20942 SEPT. 2000 THE DETERMINANTS OF ENTERPRISE RESTRUCTURING IN TRANSITION An Assessment of the Evidence A WORLD FREE OF POVERTY The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence Simeon Djankov and Peter Murrell The World Bank Washington, D.C. Copyright © 2000 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, USA All rights reserved Manufactured in the United States of America First printing September 2000 The findings, interpretations, and conclusions expressed in this book are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organiza- tions, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use is granted by the World Bank, provided that the appropriate fee is paid directly to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax a request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the Office of the Publisher, World Bank, at the address above or faxed to 202-522-2422. Cover design by Tomoko Hirata, Printing, Graphics, and Map Design, World Bank. ISBN 0-8213-4815-9 Library of Congress Cataloging-in-Publication Data has been applied for. 11 Table of Contents ACKNOW LEDGM ENTS .........................................................................iv 1. INTRODUCTION ...........................................................................1 2. ENTERPRISE RESTRUCTURING IN THE EMPIRICAL LITERATURE .............................. .3 3. STATE VERSUS PRIVATE OWNERSHIP .........................................................6 4. THE EFFECTS OF DIFFERENT TYPES OF OWNERS .............................................10 5. THE ROLE OF MANAGERS IN ENTERPRISE RESTRUCTURING ................................ 15 6. ENTERPRISE RESTRUCTURING AND HARDENED BUDGETS .................................. 17 7. PRODUCT MARKET COMPETITION .........................................................20 8. THE ROLE OF INSTITUTIONS IN ENTERPRISE RESTRUCTURING ............................. 23 9. SUM M ARY ................................................................................25 REFEREN CES..................................................................................27 FIGURES 1. How Ownership Affects Firm Performance after Privatization: Relative Effects of Changing from Traditional State Ownership to Different PrivateOwners ..........................................................11 2. Regional Variations in the Effects of Different Types of Owners: Comparing Ownership Effects in the CIS to those in the N on-CIS Countries ......................................................................14 TABLES 1. Assessment of Whether Private Ownership or State Ownership Leads to More Restructuring: Assessments of the Composite Implications of 31 Empirical Studies Analyzing the Experience of Transition Countries ........... .7 2. Comparing the Size of Privatization Effects across Regions: Assessments of the Composite Implications of 31 Empirical Studies Analyzing the Experience of Transition Countries .................................... .9 3. Differences between the Effects of Different Owners on Enterprise Restructuring: Assessments of the Composite Implications of 23 Empirical Studies Analyzing the Experience of Transition Countries .................... 12 4. The Role of Managerial Turnover and Managerial Incentives in Enterprise Restructuring: Assessments of the Composite Implications of Six Empirical Studies Analyzing the Experience of Transition Countries .......... 15 5. The Importance of Hardening Budget Constraints in Enterprise Restructuring: Assessments of the Composite Implications of Seven Empirical Studies Analyzing the Experience of Transition Countries ................. 18 6. The Effect of Competition on Enterprise Restructuring: Assessments of the Composite Implications of 13 Empirical Studies Analyzing the Experience of Transition Countries ........................................... 21 7. The Relative Effect of Different Types of Competition in Different Regions: Assessments of the Composite Implications of 13 Empirical Studies Analyzing the Experience of Transition Countries .................... 21 iii ACKNOWLEDGMENTS Simeon Djankov is a financial economist at the World Bank. Peter Murrell is a professor of Economics and the chair of the Academic Council of the Institutional Reform and the Informal Sector (IRIS) Center, University of Maryland. We would like to thank Judy Hellerstein, John McMillan, John Nellis, and Jan Svejnar for their helpful advice and Wooyoung Kim and Tatiana Nenova for their research assistance. This research was made possible through support provided by the World Bank and the U.S. Agency for International Development (USAID) under Cooperative Agreement No. DHR-00 1 5-A-00-003 1-00 to the Center for Institutional Reform and the Informal Sector. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the IRIS Center, USAID, the World Bank, its Executive Directors, or the countries they represent. Simeon Djankov Peter Murrell The World Bank Department of Economics 1818 H Street, N.W University of Maryland Washington, DC 20433 College Park, MD 20742 sdjankov@worldbank.org murrell@econ.umd.edu June 29, 2000 iv 1. Introduction Over the last decade, more than 150,000 large enterprises in 27 transition countries have encountered revolutionary changes in every aspect of their political and economic environ- ments. Some enterprises have responded to the challenge, entering world markets with great dynamism and becoming indistinguishable from their competitors in mature market economies. Others remain mired in their past, undergoing protracted deaths, delayed at times by their slippage into a netherworld of barter and ersatz money. Thus the revolutionary changes in transition countries have been matched by great variation in the degree to which tage of the enormous amount of available evidence on the enterprises have responded successfully to events. effects of privatization. A comprehensive analysis of avail- Our understanding of economic processes can greatly able evidence is necessary to assess the relative strength of benefit from analysis of these changes and the responses of the various positions in these debates on the effectiveness enterprises to them. Such analysis addresses age-old ques- of different reform and privatization strategies. tions of economics and also poses new ones. What are the The most intensive area of empirical research on tran- relative productivities of state and private enterprises? sition countries has been the examination of enterprise-level Does mass privatization work? What is the efficiency cost data to ascertain whether enterprises have responded pro- of diffuse share ownership relative to concentrated own- ductively to changes in ownership and to other reform ership? Which of the many new private owners are most measures, such as the opening of foreign and domestic effective-managers, workers, banks, or investment funds? markets. We follow common parlance and refer to such To what degree do governmental subsidies to loss-making enterprise responses as restructuring, a notion that we enterprises dull performance? Is a strengthening of man- examine closely in the second section of this paper. We have agerial incentives sufficient to inspire turnaround or is identified more than 125 empirical studies that examine the replacement of managers necessary for revitalization? determinants of enterprise restructuring using sound method- Does competition promote productivity change? Which ologies applied to data generated at the enterprise level. The institutions are necessary to complement other mecha- principal objective of this paper is to provide an overall nisms of change? assessment of the evidence generated by these studies. Although these are questions of general importance to This paper is aimed at the reader who wishes to gain economic policy everywhere, they are especially important an overview of the current state of the evidence, but who when analyzing the experience of transition countries. does not have any interest in examining the techniques of Analysis of the way that enterprises respond to different analysis that generated the evidence. Therefore, we adopt policy measures is central in any effort to gain an under- a relatively informal approach to the presentation of sta- standing of the effects of reform measures. This is especially tistical results, eschewing the precision of statement that the case in undcrstanding the consequences of privatiza- would inevitably require lengthy descriptions of method- tion. While rapid, mass privatization was an early empha- ology. Nevertheless, our informal statements do corre- sis of transition (Lipton and Sachs 1990; Boycko, Shleifer, spond exactly to precise statistical results. A companion and Vishny 1995), this is now subject to intense criticism paper (Djankov and Murrell 2000) contains an extensive (Stiglitz 1999; Black, Kraakman, and Tarassova 1999). discussion of methodology and presents results in a quan- But the formulation of this criticism has not taken advan- titative statistical framework. Readers should refer to that 1 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence paper if they are interested in precisely how we came to the prise restructuring. Section 8 examines the role of the insti- judgments that appear here. tutional and legal framework. The presentation of the evidence is organized as fol- Our intention in this paper is simply to summarize the lows. Section 2 discusses the general context of the studies existing evidence. We do not venture into speculation of that are included in our assessment and defines terms that what this evidence implies for evaluation of policy, either we use throughout the paper. Section 3 assesses the evi- in retrospect or prospect. Therefore, the concluding section dence on whether state-owned or privatized firms under- of this paper does not provide an interpretative conclusion, take more economic restructuring. Section 4 studies the but rather a summary of what has gone before. This sum- effects on restructuring of different types of owners (such mary is phrased in terms that make it self-contained. as foreigners or workers). Section 5 focuses on the role of Therefore the harried reader can simply jump to that sec- managers, analyzing whether the strengthening of man- tion to obtain the briefest summary of what we think the agerial incentives is sufficient to produce turnaround or evidence implies about the determinants of enterprise whether management turnover is required. Section 6 ana- restructuring. Nevertheless, we do not recommend this lyzes the effect of greater discipline in the government's approach, since there is much in the intervening sections reaction to enterprises in distress (usually referred to as the that serves to qualify the evidence and amplify the various hardening of soft-budget constraints). Section 7 examines findings. how variations in product market competition affect enter- 2 2. Enterprise Restructuring in the Empirical Literature What is enterprise restructuring and what changes might induce it in transition countries? The answer to this question lies in the characteristics of the socialist economy and its enterprises. These have been widely discussed in many contexts and we do not need to reiterate any- thing but a few central issues here. (See Berliner 1976, Murrell 1990, and Kornai 1992 for details.) The classic socialist enterprise received a plan on output levels and on inputs to be used in the pro- duction process. Meeting this plan was of prime importance, and the plan was normally an ambitious one. Therefore, production issues dominated entrepre- marketing personnel or developers of new products. neurship, marketing, and cost minimization in manageri- Workers had virtually no role in enterprise decisionmaking, al concerns. Consistently, the typical manager was a pro- except in the limited sphere of personnel policy, where a duction engineer and not a businessperson. These managers variety of factors led to firing rates that were extremely low responded to a complex mix of monetary and career-based by any standard (Granick 1987). One such factor was the incentives, which were a function of fulfillment of the role that the enterprise played as provider of social welfare plan, enterprise performance, and political loyalty. The benefits. Hence, efficiency considerations were often a sec- crucial point here is simply that enterprise profits and ondary consideration in determining the size of an enter- enterprise efficiency were much less important to a social- prise's work force. ist manager than to any manager of a capitalist firm, no Pretransition reforms did change this standard picture matter how remote the manager was from shareholders. in some countries, notably Yugoslavia, Hungary, and A labyrinthine bureaucracy replaced the institutions Poland (Balcerowicz 1995 and Kornai 1986). In these and the markets of capitalism. It found customers and countries, enterprises came closer to ultimate consumers, determined