The Choice of Privatization Method and the Financial Performance of Newly Privatized Firms in Transition Economies

Transcription

1 Journal of Business Finance & Accounting, 30(7) & (8), September/October 2003, X The Choice of Privatization Method and the Financial Performance of Newly Privatized Firms in Transition Economies RANKO JELIC, RICHARD BRISTON AND WOLFGANG AUSSENEGG* 1. INTRODUCTION The commitment problem in socialism is related to state ownership and the state s paternalistic relationship with enterprises (Kornai, 1992). In spite of some well-documented market failures, it is generally believed that a system based on private ownership and clearly defined property rights provides better incentives for efficient decision-making than one based on central planning (Alchian and Demsetz, 1972; Grossman and Hart, 1986; Laffont and Tirole, 1993; and Boycko et al., 1994). Privatization programs therefore became a central part of the economic and political reforms in Central and Eastern European * The authors are respectively from The University of Birmingham Business School, The University of Hull Business School and Vienna University of Technology Department of Finance. They would like to thank an anonymous referee, Laura Casares Field, Mark Schaffer, participants at the Financial Management Association International Annual Meeting in Seattle (2000) and the European Financial Management Association Annual Meeting in Lugano (2001) for their helpful comments and suggestions. Wolfgang Aussenegg acknowledges the research project funding by the Austrian National Bank (OeNB-Jubiläumsfondsprojekt Nr. 8826). He also wishes to thank Reuters GesmbH Vienna for providing data. (Paper received July 2002, revised and accepted November 2002) Address for correspondence: Ranko Jelic, University of Birmingham, Birmingham Business School, Department of Accounting and Finance, Birmingham B15 2TT, UK. R.Jelic@bham.ac.uk # Blackwell Publishing Ltd. 2003, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA. 905

2 906 JELIC, BRISTON AND AUSSENEGG Countries (CEEC) initiated in the late eighties. However, neither academics nor policy makers were certain how to design and conduct large-scale privatization programs in the absence of a well functioning financial infrastructure (Lipton and Sachs, 1990; Bolton and Roland, 1992; Frydman and Rapaczynski, 1994; Brada, 1996; Cornelli and Li, 1997; and Estrin, 1998). As a consequence, policy makers in transition economies adopted radically different privatization techniques. The most popular were mass give-away privatization programs, followed by sales to insiders (employees and managers) and sales to outsiders (domestic and foreign investors). 1 Moreover, these programs were very different from those implemented in the OECD countries where a financial infrastructure was already in place (Estrin, 1991; and Laban and Wolf, 1993). While the design of privatization programs in CEEC has been studied extensively, little attention has been paid to the financial performance of privatized enterprises in the context of different privatization methods. This paper extends the debate on the design of privatization programs by comparing the different strategies adopted by the Polish, Hungarian, and Czech governments and the subsequent financial performance of the privatized enterprises. An investigation into the determinants of the financial performance of privatized firms is particularly important for other economies in transition, which are about to start their privatization programs. In some transition economies the state has continued to hold shares in privatized companies long after privatization. This raises some interesting questions with regard to property rights theory. Firstly, if governments wish to improve enterprise efficiency by preventing state interference in management they should not remain shareholders in privatized enterprises in the long run. Secondly, how does the performance of enterprises with mixed ownership compare with the performance of state and private enterprises? This study provides new evidence on the performance of enterprises with mixed ownership in the context of different types of privatization programs. It also makes a methodological contribution to the literature by using an international benchmark of Emerging Market privatizations outside the CEEC in the measurement of long-term abnormal returns in US dollars.

3 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 907 Finally, privatized firms represent an important part of total world investment. The total market value of privatized firms reached over $3.5 trillion in 2000, representing about 10% of the world s aggregate market capitalization. 2 Many of these privatizations are concentrated in the emerging markets of CEEC. Expressed as a percentage of GDP, CEEC has witnessed the most dramatic impact from privatization of any region in the World. 3 For example, the targeted percentages of state assets to be privatized in Hungary, Poland and the Czech Republic in the early nineties were 35, 50 and 60 respectively (Table 1A). At the same time the empirical literature demonstrates conclusively that country selection is a critical aspect of investing in emerging markets (Errunza, 1994). Our comparison of long-term returns for foreign investors in CEEC companies is, therefore, of interest both to investors in emerging markets in general and to those in privatization stocks in particular. Our results suggest positive average buy-and-hold returns in all countries. For example, foreign investors who invested in Polish, Hungarian and Czech newly privatized companies on the first trading day and held those shares for three years could have earned positive average buy-and-hold returns, reaching 144% in Poland, 81% in Hungary and 69% in the Czech Republic. However, only newly privatized companies in Poland significantly outperform various benchmarks whereas their Hungarian and Czech counterparts achieve positive, but not statistically significant average buy and hold abnormal returns. Our results remain robust after adjustment for well-documented biases in the measurement of long term returns. The longterm performance seems to be influenced by retained state ownership, firm size and the type of privatization method used (case-by-case as in Poland and Hungary versus mass privatization as in the Czech Republic). The remainder of the paper is organized as follows. In Section 2 we compare privatization programs in Poland, the Czech Republic and Hungary. We summarize the relevant literature and formulate hypotheses in Section 3, and outline our methodology and present our data in Section 4. In Section 5 we present the results relating to long run financial returns and their determinants. Finally, we summarize our conclusions in Section 6.

