Privatization through an overseas listing: evidence from China's H-Share firms.
We study the partial privatization of 53 Chinese state-owned enterprises (by their listings on the Hong Kong Exchange over the period July 1993 to December 2002. We find that listing has led to a median increase of 70% in real net profits, 80% in real sales, 50% in capital spending, and a mild
**********
Ever since Margaret Thatcher's government implemented a program of privatization in the United Kingdom in the late 1970s, the privatization of state-owned enterprises (SOEs) has become a significant economic phenomenon that has attracted a great deal of interest from academics. Although there are many studies on western SOEs, this is not the case for China, where a large-scale privatization program was only examined recently by Sun and Tong (2003). Using a sample of 634 SOEs in China that went through share issue privatization (SIP) in 19941998, they found limited success in the privatization program. They argued that selling too few government shares to private owners means that the governing structure of the partially privatized SOEs cannot undergo fundamental changes. Boubakri, Cosset, and Guedhami (2002) examined 201 cases of privatization in 32 developing countries and found that the relinquishment of control by the government is one key determinant of changes to profitability. If the control on privatized SOEs is not contestable, which is the case in China, Sun and Tong's (2003) findings may not be too surprising after all. However, given the fact that privatization is typically sequential, whereby governments sell their shares bit by bit, it may not be that critical to shift controlling shares to private hands to obtain improvements in the performance of privatized SOEs.
Indeed, the Singapore government still maintains controlling power over privatized SOEs (called government-linked corporations) but these firms seem to operate quite efficiently, according to a study by Feng, Sun, and Tong (2004). Another recent paper by Gupta (2005) showed that in India, the government privatized only a very small portion of the equity in SOEs, but the operating performance of such firms still improved significantly after privatization. Gupta argued that this was because India has a well-established stock market that long predates privatization. Boubakri et al. (2002) also found more improvements in efficiency and output for firms in countries where stock markets are more developed and where property rights are better protected and enforced. Economic growth and foreign ownership were found to be other key determinants.
In this article, we examine 53 Chinese state-owned enterprises (SOEs) that were partially privatized through SIP in a well-established foreign market, the Stock Exchange of Hong Kong (SEHK) during 1993-2002 by listing a relatively small amount of "H-shares." (1) If foreign (i.e., non-Chinese) ownership and a sophisticated stock market are important to the degree to which privatization succeeds, then Chinese SOEs that have privatized by issuing shares in Hong Kong should show significant performance improvements after privatizing.
However, Chinese SOEs that listed in Hong Kong went through a process that was not necessarily based on the economic merits of the firms, especially in the early days. There were stories circulated about favoritism and the misuse of funds raised by the listed firms. In fact, H-shares are issued at discount relative to the local shares. Hence, we wish to see if China's privatization program, as a whole, can leverage on the Hong Kong market to improve the SOE performance. Our study provides evidence on this.
Since the Hong Kong stock market houses most of China's overseas listings, our study also has a direct bearing on the effectiveness of this foreign-listing strategy as a way to vitalize SOEs. On the other hand, as H-share firms already account for about 20% of the market capitalization in Hong Kong, our study is also important to international investors interested in the Hong Kong market.
We compare changes in the performance of sample SOEs before and after an SIP by using Wilcoxon tests and cross-sectional regressions, by looking at the stock returns of these SOEs after listing, and finally by a panel data analysis. We find that our sample SOEs show a median increase of 70% in real net profits, 80% in real sales, 50% in capital spending, and a mild but nonsignificant improvement in coverage ratios, although the return on sales shows no increase after privatization and the return performance is worse than several market index benchmarks. We find that the remaining state ownership in SIP firms is negatively related to profitability measures both across firms and over time, but that legal persons have a positive impact on firm performance in some circumstances. More importantly, foreign ownership plays a significantly positive role on H-share performance. Thus, our overall results suggest that doing a SIP in a mature stock market and allowing foreign ownership are helpful in changing the performance of SOEs.
Section I of our article briefly describes the process of listing Chinese SOEs in Hong Kong, their share structures and the implications for governance. In Section II, we compare firm performance before and after privatization. In Section III, we analyze the impact of various types of ownership on privatized firms after privatization. Section IV concludes.
I. H-Share Issuance, Ownership, and Governance
It is not unusual for governments to sell SOEs abroad. Bortolotti, Fantini, and Scarpa (2000) explain that governments in developing countries might want to sell privatized firms abroad because economically, limited bank financing and small debt markets could drive them to seek external financing. Politically, selling firms to foreign investors could signal that a government is committed to supporting market-oriented policies. Biais and Perotti (1997) suggest that right-wing governments tend to issue shares to domestic investors instead of to foreign ones because doing so can increase domestic political support for market-oriented policies.
China's case is different. The Chinese government is hardly concerned about sending signals to the market or gaining political support. Its main target at the early stage was to raise foreign capital. "Crossing the river through touching stones", which is a Chinese saying that means doing things by trial and error, is the way the Chinese government is approaching economic reforms. This is also the approach it is following on overseas listings. However, its moves to tap the international market through foreign listings do not seem to have been well planned. For instance, the Chinese government was rumored to have been thinking of confining primary overseas listings to New York before testing other markets. Subsequently, it concentrated on Hong Kong (Euroweek, 1994). Also, the government published the "Special Regulation of the State Council on Raising Capital and Listing Overseas by a Joint-Stock Company Limited" on April 9, 1993, only few months before the first H-share listing in July 1993.
Later, the Chinese government wished to use foreign listings as means of improving the quality of SOEs and of making them role models for locally listed SOEs. On one occasion, when the former chairman of the CSRC, Zhou Daojong, talked with the CEOs of foreign-listed SOEs, he said that:
"Overseas-listed companies are all outstanding enterprises that are representatives of their respective industries to an extent. I hope you can also be the models of listed companies. The behavior of an overseas-listed company is not only the company's own business, it relates to our country's image of reform and openness ... I hope you can solidly turn all your companies into world-class enterprises, bringing honor to the country." (May 26, 1995, CSRC web news)
On another occasion, when he talked with some newspaper reporters, Zhou said that:
"Recommending medium-to-large SOEs for overseas listing is useful in raising necessary foreign capital; but more importantly, it prods SOEs to learn from the successful experiences of overseas companies, helping them to match international standards, and making it possible for them compete in the international market." (July 26, 1995, CSRC web news)
However, despite Zhou's remarks, to keep up with the socialist doctrine of public ownership, the government has maintained control in the SOEs by holding a significant portion of ownership in the firms in the form of state shares.
From the listing of Qingdao Beer on the SEHK on July 15, 1993, to the end of 2002, 53 Chinese companies issued H-shares and 27 also issued A-shares, many within a year after their Hong Kong IPO. Hence, the A-share offerings were probably planned at the same time as the H-share offerings. As Table I shows, 41 firms were listed before 1998, when the Asian Economic Crisis hit Hong Kong. Up to 2002, Chinese overseas listed companies raised a total of about US$18.25 billion. The IPO proceeds were about $17.08 billion, with the rest being convertible bond issues (We exclude from our sample 22 H-share firms that listed during the period 2001-2002 in the Hong Kong Growth Enterprise Market, which was established at the end of 1999, because they are either private or small high tech firms). Except for the bond issues, no H-share firm had any seasoned offerings in Hong Kong between 1993 and 2002.
A. H-Share Listing Process
The China Securities and Regulatory Commission (CSRC) was responsible for selecting the SOEs and approving their applications for an overseas listing. The CSRC required every listing to go through a "pre-selection" process. Although the selection criteria were not exactly clear and were changing, we can note some general principles. For instance, firms had to be a certain size and had to show good performance with high growth potential. Firms also had to demonstrate a genuine need for capital: the net assets of a listed company had to be higher than RMB400 million, the pre-listing ROE above 10%, and the proceeds of the offering above US$50 million. The Chinese government had to have a controlling stake in these companies and retained absolute control in critical industries or enterprises. (2)
However, such a tightly controlled selection process induced keen competition for IPOs, so the process was subjected to lobbying and political pressure. Selection was often based on noneconomic factors such as political connections, geographical quotas, "protected" industry status, etc. Thus, it is conceivable that the firms selected to list in Hong Kong, at least in the early stage, might not necessarily have been the "best." When the prices of H-shares plunged during the Asian financial crisis, the CSRC admitted that: "Quite a number of enterprises ... failed to demonstrate sustained growth. Some companies have made wrong moves facing global competition and consequently incurred great losses. Most of the companies have been vulnerable to various risks" (Department of International Cooperation, CSRC, 2003).
In recent years, the CSRC has smoothed the foreign listing process. In July 1999, the CSRC issued "A Notice about Questions Related to Enterprises Applying for Overseas Listings." When Hong Kong's Growth Enterprise Market (GEM) was opened on November 25, 1999, the CSRC issued the "Selection and Monitoring Guidelines for Listing Applications of Local Enterprises in the Hong Kong Growth Enterprise Market" on September 21, 1999. Since then, all SOEs that have undergone restructuring to become shareholding companies, and that meet overseas listing requirements, have been allowed to voluntarily apply for an overseas listing through the CSRC.
B. Ownership and Structure of Governance
H-share firms also have state shares and legal-person shares. State shares were created in China to designate holdings in SOEs by the central government, provincial governments, or enterprises owned solely by the government. Our sample of H-share companies shows that the level of state ownership is about 40%, which is higher than the average of 34% for all domestically listed firms. (3)
State ownership is typically viewed as being the cause of immense agency problems in SOEs. However, by arguing that government bureaucrats may play a positive role in promoting private activities, Walder (1995) suggests a "helping-hand model." Frye and Shleifer (1997) also argue in their study on Russia and Poland that many different relationships can exist between government and business. Whether or not state ownership in China's SIP firms acts as a helping hand is an empirical issue.