prices, with bureaucratic pressure substituting including foreign ones. Decentralizing reforms reduced for competition. The state interceded between producer the scope of bureaucratic decisionmaking. Markets and and buyer, most notably in isolating enterprises from competition increased in importance. Paradoxically, how- domestic consumers and foreign markets. The bureaucra- ever, abandonment of formal planning led to increased cy acted as a contract-generating and a contract-enforcing bargaining between the bureaucracy and the enterprise, agency. Its one-year plans were an immediate guarantee of perhaps even resulting in a further softening of budget short-term working capital. A centrally determined invest- constraints. Notably also, workers gained more power ment project would automatically receive long-term cred- within enterprises, acquiring experience at being informal its. Given the ubiquitous role of the state, much would be owners. decided by negotiations, which were a major concern of Restructuring, then, is change in the enterprise behav- top managers and a key element of their expertise. One iors described above, particularly in levels of enterprise effi- consequence of the frequency of these negotiations was the ciency. To produce the empirical literature that we study, it universal presence of easy financing, which further turned has been necessary to construct measures of restructuring. managers' attention away from profits and efficiency. Obviously there is great variation among authors on how Internally, the enterprise was organized along very to define this concept. Many papers focus on the end result hierarchical lines. One-person rule was in place, and that and simply define enterprise restructuring as an improve- one person was surrounded by process engineers, not by ment in performance (measured by growth in sales or level 3 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence of productivity, for example). Other studies look at the are the phenomena of primary interest, the set of vari- internal operations of the enterprise and focus on features ables that measure reforms as they impinge on the partic- that differ greatly under capitalism and socialism, mea- ular enterprise, for example, the proportion of the enter- suring restructuring by whether these features have prise that has been privatized or the intensity of changed. Thus, for example, empirical studies have exam- competition in the product market that the enterprise faces. ined which enterprises have introduced marketing depart- We will discuss these variables in much detail in the sections ments since reforms began. of the paper that follow, devoting each section to an impor- One broad category of restructuring measures com- tant category of explanatory variable. prises quantitative indicators that are based on accounting The second category of explanatory variables includes information and that measure actual enterprise perfor- enterprise characteristics in which we have little interest mance. The most common items in this set are indices here. Examples are enterprise size, sector of operation, that reflect the productivity of the enterprise or its rate of and region of country in which the enterprise is located. growth of production. We will use the term quantitative to Given the lack of interest in these variables, why are they refer to these indicators. Other indicators of restructuring included in the empirical studies and why do we mention depend less on quantitative accounting information. They them here? The simple answer is that omission of these are measured somewhat more loosely, perhaps derived variables in the empirical studies would lead to biases in the from survey questions on economic performance that are results generated. Thus, it is important to include such posed to managers (such as forecasts of sales in the sur- variables (control variables) in statistical studies, precisely veyed year) or from information collected about reorgani- because their inclusion enables one to obtain more accurate zation (for example, whether the enterprise has introduced results. new products), or perhaps reflecting operational factors A different set of issues arises in the case of selection further removed from current performance (for example, bias, the thorniest problem encountered in estimating the the extent of wage arrears). These indicators will be effects of reform measures on restructuring. Selection bias referred to as qualitative. might occur when the decision on how a reform measure The prevailing sentiment among researchers is that the applies to an enterprise reflects some unmeasured phe- quantitative variables are to be trusted more. They certainly nomenon that also affects the amount of post-reform do measure directly the prime objective of enterprise restructuring. If the most standard statistical techniques are restructuring: an improvement in economic performance. used, the estimate of the effect of reform will be contami- However, there is also the view that quantitative perfor- nated. There are statistical techniques that can reduce the mance might suffer when an enterprise is undertaking fun- likelihood of problems arising from selection bias. damental efforts to reorganize and that these efforts might However, these are often not easy to implement, and be observed earliest in the qualitative variables. Qualitative attempts to counter selection bias vary in quality a great measures might therefore be leading indicators of enterprise deal between studies. performance. We focus primarily on the quantitative indi- We have mentioned these methodological problems in cators in this paper. This focus results primarily from our order to give the reader a flavor of the hurdles that confront own judgment, derived from our own empirical work and researchers in endeavoring to understand the determinants from an examination of the details of the papers surveyed of enterprise restructuring. Given the fact that this paper here, that the reliability of the statistical studies of quanti- omits any precise description of the methodology of the tative indicators is greater than that of the qualitative ones. pertinent empirical studies, it would be inappropriate for Nevertheless, when sufficient analyses are available, we us to leave the reader with the impression that these stud- examine both types of indicators, finding that they gener- ies are purely mechanical exercises, in which judgment ally lead to the same basic conclusions. and effort do not count. Rather, there are difficult problems The standard study that we examine focuses on the to be solved and some studies do a much better job of solv- amount of restructuring in an enterprise as the phenome- ing these problems than others do. The papers we exam- non to be explained (that is, as the dependent variable). ined vary greatly in quality. Using statistical techniques, which we shall not detail here, We have identified two important factors (addition of researchers employ enterprise-level data to investigate how control variables and removing selection bias) that capture the degree of enterprise restructuring varies with the char- elements of a paper's quality. There are other factors as acteristics of the enterprises. Those characteristics, or well. One is the number of enterprises included in the explanatory variables, fall into two categories. First, there study, since statistical precision varies with sample size. 4 Enterprise Restructuring in the Empirical Literature Another is the number of years of reform captured in the First, we use our quality rating of the papers to determine data, since one would expect the effects of reforms to the relative influence of each paper on the combined result. occur cumulatively over time. There are also intangible ele- If we judge one paper to be twice as good as another, our ments of the strength of a research exercise. However, a aggregation attributes twice as much importance to the scholar familiar with a particular literature usually has an results of the higher-quality paper. Second, we combine the ability to judge the overall strength of analysis after exam- results of all of the papers without using any information ining carefully the methods used in a paper, reaching a sub- about the quality of methodology: each paper counts equal- jective judgment of quality that reflects a sense of those ly in contributing information to our aggregate results. intangibles. An essential part of our assessment of the Using these two approaches we are able to give the empirical literature involves our reaching such a subjective reader two differing assessments of the evidence, one based judgment on each paper. on our sense of the strength of the evidence in each of the For each of the papers examined here, we arrive at a individual papers and one based purely on a mechanical rating of the overall quality of the analysis in the paper. aggregation of the individual results. Obviously, we think This quality rating reflects three items of information. that the most reliable aggregation is the one that uses our First, there are the objective factors listed above. Second, quality assessment. We provide the alternative evidence for there is our subjective judgment of the overall strength of the reader who is skeptical about our judgments. analysis. Third, our quality rating reflects the relative All of the previous remarks are somewhat general, standing of the scientific journal in which the paper is omitting discussion of exactly how we will present the published, if it has been published. Thus, our ranking of evidence to the reader. Such discussion is most easily pre- quality reflects not only our own assessments, but also sented in context. We do so in the next section, examining those of the economics profession. state versus private ownership, which is the issue that has In the companion paper to this one, we use statistical been examined most often in the empirical literature on methods to aggregate the individual results of all papers on enterprise restructuring. one topic into an overall result. We do this in two ways. 5 3. State Versus Private Ownership tate ownership is the staple of a traditional socialist economy, and private ownership is the essence of capitalism. In the early debates on transition policy, there was no disagreement about the desir- ability of creating an economy dominated by private ownership, but there were rather conflicting views on how this could be done most effectively, through fast privatization (Lipton and Sachs 1990; Boycko, Shleifer, and Vishny 1995) or through efforts concentrating on stimulating a nascent private sector (Kornai 1990; Murrell 1992). The emphasis on speedy privatization has waxed and waned with events. With Eastern Europe in deep crisis in the early low estimated probability means rejecting the assumption 1990s, fast privatization seemed to gain urgency. However, that state and private ownership are equivalent. Given an with the recovery of Poland, a relatively slow privatizer, estimated probability in each paper, it is a simple step to that perceived urgency declined somewhat (Pinto, Belka, combine all the estimated probabilities into one composite and Krajewski 1993). But Poland is only one of many tran- probability for all papers combined. This composite prob- sition countries, an outlier at that. The latter half of the ability is what we derive from the set of analyses under con- 1990s has offered examples of fast privatizers performing sideration. well and fast privatizers performing badly, with similar Our method of combining the estimates of many variation across slow privatizers, giving sustenance for a studies leads to much sharper stronger results than the variety of opinions about the results of privatization (Pohl individual studies themselves. It is easy to demonstrate and others 1997; Havrylyshyn and McGettigan 1999; Nellis this. For example, toss a coin four times and obtain three 1999; Stightz 1999; Black, Kraakman, and Tarassova 1999). heads, and one harbors no thoughts that the coin is unbal- We have identified 31 distinct studies that contain anced. However, if one repeats this experiment 20 times results on how private ownership affects economic restruc- and obtains a succession of two, three, and four heads in turing. However, some of these studies contain results for the four tosses, then one might have strong evidence of an several countries or present several conceptually distinct unbalanced coin. The example carries over into research. results for the same countries. Thus the following reflects Because the data are very rough and the statistical methods the combined information from 82 different analyses of the hardly perfect, researchers often obtain only weak results effects of private ownership. In interpreting the results below, in individual studies. But many weak results, all indicating one should remember that these analyses focus on the effects the same phenomenon, can combine to produce one very of private ownership of shares of firms, since many firms in strong result. We find this to be true in many instances in the postsocialist world have mixed ownership. In this section, this paper. when we refer to state ownership, it refers to ownership of Using the composite evidence that reflects the results shares in 100 percent state-owned enterprises or in enter- of the individual papers, one can make a judgment on prises that have been partially privatized. acceptance or rejection of the assumption that state and pri- Following standard statistical methodology, each of vate ownership are equivalent. If rejection is the conclusion, the studies begins with the assumption that state and pri- then the evidence also indicates which ownership form is vate ownership actually have the same effects. The statis- the better one. Therefore, the composite information that tical analysis then produces a probability that the data on we derive from the studies also indicates which of the two enterprise behavior are consistent with this assumption. A ownership types leads to more restructuring. 