4 908 JELIC, BRISTON AND AUSSENEGG 2. PRIVATIZATION PROGRAMS AND THE DEVELOPMENT OF STOCK EXCHANGES IN POLAND, THE CZECH REPUBLIC AND HUNGARY (i) Privatization Programs The Hungarian government rejected voucher privatization schemes in favour of firm-by-firm privatizations at the outset of its economic reforms. Consequently, direct sale and privatization initial public offerings (PIPOs) have been dominant privatization methods since the creation of the State Privatization Agency (SPA) in Hungarian Law on Investment gave foreigners the right to own up to 100% of enterprises with favourable rules for the repatriation of profits, while the Hungarian Privatization Law allowed employees to buy up to 25% of shares offered at preferential terms. The Privatization Law of 1990 provided a legal framework for privatization in Poland. The Law was a compromise between various conflicting groups, which is probably why a variety of privatization methods were introduced (e.g. direct sales, PIPOs, voucher privatization). By the end of 1996 there were 184 completed capital privatization projects involving medium and large enterprises which required the organization of competitive public tenders and PIPOs (Privatisation Yearbook, 1998; p.93). PIPOs and a well- regulated stock exchange increased the attractiveness of privatization to workers and managers who would normally get shares in their companies on preferential terms. After many delays the voucher mass privatization program started in The program was designed for small to medium companies and was facilitated through 15 investment funds. The Czech Republic undertook a mass privatization via a voucher scheme and direct sales. The majority of companies (about 1,600) were privatized in two waves of the voucher scheme (Table 1A). The Czech government consciously discouraged equity ownership by managers and foreigners on the grounds that both groups would cannibalize companies with valuable assets (Hingorani et al., 1997). The Czech Ministry of Privatization started encouraging privatizations through direct sales only in the second wave of the programme. The different privatization methods favoured by Hungary, Poland and the Czech Republic reflect the rather different

5 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 909 goals set by these countries at the beginning of their privatization programs: Hungary emphasized efficiency, the Czech Republic speed, while Poland was trying to achieve both goals simultaneously (Table 1A). The Government in each of the countries allocated shares in a politically motivated manner; however, while the Czech government focused on the free distribution of shares to citizens, the Polish and the Hungarian governments preferred to sell shares to insiders (workers and managers) at large discounts. Consequently, insider ownership is important in Poland and Hungary and of little importance in the Czech Republic. 4 All governments gave up day-to-day control in the majority of privatized enterprises and did not often exercise an option to retain some form of veto power (e.g. golden share). 5 (ii) Development of Stock Exchanges Differences in privatization programs were reflected in different paths to the development of stock exchanges in CEECs (Table 1B). 6 For example, the mass voucher privatization program in the Czech Republic contributed to the flotation of practically all companies, which resulted in a record number of companies listed in the PSE (Prague Stock Exchange). The comparatively small number of traded stocks in the BSE (Budapest Stock Exchange) and the WSE (Warsaw Stock Exchange) reflects the preference for standard (firm-by-firm) privatization methods in Hungary and Poland, and the much higher threshold requirements for listings on these markets, which prevented a flood of relatively small companies. Liquidity, measured by the turnover/capitalization ratio, seems to be much higher in Hungary and Poland than in the Czech Republic. In late 1998, daily turnover in Budapest and Warsaw was about $30 m, compared with $15 m for the Czech Republic. 7 Poor liquidity and lack of market discipline and transparency forced the Czech government to de-list about 1,300 stocks in The leading sectors in all three markets (measured by market capitalization) are the telecommunication and banking sectors. According to the IFC classification of emerging markets, Poland and Hungary are considered to be 100% open to foreigners. The Czech market is open to

6 910 JELIC, BRISTON AND AUSSENEGG Table 1A Comparison of Privatization Strategies and Methods in CEEC Characteristics Hungary Poland Czech Republic Experiments with reforms in 1980s Yes; self-management and working councils; some elements of western legal code; some large enterprises broken into smaller units. Yes; working councils; some elements of western legal code. Main priorities Efficiency Efficiency and speed Speed, equality Targets in terms of % of assets to be privatized Private sector share in GDP by mid 1995 (1998) Main privatization methods Continuous or clustered sales Privatization before or after restructuring (organizational changes) 30 35% of total state assets; 2,200 state enterprises ( ) 50% of all enterprises ( ) 60% (80% by 1998) 60% (65% by 1998) 85% Direct sales; PIPOs Direct sales; PIPOs; vouchermass privatization (through 15 national investment funds) Continuous Continuous; voucher wave in 1995 Organizational changes before privatization Restitutions About 1.2 million Hungarians received tradable restitution coupons (US$ 1.1 billion total face value). Secondary market for coupons introduced in the BSE in Around 75% of restitution coupons exchanged for stakes in various privatized companies. Organizational changes before privatization Around 200,000 individuals qualified for restitution coupons. The coupons are not tradable. No data are available as to how many coupons have been traded for shares. No 60% of assets of state enterprises (in two waves of mass privatization; over 5 years) Voucher-mass privatization; direct sales Two waves, 1992 and Privatization before restructuring; except for steel industry, mining, telecommunications Large scale restitution law; more than 20,000 demands for restitution have been met.