Legal-person shares are shares owned by domestic institutions, most of which are themselves partially owned by the central or provincial government. On average, legal persons hold about 20% of the shares of H-share firms. (4)
Legal persons are typically business agencies or local-government enterprises that have helped to start up the public companies. They do so either by giving the companies the permission to operate or by allowing resources under their control to be used for the startup. In addition, both state shares and legal-person shares are nontradable. Therefore, some economists argue that legal persons are similar to state shareholders. (5)
However, there may be subtle differences between the two. Che and Qian (1998) argue that in an environment without secure property rights against state encroachment, private ownership leads to excessive hiding of revenue. State ownership fails to provide incentives for managers, but local government ownership limits state predation and reduces costly revenue hiding. Xu and Wang (1997) find that legal-person ownership has a positive impact on firms. They suggest that legal persons in China are similar to institutional investors in market economies. Because they can monitor managers more actively, legal persons enhance firm performance. They often have direct business and financial links with SIP firms.
Some legal-person shares result from debt-equity swaps. Legal persons benefit directly from good firm performance. Likewise, since legal-person shares are not tradable, if a firm's performance is poor and the value of its shares keeps declining, legal persons will suffer directly. Hence, they have more incentive than do holders of tradable shares to see to it that the SOEs are involved in value-increasing activities.
A shares are tradable only for Chinese domestic investors. The CSRC has mandated that when a company issues its IPO, A shares must account for no less than 25% of total outstanding shares. This requirement does not apply to H-share firms. Only 27 firms in our sample, representing an average firm ownership of less than 6%, also issue A-shares. A-share investors are typically not major shareholders, because normally, the top ten shareholders are the state and legal persons.
On average, H-shares account for 35% of the total shares in our sample firms. Boycko, Shleifer, and Vishny (1996) argue that privatization is more effective when both ownership and control are transferred to private foreign hands. The majority of outstanding H-shares are registered with the Hong Kong Securities Clearing Company Nominees Limited, which often appears as a major shareholder of H-share companies. Consequently, many H-share firms include on their boards independent nonexecutive directors from Hong Kong. The managers of such firms may then be subject to better monitoring.
H-share ownership is expected to have a positive impact on firm performance. However, the impact may be limited for several reasons. First, in H-share firms, H-share ownership is mostly below 50% of the total ownership (see Appendix). Second, although some strategic institutional investors hold H-shares, H-shares are widely held by individual investors. Last, Hong Kong directors are unfamiliar with the way Chinese firms do business, and thus are often not active on the board.
II. Tests of Changes in Performance
We present the test results of performance changes around the SIP in this section.
A. Wilcoxon Test Results
We first follow studies by Megginson, Nash and Van Randenborgh (1994), Boubakri and Cosset (1998), and D'Souza and Megginson (1999), among others, to compare changes in performance before and after privatization. We focus on changes in profitability, changes in output, and changes in leverage, using similar testing proxies and methods. (6)
We obtain our data from the prospectuses and annual reports of 53 SOEs that had held their IPOs in Hong Kong over the period July 1993 to December 2002. The prospectuses of the firms list financial information going back up to three years before the IPO. Data after the IPO come from the China Securities Market and Accounting Research (CSMAR) Database and the Taiwan Economic Journal (TEJ) database. We supplement this data with information from company annual reports, the Company Analysis database of Thomson Financial, and the Pacific Basin Capital Market (PACAP) database.
We pay particular attention to one special feature of China's SIPs, because this feature may have had a significant impact on our measures of profitability. Unlike many other cases of SIPs in which the issuing proceeds went to the government (the so-called "secondary issue"), in all of China's SIPs, the proceeds were plowed back to the privatized SOEs (the so-called "primary issue"). Thus, the typical profitability measures of return on assets (ROA) and return on equity (ROE) may have been affected in at least two ways. First, the net income could have been distorted by the profits and losses from the short-term investment of the proceeds of the issue on marketable securities. Since it was difficult to identify the exact amounts, we assumed that they were not significant. The second impact could have affected the computation of total assets and total equities. The one-time increase in equity capital would automatically exert a downward bias on the ROE and ROA figures. By the same token, the commonly used leverage measures of debt-equity ratio (DE) and debt-assets ratio (DA) would also be downwardly biased. This bias would be especially acute in the year of privatization.
We follow Sun and Tong (2003) in tackling the problem by excluding the year of privatization so that the immediate impact on total equity and total assets would not be an issue. Furthermore, we use inflation-adjusted real net profit (RNP) and return on sales (ROS) as the measures of profitability, and times-interest-earned (TIE) and operating cash flow per unit of debt (OCF/TD) as alternative measures of leverage for the comparisons.
We define TIE as the ratio of EBIT to interest expense, which indicates a firm's long-term debt-paying ability. OCF/TD indicates a firm's ability to cover its total debt with its yearly cash flow. Strictly speaking, these are coverage measures, rather than leverage measures. The purpose of using these ratios instead of other, more commonly used leverage ratios is to avoid including total assets or total equities in the measures. We note that contrary to the interpretation of DE and DA, the higher the OCF/TD and TIE ratios, the better the firm's leverage condition.
We compute yearly empirical proxies for every firm for a seven-year period, counting from three years before the privatization to three years after. We also compute the mean of each proxy for each firm over the pre- and post-privatization periods. We exclude Year 0 from the calculations because it includes both the public and private ownership phases of the enterprises. We use the two-tailed Wilcoxon signed-rank test to test for any significant differences in the median changes of the proxy for the pre- and post-privatization periods.
The first major row in Table II shows the contrasts between the three-year profitability averages of the SOEs before and after privatization. The median (mean) RNP, which we normalize to one in the year in which the firm was privatized, changes from 0.56 (0.74) before privatization to 0.95 (l.12) after privatization. The real net profit shows a median increase of 70% after privatization, although the Wilcoxon Z value of 1.72 is only marginally significant at the 10% level. We note that in the last column of the table, the number of firms with a higher real net profit after privatization is 33. This figure also indicates that the real net profit of more firms improved after privatization.
When we measure by return on sales (ROS), profitability tends to decline after privatization. The median (mean) ROS changes from 0.1 (0.17) to 0.09 (0.15) after privatization, but the Wilcoxon signed-rank test of the median change of 1.54 is too low to be statistically significant. Hence, the ROS decline is mild, although only 18 firms show an increase in ROS after privatization.
We perform the above tests on various groupings. We use the two-tailed Wilcoxon Rank Sum test to examine if the median changes of these measures before and after privatization across the two subgroups are- the same. We base the first grouping on whether a firm also issues A-shares in China. It is possible that A-shares further diffuse the ownership of public investors, which in turn further exacerbates the agency problem, and thus hinders a firm's ability to improve its profitability. Indeed, as we see in the second major row of the table, the drop in ROS for H-firms that also issue A-shares is significant at the 10% level (Wilcoxon value of 1.82), but the firms that issue only H-shares have a statistically nonsignificant increase in mean ROS and a statistically nonsignificant decrease in median ROS. On the other hand, the increase in RNP of the two groups of firms has no statistical significance, and neither does the difference in the increases between them.
We base the second grouping on whether a firm has a Hong Kong director on its board. We expect firms with Hong Kong directors to perform better than those without. However, our results do not show much support for the conjecture, as there is a general lack of statistical significance in both the RNP and ROS comparison.
We base our third grouping on whether a firm changes its CEO after listing. La Porta and Lopez-de-Silanes (1997) find that if the government fires its CEO when it restructures -the firm, then the performance of an SOE improves dramatically. Hence, firms that change their CEO should show more increase in profitability than firms that do not. However, we do not find much supporting evidence for this conjecture. If anything, changing the CEO worsens the situation. For instance, firms that do not change their CEO show a significant increase in RNP of 0.39 (Wilcoxon value of 2.49), and firms that do change their CEO have a significant drop in ROS of 0.04 (Wilcoxon value of 1.66) after privatization.
In unreported results, we also conducted tests on groupings based on whether the firms are in regulated or non-regulated industries and whether it was a "control privatization" (in which the government sells more than 50% of the shares) or a "revenue privatization" (in which government sells less than 50% of the shares). The results all showed no significant differences across groups.
Overall, we find that the SIP of Chinese firms in Hong Kong show-some significant improvement in real net profit, and a mild but nonsignificant drop in return on sales after the SIP.
However, for output, the theoretical relationship is not clear. Better incentives, more flexible financing opportunities, increased competition, and a greater scope for entrepreneurial initiative should all work to increase real sales after privatization. On the other hand, effective privatization may lead to a reduction in output, since the government can no longer use subsidies to entice managers to maintain inefficiently high levels of output. Thus, output levels can go either up or down after privatization.
Table III presents the changes in real output associated with SIP. We note that the real output is normalized to one in Year 0, the listing year. For the whole sample, the median value increased from 0.71 to 1.28, an 80% increase after privatization, with a highly significant Wilcoxon test statistic of 7.17. In fact, real output increased throughout all subsamples based on various groupings. All increases are statistically significant. Second, 51 firms show a positive change in output, while only two firms show a negative change. The increase in output for firms that do not change their CEO is larger than that for firms that do change CEOs. The difference between the two is statistically significant at the 5% level.
If we assume that privatized firms should increase output, then the fact that a change in CEO is associated with a decrease in output is a bit surprising. However, it is possible that reverse causality is responsible, i.e., poor output leads to the replacement of a CEO because, due to the data available, our grouping is based on a change in CEO after the listing, not before. In any case, the increase in real output after privatization is largely consistent with previous findings. Partial privatization in China leads to increases in real output.