6 State Versus Private Ownership To summarize the evidence, we have converted the Our summary of the evidence on state versus private composite probabilities into simple phrases. These simple ownership appears in table 1. Two different ways of group- phrases are on a five-point scale of the following ratings: ing studies lead to the different rows of the table. First, there is the quantitative-qualitative division of dependent 1. Extremely likely that private ownership produces variables, to which we have already referred in section 2. more enterprise restructuring than state ownership Second, there are regional groupings. Corresponding to 2.Probable that private ownership produces more much of the rest of the literature (for example, EBRD enterprise restructuring than state ownership 1999), the basic split is between the non-Baltic former 3. No evidence that private and state ownership differ Soviet Union (the Commonwealth of Independent States 4.Probable that state ownership produces more enter- [CIS]) countries and the rest of the transition countries. In prise restructuring than private ownership the set of papers under consideration, there are two stud- 5.Extremely likely that state ownership produces more ies of Mongolia. Since this country looks like a typical enterprise restructuring than private ownership member of the CIS (Korsun and Murrell 1995), Mongolia is included in that grouping. The non-CIS group compris- Perhaps a few extra words are useful in interpreting es Eastern Europe and the Baltic states (with one study of our phrases "extremely likely," "probable," and "no evi- China). Interestingly, once we seek a criterion that corre- dence." Extremely likely means that we feel that there is a sponds exactly to our split of countries, we find that the cri- large preponderance of evidence in favor of the stated terion is the length of time that the countries labored under effect. When we use this phrase, we do not think it is a pos- communism, 70 years for each CIS country and less than sibility that evidence will arise in the future that will cause 50 years in the non-CIS grouping. The reader therefore us to change our views. Probable means that there is some might like to think of our regional groups as "two gener- evidence in favor of the stated effect and we have a con- ations" and "three generations," indicating the length of siderable degree of confidence that the stated effect is cor- time under communism. rect. The stated effect is a very good bet, but there is some There are two columns stating the composite evi- residual possibility that this bet would lose. "No evidence" dence, corresponding to the two ways we have aggre- means that we are not able to distinguish between the gated the individual items of evidence, discussed in section effects of the different owners on the basis of what we have 2. Column 1 presents the composite result that is con- read in the papers under consideration. structed without using any of our judgments about the TABLE 1. ASSESSMENT OF WHETHER PRIVATE OWNERSHIP OR STATE OWNERSHIP LEADS TO MORE RESTRUCTURING Assessments of the Composite Implications of 31 Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate: Assessment of the likelihood that there is a difference between the restructuring effectiveness of state and private ownership (1) (2) Geographical location Restructuring variable When there is no attempt to weight the When the relative importance attached to each of countries studied: examined: quantitative importance of the studies' results by the study's results reflects the methodological CIS or non-CIS or qualitative quality of their methodologies quality of the study 1. All countries Both types of variables Extremely likely that P > S Extremely likely that P > S 2. Non-CIS Both types of variables Extremely likely that P > S Extremely likely that P > S 3. CIS Both types of variables Probable that P > S Probable that P > S 4. All countries Quantitative Extremely likely that P > S Extremely likely that P > S 5. Non-CIS Quantitative Extremely likely that P > S Extremely likely that P > S 6. CIS Quantitative No evidence that private and state differ Probable that S > P 7. All countries Qualitative Extremely likely that P > S Extremely likely that P > S 8. Non-CIS Qualitative Probable that P > S Probable that P > S 9. CIS Qualitative Extremely likely that P > S Extremely likely that P > S Note: P > S is shorthand for "private ownership produces more enterprise restructuring than does state ownership." S > P is shorthand for "state ownership produces more enterprise restructuring than does private ownership." Source: All figures and tables are from the author unless otherwise noted. 7 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence methodological quality of the research of individual papers. Table 2 has many features similar to those of table 1. Column 2 presents the composite result when more weight We do the analysis separately for the two types of indica- is attributed to studies that we believe are conceptually tors, quantitative and qualitative. We present information stronger. Only in one case does the evidence differ between for the case when we weight the existing results by our per- the two columns. ception of methodological quality (column 2) and also for Taking all countries together, the evidence is extreme- the case when we do not weight for quality (column 1). We ly strong that the move from state to private ownership has offer a verbal summary of the composite statistical results. resulted in greater amounts of restructuring. For non-CIS Let us use column 2 of table 2 to discuss the results. countries taken separately, this is also the case. (The some- The first two rows examine the conclusions when both what weaker results on row 8 of the table simply reflect a qualitative and quantitative restructuring indicators are much smaller number of studies in that category than in examined together. These rows indicate that a shift of other categories included in the table.) The effects of pri- ownership in the non-CIS countries has 5.5 times as much vatization in that region are indubitable. For the CIS coun- effect on restructuring as does a similar shift of ownership tries, the picture is more complex. This is because the in the CIS countries. Thus, for example, if we found that quantitative and qualitative indicators offer a different privatized firms in the non-CIS countries had a growth rate picture. For the quantitative measures, the evidence, in of output that was 5.5 percent higher than that of non-CIS our judgment (see column 2), shows some indication that state firms, we would expect that CIS privatized firms state enterprises are more productive than private enter- would have a growth of output that was 1 percent higher prises. For the qualitative indicators, the evidence is the than CIS state firms. The second row of the table contains reverse. Row 3, which shows positive effects of privatiza- our verbal summary of the statistical evidence on whether tion in the CIS, must reflect the fact that the positive results the difference between the regions is likely to have arisen by on the qualitative indicators dominate the somewhat more chance, or whether it is likely to represent some aspect of equivocal results on the quantitative indicators. We do reality. The interpretation of these verbal statements is the not have any other information that would allow us to same as in table 1. interpret the mixed picture from quantitative and qualita- The results of table 2 should leave the reader in no tive indicators and therefore leave it to readers to provide doubt that the move from state to private ownership has a their own interpretations. much stronger effect in the non-CIS countries than in the It is tempting to conclude from table 1 that the effect CIS countries. In all cases, the privatization effect in the of change in ownership on the quantitative variables in the non-CIS countries is more than twice the size of that in the non-CIS countries is greater than the effect in the CIS. CIS countries. In all cases but one, the statistical evidence This is not appropriate, since table 1 does not provide an indicates that these differences are not likely to have arisen explicit statistical comparison of results for the CIS coun- by chance. Certainly, we have no doubt in the soundness of tries versus those for the non-CIS countries. However, the conclusion that privatization has been less effective in using the published results of existing studies, one can the CIS than in other countries. examine whether privatization effects in the two regions are There remains the issue of the economic size of the numerically different. This is the task of the remainder of privatization effect. One way to address this issue is to take this section. the numbers underlying tables 1 and 2 and see what they We can calculate a numerical score for the restruc- imply in a very simple case. Suppose that our measure of turing effect of moving ownership from the state to the pri- restructuring is a very crude one, whether or not an enter- vate sector. This can be done for each region separately, in prise is growing or declining. Then, by using simple and a manner that makes the numerical scores for each region reasonable assumptions one can show that the numbers comparable in the sense that their units of measurement are leading to row 1 and column 2 of the table imply that com- the same. Since an understanding of the absolute value of plete privatization in the non-CIS countries would result in these scores requires insight into the statistical methodol- 16 percent more firms growing, while complete privatiza- ogy, we do not present these absolute values here. Rather, tion in the CIS would result in only 2 percent more firms we present relative values, comparing the score in one growing. region relative to the score in the other. Table 2 presents One can also examine individual studies to get an idea estimates of the effect on restructuring of moving owner- of the economic size of privatization effects. One typical ship from state to private in the non-CIS countries divided restructuring measure used is rate of growth of output. by the same effect for the CIS countries. Studies on Eastern Europe indicate that the annual rate of 8 State Versus Private Ownership TABLE 2. COMPARING THE SIZE OF PRIVATIZATION EFFECTS ACROSS REGIONS Assessments of the Composite Implications of 31 Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate (1) (2) Type of When there is no attempt to weight When the relative importance restructuring the importance of the studies' attached to each study's results variable results by the quality of reflects the methodological examined their methodologies quality of the study 1. Both qualitative Effect on restructuring of moving from 2.3 5.5 and quantitative state to private ownership in non-CIS countries divided by the same effect for CIS countries 2. Both qualitative Assessment of the likelihood that there is a Extremely likely that privatization Extremely likely that privatization and quantitative difference between the two regions in the effect of effect in non-CIS countries is effect in non-CIS countries is moving from state to private ownership larger than in the CIS larger than in the CIS 3. Quantitative Effect on restructuring of moving from state to 5.2 Privatization effect in the CIS private ownership in non-CIS countries is negative (see note) divided by the same effect for CIS countries 4. Quantitative Assessment of the likelihood that there is a Extremely likely that privatization Extremely likely that privatization difference between the two regions in the effect of effect in non-CIS countries is effect in non-CIS countries moving from state to private ownership larger than in the CIS is larger than in the CIS 5. Qualitative Effect on restructuring of moving from state to 2.4 2.3 private ownership in non-CIS countries divided by the same effect for CIS countries 6. Qualitative Assessment of the likelihood that there is a Probable that privatization effect in No evidence that privatization difference between the two regions in the effect of non-CIS countries is larger effect in non-CIS countries moving from state to private ownership than in the CIS differs from that in the CIS Note: When the privatization effect in the CIS is negative, it is not appropriate to calculate the ratio of the effects in two regions growth of output in privatized firms minus the annual period studied. For the CIS, the range is much wider, due rate of growth of output in state firms ranges from 2 per- perhaps to greater variability in data and the wider variety cent to 8 percent, depending on the country and the time of experience in that region. 