for example, foreign assets or government bonds. The basic approach of the rest of the chapter is the same as in Chapters 3 and 4 as far as estimation, within-sample forecasts and partial policy simulations are concerned.

Chapter 6 builds a complete version of the model. For this model we bring together the three sub-models developed in Chapters 3, 4 and 5 and supplement them with the behaviour of the other agents identified above. More specifically, we deal first with the behaviour of the Polish central bank (NBP). Our basic approach in this case also remains the same. That is, we start with its balance sheet and identify the main assets and liabilities. As a preliminary, we examine recent changes in the balance sheet. We then look at some of the factors underlying the observed changes. Next, we look at the behaviour of the central government. In this case, we proceed by examining the behaviour of the government with respect to the financing of its expenditures. Since the Polish government, like most others, systematically runs budget deficits, it must find means of financing it. The major issues that have characterized the debate in Poland on this issue have revolved around the role of the central bank and that of the commercial banks in financing budget deficits. At another level, it has revolved around financing by shortterm T-bills or long-term bonds, regardless of who buys them. We then specify the budget constraint of the government. Since the assets and liabilities in this case appear elsewhere in the four other balance sheets, we specify appropriate consistency conditions. Assuming constant prices (fixed price model) and ignoring the foreign sector, this gives us a self-contained model that can be solved for all the endogenous variables.

But we do not assume away either the external sector or the supply side. The external sector is modelled quite simply. The exchange rate in Poland is based on a “crawling peg” system and we model its behaviour relatively simply. We do the same thing for the net trade balance. Modelling supply side was quite a bit problematic, because of the difficulties in constructing data on capital stock and the rates of capital utilization. So we use a Cobb-Douglas production function with constant returns to scale and derive the values of its coefficients from observed factor shares. We assume that the goods market is continuously cleared by prices, thus assuming perfect price flexibility. However, we impose some nominal wage rigidity. Therefore, we model the demand for labour by the usual marginal productivity condition and the behaviour of nominal wages via a simple Phillips type curve. Labour supply is assumed to be exogenous.

The entire model is checked for consistency in the sense that the various balance sheets are consistent with each other. In order to see whether the estimated model fits the data adequately, we report a few within-sample dynamic forecasts.

Chapter 7 addresses the issue of policy applications of the model. Since our main purpose in this study is to examine the implications of various financial policies, we confine ourselves to the effects of those instruments which are used most frequently by the central bank and the commercial banks. The central bank uses the reference rate or the discount rate and the various required reserve ratios as the most common instruments of monetary policy. It is well known that high required reserve ratios drain liquidity from the system and can adversely affect the supply of loanable funds to the household and the firms which can be quite detrimental to consumption, savings and investment, particularly if both households and firms are credit-constrained as would seem to be the case in Poland. Since the NBP imposes differential required ratios on the four types of deposits, two zloty and two types denominated in foreign currency, we examine how changes in various required ratios affect the major macro variables, like, inflation, GDP, balance of payments and unemployment rate. The rationale for the particular ratios used is not clear, so our simulations might shed some light on the relative ratios imposed.

The second set of policy simulations relate to the effect of changes in the reference rate. Here we proceed in two alternate ways. First, we assume that the other rates do not change immediately in response to the central bank rate changes. There is some evidence to support this point. The alternative assumption is that they do, but with a lag. The simulations in this case are meant to shed light on the effectiveness of discount rate policy. Of course, a comparison of the simulations with the required reserve ratio changes could shed light on the relative effectiveness of the two instruments.

Turning to the effects of policy changes induced by the commercial banks, we consider two sets of rates: those on banking loans to the firms and the rate on zloty time deposits as a representative of the rates on four types of deposits. As for the effects of changes in the deposit rates, there are various possibilities here. For example, do identical changes in the four rates have different effects? If so, how different are they and in what sense? Would it be better for the economy if the banks attracted zloty time deposits rather than the other three types of deposits? How significant is the role of foreign currencydenominated deposits in the transmission mechanism of changes in the domestic deposit rates?