We expect SOEs to have high leverage for two reasons. Their cost of borrowing is lower, because SOEs carry (implicit or explicit) government guarantees. Also, borrowing is the only avenue by which SOEs can raise funds (apart from equity injections by the government and retained earnings), since they do not have access to private investors. Hence, after privatization, we expect the leverage of former SOEs to drop, since the state's withdrawal of debt guarantees will increase their cost of borrowing. Also, former SOEs will now have access to public equity markets.
However, since the H-share firms are strongly backed by the government, and since the government does not want to see the H-share firms fail, it may still provide a de facto guarantee for the firm's loans. With access to the equity market, these firms may actually increase their borrowing capacity. Hence, we may observe an increase in leverage after privatization. This is an empirical issue.
The results in Table IV indicate that for the whole sample, the median TIE ratio increases from 4.09 before privatization to 4.55 after privatization, while the OCF/TD ratio increases from 0.21 to 0.24. However, both measures lack statistical significance. This is generally the case across various subgroups. Hence, there is a tendency for the coverage ratios to improve after SIP.
For the primary-issue reason discussed earlier, the "standard" debt ratios are unreliable leverage measures. However, we check the total debt ratio (i.e., total liabilities/total assets, TL/TA) and the debt-equity ratio (i.e., total liabilities/total equity, TL/TE). Although we do not report them here, we find that TL/TA have no significant change and that TL/TE decrease significantly after privatization. Given that TE would increase due to the plowing back in of the SIP proceeds, such results suggest that SOEs might increase their long-term investments after privatization and finance partly through debt. (7) To look into such a possibility, we compare the investments on plants, properties, and equipments (PPE) and PPE/TA before and after privatization.
Table V shows that for the full sample, PPE indeed shows an increase in median (mean) value from 0.85 (1.07) to 1.28 (2.13), a median increase of 50%. The Wilcoxon test statistic of 1.74 suggests that the increase is statistically significant at the 10% level. The subgroup analysis suggests that the increase is particularly important for the pure H-share firms that do not issue A-shares and for the firms that do not change their CEOs. On the other hand, the PPE/TA ratio comparison shows a nonsignificant decline after privatization for both the full sample and across various subgroups. This mean indicates that not all increases in TA come from the increase in PPE.
B. Cross-Sectional Regression Analysis
Although the Wilcoxon test results are robust across various groupings, they do not control for factors that may affect comparisons of performance. Therefore, we conduct a cross-sectional regression analysis on the changes in performance of various proxies.
[DELTA][PP.sub.i] = [[alpha].sub.i] + [[beta].sub.i] [DELTA][GDPGR.sub.i] + [[beta].sub.2][CTRL.sub.i] + [[beta].sub.][REG.sub.i] + [[beta].sub.4][ASHARE.sub.i] + [[beta].sub.5][FD.sub.i] + [[beta].sub.6].sub.][CEO.sub.i] + [[epsilon].sub.i] (1)
The sign "[DELTA]" is the difference in the three-year averages of the variable in question before and after privatization. PP is the three-year average of the performance proxy and GDPGR is the three-year average of real GDP growth. Putting in GDPGR allows us to capture the impact of general economic conditions on firm profitability, output, and leverage.
We construct several dummy variables to control for various factors. CTRL takes a value of one if state shares occupy less than 50% of the total outstanding shares after privatization, and zero otherwise. REG takes a value of one for firms in regulated industries (utilities, energy, and basic materials), and zero otherwise. ASHARE takes a value of one for firms that also issue A-shares, and zero otherwise. CEO takes a value of one if a firm changes its CEO after privatization, and zero otherwise. FD takes a value of one if a firm has any Hong Kong directors on its board, and zero otherwise.
Since we have only 53 observations for the cross-sectional regression, the degree of freedom could be a concern. Hence, we run our regressions both with and without control dummy variables to see how the performance variables change after privatization. If the performance before and after privatization is different, the regression constant should differ significantly from zero.
The results in Table VI, Panel A, confirm the Wilcoxon test results in Tables II through IV. The intercept terms for changes in real net profits (RNP), real sales (RSALES), and capital spending (PPE) are positive and significant. RNP has an intercept of 0.56 with a t-value of 2.42 and RSALES has an intercept of 0.68 with a t-value of 5.28. Both t-values are statistically significant at the 1% level. PPE has an intercept of 1.41 with a t-value of 1.69, which is statistically significant at the 10% level.
These results suggest that controlling GDP growth before and after privatization does not change the conclusion obtained from Tables II through V.
In Panel B, the results, which included the control dummy variables in the regression, showed essentially the same picture, but the intercept term in the [DELTA]PPE regression becomes statistically insignificant with a coefficient of 0.29 and a t-value of 0.23.
C. Performance after Privatization
We examine changes in the share price of the privatized firms up to five years after privatization. Our idea is that it may take some time before privatized firms can improve their performance, yet accounting measures do not capture the market's view of the future prospects of privatized firms. If the market believes that privatization is good for the firm, then this should be reflected in the firm's share prices. Following Dewenter and Malatesta (2001) and Megginson, Nash, Netter, and Schwartz (2000), we looked at the adjusted yearly stock returns of firms after privatization to determine whether the privatized firms outperformed the market. We calculate the market-adjusted buy-and-hold return for each firm using the following formulation:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
where ER is the market-adjusted cumulative return, R is a firm's daily stock return, MR is the relevant daily market return, and t = (a to b) is the timeframe of one to five years.
The market proxies we use for comparison are the Hong Kong Hang Seng Index, the Shenzhen B-share Index, the Shanghai B-share index, and the Hang Seng China Affiliated Corporations (red-chip) index. (8) Red chip companies are essentially Chinese conglomerates that operate in China, although some also have businesses in Hong Kong. They are controlled by, and many have direct ownership ties with, political and economic decision-makers at the provincial or central government levels, but they are registered under Hong Kong's corporate ordinances as Hong Kong companies. Contrasting H-share SOEs with these red-chip companies can be illuminating, as the results reflect the market's view of the performances of China-registered SOEs against non-China registered SOEs listed in Hong Kong.
Table VII-presents the mean and median market-adjusted cumulative returns and the corresponding t- and Wilcoxon statistics.
The results indicate that the privatized SOEs underperform -the Hang Seng Index up to five years after their listing. In the first year, H-share firms underperform -the Hong Kong market by an average of 10%. Such underperformance increases yearly until it reached 51% in the fifth year. Except for the first year, the corresponding t-statistics for the differences are all significant at 1%.
Using the difference in median return yields a similar pattern. Even compared to B-share firms, H-share firms underperform significantly and the difference grows in Years 4 and 5. The larger Years 4 and 5 underperformance may be due to the big jump in B-share prices in 2001 after the B-share market was partially opened to domestic investors. However, compared to the red-chip index, the underperformance in H-shares is not obvious except in the third and fourth years.
The long-run underperformance of H-share stock returns indicates that investors do not particularly favor China's privatized firms. Hence, the result is different from that in Megginson et al. (2000), who finds that the returns of SIP firms topped industrial, country, and world benchmarks from the first to the fifth years after listing. Foerster and Karolyi (2000) study the long-term return performance of 333 global equity offerings in the US in the form of American depository receipts (ADRs) from 35 countries. On average, the three-year return after the issuance of the ADRs, under-performs local market benchmarks of comparable firms by 8% to 15%. This underperformance is greater for offerings from emerging markets than from developed markets.
Even for the subsample of 68 ADRs resulting from government privatization, the three-year returns also tend to be lower than both the local market indexes and the matched firms, although there is statistical significance only with the seasoned offerings. Schaub (2003) finds that the long-term underperformance of ADRs is greater for IPO issues than for SEO issues, and for emerging market issues than for developed market issues. Aybar (2001) focuses on privatization-related ADRs and confirmed that the long-term return performance of emerging market issues is poor. Our findings are consistent with all of these studies.
We could argue that IPO firms typically underperform in the long run (Ritter, 1991; Aggarwal, Leal, and Hernandez, 1993; Loughran and Ritter, 1995 and 1997). But recent studies indicate that such documentation should be interpreted with care (see Bray and Gompers, 1997; Ritter and Welch, 2002; and Gompers and Lerner, 2003; among others). Furthermore, our underperformance results are greater than those documented elsewhere, even after taking into consideration the long-run IPO performance phenomenon.
We could also argue that the SOEs managed their earnings prior to listings (Yeoh, Welch, and Wong, 1998a and 1998b; DuCharme, Malatesta, and Sefcik, 2001) but the continuous decline of stock prices year after year cannot be due solely to earnings management before the IPO. Hence, we believe that negative returns reflect the market's negative view. (9)
III. Ownership Impact on Firm Performance after an SIP
Here, we use the pooled regression technique to further investigate the situation of the SOEs after privatization. We want to examine the possible differential impact of state ownership, legal-person ownership, and H-share ownership on the profitability of the partially privatized SOEs. It is important to note that since we are not comparing performance before and after the listing, which requires up to three years after listing data, we extend our sample period to 2003. Thus, the panel data cover the same 53 firms for the period 1993-2003.