9 4. The Effects of Different Types of Owners ne of the reasons that changes of ownership might have had different effects across regions is that differences in privatization processes resulted in different mixes of owners across coun- tries. The hoped-for quick retrading of shares to the most effective owners has not happened and therefore the owners created initially by the privatization process will have more than a short-term effect on enterprise performance. This is important, of course, only if the type of ownership makes a difference. As it happens, transition experience offers unusually comprehensive evidence on whether the type of private ownership matters. Just as the papers under review offer evidence on the that have been legally separated from the state, effects of state ownership versus private ownership, they that are treated as private enterprises under corpo- also offer evidence on state ownership versus particular rate laws, and that have usually been part of a pri- types of owners (for example, foreigners or managers) and vatization program. In practice, this type of state indeed on specific types of private owners versus other ownership almost always occurs in firms that are types (for example, foreigners versus managers). Thus, partially privatized. just as before, we can combine all of this evidence and pro- 3. Enterprise insiders (a composite group, where duce estimates of the restructuring effectiveness of each workers and managers were not differentiated). type of owner relative to each other type. The only added 4. Enterprise outsiders (a composite group consisting complication is that we now have a multilateral compari- of all nonemployee, nonstate owners). son instead of a dichotomous one. But this complication 5. Workers (nonmanagement employees). only adds to the methodological problems, with which 6. Managers (managerial employees). we are not concerned here. It is as easy to imagine a com- 7. Banks. parison of manager-owners versus worker-owners as it is 8. Investment funds (other than those owned by banks state ownership versus private ownership. or the state). The first task is to determine the set of owners to be 9. Foreign owners. compared to each other, which is shown below. It must be 10. Blockholders: outsider ownership that has been emphasized once again that we are referring to ownership concentrated in the hands of large individual own- of shares of firms. We followed the empirical literature in ers (such as individual entrepreneurs or domestic identifying the ownership categories that we analyze, dis- firms) other than those listed above. tinguishing 11 in all. The first two are types of ownership 11. Diffuse outsiders: the residual outsider ownership that were included in the state category in the previous sec- category, when outsider owners are not identified as tion. The remaining nine are private owners. belonging to the categories above. This category is dominated by individual outsider ownership that 1. Traditional state ownership: state ownership in remains diffused across large numbers of individual enterprises that are 100 percent state-owned and owners. that have not been part of a privatization program. 2. State ownership in commercialized (or "corpora- The reader will immediately notice that some of these tized") enterprises: state ownership in enterprises categories overlap: for example, workers and managers 10 The Effects of Different Types of Owners together are insiders. However, one should not assume the data to be the most effective, and therefore are assigned that the collective entity is the sum of its parts. We will a score of 100. A score of 50 on the scale for owner X would examine this point when the results are presented. then indicate that privatization to owner X produces only From 23 studies, we have compiled a data set that half as much restructuring as privatization to foreigners. allows us to build a picture of the effects on restructuring The results are presented in figure 1, which suggests of different types of owners. We have restricted our analysis that differences between owners are of great economic to examination of the quantitative indicators only, since there importance. Before proceeding to discuss the differences are not enough observations from studies that use qualitative between owners, it is useful to present information on indicators to undertake a separate analysis for these. whether statistical tests indicate these differences to be sig- The information produced by the individual studies nificant, or whether they could have simply arisen by only allows one to ascertain the relative effects of different chance. We adopt essentially the same procedure as that for owners, not the absolute effect of any single owner. For this tables 1 and 2, that is, we combine the probability infor- reason, we can only arrange the effectiveness of the differ- mation from many papers to obtain a much stronger com- ent owners on a relative scale. Thus, to present the results posite result. Then, we summarize the composite proba- in a simple fashion, we assign a score of 0 to the least effec- bility judgment with the types of phrases that we have tive owner (according to the data) and a score of 100 to the used in tables 1 and 2. The results appear in table 3. most effective. It is traditional state ownership that turns At a rough approximation, there are three groupings out to be the least effective owner, and is therefore made the of owners. At the bottom, traditional state ownership and 0 point. This is convenient, because the ordering on the diffuse individual ownership do not have significantly dif- scale then shows the relative effect of privatizing to the dif- ferent effects. In the middle, insiders, outsiders, workers, ferent types of private owners. Foreigners are revealed by banks, and commercialized state ownership are clustered. FIGURE 1. HOW OWNERSHIP AFFECTS FIRM PERFORMANCE AFTER PRIVATIZATION Relative Effects of Changing from Traditional State Ownership to Different Private Owners Relative effect of privatizing to specific type of owner 120 98.1 100.0 100006 9- 80 64.9 6047.9 48.6 52.9 40 31.4 20 0 Category of Owners 11 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence TABLE 3. DIFFERENCES BETWEEN THE EFFECTS OF DIFFERENT OWNERS ON ENTERPRISE RESTRUCTURING Assessments of the Composite Implications of 23 Empirical Studies Analyzing the Experience of Transition Countries Composite judgment on whether owner listed on row is more effective than owner listed on column Category of Traditional Diffuse Commercialized Investment owner state individual Insiders Outsiders Workers Banks state Managers Blockholder funds Diffuse Individual No Insiders Probably Probably Outsiders Probably Extremely No likely Workers Probably Probably No No Banks Extremely Probably No No No likely Commercialized Extremely Extremely Extremely No No No state likely likely likely Managers Extremely Extremely Extremely Probably Probably Probably Probably likely likely likely Blockholder Extremely Extremely Extremely Extremely Probably Probably Probably No likely likely likely likely Investment funds Extremely Extremely Extremely Extremely Probably Probably Probably No No likely likely likely likely Foreign Extremely Extremely Extremely Extremely Extremely Probably Probably No No No likely likely likely likely likely Note: Interpreting the table: To compare owners A and B, first find the nonempty cell that corresponds to the row of one of the entities and the column of the other. Suppose it is the cell corresponding to A's colunm and B's row. Then the phrase in that cell indicates the authors' degree of confidence in the conclusion that B's effect on restructuring is greater than A's effect. The table legend is defined as follows: Extremely likely that owner identified on row is more effective in undertaking enterprise restructuring than owner identified on column; Probably correct to conclude that owner identified on row is more effective in restructuring than owner identified on column; No evidence that owner identified on row is more effective in restructuring than owner identified on column. The most effective owners, none of which have statistically and managers, who have diverse interests. Such struggles different effects from any other, are managers, concentrated will delay adjustment. In country B, the worker-owned and individual ownership, investment funds, and foreigners. manager-owned enterprises might each quickly implement One result seems, on first glance, to be paradoxical, their own forms of adjustment, since separately they each but, on reflection, reveals important information. Insiders can reach decisions more quickly. Given the way that pri- are less productive owners than both managers and work- vatization studies are carried out, researchers on country A ers, even though these two groups of employees constitute would be more likely to report results for the group of the full set of insiders. How can this come about? The prin- insiders, while those on B would show results separately for cipal reason is that studies treating insiders as one group managers and workers. Then, our synthesis would show will usually be derived from different countries than stud- that undifferentiated insiders are not as good owners as ies treating workers and managers separately. Consider managers and workers separately. Moreover, this would the following example: country A privatizes all of its enter- not be an artifact of the research process, but would con- prises by giving generalized concessions to insiders, where- vey something very important about reality: privatizing to as country B privatizes half of its enterprises to managers heterogeneous groups might be worse than privatizing to and half of its enterprises to workers. In country A, priva- homogeneous groups. The whole is less than the aggrega- tization might be followed by struggles between workers tion of its separate parts. 12 The Effects of Different Types of Owners By and large, the rest of the results are in accordance Also, the effects of different types of owners could with expectations, but there are some surprises. Foreigners vary between regions because different types of owners were expected to make productive changes and they are require different levels of institutional support, and insti- unsurprisingly the best owners. But it is notable that three tutional quality varies across countries. We follow North other ownership types are very close in effectiveness to for- (1990) in defining institutions as the rules that constrain eigners. Certainly, the estimated productiveness of man- economic agents and the incentives to follow these rules. In agers and investment funds was not uniformly expected. the present context, we particularly refer to the set of insti- Similarly, diffuse individual ownership was not expected to tutions pertinent to the governance of large enterprises: cor- be very effective, but it is perhaps surprising that it is sta- porate governance laws and their enforcement; securities tistically indistinguishable from traditional state ownership. laws and their enforcement; and the elements of civil and Perhaps the most notable and unexpected result is the criminal law and their enforcement that help to protect share- place of state ownership in commercialized enterprises. holders from the malfeasance of managers and directors. This is not some artifact, but depends rather on results that Figure 2 presents estimates of the effects of the dif- appear across a wide range of studies, from Mongolia ferent types of owners in the two regions. All of these (Anderson, Lee, and Murrell, forthcoming) to Central effects are relative ones, in that we have adopted the same Europe (Frydman and others 1999). In reflecting on this conventions as in constructing figure 1: we have assigned result, it is important to remember that this type of own- an effectiveness of 0 in each region to traditional state ership usually occurs in enterprises that are partially