7 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 911 Body in charge of privatization process State Privatization Agency and the State Holding Company Golden share policy Yes (e.g. Pick, Matav, Mol, OTP Bank, Regional gas and electricity companies); Employees Medium presence in large companies (6%); bigger presence in small and medium companies Ministry of privatization and State Treasury Ministry of privatization separated from State Treasury No Yes; (e.g. Becherovka, Beer and Porcelain industries) Very strong presence; working councils important; 20% shares (10% for free) went to employees Limited presence. Discount to employees Yes Yes n.a. Role of managers Very important role in privatization; Important role Important role in prepara- controlled after bad experience with spontaneous privatization tion of privatization plans and choice of privatization methods Participation of foreign investors More than 70% cases More than 10% cases More than 20% cases Participation of domestic private Small Significant Significant investors PIPO vs. Direct sale 1:2 1:2.8 n.a. Tradability of vouchers n.a. Vouchers tradable Vouchers not tradable Participation of investment funds n.a. Compulsory Encouraged in voucher privatization Notes: PIPOs are privatization share offerings; Direct sales are defined as the government s divestment through a direct sale to one or more investors; Joint ventures are included in sales; PIPOs include domestic and international issues, and private placements; Average percentage sold to employees excluding management; Results for standard methods (PIPOs and Sales) only. Source: Bolton and Roland (1992), Dewenter and Malatesta (1997), Earle and Estrin (1996), Estrin (1998), Hingorani et al. (1997), Jelic and Briston (1999), Jelic and Briston (2003), Privatisation Yearbook (1998) and Zloch-Christy (1995).

8 912 JELIC, BRISTON AND AUSSENEGG Table 1B Stock Exchanges in Poland, Czech Republic and Hungary WSE PSE BSE Opened Index/since WIG/1991;value weighted PX/1993; value weighted BUX/1991; value weighted Fin. Derivatives Index futures/1998 n.a. Index, currency, T-bill futures/ Companies (1998) (1400 before delistings in 1997) Biggest company TPSA (30% of total market capitalisation) SPT Matav (20% of total market capitalisation) Daily turnover (1998) $30m $15m $30m >90% 50% >90% % of total volume of transactions traded on the stock exchange Dual listing (Germany; %) Price ceiling 10% n.a. No Minority shareholders protection Weak Weak Weak Investors protection rules Good Weak (until 1996) Good Openness (IFC) Yes; fully Yes; except for banks Yes; fully Pearson r WSE-BSE ¼ 0.43 PSE-WSE ¼ 0.46 BSE-WSE ¼ 0.43 WSE-PSE ¼ 0.46 PSE-BSE ¼ 0.32 BSE-PSE ¼ 0.32 WSE-FTSE100 ¼ 0.20 PSE-FTSE100 ¼ 0.22 BSE-FTSE100 ¼ 0.27 WSE-S&P500 ¼ 0.15 PSE-S&P ¼ 0.09 BSE-S&P ¼ 0.31 WSE IFCICom. ¼ 0.46 WSE IFCICom. ¼ 0.34 BSE IFCICom. ¼ 0.46 Source: Poland (Warsaw Stock Exchange, WSE): Number of companies from IFC Emerging Stock Markets Factbook, Euromoney, The 1999 Guide to Poland, April 1999; Data for M.Cap/GDP ratio, turnover, and number of companies traded on German exchanges from Pistor (2001); Czech Republic (Prague Stock Exchange, PSE): Number of companies from Burza Ceennych Papiru Praha website; Data for M.Cap/GDP ratio, turnover, and number of companies traded on German exchanges from Pistor (2001); Hungary (Budapest Stock Exchange, BSE): BSE reports, as cited in The 1997 Guide to Hungary, Euromoney, September 1997; The 1998 Guide to Central and Eastern European Equities, Euromoney, February 1998; Data for turnover in 1998, number of companies in 1998, percentage of companies listed on German exchanges in 1998, minority shareholders protection and investors protection rules from Pistor (2001); shift in regulatory environment for PSE by authors; Data on Pearson r ( ) adopted from IFC (1998).

9 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 913 foreigners except for trading in bank stocks, which is subject to prior approval of the Czech National Bank. 8 Among these three countries Poland seems to have the strongest popular equity culture. This is in sharp contrast to Hungary where domestic investors own only about 30% of equities. 9 Another important characteristic of the WSE is a rather strict and detailed regulatory system designed to increase transparency and to provide security of trading. Despite some complaints about the extensive disclosure requirements, the regulation of the WSE is judged to be the best in Central Europe. 10 In all three countries the protection of minority shareholder rights was relatively weak after the privatization programmes (Pistor, 2001). 11 The nature of investor protection rules, however, differs considerably. The Czech Republic had the weakest investor protection rules at the outset of the reform (Pistor, 2001). 12 The main differences, however, largely disappeared with the introduction of the PSE new segmentation rules (1 September, 1995) and amendments to Commercial Code(1996). From the above summary, two different methods of privatization and two distinctive sub periods, in terms of differences in the investment environment, are apparent: First, the mass give-away privatization program in the Czech Republic associated with the formation of a large number of listed firms and a laissez faire approach towards the regulation of trading activities and investor protection rules until the mid-nineties. Second, the gradual (firm-by-firm) approach towards privatization with much slower expansion, and strict regulation, of the stock exchange followed in Hungary and Poland throughout the 1990s. 3. LITERATURE REVIEW AND HYPOTHESES The predominance of partial sales in selected CEEC resulted in numerous companies with a mixed ownership. While property rights theory clearly demonstrates that private enterprises should perform better than either enterprises in the state sector or enterprises with a mixed ownership, empirical evidence on the performance of enterprises with a mixed ownership is scarce and tends to concentrate on companies from regulated industries. In one of the rare studies, Boardman and Vinning (1989) analyze a