We begin with the following common-intercept specification:
[MBR.sub.it] = [[alpha].sub.it] + [[beta].sub.1] [GDP.sub.it] + [[beta].sub.2] [ST.sub.it]/[LP.sub.it] + [[beta].sub.3] [SALE.sub.it] + [[beta].sub.4] [DA.sub.it] + [[beta].sub.5] [LOC.sub.i] + [[beta].sub.6] [ASHARE.sub.it] + [[epsilon].sub. it] (3)
Since we use only post-listing data, MBR serves as our performance proxy. GDP, ST, LP, and H are as previously defined. SALE is the natural logarithm of annual real sales revenue, a proxy for size. Large SOEs tend to have a larger market share and more market power, but they may encounter more government bureaucracies and bigger agency problems. Therefore, the overall size impact on performance is unclear. DA is the ratio of total debts to total assets, and controls for any possible leverage effect. We expect to see a negative relation between firm performance and leverage. LOC is a dummy variable that takes a value of one if a firm is located in the coastal provinces, and zero otherwise. Historically, China's coastal area has been more developed and attracted more than 85% of the total foreign direct investment in China. To reduce regional differences, the Chinese government intentionally selected some firms located in inland areas to list in Hong Kong. ASHARE is a dummy that takes a value of one if the firm also issues A shares, and zero otherwise. We also run a fixed-effect formulation to control for industry-specific effects. (10)
We have a third, firm-specific, fixed-effect formulation. Himmelberg, Hubbard, and Palia (1999) show that if the unobserved, exogenous variables are correlated to ownership and performance, then- unobserved heterogeneity across firms can generate a spurious correlation between ownership and performance. Our specification, with the GDP variable, controls for both year-variant but firm-invariant omitted variables, and firm-variant but time-invariant omitted variables. Also, the Himmelberg et al. (1999) concern about managerial ownership being endogenously determined should not be serious here. First, the proportion of STor LPs is determined based on government policy and a quota system, which are not endogenous to the SOE. Second, when H-share companies later issue A-shares, many do it within a year, so the issue amounts are likely to be jointly determined. Also, such issues are subject to government control, not at the managers' discretion. Third, there were no seasoned issues of H-shares in the SEHK during our sample period, so there should not be a serious endogenous problem of the SOE sequentially privatizing the rest of the state-owned and legal-person shares conditional on the performance of the existing share price. As individual (or firm-specific) time-invariant variables are controlled, the location, A-share, and industrial dummies become unnecessary as their possible effects are controlled.
Table VIII presents the results with MBR as the measure of performance. Since ST, LP, and H add up to almost one for most firms, and because ST and LP are highly correlated, we run the three ownership variables separately in the pooled regressions and place the results in three major columns. We present the results under three different formulations within each column. We allow a possible nonlinear effect of government ownership on a firm's MBR. ST1 and LP1 respectively refer to state and legal-person ownerships that are less than 50% of the total share ownership of the firm. "ST2" and "LP2" refer to ownerships above 50% of a firm's total share ownership. Since H-share ownership is typically lower than 50% of the total, we do not further split H shares.
The pooled regression results in Table VIII indicate that the state-ownership variable ST tends to have less negative impact on a firm's MBR for firm ownerships of less than 50%. For the firm-specific fixed-effect regression, STI enters positively into the regression with a t-value of 1.71. ST2, which captures state ownerships of more than 50%, has a negative impact. The coefficients of ST2 uniformly show significant negative t-values in all three specifications. This suggests the existence of nonlinearity. However, too much ST ownership has a negative effect on a firm's MBR.
The regression results with LP that we present in the second major column show that legal persons have a significant, positive impact on a firm's MBR in all three specifications. These results are consistent with the previous findings that legal-person ownership is "better" than state-share ownership. However, we do not find much nonlinearity relationship in the results.
More interesting is the effect of the H-shares shown in the last major column of Table VIII. The coefficient of H is 0.01 with a highly significant t-value of 5.41. Such a significant, positive impact on a firm's MBR holds in the other specifications. This finding suggests that as a whole, H shares investors -play a positive role in corporate governance. H-share investors are more sophisticated than are holders of A-shares. Also, there are independent institutional investors among the holders of H-shares, and holders of H-shares are represented on the boards of many H-share firms.
In unreported results, we also look at ROS, ROA, and ROE. The results are largely consistent with the MBR results in Table VIII. Although all of these results indicate that more STs (LPs) cause lower (higher) performance and profitability, reverse causality is possible. That is, government (legal persons) may retain fewer (more) shares for good firms in an SIP. We examined such a possibility by running the following regression:
[ST.sub.i] ([LP.sub.i]) = [[alpha].sub.i] + [[beta].sub.1][Size.sub.i] + [[beta].sub.2][Leverage.sub.i] + [[beta].sub.3][Profitability.sub.i] + [SIGMA][[gamma].sub.ij][IND.sub.ij] + [[epsilon].sub.i] (4)
where ST (LP) is the percentage of firm shares owned by the government (legal persons) when the firm is publicly listed. Our proxy for Size is the three-year average of log real sales before listing. We use the three-year average debt-equity ratio before listing as our proxy for leverage. IND are industry dummy variables. The key variable is "Profitability" Our proxy for profitability is the three-year average of the accounting profits, i.e., ROS, ROA, and ROE, before listing. If our conjecture of more government retention of the shares of non-performing firms upon privatization is correct, [b.sub.3] should be significant and negative. Our results show that the t-values of [b.sub.3], whether for ROS, ROA, or ROE, are all nonsignificant.
Gupta (2005) suggests- that a well-established stock market that long predates privatization could contribute to the success of share issue privatization programs. Our findings are consistent with her argument. The Hong Kong stock market is well established and has a long history. SIP in Hong Kong brings positive impacts to the SOEs. However, it is true that our documented improvements are not as strong as shown in her study and in some other literature. One possible reason is that the Hong Kong market is a market "external" to China, which is not exactly the same as the situation in India. In China's case, the possible role of a stock market may be weakened. Another possible reason may be that H-share firms are actually not particularly good, because, as discussed, the companies chosen for listing in Hong Kong, especially in the earlier days. were based on political considerations rather than on economic merits. Thus, they might perform worse than the average.
IV. Conclusion
Based on the results of our tests on 53 SOEs with IPOs in Hong Kong over the period July 1993 to December 2002, we find that the partial privatization of Chinese firms via a listing on the Hong Kong Exchange leads to a median increase of 70% in real net profits, 80% in real sales, 50% in capital spending, and a mild but nonsignificant improvement in coverage ratios. On the other hand, there is no improvement in ROS and a significant underperformance of returns against several market index benchmarks.
When we further investigate the cross-sectional relationship between ownership distribution and firm performance over time after privatization, we find that state ownership is negatively related to firm performance. This negative relation is especially strong for firms in which 50% or more of the shares are owned by the state. We note that despite the tact that legal-person ownership appears simply as indirect state ownership, it demonstrates some positive impact. More importantly, H-share ownership has a significant, positive impact on a firm's performance. An H-share listing can be the right direction for share-issue privatizations. However, the positive effect of H-share ownership seems to be limited, probably because ownership percentages are not large enough to allow holders of H-shares to exert effective control and governance over a firm.
In 2002. the Chinese government announced three policies in succession: guidelines for the takeover and merger of listed companies, permission for foreign investors to buy nontradable state-owned shares and legal-person shares in listed SOEs (Hong Kong Economic Journal, Nov. 4. 2002, p. 5), and rules for qualified foreign institutional investors (QFII) to invest in "A" shares and in government, corporate, and convertible bonds (Asian Wall Street Journal, Nov. 8, 2002, p. 1). The CSRC also issued the "Code of Corporate Governance for Listed Companies in China." In view of our analysis these are the right moves to advance SOE reform in China.
Appendix. Summary Statistics of Ownership Distribution and Other
Pooled Regression Variables
This appendix presents the summary statistics of the pooled regression
variables. MBR is the market value of equity divided by the book value
of net assets. ROS, ROA, and ROE are net income divided by sales,
total assets and owner's equity, respectively. ST, LP, A, and HSHARE
are the shares (in fractions) of equity owned by the state, legal
persons, domestic, and Hong Kong investors, respectively. SIZE is the
real annual sales in millions of RMB. LEV is the book value of debt
divided by the book value of owner's equity.