pri- ownership and an effectiveness of 100 to foreign ownership vatized. It might well be that the private part-owners are in each region. Then, all other types of owners are placed playing an important role in enterprise affairs (Frydman on this scale, separately for each region. The appropriate and others 1999). One must remember also that this result use of figure 2 is to examine differences in the relative is not for economies in which real ownership has been ranking of different owners between the two regions. There developed organically for decades, but rather for situations are some dramatic and obvious divergences (such as banks in which ownership has been artificially transferred, some- and workers), but in about half of the cases the rankings times to private owners who are creatures of the state. are rather close (for example, commercialized state and Then, if shareholders are weak, share retrading is sluggish, managers). and the state is focused on solving economic problems, it There is the "germ" of an institutional story in these is not surprising that state ownership can be superior to results, although our discussion here turns much more some types of ownership (Anderson, Lee, and Murrell, interpretive than it has been up to now. For some owners, forthcoming). In addition, the very act of commercializa- it is important that the mechanisms of corporate gover- tion could change the incentives facing the state (Shleifer nance function well and function continuously, while other and Vishny 1994). owners are not so dependent on these mechanisms. When the institutions of corporate governance are weak, the Comparing Owners across Regions effectiveness of manager-owners and powerful blockhold- We have found that privatization has stronger effects in ers (including banks and investment funds) would not be non-CIS countries than in the CIS and that different types so greatly diminished because of their direct access to of owners have different effects. This immediately raises the power, for example, by blockholders quickly installing question of whether the latter finding could explain the for- their own managers (Barberis and others 1996). The own- mer. One could directly address this question by using ers dependent on institutional help are diffuse individual data on ownership in different countries, but there is no owners, outsiders when there are a number of different systematic collection of such data. Nevertheless, the papers blockholders, and perhaps even workers. Given these used for this study do contain some evidence on ownership. observations, the pattern of ownership effects in figure 2 is The strong impression gained from this evidence is that broadly consistent with the argument, most forcefully pro- worker and diffuse individual ownership are more preva- posed by Fox, Merritt, and Heller (1999) and Coffee lent in the CIS than in non-CIS countries, while foreign, (1999), that corporate governance institutions functioned investment fund, concentrated individual, and bank own- less well in the CIS than elsewhere. ership is less prevalent. Thus, since the CIS has an ownership Thus, we conclude that the effectiveness of privatiza- portfolio that contains a greater share of less effective own- tion in the CIS, relative to non-CIS countries, has been ers, structure of ownership is a strong candidate to explain diminished by two factors. First, ownership in the CIS is differences in the effects of privatization between regions. higher among those types of owners who are less effective 13 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence FIGURE 2. REGIONAL VARIATIONS IN THE EFFECTS OF DIFFERENT TYPES OF OWNERS Comparing Ownership Effects in the CIS to those in the Non-CIS Countries Relative effect ot privatizing to specific type of owner 300- 275.1 250 CIS Non-CIS 200 150 125.0 115.7 101.5 103.7 100.0100.0 100 - - l A 74 5 60.6 630 59.3 3322. 5 13.7197 3 2-5. Category of Owners everywhere. Second, the types of owners that need institu- non-CIS countries, perhaps because they have received tional help (diffuse individual, outsiders, and workers) less assistance from institutions in the CIS than elsewhere. seem to have fared relatively worse in the CIS than in the 14 5. The Role of Managers in Enterprise Restructuring T he previous two sections have documented where privatization enhances enterprise restructur- ing. They have shed less light, however, on the precise mechanisms by which privatization yields greater efficiency. One explanation, discussed in the transition literature, is that private owners are better at selecting managers who can run the firm efficiently. The hypothesis that management turnover, or more broadly, bringing in new human capital, is important in improving enterprise performance was first put forward and tested by Barberis and others (1996) for a sample of privatized Russian shops. An alternative hypothesis states that what matters for od once a private sector emerged (Pinto, Belka, and the performance of managers is the correct incentive struc- Krajewski, 1993). Sections 3 and 4 present equivocal evi- ture. This includes both "sticks" and "carrots": if managers dence on this hypothesis, since traditional state-owned do not perform well they are dismissed, and if they run the firms have performed poorly, but commercialized firms firm well they receive better remuneration. A corollary to have performed somewhat better. this hypothesis is that management turnover is not neces- We have identified six studies that test the impor- sary to enhance restructuring efforts, except perhaps as a tance of managerial turnover and managerial incentives in signaling device to managers who may want to shirk restructuring. To understand the aggregate implications responsibility. This hypothesis has been illustrated in the of these studies, we present the composite information in case of Poland, where managers of state-owned enterpris- table 4 in exactly the same manner as that adopted in sec- es initiated restructuring efforts in the early transition peri- tion 3. The construction of the information in panel A TABLE 4. THE ROLE OF MANAGERIAL TURNOVER AND MANAGERIAL INCENTIVES IN ENTERPRISE RESTRUCTURING Assessments of the Composite Implications of Six Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate (1) (2) When there is no attempt to weight the When the relative importance importance of the studies' results by attached to each study's results reflects the quality of their methodologies the methodological quality of the study A. Do polices affecting managers matter? Assessment of the likelihood that policies affecting managers are effective in promoting restructuring. 1. Mixture of management turnover Extremely likely that changes in turnover Extremely likely that changes in turnover and incentives and in incentives, taken together, work and changes in incentives, taken together, work 2. Management turnover Extremely likely that changes in turnover work Extremely likely that changes in turnover work separately 3. Management incentives No evidence that changes in incentives work No evidence that changes in incentives work separately B. How do policies compare? Assessment of the likelihood that changes in turnover have a stronger effect on restructuring than changes in incentives 4. Comparison of the effectiveness Extremely likely that changes in turnover Extremely likely that changes in turnover of turnover and incentives are more effective than changes in incentives are more effective than changes in incentives 15 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence employs exactly the same methods as those used for table is further bolstered by the findings that management 1. We find that turnover and incentives, considered togeth- turnover also contributes to enterprise restructuring in state- er, are an important determinant of restructuring. owned enterprises, that is, it is not dependent on the strong Management turnover on its own also has a significant monetary incentives that come with private ownership. effect on restructuring. But manager incentives are not What is the economic significance of management significant on their own. turnover? Barberis and others (1996) find that manage- Panel B directly compares the effect of turnover to the ment turnover more than doubles the likelihood of reno- effect of incentives. The methods employed are exactly vation occurring in Russian retail shops. It increases the the same as those used to derive lines 2, 4, and 6 of table amount of extra hours worked by 80 percent, and induces 2, when we were comparing the effects of privatization in 50 percent more change in suppliers. Claessens and two different regions. The results leave us in no doubt Djankov (1999) find that management turnover in state- whatsoever that turnover is much more effective in pro- owned and privatized enterprises results in 1.9 percent ducing restructuring than are changes in management and 6.2 percent higher labor productivity. Frydman, incentives. What explains the great importance of man- Hessel, and Rapaczynski (1998) find an even larger effect agement turnover? This points to the importance of human on labor productivity, 7.3 percent, in their sample of capital that is new to the enterprise, an interpretation that Central European firms. 16 6. Enterprise Restructuring and Hardened Budgets S oft budgets occur when enterprises have the expectation that the state (or other economic actors) will come to their aid when they are in financial trouble. When such aid is expected, incentives to perform efficiently are muted. Soft budgets were pervasive under the old socialist system, and, not surprisingly, there was much emphasis on hardening budget- constraints in the transition period. However, this emphasis begs the question of why soft budgets occur. If state ownership is an essential causal mechanism in producing soft budgets, then a change in the amount of soft-budget aid to enter- prises is one channel for privatization to have an effect. Three alternative theories exploring the causes of These three theories of the causes of soft budgets dif- soft-budget constraints have been suggested in the transi- fer significantly. The first explains accommodating lending tion literature. First, Janos Kornai (1979, 1998) relates the behavior induced by a benign government's paternalism, softness of budget constraints to the paternalistic attitude while the second suggests that soft budgets arise from of the government in socialist economies, which results in politicians' self-interest. In both, soft budgets compensate the accommodation of enterprise requests for extra financ- the enterprise for keeping surplus employment. The pre- ing. Firms are financed even when the expected return is dicted effect on enterprise restructuring from soft budgets below the real interest rate. The government's goal is to pre- is the same in both cases: lack of productivity improve- vent threats of job losses and to provide auxiliary services ments and continuation of unprofitable production (and (such as kindergartens, schools, hospitals, and recreation nonproduction) activities. The third explains an undesir- facilities) at the enterprise level. That is, soft budgets are a able outcome of optimal decisions by a financial institution substitute for a functioning social safety net. in a situation of imperfect information. The prediction on A second, complementary reason for the existence of enterprise restructuring is improved performance over time soft budgets has been advanced in Shleifer and Vishny as the investment enters the production process. (1994). They model the bargaining between politicians The predictions on the channels of soft budgets also and managers. Politicians pursue noneconomic objectives differ among the three theories. The first theory suggests in order to enlarge their political constituency, for example, that the central government will be the main source of soft by keeping enterprise employment high. Managers provide financing. The second supports the notion that local politi- the higher levels of employment, while politicians use cians provide soft budgets through direct subsidies, tax the state treasury to pay subsidies to sustain the extra exemptions, or arrears. Finally, the third hypothesis employment. identifies banks (or financial intermediaries more general- A third analysis views soft budgets as the continued ly) and suppliers of trade credit as the main channel of soft extension of credit even when the substandard perfor- financing. mance of an already-financed investment project has been Most of the literature that documents the use of dif- revealed (Maskin 1999). Poor information will lead bad ferent channels of soft budgets during early transition sup- projects to be initially financed. Then, by the time creditors ports the first hypothesis. Schaffer (1998) finds that bank can observe project quality, they will continue to lend lending is the