10 914 JELIC, BRISTON AND AUSSENEGG sample of the 500 largest non-us industrial firms and demonstrate that private enterprises outperform both state-owned enterprises and enterprises with a mixed ownership. They explain this result by the conflict between private and public shareholders in mixed enterprises, which inhibits the monitoring of management. Consequently, partial privatizations may be worse than complete privatization or continued state ownership. Boycko et al. (1996), for example, do not see any benefits from the involvement of government in the management of stateowned enterprises (SOEs). They show that the higher the fraction of SOEs sold, the lower is the possibility that politicians directly interfere. They conclude that the relationship between corporate restructuring activities and the fraction sold is likely to be positive, resulting in a positive relationship between the fraction sold and the long run performance of privatized companies. We therefore test the following hypothesis: H 1 : The fraction sold is positively related to the long-term performance of privatized companies. Anderson et al. (1997) study foreign participation in the Czech mass privatization program and find that foreigners prefer profitable firms in which they can obtain major shareholdings and can have undisputed control. It is suggested that this can be explained by lower agency costs and better control of political risks. Significant concentration of ownership might avoid the agency costs of managerial control, which tend to occur in privatizations without large investors. On the other hand, governments may be less willing to sell highly profitable enterprises to foreigners. 13 Finally, the large blocks typically owned by foreigners signal lower agency costs in the spirit of Shleifer and Vishny (1986). In CEEC this further implies lower resistance to and better chances for restructuring. 14 We, therefore, expect a positive relationship between foreign ownership and the long run performance of privatized companies: H 2 : There is a positive relationship between foreign ownership and the financial performance of privatized companies. In all transition economies governments remained influential in telecom, banking, and natural resources (water, gas, and electric) utilities. The government s involvement and monitoring

11 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 915 of companies in these industries after privatization can be explained by their importance for social welfare (Bös, 1991). The temptation for government to interfere is particularly strong in areas of traditional public ownership (e.g. utilities and infrastructure) where profits represent substantial rents and, therefore, may create political opposition to further market reforms. 15 In all transition economies, for example, governments interfered in pricing policy (e.g. by imposing price caps) for water, gas, and electricity after privatization. Government regulation of banking is due to the fact that banks in transition economies inherited non-performing loans from the pre-reform period (see Jelic et al., 1999). This required implementation of bank rehabilitation programs and regulation of banking activities during the post-privatization phase. In this sense, companies in these sectors are regarded as regulated as opposed to companies in other sectors in which government would not be directly involved in monitoring. Due to government interference, these companies are expected to have lower returns than their counterparts from non-regulated industries. The expected lower long term returns would also be consistent with findings that restructuring and privatization of regulated firms, such as natural monopolies, seem to be more complex and timeconsuming than in their counterparts (Galal et al., 1994; and D Souza and Megginson, 1999). On the other hand, Boardman and Laurin (1998) find support for the hypothesis that companies in regulated industries in developing countries are more likely to have higher returns than their counterparts in competitive industries. Some industries such as banking, telecommunication and utilities may enjoy a monopoly position in transition economies for a number of years. Good examples are the banking sector in the Czech Republic, and the Polish telecommunication sector, where government has only recently allowed competition and where high monopoly rents are expected. As there are two competing theories regarding the long-term performance of companies in regulated industries, we test the following hypothesis: H 3 : Companies in regulated industries will have the same long-run returns as those in unregulated industries.

12 916 JELIC, BRISTON AND AUSSENEGG Perotti (1995) and Dewatripont and Roland (1992) prefer a gradual approach to privatization, arguing that a speedy privatization may lead to premature restructuring and partial re-nationalization. For example, Perotti (1995) sees gradual privatizations as a signal of a government s commitment to its privatization program and a willingness to share political risk with investors after privatization. Gradual privatizations and underpricing in their model serve to build investors confidence in a stable policy towards privatized companies, which should further enhance revenues. On the other hand, populist governments prefer complete and rapid sales since they are expected to extract rents from private shareholders by changing policies (e.g. taxation, price regulation) after the sale (Perroti, 1995). Perotti and Guney (1993) present evidence for extensive underpricing and gradual sale programs in twelve countries. They also find that proceeds from privatization increase over time, suggesting that gradual selling contributed to investors confidence. It is important, however, to stress the difference between ownership retained by the state during gradual (firm-by-firm) privatization programs, where the government is explicitly committed to sell the retained shares in the near future (Poland and Hungary), and retained ownership in the context of mass giveaway privatization programs where the government emphasizes the quick break up between the state and the private sector (Czech Republic). State shareholdings in privatized companies seem to be difficult to justify in the context of mass give-away privatization programs with no definite sequencing. Gradual sales, increase in proceeds over time and extensive underpricing in Poland and Hungary are characteristics of credible privatizations as defined in Perotti (1995). On the other hand, the Czech voucher scheme seems to have led to the privatization of cash flow claims but not to the privatization of control rights. We, therefore, expect a better performance of newly privatized companies in Poland and Hungary, due to superiority of firm-by-firm privatization methods: 16 H 4 : Polish and Hungarian privatized companies outperform their Czech counterparts. The importance of solid regulatory and legal environments and the protection of minority investors for the development