Panel A. 1993
Variables Mean Median Std. Dev. Minimum Maximum
MBR 2.4510 2.2974 0.6985 1.6784 3.4623
ROS 0.3188 0.1656 0.3828 0.0966 1.0885
ROA 0.0666 0.0589 0.0205 0.0511 0.1065
ROE 0.1009 0.0896 0.0269 0.0782 0.1508
ST 42.3033 43.5150 22.8483 0.0000 62.5000
LP 14.9133 3.6450 29.5948 0.0000 75.0000
A 13.8267 11.8050 9.7212 0.0000 25.5700
HSHARE 28.9567 26.6900 5.3231 25.0000 38.5400
SIZE 503 149 691 39 1765
LEV 0.5427 0.4729 0.2557 0.2493 9719
Panel B. 1994
MBR 1.3568 1.3838 0.6525 0.0000 2.6785
ROS 0.1605 0.1573 0.1090 0.0236 0.4855
ROA 0.0619 0.0589 0.0329 0.0095 0.1176
ROE 0.1029 0.0969 0.0472 0.0152 0.1662
ST 32.1493 42.6100 28.0304 0.0000 63.0800
LP 27.5227 5.9300 32.9957 0.0000 75.0000
A 6.9833 0.0000 9.0132 0.0000 25.5700
HSHARE 33.3440 31.8200 7.8418 25.0000 49.0000
SIZE 380 183 413 42.70 1502
LEV 0.7787 0.7148 0.4507 0.1836 1.9771
Panel C. 1995
MBR 0.9836 0.9521 0.2232 0.5479 1.4298
ROS 0.1017 0.0793 0.0762 0.0069 0.2804
ROA 0.0478 0.0377 0.0361 0.0024 0.1261
ROE 0.0791 0.0763 0.0558 0.0036 0.2029
ST 35.6894 44.4200 28.3119 0.0000 72.9000
LP 24.4282 2.8900 32.0798 0.0000 75.0000
A 7.1288 0.0000 8.9274 0.0000 25.5700
HSHARE 32.7529 29.5400 7.5350 25.0000 49.0000
SIZE 352 122 556 52.3 2132
LEV 0.7620 0.6007 0.4453 0.1775 1.7853
Panel D. 1996
MBR 0.7715 0.7076 0.4166 0.0151 1.5219
ROS 0.1197 0.0482 0.1846 -0.0489 0.6164
ROA 0.0477 0.0208 0.0756 -0.0159 0.3421
ROE 0.0778 0.0481 0.1216 -0.0402 0.5605
ST 38.9965 51.5800 27.4368 0.0000 72.9000
LP 20.0543 2.4400 29.4935 0.0000 75.0000
A 6.6913 5.4300 7.9493 0.0000 25.5700
HSHARE 34.2570 35.0000 7.4051 25.0000 49.0000
SIZE 220 83 316 57.10 1081
LEV 0.8212 0.6676 0.4944 0.0713 1.7826
Panel E. 1997
MBR 1.0426 0.9897 0.5017 0.0000 2.3434
ROS 0.1409 0.0816 0.2665 -0.519 1.1607
ROA 0.0317 0.0263 0.0428 -0.0894 0.1158
ROE 0.0473 0.0490 0.0882 -0.3226 0.1880
ST 38.8449 52.1000 28.6758 0.0000 72.9000
LP 22.6885 1.3600 30.6646 0.0000 75.0000
A 4.4108 0.0000 6.9628 0.0000 25.5700
HSHARE 34.0551 35.0000 6.4390 25.0000 49.0000
SIZE 262 101 362 72.30 1402
LEV 0.8724 0.5858 0.7480 0.0720 2.8879
Panel F 1998
MBR 0.5760 0.5179 0.3219 0.2002 1.7258
ROS 0.0979 0.0394 0.3155 -0.7081 1.2713
ROA 0.0164 0.0172 0.0520 -0.1481 0.1147
ROE -0.0002 0.0346 0.1794 -0.9302 0.1815
ST 38.5166 52.1000 28.8948 0.0000 72.9000
LP 23.3732 1.3600 31.1466 0.0000 75.0000
A 4.2707 0.0000 6.8240 0.0000 25.5700
HSHARE 33.8388 34.8000 6.3871 25.0000 49.0000
SIZE 161 75 445 47.70 1838
LEV 1.0408 0.7141 1.0503 0.1103 5.2789
Panel G. 1999
MBR 1.0161 0.5338 2.7357 0.1832 17.2817
ROS 0.1265 0.0525 0.2281 -0.3017 0.9908
ROA 0.0273 0.0283 0.0367 -0.0829 0.1086
ROE -0.1221 0.0478 1.0750 -6.9915 0.1563
ST 38.7067 52.1500 29.3340 0.0000 72.9000
LP 23.8000 0.0000 31.6304 0.0000 75.0000
A 3.5026 0.0000 5.9426 0.0000 25.5700
HSHARE 33.9435 34.8000 6.2896 25.0000 49.0000
SIZE 288 126 417 22.53 1868
LEV 3.0002 0.8863 12.5771 0.0917 83.2975
Panel H. 2000
MBR 0.5421 0.3959 0.4274 0.1572 2.2690
ROS 0.1114 0.0696 0.2257 -0.3275 0.8690
ROA 0.0216 0.0374 0.0681 -0.2636 0.1356
ROE 0.0171 0.0620 0.2168 -1.1681 0.2563
ST 40.7144 53.1750 30.2311 0.0000 90.0000
LP 22.6947 0.6800 30.9840 0.0000 75.0000
A 3.2279 0.0000 5.8863 0.0000 25.5700
HSHARE 33.3062 34.9000 7.5998 10.0000 49.0000
SIZE 2027 181 9009 38.57 55066
LEV 1.0611 0.7070 0.8459 0.8950 3.4308
Panel I. 2001
MBR 0.7461 0.6819 0.3845 0.0000 2.0928
ROS 0.1063 0.0799 0.2580 -0.7095 0.8666
ROA 0.0248 0.0301 0.0720 -0.2398 0.1481
ROE 0.0246 0.0535 0.4772 -2.2742 2.0295
ST 54.2481 61.0934 20.0647 0.0000 90.0000
LP 6.4352 0.0000 18.2386 0.0000 73.8900
A 6.1927 0.0000 9.0988 0.0000 33.6200
HSHARE 32.5527 34.6900 7.9784 10.0000 49.6809
SIZE 1645 207 7286 10.18 46404
LEV 1.5670 0.7999 2.3212 0.0664 12.7013
Panel J. 2002
MBR 0.8006 0.7413 0.4425 0.2976 3.1467
ROS 0.1382 0.0777 0.1782 -0.3159 0.6979
ROA 0.0418 0.0419 0.0498 -0.106 0.2199
ROE 0.0686 0.0687 0.1195 -0.4939 0.3905
ST 51.7388 55.2180 27.5050 0.0000 90.0000
LP 8.0386 0.0000 25.7598 0.0000 72.7739
A 5.9066 0.0000 7.9997 0.0000 33.6204
HSHARE 34.2073 34.6850 11.7900 10.0000 100.0000
SIZE 1816 209 7363 16.71 47009
LEV 1.1147 0.6867 1.1631 0.0656 5.5604
Panel K. 2003
MBR 0.9435 0.7812 0.5479 0.3983 2.9502
ROS 0.1607 0.0938 0.3220 -0.3482 2.1490
ROA 0.0528 0.0526 0.0598 -0.152 0.1600
ROE 0.0833 0.0984 0.1236 -0.4304 0.2721
ST 52.4639 53.5600 26.7705 0.0000 90.0000
LP 7.3797 0.0000 25.2037 0.0000 72.7739
A 6.4482 3.2294 8.1688 0.0000 33.6204
HSHARE 34.2620 34.6850 12.2200 10.0000 100.0000
SIZE 2253 413 7117 19.32 41929
LEV 0.9633 0.00 0.9421 0.0728 4.8251
Panel L. 1993-2003
MBR 0.8524 0.6980 1.0515 0.0000 17.2817
ROS 0.1299 0.0749 0.2470 -0.7095 2.1490
ROA 0.0344 0.0376 0.0563 -0.2636 0.3421
ROE 0.0283 0.0579 0.4285 -6.9915 2.0295
ST 45.0346 55.3868 26.0174 0.0000 90.0000
LP 15.1494 0.0000 25.3831 0.0000 73.8900
A 5.8815 0.0022 7.7180 0.0000 33.6204
HSHARE 33.9043 34.8000 8.5263 10.0000 100.0000
SIZE 1101 253 4407 10.18 55066
LEV 1.2863 0.6888 4.5093 0.0656 83.2975
We would like to thank the anonymous referee for his helpful comments and suggestions, which significantly improved the quality of the paper. All remaining errors are ours. The corresponding author acknowledges the RGC Grant (No. PolyU 5334/03H) from the Hong Kong SAR Government.
References
Aggarwal, R., R. Leal, and L. Hernandez, 1993, "The Aftermarket Performance of Initial Public Offerings in Latin America," Financial Management 22, 42-53.
Aybar, C.B., 2001, "The Long Run Performance of Privatization Related ADR Issues," Southern New Hampshire University Working Paper.
Biais, B. and E. Perotti, 1997, "Machiavellian Privatization," American Economic Review 92, 240-258.
Bortolotti, B., M. Fantini, and C. Scarpa, 2000, "Why Do Governments Sell Privatized Companies Abroad?" FEEN Working Paper.
Boubakri, N. and J.C. Cosset, 1998, "The Financial and Operating Performance of Newly Privatized Firms: Evidence from Developing Countries," Journal of Finance 53, 1081-1110.
Boubakri, N., J.C. Cosset, and O. Guedhami, 2002, "Liberalization, Corporate Governance and the Performance of Newly Privatized Firms," Universite Laval Working Paper.
Boycko, M., A. Shleifer, and R.W. Vishny, 1996, "A Theory of Privatization," Economic Journal 106, 309-319.
Bray, A. and P. Gompers, 1997, "Myth or Reality? The Long-Run Underperformance of Initial Public Offerings: Evidence from Venture and Nonventure Capital-Backed Companies," Journal of Finance 52, 1791-1821.
Che, J. and Y. Qian, 1998, "Insecure Property Rights and Government Ownership of Firms," Quarterly Journal of Economics 113, 467-496.
Chen, Y.M., 1998, "Ownership Structure and Corporate Performance: Some Chinese Evidence," San Francisco State University Working Paper.
China National State-Asset Administration Bureau, Handbook for State-owned Shareholders, 1999, Northeast University of Finance and Economics Press.
China Securities Regulatory Commission, 1999, "A Notice about Questions Related to Enterprises' Applying for Overseas Listing," July 14.
China Securities Regulatory Commission Web News, http://www.csrc.gov.cn/cn/homepage/index.jsp
Dewenter, K. and P.H. Malatesta, 2001, "State-Owned and Privately-Owned Firms: An Empirical Analysis of Profitability, Leverage, and Labour Intensity," American Economic Review 91, 320-334.
Djankov, S. and P. Murrell, 2000, "Enterprise Restructuring in Transition: A Quantitative Survey," Working Paper.
D'Souza, J. and W.L. Megginson, 1999, "The Financial and Operating Performance of Newly Privatized Firms in the 1990s," Journal of Finance 54, 1397-1438.
DuCharme, L., P. Malatesta, and S. Sefcik, 2001, "Earnings Management: IPO Valuation and Subsequent Performance," Journal of Accounting, Auditing and Finance 16, 369-396.
Euroweek, 1994, May 13.
Foerster, S.R. and G.A. Karolyi, 2000, "The Long-Run Performance of Global Equity Offerings, Journal of Financial and Quantitative Analysis 35, 499-528.