primary source of soft budgets in transition because refinancing may be better than canceling a project countries, where the banking sector is in central state that is under way. hands. Tax arrears to the central government are the main 17 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence source of soft financing in Hungary and Poland. Anderson, focus on the question of whether hardening budget con- Korsun, and Murrell (1999), using a survey of 250 straints would entail improvements in enterprise perfor- Mongolian enterprises, show that central government own- mance and what types of restructuring would be most ership is the prime determinant of soft budgets. In contrast, likely. However, we have identified seven papers that use the McKinsey Global Institute (1999) shows that tax statistical analysis to examine the link between restructur- exemptions by the local government are the main channel ing and soft budgets. Again, the methodology used is iden- of soft financing in Russia. Claessens and Djankov (1998) tical to that of section 3. The results appear in table 5, use a sample of more than 6,000 enterprises in seven whose construction is identical to that of table 4, except Central and East European countries to show that the that panel B now examines differences between regions. availability of bank credit to nonviable enterprises is asso- The effect of hardened budgets on enterprise restructuring ciated with the importance of politicians in regulating the (defined as sales growth or productivity growth) is seen particular industry and the corruptibility of politicians. clearly in non-CIS countries, but does not appear for CIS They conclude that the evidence provides significant sup- countries. A possible explanation for this result is docu- port for the Shleifer-Vishny model. mented by the McKinsey Global Institute (1999): politi- Transition experience provides little evidence that cians often complement soft budgets with barriers to com- points specifically to the third hypothesis. Schaffer's (1998) petition from imports or new local entry. In such evidence on bank lending and soft budgets suggests that the circumstances, an enterprise can show artificially higher critical factor is that the banking sector is in central state labor productivity as it captures or keeps a large share of hands. He finds that trade arrears are not a major channel the market. of soft financing, since on average they comprise a payment The final row of table 5 compares the size of the period of 2 months. This finding compares favorably to the hardened budget effect across the two regions. The studies level of trade arrears in mature market economies. on non-CIS and CIS countries show effects that are of McMillan and Woodruff (1999a) show that trade creditors similar magnitude, which are judged, in a statistical sense, in Vietnam stop financing enterprises once their payments to be the same size as each other. We are left with the are two months in arrears. slightly paradoxical set of conclusions that there is evidence Most of the empirical studies of soft budgets to date for the effects of hard budgets in the non-CIS countries, but focus on causes and the channels of transfer. There is less not in the CIS countries, while there is no evidence that the TABLE 5. THE IMPORTANCE OF HARDENING BUDGET CONSTRAINTS IN ENTERPRISE RESTRUCTURING Assessments of the Composite Implications of Seven Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate (1) (2) When there is no attempt to weight the When the relative importance attached importance of the studies' results by the to each study's results reflects quality of their methodologies the methodological quality of the study A. Does hardening of budget constraints matter? Assessment of the likelihood that a hardening of the budget constraint is effective in promoting restructuring 1. For all countries Extremely likely that a hardening of the Extremely likely that a hardening of the budget constraint leads to restructuring budget constraint leads to restructuring 2. For Non-CIS countries Extremely likely that a hardening of the Extremely likely that a hardening of the budget constraint leads to restructuring budget constraint leads to restructuring 3. For CIS countries No evidence that a hardening of the No evidence that a hardening of budget constraint works the budget constraint works B. How do regions compare? Assessment of the likelihood that a hardening of the budget constraint has more effect in the non-CIS countries than in the CIS 4. Comparison of the No evidence that a hardening of No evidence that a hardening of effectiveness of hard the budget constraint has an effect of the budget constraint has an effect of budgets in the two regions a different magnitude in the two regions a different magnitude in the two regions 18 Enterprise Restructuring and Hardened Budgets size of the effects of hard budgets differs between the two gets reduce the amount of labor shedding by 4 percent regions. Such a paradox is always possible in statistical annually in Eastern Europe and the amount of productiv- analysis duc to the effects of random variation in the data. ity growth by 3 percent. In the CIS, one study has found In this case, it probably results from the fact that the esti- that soft budgets can diminish labor productivity growth mated effects of hard budgets in the CIS vary greatly in size. by as much as 6 percent a year. But as the previous para- What is the economic significance of soft budgets on graph indicates, the effects of hard budgets in the CIS are enterprise restructuring? Studies have found that soft bud- often insignificant. 19 7. Product Market Competition here is substantial theoretical literature that studies the relationship between competition and corporate efficiency. The initial period of transition from central planning to capitalism pro- vides a unique opportunity to test the importance of product market competition on the sub- sequent performance of enterprises. This is because the majority of transition economies liberalized their trade regimes relatively fast. Some went on to de-monopolize their industrial sectors through breakups of conglomerates and spin-offs of individual production units and by allowing entry of new private firms. The short period in which these changes took place from abroad as are CIS countries. This might be because of allows the researcher to identify the timing of the policy the underdeveloped transport infrastructure in CIS coun- change and control for other economic or firm-specific tries or because their regional governments shield produc- variables. ers from foreign competition. Putting barriers on import We have identified 13 studies that explicitly investi- competition is a cheap way for regional governors to sub- gate the effect of product market competition on enterprise sidize inefficient local producers. Finally, a number of CIS restructuring. These studies provide results sufficient to countries, particularly in Central Asia and the Caucasus, examine the effects of competition separately in the two dif- have an industrial sector geared toward extracting and ferent regions and to analyze the effects of two rather dif- processing industries, while imports comprise the majori- ferent measures of competition-import and domestic. ty of consumer goods. In such countries, while the average The results appear in table 6. The analysis leading to this import penetration may be high, there is little direct com- table and its structure are identical to that of table 1. petition within industries. Changes in domestic market Overall, the analyses indicate that product market com- structure are important in explaining enterprise restruc- petition has been a major force behind improvements in turing in both the CIS and the non-CIS samples. enterprise productivity in transition economies. When we Table 7, based on the methodology used for table 2, divide the sample into analyses based on import competi- directly examines whether different types of competition tion versus domestic market structure (for all countries have effects of different magnitudes, and whether there are together), we find that each analysis is significant in regional differences in the effects of competition. In the explaining enterprise performance. Examining the effects combined results for CIS and non-CIS countries, there is no of competition in each of the regions, table 6 shows that evidence that competition from local producers has a there is clear evidence of the effects of competition for the stronger effect than import competition. When combining non-CIS countries, but the evidence is much more equivo- the results for all types of competition, however, we find that cal for the CIS region. competition has a stronger effect in explaining enterprise A further subdivision of the studies shows a very restructuring in non-CIS countries than it does in the CIS interesting pattern: while import competition in the CIS countries. This regional difference is primarily due to the countries does not have a significant effect on enterprise effects of import competition, which has a larger effect in restructuring, it is always very significant in explaining the non-CIS countries than it does in the CIS countries. The enterprise restructuring in the non-CIS sample. What could last comparison shows that there are no discernible patterns explain this difference? EBRD (1998) shows that, on aver- in the way in which the effects of domestic competition dif- age, non-CIS countries are twice as open to competition fer between the CIS and non-CIS countries. 20 Product Market Competition TABLE 6. THE EFFECT OF COMPETITION ON ENTERPRISE RESTRUCTURING Assessments of the Composite Implications of 13 Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate: Assessment of the likelihood that competition is effective in promoting restructuring When there is no attempt to weight the When the relative importance attached Countries Type of importance of the studies' results by the to each study's results reflects the included competition examined quality of their methodologies methodological quality of the study 1. All All Extremely likely that an increase in competition Extremely likely that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 2. Non-CIS All Extremely likely that an increase in competition Extremely likely that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 3. CIS All No evidence that an increase in competition Probable that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 4. All Import competition Probable that an increase in competition Extremely likely that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 5. All Domestic market structure Probable that an increase in competition Extremely likely that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 6. Non-CIS Import competition Extremely likely that an increase in competition Extremely likely that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 7. CIS Import competition No evidence that an increase in competition No evidence that an increase in competition leads to more enterprise restructuring leads to more enterprise restructuring 8. Non-CIS Domestic market Probable that an increase in competition Probable that an increase in competition structure leads to more enterprise restructuring leads to more enterprise restructuring 9. CIS Domestic market Probable that an increase in competition Probable that an increase in competition structure leads to more enterprise restructuring leads to more enterprise restructuring TABLE 7. THE RELATIVE EFFECT OF DIFFERENT TYPES OF COMPETITION IN DIFFERENT REGIONS Assessments of the Composite Implications of 13 Empirical Studies Analyzing the Experience of Transition Countries Implications of the studies taken in the aggregate (1) (2) When there is no attempt to weight the When the relative importance attached to importance of the studies' results each study's results reflects the Type of comparison being made by the quality of their methodologies methodological quality of the study 1. Import competition versus domestic No evidence that the effect of import No evidence that the effect of import market structure competition is different from that of competition is different from that of domestic market structure domestic market structure 2. All types of competition in the CIS Extremely likely that competition in the Probable that competition in the non-CIS versus all types in the non-CIS countries non-CIS countries has a stronger effect on countries has a stronger effect on restructuring than competition in the CIS restructuring than competition in the CIS 3. Import competition in the CIS Probable that import competition in the Probable that import competition in the versus import competition in the non-CIS countries has a stronger effect on non-CIS countries has a stronger effect on non-CIS countries restructuring than import competition in the CIS restructuring