13 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 917 and performance of capital markets are well-documented in the literature (Modigliani and Perotti, 1997; LaPorta et al., 1997; and Demirguc-Kunt and Maksimovic, 1998). Evidence is now emerging that stock markets in countries with weaker legal protection for minority shareholders tend to experience more severe declines during a crisis than countries with a better regulatory environment and protection for minority shareholders (Johnson et al., 2000). This further implies that previously noted differences in investment environment across CEEC, rather than different privatization methods, can drive the performance of newly privatized companies. We, therefore, control for the effect of different investment environment in CEEC when testing hypothesis H (i) Data 4. DATA AND METHODOLOGY We identify Hungarian and Polish PIPOs as well as privatizations in the Czech Republic from various issues of Privatization International, stock exchange fact books, REUTERS Business Briefing Archives, 18 and stock market databases bought (in the case of Poland and the Czech Republic) or received for free (in the case of Hungary) from national stock exchanges. Our investigation period starts with the resumption of the national stock exchanges (WSE: 16 April, 1991; BSE: 21 June, 1990; PSE: 6 April, 1993) and ends on 5 April, The total sample consists of all privatized companies first listed in the first or second market segment of the three exchanges. In order to measure the long-run performance up to the third anniversary of aftermarket trading, only firms that went public before 5 April, 1998 are included (inclusive of all delistings). Therefore, the final sample comprises 143 firms, 47 Polish PIPOs, Hungarian PIPOs and 65 Czech privatizations. 21 If a company is delisted before the third anniversary of aftermarket trading it is re-invested in the benchmark (until the third anniversary). Data on the percentage of shares sold and foreign participation were obtained from the financial press, Privatisation International Yearbook (1998), Dewenter and Malatesta (1997; unpublished

14 918 JELIC, BRISTON AND AUSSENEGG Appendix A), stock exchange fact books and REUTERS Business Briefing Archives. Information about daily stock prices, first trading day, seasoned equity offerings, stock splits, dividends and other company-related information as well as daily values for national and international stock indices, are extracted from databases of the three local stock exchanges, Thomson Financial Datastream and Reuters 3000 Equities History. Table 2 reports summary statistics for the total sample as well as the three national sub-samples. The mean size (market value on the first trading day) of all 143 privatized companies amounts to U$ million (see Table 2, Panel A). Polish and Hungarian PIPOs as well as Czech privatizations belong to the same size class. Mean as well as median market values of the three national sub-samples are not significantly different from each other (see Table 2, Panel B). Governments in the three transition economies sold on average more than two-thirds (68.9%) of the total share capital at the initial offer. The three countries significantly differ from each other in this respect. The average fraction sold is significantly higher in the Czech Republic (81.3%) than in Poland (67.3%) and significantly higher in Poland than in Hungary (45.3%) (Table 2, Panel B). The average fraction of privatized companies with foreign participation is highest in Hungary (74.2%) and significantly lower in Poland (12.3%) and the Czech Republic (23.1%). About a quarter of all privatizations in the total sample of 143 companies belong to regulated industries. The differences between the three national sub-samples are not statistically significant with respect to representation of regulated industries. (ii) Methodology The long-run post-issue performance is examined by analyzing buy-and-hold returns (BHRs) and buy-and-hold abnormal returns (BHARs). The buy-and-hold return for company i is defined as: " BHR i;t ¼ YT # 1 þ R i;t 1 ð1þ t¼1

15 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 919 where R i,t is the return of company i in period t, t ¼ 1 indicates the first trading day in the aftermarket and T is the aftermarket trading day at the third anniversary of aftermarket trading. The Table 2 Sample Characteristics Panel A: Summary Statistics* Total Poland Hungary Czech Republic Number of firms Mean market value c Median market value c Fraction sold in offer (%) Fraction of firms with foreign participation (%) Fraction of firms in regulated industries (%) Panel B: Mean and Median Differences Between the Three Sub-samples** Poland minus Hungary Poland minus Czech Rep. Hungary minus Czech Rep. Mean Median Mean Median Mean Median Market value c ( 0.225) (0.424) ( 0.172) ( 0.598) (0.117) ( 1.030) Fraction 22.0 a 33.3 a 14.0 a 29.7 a 36.0 a 60.0 a sold d (3.648) (3.237) ( 2.707) ( 2.960) ( 6.672) ( 5.324) Foreign 61.4 a a a a part. e ( 6.547) ( 4.569) ( 1.431) ( 0.929) (5.342) (4.035) Regulated ind. f (0.340) (0.822) ( 1.598) (0.236) ( 1.817) ( 1.277) Notes: * Summary statistics (period: 21 June, 1990 till 5 April, 1998) for four samples: (i) all 143 privatizations (Total), (ii) 47 Polish privatization initial public offerings (PIPOs) (Poland), (iii) 31 Hungarian PIPOs (Hungary) and (iv) 65 Czech privatizations (Czech Republic). ** Mean and median differences between the three sub-samples (Poland minus Hungary; Poland minus Czech Republic and Hungary minus Czech Republic) for several characteristics. It is tested whether (i) the differences in the mean and (ii) the differences in the median values are significantly different from zero. For the means a t-test and for the medians a Wilcoxon-Signed-Rank-Test is used. Test statistics in parentheses. a Significant at the 1% level. b Significant at the 5% level. c Market value in million U$ on the first trading day. d Fraction of the share capital sold in offer (%). e Fraction of firms with foreign participation (%). f Fraction of firms in regulated industries (%).