Feng, F., Q. Sun, and W. Tong, 2004, "Do Government-Linked Companies Necessarily Underperform?" Journal of Banking and Finance 28, 2461-2492.
Frye, T. and A. Shleifer, 1997, "The Invisible Hand and the Grabbing Hand," American Economic Review 87, 354-358.
Gompers, P. and J. Lerner, 2003, "The Really Long-Run Performance and Initial Public Offerings: The Pre-Nasdaq Evidence," Journal of Finance 58, 1355-1392.
Gupta, N., 2005, "Partial Privatization and Firm Performance," Journal of Finance (Forthcoming).
Himmelberg, C.P., G.R. Hubbard, and D. Palia, 1999, "Understanding the Determinants of Managerial Ownership and the Link between Ownership and Performance," Journal of Financial Economics 53, 353-384.
La Porta, R. and F. Lopez-de-Silanes, 1997, Benefits" of Privatization--Evidence from Mexico, Private Sector (World Bank, Washington D.C., June), 21-24.
Loughran, T. and J. Ritter, 1997, "The Operating Performance of Firms Conducting Seasoned Equity Offerings, Journal of Finance 52, 1823-1850.
Loughran, T. and J. Ritter, 1995, "The New Issues Puzzle," Journal of Finance 50, 23-51.
McConnell, J.J. and H. Servaes, 1990, "Additional Evidence on Equity Ownership and Corporate Value," Journal of Financial Economics" 27, 595-612.
Megginson, W.L., R.C. Nash and M. Van Randenborgh, 1994, "The Financial and Operating Performance of Newly Privatized Firms: An International Empirical Analysis," Journal of Finance 49, 403-452.
Megginson, W.L., R.C. Nash, J.M. Netter and A.L. Schwartz, 2000, "The Long Run Return to Investors in Share Issue Privatizations," Financial Management 29, 67
Ritter, J. and 1. Welch, 2002, "A Review of IPO Activity, Pricing, and Allocations," Journal of Finance 57, 1795-1828.
Ritter, J. R. 1991, "The Long-Run Performance of Initial Public Offerings," Journal of Finance 46, 3-27.
Schaub, M., 2003, "Investment Performance of American Depository Receipts listed on the New York Stock Exchange: Long and Short," Journal of Business & Economics Studies" 9, 1-19.
Sun, Q. and W. Tong, 2003, "China Share issue Privatization: The Extent of Its Success," Journal of Financial Economies 70, 183-222.
Sun, Q. and W. Tong, 2000, "The Effect of Market Segmentation on Stock Prices: the China Syndrome," Journal of Banking and Finance 24, 1875-1902.
Sun, Q., W. Tong, and Q. Yu, 2002, "Determinants of Foreign Direct Investment across China," Journal of International Money and Finance 21, 79-113.
Teoh, S.H., I. Welch, and T.J. Wong, 1998a, "Earnings Management and the Under-Performance of Seasoned Equity Offerings," Journal of Financial Economics 50, 63-99.
Teoh, S.H., I. Welch, and T.J. Wong, 1998b, "Earnings Management and Long-Term Under-Performance of Initial Public Offerings," Journal of Finance 53, 1935-1974.
Tian, G., 2000, "Performance of Mixed Enterprises, State Shareholding and Agency Cost," London Business School Working Paper.
Walder, A., 1995, "China's Transitional Economy: Interpreting its Significance," China Quarterly 144, 963-979.
Xu, X. and Y. Wang, 1997, "Ownership Structure, Corporate Governance, and Corporate Performance," The World Bank, Washington, DC, Policy Research Working Paper 1794.
Zhou, X., 2001, "Understanding the Determinants of Managerial Ownership and the Link between Ownership and Performance: Comment," Journal of Financial Economics 62, 559-571.
(1) Some SOEs have also issued shares in New York as "N-shares," mostly in the form of ADRs, which are thinly traded, and the underlying shares are mostly H-shares. Listings made elsewhere, for instance, London and Singapore, are very few during our sample period and cannot be compared to those made in Hong Kong in terms of proceeds. There are also "B-shares", which are open to foreign investors, but these shares are listed locally in China on either the Shanghai Stock Exchange (SHSE) or the Shenzhen Stock Exchange (SZSE).
(2) See the Handbook of State-owned Shareholders (1999) for details.
(3) See Appendix for ownership distribution over the years.
(4) For comparison, we note that the corresponding figure for domestically listed firms is about 27%.
(5) See Chen (1998). Also, a well-known Chinese economist has said that the fundamental problem with China's stock market is the split between tradable and nontradable shares. This economist argues that the interests of the two groups of shareholders are incompatible. When nontradable shareholders control a firm, they do not maximize the interests of circulating shareholders (Liberation Daily, September 28, 2001).
(6) Previous studies also examine dividend payouts and employment. However, because of limitations to the data, we do not examine these variables. Data on dividends are not available for the years before the listing, and employment data are only available for some firms in the year before the listing.
(7) We thank the referee for suggesting this to us.
(8) The reason we did not look at A-share indices is because A-shares are open to Chinese local investors only and B-shares are open to foreign investors only, although the B-share market has been open up to Chinese local investors since March 2001.
(9) Theoretically, if there is a general decline in leverage after the SIP, a firm's financial risk would be lowered, leading to lower shareholder demand for return on equity. In that case, an ex post return decline might be due to an ex ante decline in required return. We thank the referee for pointing this out to us.
(10) See Table I for our sample firm distribution across eight industries.
Jin Jia, Qian Sun, and Wilson H.S. Tong *
* Jin Jia is with the Motorola Innovation Centre, Singapore. Qian Sun is an Associate Professor at the Nanyang Technological University, Singapore. Wilson Tong is an Associate Professor at the Hong Kong Polytechnic University in Kowloon, Hong Kong.
Table I. Distribution of Sample Firms across Industries and
Incorporation Place
H-share companies listed on the Hongkong Exchange Mainboard
between 1993 and 2002
Panel A Industrial Distribution
1993 1994 1995 1996 1997
Manufacturing 2 5 1 3 2
Petrochemical 1 2 1 1
Transportation 1 1 2 6
Utilities 1
Agricultural 1
Basic Materials 1 1 4
Technology
Health 1 1
Property 1
Total 6 9 2 6 16
Panel B Industrial Distribution
1993 1994 1995 1996 1997
Coastal 4 4 1 4 11
Inland 2 5 1 2 5
Panel A Industrial Distribution
Manufacturing 1998 1999 2000 2001 2002
Petrochemical
Transportation 1 1
Utilities 2 1
Agricultural 1 1
Basic Materials 1 2
Technology
Health 1
Property 1 1 1 1
Total
2 3 3 2 4
Panel B Industrial Distribution
Coastal 1998 1999 2000 2001 2002
Inland
1 3 3 1 3
1 0 0 1 1
Note: The following provinces (municipalities) would be considered
coastal area: Liaoning, Beijing, Tianjin, Shandong, Jiangsu,
Zhejiang, Fujian, Guangdong, and Hainan. The rest are considered
inland areas.
Table II. Results for the Profitability of H-Share Companies
The table presents the number of observations, the mean and median
values of the profitability proxies for the average of three years
before and after the listing period, the mean and median changes in the
profitability proxies (after-before), and the tests of significance of
the median change for each empirical proxy in various samples. The real
net profit (RNP) is standardized to one in year 0 (the year of
privatization). We use the Wilcoxon Z-test to test for any significant
changes in the median value (paired observation), and for any
significant differences in changes between any two groups. The table
also shows the number of positive compared to negative changes.
Median Median
Sample (Mean) (Mean)
(-3 To +3) Var. N Before After
Whole RNP 53 0.5636 0.9491
Sample (0.7406) (1.1210)
ROS 53 0.1055 0.0904
(0.1768) (0.1559)
A Share RNP With 27 0.4854 0.6038
Dual-listings (0.6343) (0.9103)
Without 26 0.7096 1.0274
(0.8403) (1.3304)
ROS With 27 0.0970 0.0627
(0.1275) (0.1187)
Without 26 0.1553 0.1659
(0.2323) (0.1988)
Foreign RNP With 28 0.6962 0.9906
Director (0.8685) (1.2620)
Without 25 0.5158 0.6927
(0.5862) (0.9534)
ROS With 28 0.1433 0.1086
(0.2349) (0.2155)
Without 25 0.0855 0.0564
(0.1162) (0.0936)
Change of RNP With 23 0.5179 0.5213
CEO (0.6692) (0.6775)
Without 30 0.6762 1.0732
(0.7861) (1.4529)
ROS With 23 0.0806 0.0382
(0.1128) (0.0845)
Without 30 0.1748 0.1375
(0.2296) (0.2144)
Median Wilcoxon Test
Sample (Mean) Paired Between +ve/-ve
(-3 To +3) Change Obs. Groups Ratio
Whole 0.3855 1.7235 * 33/20
Sample (0.3804)
-0.0151 1.5453 18/35
(-0.021)
A Share 0.1184 0.8275 0.5249 18/9
Dual-listings (0.2760)
0.3178 1.6435 15/11
(0.4902)
-0.0343 1.8271 * 0.1690 9/18
(-0.008)
0.0106 0.8131 9/17
(-0.033)
Foreign 0.2944 1.3996 0.2227 16/12
Director (0.3935)
0.1769 0.9608 17/8
(0.3672)
-0.0347 1.0575 0.4722 10/18
(-0.019)
-0.0291 1.3634 8/17
(-0.023)
Change of 0.0034 0.2784 1.5703 12/11
CEO (0.0083)
0.3970 2.4919 ** 21/9
(0.6668)
-0.0424 1.6599 * 0.0090 7/16
(-0.028)
-0.0373 1.0981 11/19
(-0.015)
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table III. Results for the Output of H-Share Companies
The table presents the number of observations, the mean and median
values of the real output for the average of three years before and
after the listing period, the mean and median changes in the
profitability proxies (after-before), and the tests of significance of
the median change for real output in various samples. The output is
standardized to one in year 0 (the year of privatization). We use the
Wilcoxon Z-test to test for any significant changes in the median value
(paired observation), and for any significant differences in changes
between any two groups. The table also shows the number of positive
changes compared to negative changes.