than import competition in the CIS 4. Domestic market structure in No evidence that domestic market structure No evidence that domestic market structure the CIS versus domestic market in the non-CIS countries has a stronger effect in the non-CIS countries has a stronger effect structure in the non-CIS countries on restructuring than domestic market on restructuring than domestic market structure in the CIS structure in the CIS 21 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence The economic effects of competition are large. The changes in enterprise restructuring in response to changes studies surveyed here imply that in CIS countries, firms that in market structure exhibit diminishing returns. Since the face near-perfect competition are 40 to 60 percent more non-CIS countries started the transition process earlier, efficient than enterprises that operate in near-monopoly the effects of additional changes in competitive pressures markets, while the efficiency gain is 30 percent in non-CIS may be smaller. countries. This difference may be due to the fact that 22 8. The Role of Institutions in Enterprise Restructuring P he beginning of transition coincided with the publication of North's (1990) influential book, with its central message that institutions provided a crucial underpinning to market capital- ism and that the process of building these institutions was fraught with difficulties. This mes- sage was not at the forefront of policy discussions during the early years of transition. Stabilization, privatization, and liberalization dominated the agenda. Gradually the focus has changed, spurred by studies showing the hefty costs of inefficient state administrations and corruption (Kaufmannl994) and by the recognition that the relatively poor perfor- Kremer hypothesis. But the results are also consistent with mance of the CIS countries was not easily explained by dif- the view that the breakdown of old relationships, that is, ferences in more standard reforms. Some scholars have also the destruction of information and relationships, might be ascribed the disappointing Czech economic performance to the critical factor rather than weak institutions. a lack of attention to corporate governance and the finan- Institutional reform can lead to improved enterprise cial system during mass privatization (Coffee 1996). Now, efficiency when legal rules are effective in structuring eco- in contrast to the early neglect, institutions are in vogue nomic transactions and resolving disputes. Economic (Johnson, Kaufmann, and Shleifer 1997; Blanchard and agents can then turn to public bodies, such as the courts Kremer 1997; Stiglitz 1999). and the police, to enforce those rules. Institutional reforms Restricting ourselves to enterprise-level empirical stud- may therefore enhance enterprise restructuring if the legal ies of the determinants of enterprise restructuring, as we do system replaces more costly private mechanisms of sup- in this paper, there is only a relatively small amount of evi- porting transactions. Focusing on private Vietnamese firms, dence on the importance of institutions. One reason for this McMillan and Woodruff (1999a, b) document the nature is that research has tended to follow policy, focusing on pri- of enforcement of trading relations when formal institu- vatization, competition, and soft budgets rather than on tions are virtually nonexistent. Trading relations depend on institutions. Thus, our review of the evidence on institu- reputation, which are built using information from business tions necessarily examines only a small number of studies. networks or prior experience, with networks used to sanc- Since these studies vary widely in methodology and focus, tion defaulting customers. But these private mechanisms we cannot synthesize the results using the methods of pre- may lead to inefficiency. Reliance on private sources of vious sections. The findings in this section are less emphat- information requires firms to continue to deal with cus- ic: the enterprise-level evidence on the link between insti- tomary trading partners, which means refusing to deal tutional reform and enterprise restructuring is still thin. with new entrants, and consequently less restructuring in An influential paper by Blanchard and Kremer (1997) procurement activities. has claimed that the absence of contract enforcement Formal business associations and informal networks mechanisms was a primary factor in causing the dramatic can also serve as repositories of information and disposers fall in output during early transition in the CIS. They of sanctions, supporting transactional activities. Such asso- hypothesize that weak contract enforcement will be more ciations have emerged spontaneously during the transi- critical for those enterprises whose input-supply relation- tion process, and there is some evidence that their members ships are more complex, a prediction that also follows are more likely to undertake restructuring activities. from the observation that the supply of information and Some commentators have argued that the absence of the coordination of decisions was a central task of the institutions can lead to a reliance on criminals as contract now-defunct planning apparatus (Murrell 1992). There are enforcement agents, perhaps even spurring the rise of such several papers that test this hypothesis using enterprise-level groups. The overall picture, obtained from the rather small data, leading to only weak support for the Blanchard- amount of evidence available, does not suggest the extreme 23 The Determinants of Enterprise Restructuring in Transition: An Assessment of the Evidence failure of formal contract enforcement institutions and The above paragraphs have focused on the direct heavy reliance on extra-legal methods of enforcement that effects of institutional reform on enterprises. But indirect had sometimes been suggested. effects might be just as important. When good institutions The more usual way in which criminal groupings are are lacking, costly substitutes might be needed. Those expected to affect businesses is when such groupings wield owners who are most effective in a world of perfectly their comparative advantage, for example, by running pro- functioning institutions might be relatively less effective tection rackets and stealing goods and cash. Such criminal when corporate governance institutions do not function activity certainly represents a failure of institutional reform, well or when contract enforcement is weak. For example, in this case of law enforcement institutions. Johnson, Hendley, Murrell, and Ryterman (forthcoming) find that McMillan, and Woodruff (1999) find remarkable variation increases in both state ownership and employee control in such activity across Eastern Europe: while less than 1 raise the effectiveness of enterprise transactions. A decrease percent of Romanian firms make payments for protec- in competition increases the success of transactions. The tion, more than 90 percent of Russian firms do so. But explanation for these results is that alternative mecha- these direct costs are only part of the picture, since crimi- nisms substitute for weak institutions. In the dire eco- nal activity also reduces the incentive for enterprise restruc- nomic conditions of Russia, the probability that the enter- turing. Using the opinion of managers on whether courts prise will survive and the probability that enterprise can enforce contracts as a measure of property rights personnel will be around to implement long-term agree- enforcement, the same authors estimate that firms per- ments are greater the smaller nonstate outsider ownership ceiving property rights to be insecure invest nearly 40 per- is. Similarly, when contracts are poorly enforced, increases cent less than firms that perceive property rights to be in competition expand the opportunities for firms to use adequate. These studies suggest that, at low levels of insti- threats of defaulting on their contracts. This analysis sug- tutional development, lack of enforceable property rights gests that institutional weaknesses can reduce the potency might be more important than the absence of external of policies that previous sections have shown to be effective. financing in determining investment in new projects or Conversely, institutional innovations can help to mod- expanded capacity. erate the deleterious effects of less-than-optimal policies. The creation of effective mechanisms of corporate Prasnikar and Svejnar (1998) show that Slovenian work- governance was at the heart of the early institutional ers in state-owned firms appropriate depreciation funds less reforms that were aimed at the firms on which this paper than other funds, because of a rule that these must be is focused: the large firms beginning the transition in the used for investment. Hence a crude institution, a rule and state sector. Surprisingly, however, there has been little sys- its enforcement, can counter deficiencies in policies else- tematic empirical work at the enterprise level on the effects where, for example, when workers might be tempted to of corporate governance institutions. While Black, decapitalize state-owned firms. This study also shows that Kraakman, and Tarassova (1999) and Fox and Heller state-enterprise managers who have their own private firms (1999) for Russia, and Stiglitz (1999) more generally, claim do not siphon off cash flows to those firms. The authors that the failure of corporate governance institutions has interpret this as evidence of a well-functioning system of been of great importance, their evidence is anecdotal. penalties for breach of management contracts. However, Anderson, Korsun, and Murrell (1999) do use systematic seemingly sensible second-best institutions fail as well, as survey evidence to show that corporate governance laws Djankov (1999) shows for the enterprise isolation pro- work poorly in Mongolia, but they present no evidence on gram in Romania. whether there is a cost in terms of foregone restructuring. This section is ample testament to the disjointedness Similarly, the evidence that we present in section 4 on the in the enterprise-level evidence on the effect of institutions effects of different owners in the CIS and Eastern Europe on restructuring. Thus, the major difference between this is consistent with greater dysfunction of corporate gover- and the preceding sections, which is the absence of tables nance institutions in the CIS, but the argument is indirect. synthesizing the major results, reflects the state of the lit- Further enterprise-level work on the effects of corporate erature. Evidently, if institutions are to deserve the promi- governance institutions is certainly of some urgency, given nence in policy deliberations that they presently have, the present policy importance of the topic and the paucity empirical work at the enterprise level is a matter of some of existing evidence. urgency. 24 The Role of Institutions in Enterprise Restructuring 9. Summary his study documents and synthesizes the empirical evidence on the determinants of enterprise restructuring in the early years of transition from central planning to a market economy. The purpose here has been to present the evidence in an unvarnished manner, rather than provid- ing interpretative commentary that emanates from the authors' own views of the transition process. Similarly, in this conclusion we refrain from making any judgments on the implications of the results for the choice of policies for the future or on decisions made in the past. We believe that there are a large number of implications that our presentation of facts tive privatization (to diffuse individual ownership). has for policy, but discussion of those implications is best Managers are more than nine times as productive as reserved for different papers. In that way, we can make the diffuse individual ownership. Privatization to out- clearest statement of what exists in the empirical evidence. siders is associated with 50 percent more restruc- In this spirit, we will not provide an overall conclu- turing than is privatization to insiders (managers sion, but rather a summary of the evidence presented above and workers). State ownership within traditional as we see it. The following are the main facts revealed by state firms is less effective than all other ownership the synthesis of the empirical evidence on restructuring in types. transition economies: 4. At a rough approximation, there are three group- ings of owners. At the bottom, traditional state 1. Privatization is strongly associated with more enter- ownership and diffuse individual ownership have prise restructuring. However, the evidence varies similar effects. In the middle, insiders, outsiders, between geographical regions. The empirical liter- workers, banks, and commercialized state owner- ature resoundingly endorses the hypothesis that ship are clustered. The most effective owners are private ownership produces more restructuring managers, concentrated individual ownership, than does state ownership in the non-CIS region. In investment funds, and foreigners. contrast, evidence is mixed for the CIS: the most 5. Undifferentiated insiders are not as good owners as reasonable interpretation of the evidence for that are managers and workers separately, implying that region is that there is no reason to conclude that privatizing to heterogeneous groups might be worse either private ownership is superior to state own- than privatizing to homogeneous groups. ership or the reverse. 