16 920 JELIC, BRISTON AND AUSSENEGG buy-and-hold return for the corresponding benchmark of company i is defined as: " BHR B;i;T ¼ YT # 1 þ R B; i;t 1 ð2þ t¼1 where R B,i,t is the return of the benchmark of company i in period t, t ¼ 1 indicates the first trading day of company i in the aftermarket, T is the aftermarket trading day at the third anniversary of aftermarket trading. Therefore, buy-and-hold returns over identical intervals are calculated for each company and its corresponding benchmark. To measure the abnormal long-run performance buy-and-hold abnormal returns are calculated for each company i as: BHAR i;t ¼ BHR i;t BHR B;i;T : Then, the average buy-and-hold abnormal return is given by: BHAR T ¼ 1 X n h i BHAR n i;t : ð4þ i¼1 A positive (negative) average buy-and-hold abnormal return can be interpreted as an outperformance (underperformance) of the sample of n companies relative to the benchmark. Four benchmarks are used to measure the abnormal performance. For domestic investors the benchmark returns are local currency returns achieved by investing in the national market indices. National market indices have the disadvantage that they also include privatized companies, which may reduce the calculated abnormal long-run performance of large privatizations. We therefore also use international benchmarks, namely the Datastream Europe Index (DS-Europe), which includes most European markets, and the Datastream World Index (DS-World) to capture world stock market performance. Finally, to avoid the problem that privatized stocks might have different risk profiles from companies included in the DS-Europe and the DS-World Indices, we construct and use a benchmark made of privatized companies in Emerging markets (excluding Poland, the Czech Republic and Hungary). To calculate the performance of a portfolio of Emerging Market privatizations the following steps are used: First, Emerging ð3þ

17 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 921 Markets are defined in accordance with the country classification of the International Finance Corporation s Emerging Stock Market Factbook Second, to identify privatized companies the following sources are used: several issues of the Privatization International Yearbook (1998), appendices in Boubakri and Cosset (1998) and Dewenter and Malatesta (1997) as well as tables in Boutchkova and Megginson (2000) and Perotti and Guney (1993). The total sample consists of 275 Emerging Market PIPOs from 24 countries. 22 Third, the portfolio of Emerging Market privatizations is rebalanced once a year on the first trading day in January. 23 The portfolio performance is based on a buy-and-hold strategy. On the rebalancing date an equal amount of money is invested in each privatization. Fourth, only privatizations with a first trading day not earlier than four years before the rebalancing date are used. This procedure ensures that sample firms of the three Eastern European transition economies are compared with companies recently privatized in other Emerging Markets. The base currency for all three international benchmarks is US dollars. Lyon et al. (1999) show that conventional tests of long-run abnormal security returns are often misspecified. They identify three main reasons for potential misspecification: first, survivalrelated biases, which occur if failing firms are excluded, second, rebalancing biases, which arise if cumulative return procedures are used and third, biases due to the fact that long-run abnormal performance measures are typically skewed. To minimize the first two sources of misspecification, all privatizations delisted on the three markets during the investigation period are included in the samples and buy-and-hold returns are used to calculate the long-run performance. In addition, to account for skewness bias, a skewness-adjusted t statistic with bootstrapped p-values (see Lyon et al., 1999) and a non parametric test are used to test the null hypotheses. We compare returns on the international indices with company returns calculated using share prices in US dollars. We thereby account for the existence of any paper profits due to changes in exchange rates between local currencies and US dollars. Finally, by using international indices we overcome problems where local firms are included in local market indices. 24

18 922 JELIC, BRISTON AND AUSSENEGG Finally, we control for the effect of different investment environments in CEEC. For the Poland and Hungarian stock exchanges, investment environments are comparable throughout the 1990s, but this is not the case with the Czech Republic where the stock exchange investment environment had been weaker compared to Poland and Hungary until the introduction of the new segmentation rules adopted by PSE (1 September, 1995) and amendments to the Commercial Code in As a result of these new stricter rules the investment environment in the Czech Republic is comparable to that in Poland and Hungary from 1995 onwards, which coincides with the second wave of the mass privatisation programme. We, therefore, control for the effect of the different investment environments in the Czech Republic by dividing our Czech sample into firms privatised in the first wave which operated in the first environmental regime (trading started in mid-1993) and the second wave firms which operated in the second environmental regime (trading started in March 1995). To test hypothesis H 4, (univariate and multivariate analysis) we use only Czech firms privatized in the second wave of the privatization program as the investment environment of firms privatized in the first wave is not comparable with the environment in Poland and Hungary. 5. ANALYSIS OF THE RESULTS Table 3 summarizes the results for the long-run aftermarket performance of privatized companies. The average and median local buy-and-hold returns (BHRs) for Polish and Hungarian PIPOs as well as for Czech privatizations are significantly positive. A domestic investor purchasing each Polish PIPO for an equal amount of money on the first aftermarket trading day would have earned a mean return of þ355% (or an average annual return of þ65.7%) holding the PIPOs until the third anniversary of public trading. The same trading strategy applied to Hungarian PIPOs would have yielded a mean return of þ208% (or an average annual return of þ45.5%), and þ74.4% for privatizations in the first and 54.5% in the second wave of the Czech Republic voucher program (or an average annual return of þ20.4% and 15.6%, respectively).