Median Median
Sample Real (Mean) (Mean)
(-3 To +3) Output N Before After
Whole 53 0.7085 1.2835
Sample (0.8128) (1.5632)
A Share With 27 0.6935 1.3056
Dual- (0.8170) (1.4830)
listings Without 26 0.7097 1.2286
(0.8127) (1.6582)
Foreign With 28 0.7482 1.3503
Director (0.8403) (1.7531)
Without 25 0.6806 1.2133
(0.7865) (1.3628)
Change of With 23 0.7128 1.1882
CEO (0.8140) (1.3471)
Without 30 0.7006 1.3777
(0.8156) (1.7391)
Median Wilcoxon Test
Sample (Mean) Paired Between +ve/-ve
(-3 To +3) Change Obs. Groups Ratio
Whole 0.5750 7.1737 *** 51/2
Sample (0.7504)
A Share 0.6121 4.8587 *** 0.1690 25/2
Dual- (0.666)
listings 0.5189 5.3630 *** 26/0
(0.8455)
Foreign 0.6021 5.0542 *** 1.2918 27/1
Director (0.9128)
0.5327 5.2616 *** 24/1
(0.5763)
Change of 0.4754 4.6291 *** 2.144 ** 21/2
CEO (0.5331)
0.6771 5.5892 *** 30/0
(0.9235)
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table IV. Results for Coverage Ratios of H-Share Companies
The table presents the number of observations, the mean and median
values of the leverage proxies for the average of three years before
and after the listing periods, the mean and median changes in the
profitability proxies (after-before), and the tests of significance of
the median change for each empirical proxy in various samples. We use
the Wilcoxon Z-test to test for any significant changes in the median
value (paired observation), and for any significant differences in
changes between any two groups. The table also shows the number of
positive changes compared to negative changes.
Median
(Mean)
Sample (-3 To +3) Variables N Before
Whole TIE 47 4.0383
Sample (9.4106)
OCF/TD 37 0.2110
(0.6243)
Firms with a Share TIE With 27 4.2224
Dual-listings (8.3611)
Vs Without 20 4.0155
Firms without a
Share Dual-listings (10.8275)
OCF/TD With 22 0.2405
(0.4019)
Without 15 0.1728
(0.9505)
TIE With 25 4.6149
Firms with (10.2533)
Foreign Directors Without 22 3.3644
Vs
Firms without (8.4529)
Foreign Directors OCF/TD With 17 0.1728
(0.4481)
Without 20 0.2405
(0.7740)
Firms with Change of TIE With 22 3.9271
CEO (10.7565)
Vs Without 25 4.2224
Firms without (8.2262)
Change of CEO OCF/TD With 17 0.1690
(0.2875)
Without 20 0.2378
(0.9105)
Median Median Wilcoxon Test +ve/-
(Mean) (Mean) Paired Between ve
Sample (-3 To +3) After Change Obs. Groups Ratio
Whole 4.5526 0.5143 0.0302 23/24
Sample (15.3077) (5.8971)
0.2456 0.0346 0.3351 15/22
(0.6108) (-0.0135)
Firms with a Share 4.0293 -0.1931 0.3114 0.6347 11/16
Dual-listings (11.3939) (3.0328)
Vs 5.9372 1.9217 0.3652 12/8
Firms without a
Share Dual-listings (20.5915) (9.7640)
0.2659 0.0254 0.2934 1.0363 8/14
(0.4071) (0.0052)
0.1636 -0.0092 0.1244 7/8
(0.9096) (-0.0409)
5.7748 1.1599 0.5433 0.9274 14/11
Firms with (24.7980) (14.5447)
Foreign Directors 3.9214 0.557 0.6220 9/13
Vs
Firms without (4.5234) (-3.9295)
Foreign Directors 0.1498 -0.023 0.4822 0.5943 6/11
(0.2111) (-0.237)
0.3249 0.0844 0.0406 9/11
(0.9506) (0.1766)
Firms with Change of 4.0228 0.0957 0.8802 0.6289 10/12
CEO (13.9061) (3.1496)
Vs 5.3021 1.0797 0.7179 13/12
Firms without (16.5412) (8.315)
Change of CEO 0.1543 -0.0147 0.1378 0.5029 6/11
(0.3269) (0.0394)
0.2659 0.0281 0.6086 9/11
(0.8521) (-0.0584)
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table V. Results for the PPE of H-Share Companies
The table presents the number of observations, the mean and median
values of the PPE measures for the average of three years before and
after the listing period, the mean and median changes in the
profitability proxies (after-before), and the tests of significance of
the median change for each empirical proxy in various samples. The real
net profit (RNP) is standardized to one in year 0 (the year of
privatization). We use the Wilcoxon Z-test to test for any significant
changes in the median value (paired observation), and for any
significant differences in changes between any two groups. The table
also shows the number of positive compared to negative changes.
Median
(Mean)
Sample Variables N Before
Whole PPE 18 0.8553
Sample
(-3 To +3) (1.0756)
PPE/TA 22 0.0711
(0.0892)
A Share PPE With 5 0.7783
Dual-Listings (1.2814)
Without 13 0.8854
(0.9964)
PPE/TA With 7 0.0806
(0.0984)
Without 15 0.0672
(0.0848)
Foreign PPE With 11 0.8253
Director
(0.9523)
Without 7 0.8854
(1.2693)
PPE/TA With 11 0.0672
(0.0827)
Without 11 0.0806
(0.0957)
Change of PPE With 6 0.7672
CEO (0.8853)
Without 12 0.9548
(1.1707)
PPE/TA With 6 0.0868
(0.0775)
Without 16 0.0676
(0.0935)
Median Median Wilcoxon Test +ve/-
(Mean) (Mean) Paired Between ve
Sample After Change Obs. Groups Ratio
Whole 1.2850 0.4297 1.7429 * 11/7
Sample
(-3 To +3) (2.1296) (1.0540)
0.0607 -0.0104 1.3623 6/16
(0.0622) (-0.027)
A Share 0.9261 0.1478 0.7856 0.7961 3/2
Dual-Listings (2.0285) (0.7471)
1.8605 0.9751 1.5286 8/5
(2.2378) (1.2414)
0.0615 -0.0191 1.2758 1.0421 1/6
(0.0621) (-0.0363)
0.0489 -0.0183 0.9954 5/10
(0.0623) (-0.0225)
Foreign 1.4377 0.6124 1.5895 0.5303 6/5
Director
(2.1232) (1.1709)
1.2850 0.3996 0.5293 5/2
(2.1437) (0.8744)
0.0614 -0.0058 1.1904 0.3534 3/8
(0.0609) (-0.0218)
0.0607 -0.0199 0.8160 3/8
(0.0637) (-0.032)
Change of 0.8106 0.0434 0.7396 0.4125 4/2
CEO (1.4979) (0.6126)
2.0700 1.1152 1.8774 * 7/5
(2.8063) (1.6356)
0.0489 -0.0379 1.6038 0.3958 2/4
(0.0554) (-0.0221)
0.0648 -0.0028 0.6218 4/12
(0.0682) (-0.0253)
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table VI. Cross-Sectional Regression of Performance Change on Ownership
Change
This table presents the empirical results of our cross-sectional
regression analysis on the full sample of H-share firms. We use the
following model:
[DELTA]P[P.SUB.I] = [[alpha].sub.I] + [[beta].sub.1]
[DELTA]AGDPG[R.SUB.i] + [[beta].sub.2]CTRL + [[beta].sub.3]CE[O.sub.I]
+ [[beta].sub.4]F[D.sub.I] + [[beta].sub.5]RE[G.sub.j] +
[[beta].sub.6]ASHAR[E.sub.j] + [[epsilon].sub.I]
PP is the performance proxy, which includes real net profit (RNP),
return on sales (ROS), real output (RSALES), times-interest-earned
(TIE), and operating cash flow over total debt (OCF/TD). GDPGR is
China's real GDP growth rate. The difference sign "_" is the average of
the three-year post-privatization data minus the average of the
three-year pre-privatization data of the variable in question, which
captures the difference in the mean of the variable before and after
the privatization. Other variables are controlling dummies. CTRL takes
a value of one if the state has less than 50% of the total share
ownership after listing. CEO takes a value of one if the firm changes
its CEO of the board of directors after listing. FD takes a value of
one if there are foreign (Hong Kong) directors on the board. REG takes
a value of one if a firm is in a regulated industry (petrochemicals,
utilities, and some of the basic materials). ASHARE takes a value of
one if a firm has A-share offerings. Panel A presents the results of
the regression without the control dummy variables. Panel B presents
the results with full specifications.