Loan rates can affect both the supply of credit and its cost to firms and the households. As mentioned above, the issue of high loan default rates is very important in Poland. What would happen to changes in the loan rates under alternative assumptions about the default rates? What would be the effect of ignoring the issue of default risk? Are cost effects more important than credit supply effects of a given change in the loan rate?

In the policy simulations suggested above, there are certain features of the model which are exploited. For example, it was pointed out above that in the case of the households we use an integrated model of portfolio selection and consumption. We examine the sensitivity of the simulation results to this assumption by assuming the more traditional approach in which such interdependence does not exist. We could reasonably argue that as Polish consumers become more sophisticated and the financial markets mature, the interdependence between the two decisions can only increase. So, our simulations are meant to shed some light on this phenomenon.

Chapter 8 offers a brief summary of the major findings and the lessons that can be derived from our modelling exercise.

2
A STATISTICAL OVERVIEW

There is substantial literature which describes the economic events in the three countries which are the subject of this chapter. Therefore, no attempt is made to be exhaustive. Instead, we highlight some of the major economic changes that have taken place in these countries over the last five years. Since the financial sector is dealt with in the next chapter, it will not be considered here.

Our story essentially starts in 1990, when economic reforms really took over in these economies. An overview of the major changes is given in Tables 2.1, 2.2 and 2.3. The first thing we notice from these tables is the well-known dramatic decline in the growth of the gross domestic product in 1991 and 1992. Much has been written on the causes of this precipitous contraction. We’ll have more to say on this later in the chapter. But we can also see a gradual recovery in all three economies since 1992. Thus, in 1995 the growth rate had climbed to 4.3 per cent in the Czech Republic, to 1.5 per cent in Hungary and to 7 per cent in Poland. Despite the differences in the rate of recovery in the three countries, the patterns are very similar and this raises the interesting question of what were the common factors characterizing all three economies? We do not pursue this question in this study, although we hope that the estimated model for Poland and the policy simulations based on it will shed some light on the underlying factors.

Structural and other changes

Since the dominance of the public sector had been one of the major defining features of the economies now in transition, it is useful to look at what has been happening to its role over the past five years. While the data are not available for the most recent years, none the less we can see from the data in Table 2.4 for the shares of the private sector in GDP and employment that its role has been shrinking in all three countries. Thus the share of the public sector in GDP had shrunk to about 40 per cent in the Czech Republic in 1993; to less than 50 per cent in Hungary and Poland by the end of 1993. A similar situation existed in terms of the share of employment, though to a slightly lesser extent. This trend toward increasing share of the private sector, of course, is the result of deliberate policies and we shall have more to say on this later in the chapter.

Table 2.1 Selected data: Czech Republic* (% change)

Table 2.2 Selected data: Hungary* (% change)

Table 2.3 Selected data: Poland* (% change)

Table 2.4 Private sector share in GDP and employment*

Another aspect of structural change is in terms of the share of industry in GDP. Here we do not have very recent data for any of the three countries, but from what we do have, the share was about 68 per cent in the Czech Republic in 1991, 25 per cent in Hungary in 1993 and about 40 per cent in Poland in 1992. In terms of the data for 1989 it would appear that the share had gone down to some extent. This deindustrialization may seem surprising at first, but when we consider the fact that state sector was shrinking but the private sector was not expanding at the compensatory rate, this outcome is not surprising.

It is also useful to look at the behaviour of inflation. In the Czech Republic—with the exception of 1991 when it reached a high of 57 per cent, which might have had something to do with the extreme contraction in the GDP—it has not shown signs of such resurgence, alth...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Figures
• Tables
• Acknowledgements
• 1 Introduction
• 2 A Statistical Overview
• 3 Commercial Banks1
• 4 Households
• 5 The Firm
• 6 The Complete Model1
• 7 Some Policy Simulations
• 8 Summary and Conclusions
• Appendix
• Notes and Appendices
• Bibliography