19 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 923 Table 3 Long Run Performance of Privatized Companies Total Poland Hungary CR 1 st Wave CR 2 nd Wave Mean Median Mean Median Mean Median Mean Median Mean Median BHR (local) a a a a a a b b (8.242) (4.432) (5.718) (3.355) (4.631) (1.616) (3.306) (1.640) (2.738) (2.236) BHAR (local) (0.756) (1.075) (0.855) ( 0.729) ( 0.306) ( 0.898) (1.367) (0.447) (1.285) (0.894) BHR ($) a a a a a a c (6.904) (2.592) (4.518) (1.021) (2.989) (0.539) (3.764) (3.177) (1.197) (1.789) BHAR ($) a a c a a DS-Europe (2.620) ( 1.791) (2.693) ( 0.729) (1.046) ( 1.257) (1.797) (0.447) ( 5.236) ( 3.578) BHAR ($) a a a b c DS-World (3.661) ( 1.254) (3.109) ( 0.729) (1.468) ( 0.898) (2.269) (0.447) ( 2.722) ( 1.789) BHAR ($) a a c EM-Priv. (3.374) ( 0.920) (3.475) (0.146) (1.324) ( 0.539) (0.918) ( 1.938) ( 0.108) (0.894) Notes: Buy-and-hold returns (BHRs) and Buy-and-hold abnormal returns (BHARs) exclusive of first day returns during the first three years of aftermarket trading for privatized companies in Poland, Hungary, the Czech Republic. The Czech Republic sample is split into firms sold in the first wave (CR 1 st wave) and the second wave (CR 2 nd wave) of the privatization program. BHARs are defined as the difference between the BHR of the issue and the BHR of the benchmark over the same period. Four benchmarks are used: (i) a domestic market index (Poland: WIG Index, Hungary: BUX Index, Czech Republic: PX50 Index), (ii) the Datastream Europe Index (DS-Europe), (iii) the Datastream World Index (DS-World) and (iv) a portfolio of Emerging Market privatizations (EM-Priv.), excluding Eastern European transition economies. The portfolio of Emerging Market privatizations is rebalanced once a year on the first trading day in January. The portfolio performance is based on a buy-and-hold strategy and on the rebalancing date an equal amount of money is invested in each PIPO. Only PIPOs are used with a first trading day not earlier than four years before the rebalancing date. Over the investigation period (21 June, 1990 till 5 April, 1998) a total of 275 privatizations are used for this portfolio. The local market indices are based on local currencies and the international indices are based on US dollars. BHRs are measured by equations (1) and (2) respectively. It is tested whether the BHRs and the BHARs are significantly different from zero. For the means a skewness-adjusted test statistic with p-values using a bootstrapping procedure is employed (see Lyon et al., 1999). For the medians a Wilcoxon-Signed-Rank-Test is used. Test statistics in parentheses. BHRs and BHARs in per cent. Significant at the 1% level. Significant at the 5% level. a b

20 924 JELIC, BRISTON AND AUSSENEGG A comparison with the local market index as benchmark shows that domestic investors were not able to earn significantly more by investing in PIPOs than in the local stock market. The average (median) BHARs for domestic investors are þ41.5% ( 14.0%) for Poland, 18.2% ( 34.2%) for Hungary, 15.3% (10.5%) for firms privatized in the first wave and 11.2% (4.6%) for firms privatized in the second wave of the Czech mass privatization programme. None of these average (median) BHARs are significantly different from zero. In the same way as domestic investors, foreign (US dollar) investors were also able to earn significant positive average BHRs (total sample: þ96.3%, Poland: þ143.8%, Hungary: þ80.9%, Czech Republic first wave: þ92.1%, Czech Republic second wave: þ17.9%). Due to the weakness of the Polish zloty and the Hungarian forint, US dollar-based BHRs are lower for these sub-samples compared to BHRs for domestic investors. This does not apply for the sample of first wave Czech privatizations. Regardless of which international benchmark is used, the average BHARs are positive for all three countries (in the Czech Republic only for first wave firms), but only significant for all three benchmarks in Poland. High and statistically significant average returns in Poland are consistent with results reported in earlier studies on Polish privatized companies (see Aussenegg, 2001; and Jelic and Briston, 2003). It is interesting that the 3-year abnormal performance is significantly negative for the subsample of firms sold in the second wave of the Czech Republic mass privatization program when the Datastream Europe or the Datastream World Index are used as benchmark. 25 To test hypothesis H 1, we compare the long-term aftermarket performance of privatized companies with a large fraction of the share capital sold with the performance of issues where the government initially sold only a small fraction of share capital (Table 4, Panel A). 26 For the total sample of 143 privatizations the difference in returns is positive but not statistically significant. The picture changes when national sub-samples are examined. Polish PIPOs in which the government sells initially a large fraction significantly outperform privatizations in which only a small fraction is sold. This applies to local and US dollar BHRs as well as BHARs for all three international benchmarks

21 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 925 Table 4 Long-run Performance Differences Between Privatized Companies Total Poland Hungary Czech Republic Mean Median Mean Median Mean Median Mean Median Panel A: Privatizations with a Large Fraction of the Share Capital Sold at the Initial Offer Minus Privatizations with a Small Fraction Sold (values above the median fraction sold are defined as large and values below the median as small) BHR (local) a a (1.093) ( 0.775) (2.697) (2.639) ( 0.898) ( 0.953) (1.172) (0.039) BHAR (local) (0.598) ( 0.218) (0.605) (0.468) ( 0.265) ( 0.198) (1.248) ( 0.394) BHR ($) b b (1.607) (0.719) (2.430) (2.341) ( 0.910) ( 0.950) (1.743) (1.443) BHAR ($) a b b b DS-Europe (1.941) (1.518) (2.571) (2.554) ( 0.801) ( 0.992) (2.144) (2.546) BHAR ($) b b b b DS-World (1.831) (1.223) (2.517) (2.341) ( 0.865) ( 0.834) (1.992) (2.191) BHAR ($) b EM-Priv. (1.011) ( 0.872) (2.228) (1.745) ( 0.914) ( 0.914) (1.224) ( 0.066) Market value b a a b b (Million $) ( 2.129) ( 3.577) ( 1.904) ( 3.831) ( 1.329) ( 1.350) ( 2.139) ( 3.451) Fraction sold (%) 51.1 a 54.4 a 48.8 a 45.1 a 39.3 a 33.5 a 38.0 a 34.1 a (21.240) (10.320) (11.802) (5.874) (7.913) (4.724) (9.329) (6.928)