Panel A. Without Control of Dummy Variables
[DELTA]RNP [DELTA]ROS [DELTA]RSALES [DELTA]TIE
C 0.5688 0.0079 0.6812 1.7621
(2.42) ** (0.31) (5.28) *** (0.24)
[DELTA]GDPGR 9.1096 1.3985 -3.5339 327.890
(1.12) (1.64) * (-0.79) (1.35)
Adjusted [R.sup.2] 0.0049 0.0294 -0.0073 0.0176
Observation 53 53 53 47
Panel B. Full Specification
C 0.8542 0.0024 0.7275 7.5199
(1.96) * (0.05) (3.19) *** (0.53)
[DELTA]GDPGR 7.0854 1.5381 -3.8326 202.8011
(0.85) (1.65) * (-0.88) (0.84)
CTRL 0.3044 -0.0109 0.2242 -19.456
(0.89) (-0.29) (1.25) (-2.02) **
REG -0.24024 -0.01509 -0.16058 16.78855
(-0.67) (0.38) (-0.86) (1.64)
ASHARE -0.10978 0.03208 -0.08427 -2.58993
(-0.33) (0.85) (-0.48) (-0.27)
FD -0.024 0.0128 0.2518 20.0944
(-0.07) (0.33) (1.38) (1.99) **
CEO -0.6883 -0.0112 -0.4102 1.9570
(-2.02) (-0.30) (-2.31) ** (0.21)
Adjusted [R.sup.2] 0.0052 -0.0507 0.0907 0.1041
Observation 53 53 53 47
Panel A. Without Control of Dummy Variables
[DELTA]OCF/TD [DELTA]PPE [DELTA]PPE/TA
C -0.0208 1.4160 0.0151
(-0.09) (1.69) * (0.51)
[DELTA]GDPGR -0.3437 16.992 -0.2880
(-0.05) (0.58) (-0.26)
Adjusted [R.sup.2] -0.0285 -0.0501 -0.0579
Observation 37 15 18
Panel B. Full Specification
C 0.1263 0.2947 0.0639
(0.26) (0.23) (1.36)
[DELTA]GDPGR -1.467 -6.2626 0.1159
(-0.18) (-0.15) (0.08)
CTRL -0.127 2.0363 -0.0228
(0.36) (1.85) (-0.60)
REG 0.07391 0.1156 -0.0662
(0.21) (0.09) (-1.56)
ASHARE -0.0087 1.0943 0.0203
(-0.03) (0.88) (0.45)
FD -0.4121 -0.1703 -0.0264
(1.16) (0.16) (-0.60)
CEO 0.0957 -1.3998 0.0118
0.28 (-1.15) (0.27)
Adjusted [R.sup.2] -0.131 -0.0813 -0.1637
Observation 37 15 18
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table VII. Excess Long-run Buy-and-Hold Returns of H-Share Firms after
Stock Issuing Privatization
We calculate the market-adjusted buy-and-hold return for each firm
using the following formulation:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
where ER is the market-adjusted cumulative return. [R.sub.i] denotes
daily stock return for firm i, MR is the relevant daily market return
and t = (a to b) is the timeframe of one to five years. We use the Hang
Seng index (HSI), Shanghai B-share index (SH-B), Shenzhen B-share index
(SZ-B), and China Affiliated Enterprise (Red-Chip) index as the market
proxies. To compare the means and medians of the buy-and-hold returns,
we provide t-statistics and Wilcoxon Z statistics, respectively.
H
Holding Pd. Share HIS Diff Test
One Year Mean -0.1574 -0.0527 -0.1047 -1.83 *
Median -0.2374 -0.0186 -0.2188 2.45 **
Obs. 53 53
Two Years Mean -0.2788 0.0475 -0.3263 -3.41 ***
Median -0.4353 -0.0658 -0.3695 5.05 ***
Obs. 49 49
Three Years Mean -0.3416 0.1780 -0.5197 -6.02 ***
Median -0.462 0.1637 -0.6257 5.60 ***
Obs. 47 47
Four Years Mean -0.3594 0.1353 -0.4947 -6.73 ***
Median -0.4616 0.0346 -0.4962 5.48 ***
Obs. 44 44
Five Years Mean -0.4336 0.0826 -0.5162 -5.30 ***
Median -0.4603 -0.0694 -0.3909 5.17 ***
Obs. 41 41
Holding Pd. SZ-B Diff Test SH-B Diff
One Year 0.0111 -0.1685 -1.79 * -0.019 -0.1384
-0.3337 0.0963 1.02 -0.203 -0.0344
53 53
Two Years 0.0750 -0.3538 -3.02 *** 0.0693 -0.3481
-0.2103 -0.225 3.08 *** -0.27 -0.1653
49 49
Three Years 0.1901 -0.5317 -5.69 *** 0.1118 -0.4535
-0.0308 -0.4312 4.92 *** -0.148 -0.314
47 47
Four Years 0.5612 -0.9206 -8.54 *** 0.5234 -0.8828
0.4012 -0.8628 6.13 *** 0.2890 -0.7506
44 44
Five Years 0.5900 -1.0236 -6.21 *** 0.5507 -0.9843
0.4240 -0.8843 5.70 *** 0.6622 -1.1225
41 41
Red
Holding Pd. Test Chip Diff Test
One Year -1.53 -0.121 -0.0364 -0.60
0.94 -0.2182 -0.0192 0.01
53
Two Years -2.41 ** -0.101 -0.1778 -1.27
2.55 ** -0.335 -0.1003 0.99
49
Three Years -4.03 *** 0.0364 -0.3781 -3.31 ***
4.23 *** -0.2148 -0.2472 3.05 ***
47
Four Years -7.31 *** -0.0175 -0.3418 -2.87 ***
5.34 *** -0.2782 -0.1834 2.13 **
44
Five Years -7.44 *** -0.297 -0.1366 -1.31
4.80 -0.4226 -0.0377 1.39
41
Note: (1.) The return computation for the red-chip index is not
available for the five-year investment horizon, as the index was not
introduced before 1995.
(2.) * , **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
Table VIII. Pooled Regression Analysis of the Impact of Government and
Legal-Person Ownership on the Market Measure of Firm Performance after
Listing
In this table, the dependent variable is MBR, the market value of
equity divided by the book value of net assets. SIZE is the natural
logarithm of annual sales in RMB. LEV is the value of total debt
divided by the owner's equity. ST, LP, and H represent the share of
equity owned by the state, legal persons, and Hong Kong investors,
respectively. STI equals ST for STs below 50%, and zero otherwise. ST2
equals ST for STs greater than 50%, and zero otherwise. We define LP1
and LP2 in the same way. LOC is a dummy variable that equals one if
the firm is incorporated in the coastal area and zero if it is located
inland. A-shares are a dummy if the firm is listed dually in China's
domestic market and zero if it is only listed in Hong Kong. The pooled
sample consists of 53 H-share companies that listed between 1993-2002.
Their post-listing data cover 1993-2003. Heteroscedasticity-consistent
t-values are in parentheses.
Fixed Effect
Common Industry- Firm-
Intercept Specific Specific
C -0.5272
(-1.44)
GDP 0.1858 0.1836 0.1655
(7.63) *** (8.23) *** (7.44) ***
ST1 -0.0028 -0.0031 0.0043
(-1.59) (-1.65) * (1.71) *
ST2 -0.0037 -0.0035 -0.0034
(-2.19) ** (-2.14) *** (-2.23) ***
LP1
LP2
H
SALE -0.021 -0.0407 -0.2106
(1.02) (-2.50) ** (-3.14) ***
LEV -0.0120 -0.0680 -0.092
(-0.0743) (-0.4218) (-0.5346)
LOC 0.3608 0.3474
(5.90) *** (4.59) ***
A-Share -0.1919 -0.1727
(-3.30) *** (-2.49) **
Obs 323 323 323
Adj. [R.sup.2] 0.7774 0.8726 0.9278
Fixed Effect
Common Industry- Firm-
Intercept Specific Specific
C -0.6983
(-1.92) *
GDP 0.1836 0.1677 0.1755
(7.52) *** (7.59) *** (7.82) ***
ST1
ST2
LP1 0.0097 0.0095 0.0018
(2.70) *** (2.54) *** (1.70) *
LP2 0.0023 0.0020 -0.0065
(2.05) ** (1.62) (-0.78)
H
SALE -0.0228 -0.0618 -0.2354
0 (-3.62) *** (-3.70) ***
LEV 0.0834 0.0877 -0.109
-0.6934 (0.8455) (-0.3692)
LOC 0.3408 0.3595
(5.54) *** (4.79) ***
A-Share -0.2023 -0.2088
(-3.50) *** (-3.11) ***
Obs 323 323 323
Adj. [R.sup.2] 0.7751 0.8699 0.9253
Fixed Effect
Common Industry- Firm-
Intercept Specific Specific
C -1.8111
(-4.42) ***
GDP 0.1854 0.1652 0.1744
(7.83) *** (7.86) *** (7.79) ***
ST1
ST2
LP1
LP2
H 0.0177 0.0209 0.0087
(5.41) *** (6.33) *** (2.15) **
SALE 0.0114 -0.0672 -0.2375
(0.56) (-4.27) *** (-3.69) ***
LEV -0.0836 0.0126 -0.0709
(-0.6417) -0.0435 (-0.5933)
LOC 0.3543 0.3370
(6.00) *** (4.89) ***
A-Share -0.1311 -0.0872
(-2.31) ** (-1.35)
Obs 323 323 323
Adj. [R.sup.2] 0.7883 0.8822 0.9251
Note: *, **, and *** denote significance at the 0.1, 0.05, and 0.01
levels, respectively (two-tails).
In addition, make sure to read these articles:
- From Wall Street to Beijing: global finance has new rules and new players.
- The deregulation of state-owned enterprises in India and China.
- FDI in India.
- Jones Day office authorized to practice as local solicitors in Hong Kong; a first for U.S.
- Hong Kong: Favorable turn in Hong Kong apparel exports
- Hong Kong Firm Awarded for Inventive Design
- Hong Kong: Synthetic fiber market weakens slightly
- Hong Kong shakes off Sars effect
- Hong Kong: Highly reevaluated business base
- Sidley Austin Brown & Wood Expands Korean Practice; Five Corporate Attorneys Join Firm...