6. A notable result is that state ownership in com- 2. Not surprisingly, given the previous point, the move mercialized enterprises is quite effective. This result from state to private ownership has a much appears across a wide range of studies, from stronger effect in the non-CIS countries than in the Mongolia to Central Europe. CIS countries. The privatization effect in the 7. The relative effects of different owners vary non-CIS countries is more than twice the size of that between regions. Workers and outsiders are rela- in the CIS countries. tively better owners outside the CIS than in the 3. Different types of private owners have very differ- CIS, while banks and concentrated individual own- ent effects. The most effective privatization (to for- ership are more effective in the CIS than elsewhere. eigners) is 10 times as productive as the least effec- Indirect evidence suggests that these differences are 25 at least in part due to regional variations in the in CIS countries. This is due to import competition, strength of the legal and institutional environment. which has a larger effect in the non-CIS than in the 8. The effectiveness of privatization in the CIS, relative CIS countries. There are no discernible patterns in to non-CIS countries, has been diminished by two the way in which the restructuring effects of domes- factors. First, ownership in the CIS is higher among tic market structure differ between the CIS and those types of owners who are less effective every- non-CIS countries. where. Second, the types of owners that need insti- 14. Restricting ourselves to enterprise-level empirical tutional help have received less assistance from studies of the determinants of enterprise restruc- institutions in the CIS than elsewhere. turing, as we do in this paper, there is only a rela- 9. Management turnover is associated with improved tively small amount of evidence on the importance enterprise performance, in both the CIS and non- of institutions. CIS countries. We find no evidence that the 15. The literature suggests that when effective institu- strengthening of managerial incentives, on its own, tions are lacking, costly substitutes emerge in their leads to more restructuring. Management turnover place. This, in turn, implies that benefits could flow is much more effective in producing restructuring from second-best measures in other policy areas. than are changes in incentives. Institutional development can foster progress in 10. The hardening of budget constraints has had a ben- two ways, by helping to moderate the deleterious eficial effect on enterprise restructuring in the non- effects of suboptimal policies and by creating fertile CIS countries. For the CIS, however, the effects of territory for the implementation of first-best poli- hardened budget constraints do not appear in the cies. data on enterprise performance. Nevertheless, dif- 16. A central finding of the paper is that transition ferences between regions are not clear: the non-CIS policies have had similar effects on the restructur- countries and the CIS show effects of similar mag- ing process in CIS and non-CIS countries in terms nitude, which are not significantly different from of direction, but not in terms of economic or sta- each other. tistical significance. In particular, privatization, 11. Product market competition has been a major force hardened budget constraints, and product market behind improvements in enterprise productivity in competition all appear to be important determi- transition economies. nants of enterprise restructuring in non-CIS coun- 12. There is no evidence that competition from local tries, while they are less obviously so in the CIS. The producers has a stronger effect than does import evidence suggests, but not with any great certainty, competition. that the difference in impact is due to the varying 13. Competition has a stronger effect in explaining degree of institutional development between the enterprise restructuring in non-CIS countries than two regions 26 Summary References Anderson, James H., Georges Korsun, and Peter Murrell. Governance in Central Europe and Russia. Budapest: 1999. "Ownership, Exit, and Voice after Mass Central European Press. Privatization: Evidence from Mongolia." Economics - . 1999. "Privatization and Corporate Governance: of Transition 7(1): 215-43. The Lessons from Securities Market Failure." Journal Anderson, James H., Young Lee, and Peter Murrell. of Corporation Law 25: 1-39. Forthcoming. "Competition and Privatization Amidst Djankov, Simeon. 1999. "The Enterprise Isolation Program Weak Institutions: Evidence from Mongolia." in Romania." Journal of Comparative Economics Economic Inquiry. 27(2): 281-293. Balcerowicz, Leszek. 1995. Socialism, Capitalism, Djankov, Simeon, and Peter Murrell. 2000. "Enterprise Transformation. Budapest: Central European Restructuring in Transition: A Quantitative Survey." University Press. http://www.bsos.umd.edulecon/murrell/. Barberis, Nicholas, Maxim Boycko, Andrei Shleifer, and European Bank for Reconstruction and Development Natalia Tsukanova. 1996. "How Does Privatization (EBRD). 1998. Enterprise Performance and Growth: Work? Evidence from the Russian Shops." Journal of Transition Report. London. Political Economy 104 (4): 764-790. - . 1999. Ten Years of Transition: Transition Report. Berliner, Joseph. 1976. The Innovation Decision in Soviet London. Industry. Cambridge, Mass.: MIT Press. Fox, Merritt, and Michael Heller. 1999. "Lessons from Black, Bernard, Reinier Kraakman, and Anna Tarassova. Fiascos in Russian Corporate Governance." Working 1999. "Russian Privatization and Corporate Paper 282. William Davidson Institute, University of Governance: What Went Wrong?" Working Paper Michigan Business School, Ann Arbor, Michigan. 178, Stanford Law School. Frydman, Roman, Cheryl Gray, Marek Hessel, and Andrej Blanchard, Olivier, and Michael Kremer. 1997. Rapaczynski. 1999. "When Does Privatization Work? "Disorganization." Quarterly Journal of Economics The Impact of Private Ownership on Corporate 112(4): 1091-1126. Performance in the Transition Economies." Quarterly Boycko, Maxim, Andrei Shleifer and Robert W. Vishny. Journal of Economics 114(4): 1153-1191. 1995. Privatizing Russia. Cambridge, Mass. and Frydman, Roman, Marek Hessel, and Andrej Rapaczynski. London: MIT Press. 1998. "Why Ownership Matters? Entrepreneurship Claessens, Stijn and Simeon Djankov. 1998. "Politicians and the Restructuring of Enterprises in Central and Firms in Seven Central and Eastern European Europe." New York University Working Paper. Countries." Policy Research Working Paper 1954. Granick, David. 1987. Job Rights in the Soviet Union: World Bank, Washington D.C. Their Consequences. Cambridge: Columbia University - . 1999. "Enterprise Performance and Management Press. Turnover in the Czech Republic." European Economic Havrylyshyn, Oleh, and Donal McGettigan. 1999. Review 43(4-6): 1115-1124. "Privatization in Transition Countries: A Sampling of Coffee, John C., Jr. 1996. "Institutional Investors in the Literature." IMF Working Paper WP/99/6. Transitional Economies: Lessons from the Czech Hendley, Kathryn, Peter Murrell, and Randi Ryterman. Experience." In Roman Frydman, Cheryl W Gray, Forthcoming. "Law Works in Russia: The Role of and Andrzej Rapaczynski, eds., Corporate Legal Institutions in the Transactions of Russian 27 Enterprises." In Peter Murrell, ed., Assessing the Value _. 1999b. "Dispute Prevention without Courts in of Law in Transition Economies. Ann Arbor, Vietnam." Journal of Law, Economics, and Michigan: University of Michigan Press. Organization 15(3): 637-658. Johnson, Simon, Daniel Kaufmann, and Andrei Shleifer. Murrell, Peter. 1990. "Behavioral Regularities of Socialist 1997. "The Unofficial Economy in Transition." Systems." The Nature of Socialist Economies: Lessons Brookings Papers on Economic Activity 0(2): from Eastern European Foreign Trade. Princeton, 159-221. N.J.: Princeton University Press. Johnson, Simon, John McMillan, and Christopher _. 1992. "Evolution in Economics and in the Woodruff. 1999. "Why Do Firms Hide? Bribes and Economic Reform of the Centrally Planned Unofficial Activity After Communism." Center for Economies." In Christopher Clague and Gordon C. Economic and Policy Research (CEPR) Discussion Rausser, eds., Emergence of Market Economies in Paper 2105, London. Eastern Europe. Cambridge, Mass. and Oxford: Kaufmann, Daniel. 1994. "Diminishing Returns to Blackwell Publishers. Administrative Controls and the Emergence of the Nellis, John. 1999. "Time to Rethink Privatization?" IFC Unofficial Economy: A Framework of Analysis and Discussion Paper 38, World Bank, Washington D.C. Applications to Ukraine." Economic Policy 9(19): North, Douglass C. 1990. Institutions, Institutional 51-69. Change, and Economic Performance. Cambridge, Kornai, Janos. 1979. "Resource-Constrained Versus Mass., New York, and Melbourne: Cambridge Demand-Constrained Systems. Econometrica 4(4). University Press. _. 1986. "The Hungarian Reform Process: Visions, Pinto, Brian, Marek Belka, and Stefan Krajewski. 1993. Hopes and Reality." Journal of Economic Literature "Transforming State Enterprises in Poland: Evidence 24(4): 1687-1737. on Adjustment by Manufacturing Firms." Brookings _. 1990. The Road to a Free Economy. New York: Papers on Economic Activity 23(1): 213-261. Norton. Pohl, Gerhard, Robert E. Anderson, Stijn Claessens, and _. 1992. The Socialist System: Political Economy of Simeon Djankov. 1997. "Privatization and Restruc- Communism. Princeton University Press. turing in Central and Eastern Europe: Evidence and _. 1998. "The Place of the Soft Budget Constraint Policy Options." World Bank Technical Paper 368. Syndrome in Economic Theory." Journal of The World Bank, Washington D.C. Comparative Economics 26(1): 11-17. Prasnikar, Janez, and Jan Svejnar. 1998. "Investment and Korsun, Georges, and Peter Murrell. 1995. "The Politics Wages during the Transition: Evidence from Slovene and Economics of Mongolia's Privatization Program." Firms." Davidson Institute Working Paper 184. Asian Survey 35(5): 472-486. Schaffer, Mark. 1998. "Do Firms in Transition Economies Lipton, David, and Jeffrey Sachs. 1990. "Creating a Market Have Soft Budget Constraints? A Reconsideration of Economy in Eastern Europe: The Case of Poland." Concepts and Evidence." Journal of Comparative Brookings Papers on Economic Activity 1: 75-148. Economics 26(1): 80-103. Maskin, Eric. 1999. "Recent Theoretical Work on the Shleifer, Andrei, and Robert Vishny. 1994. "Politicians Soft-Budget Constraint." American Economic Review and Firms." Quarterly Journal of Economics 109(4): 89(2): 421-425. 995-1025. McKinsey Global Institute. 1999. Unlocking Economic Stiglitz, Joseph. 1999. "Whither Reform? Ten Years of Growth in Russia. Washington, D.C. the Transition." Paper presented at the Annual Bank McMillan, John, and Christopher Woodruff. 1999a. Conference on Development Economics, Washington, "Interfirm Relationships and Informal Credit in D.C., April. Vietnam." Quarterly journal of Economics 114(6): 1285-1320. 28 140 ) T en years after the fall of the Berlin Wall, THE WOR LD BANK the countries in Central and Eastern Europe 1818 H Street, N.W. and the Commonwealth of Independent Washington, D.C. 20433 USA States still face major challenges: although Telephone: 202-477-1234 much has been achieved in the transition to market-based democracies, much remains to Facsimile: 202-477-6391 be done. The World Bank has been an active Telex: MCI 64145 WORLDBANK partner in helping countries design and im- MCI 248423 WORLDBANK plement their reforms. Internet: www.worldbank.org E-mail: books@wvorldbank.org The region has witnessed many successes, as well as setbacks, in the transition process. This series of publications is part of the Bank's con- tribution to the debate about the unfinished agenda and possible approaches to future chal- lenges. The 14 titles in the series cover the issues of resurgent poverty and inequality, the importance of sound corporate citizenship, strategies for better education systems, social and environmental protection, institution building, investments, and livable cities. The series was prepared to facilitate the discus- sions during the thematic seminars at the 2000 Annual Meetings of the World Bank and the International Monetary Fund in Prague, and the broad-based dialogue to follow. 0-8213-4815-9