22 926 JELIC, BRISTON AND AUSSENEGG Table 4 (Continued) Total Poland Hungary Czech Republic Mean Median Mean Median Mean Median Mean Median Panel B: Privatizations with Foreign Participation Minus Privatizations with no Foreign Participation BHR (local) b ( 0.316) (0.805) ( 0.357) (0.255) (2.399) (1.580) ( 1.425) ( 0.327) BHAR (local) a a (0.700) (1.130) ( 1.739) ( 1.084) (2.994) (2.980) ( 1.552) ( 0.607) BHR ($) b ( 0.642) (0.385) ( 0.816) (1.976) ( 1.709) ( 1.828) ( 1.752) ( 0.701) BHAR ($) b b DS-Europe ( 0.678) (0.586) ( 0.441) (0.383) (2.249) (2.300) ( 1.879) ( 1.121) BHAR ($) b b DS-World ( 0.708) ( 0.332) ( 0.440) (0.383) (2.195) (1.986) ( 1.852) ( 1.028) BHAR ($) b b EM-Priv. ( 0.367) (0.722) ( 0.099) (1.084) (2.202) (1.986) ( 1.590) ( 0.623) Market value b (Million $) (1.264) (1.785) (1.116) (1.116) (1.510) (2.212) (0.508) (1.059) Fraction sold (%) 19.9 a 31.2 a ( 4.012) ( 3.584) ( 1.262) ( 1.036) ( 0.996) ( 1.061) ( 1.191) ( 0.911)

23 PERFORMANCE OF PRIVATIZED FIRMS IN TRANSITION 927 Panel C: Privatizations in Regulated Industries Minus Privatizations in Unregulated Industries BHR (local) ( (0.249) ( 1.186) ( 0.406) (0.821) (0.859) ( 0.818) (0.786) BHAR (local) ( 0.009) (1.315) ( 0.163) (0.757) (0.429) (0.376) ( 1.045) ( 0.238) BHR ($) ( 1.267) ( 0.188) ( 0.799) (0.162) (0.806) (0.913) ( 1.452) ( 0.351) BHAR ($) DS-Europe ( 1.692) ( 0.545) ( 0.955) ( 0.514) (0.750) (0.806) ( 1.827) ( 1.122) BHAR ($) DS-World ( 1.478) ( 0.113) ( 0.881) ( 0.351) (0.783) (0.859) ( 1.673) ( 0.701) BHAR ($) EM-Priv. ( 0.850) (0.855) ( 0.721) (0.189) (0.981) (1.343) ( 0.988) (0.715) Market value a a a b a (Million $) (2.644) (4.864) (1.541) (3.380) (0.986) (0.967) (1.964) (3.507) Fraction sold (%) 16.8 a 29.2 a 18.6 b 29.6 b a 47.1 a ( 3.061) (2.970) ( 2.180) (1.987) ( 1.001) ( 1.074) ( 3.928) ( 3.521)

24 928 JELIC, BRISTON AND AUSSENEGG Table 4 (Continued) Czech Republic (2 nd Wave) Minus (Poland and Hungary) Mean Median Panel D: Privatizations in the Czech Republic (2 nd Wave) versus Privatizations in Poland and Hungary BHR ($) a 5.11 ( 3.121) ( 0.811) BHAR ($) a b Benchmark: DS-Europe ( 4.289) ( 2.239) BHAR ($) a 8.57 Benchmark: DS-World ( 3.680) (1.454) BHAR ($) a b Benchmark: EM-Priv. ( 2.625) ( 1.932) Market value (Million $) (0.011) (1.587) Fraction sold (%) a a ( 3.146) ( 2.869) Notes: Mean and median differences between five sub-samples of privatized companies: (i) Panel A (hypothesis H1): privatizations with a large fraction of the share capital sold at the initial offer versus privatizations with a small fraction sold (values above the median fraction sold are defined as large and values below the median as small), (ii) Panel B (hypothesis H 2 ): privatizations with foreign participation versus privatizations with no foreign participation, (iii) Panel C (hypothesis H 3 ): privatizations in regulated industries versus privatizations in unregulated industries, and (iv) Panel D (hypothesis H 4 ): firms privatized in the 2 nd wave of the Czech Republic mass privatization program versus privatizations in Poland and Hungary. For each privatized company the long-run performance is measured until the third anniversary of aftermarket trading using buy-and-hold returns (BHRs) and buy-and-hold abnormal returns (BHARs). Four benchmarks are used: (i) a domestic market index (Poland: WIG Index, Hungary: BUX Index, Czech Republic: PX50 Index), (ii) the Datastream Europe Index (DS-Europe), (iii) the Datastream World Index (DS- World) and (iv) a portfolio of Emerging Market privatizations (EM-Priv.), excluding Eastern European transition economies. The portfolio of Emerging Market privatizations is rebalanced once a year on the first trading day in January. The portfolio performance is based on a buy-and-hold strategy and in each PIPO an equal amount of money is invested on the rebalancing date. Only PIPOs are used with a first trading day not earlier than four years before the rebalancing date. Over the investigation period (21 June, 1990 till 5 April, 1998) a total of 275 privatizations are used for this portfolio. The local market indices are based on local currencies and the international indices are based on US dollars (USD). BHRs are measured by equations (1) and (2) respectively. It is tested whether the BHRs and the BHARs are significantly different from zero. For the means a skewness-adjusted test statistic with p-values using a bootstrapping procedure is employed (see Lyon et al., 1999). For the medians a Wilcoxon-Signed-Rank-Test is used. Test statistics in parentheses. BHRs and BHARs in percentages. a b Significant at the 1% level. Significant at the 5% level.