عدم تقارن اطلاعات و برآورد ریسک: شواهد اولیه از بازارهای سهام چین

This study discusses the implication of information asymmetry between firms and investors for the estimation risk of asset returns. We evaluate various risk measures of information asymmetry between firms and investors for China, an excellent example of a low information environment. We find a significant negative relationship between voluntary disclosure based on U.S. GAAP with certification of a credible audit firm and the variation of risk-adjusted returns. On the other hand, neither private information production nor certified voluntary disclosure is associated with any variation of asset returns in the primary markets.

The evaluation of information asymmetry between firms and investors has remained inconclusive. One difficulty lies in understanding the impact of investors making investment decisions under imperfect information in capital markets.1 Another difficulty lies in measuring the impact of information asymmetry. In this study, we adopt the Bayesian approach to explain how investor's imperfect information about the parameters of the firm's underlying cash flow or return generating process gives rise to estimation risk. Lam (1991) defines estimation risk as the incremental variation of an investor's predictive return distribution relative to the limiting case when all agents have perfect information. Following Clarkson et al. (1996) and Lewellen and Shanken (2002), we posit that both the beta estimate and the variance of returns are appropriate measures of estimation risk.2 If estimation risk increases a firm's beta estimate and variance of returns, the firm has incentives to reduce information asymmetry because estimation risk translates to a higher cost of equity. Healy and Palepu (1995) analyze at length how CUC International used costly capital structure modifications as a signal to investors. Since investors cannot observe the firm's underlying cash flow, CUC had difficulty convincing investors that its marketing outlays were profitable investments. Such anecdotal evidence suggests that financial disclosure in a high information environment like the U.S. is less likely to be effective in resolving information asymmetry, resulting in poor stock price performance in this instance. A solution is effective if, by revealing information about the firm's underlying cash flow or return generating process to investors, it reduces estimation risk (and therefore beta estimate and variance of returns).3 Mandatory disclosures, effective enforcements and well-developed analysts' followings in developed economies are expected to yield a relatively low level of estimation risk cross-sectionally in the economy. Therefore, the challenge of working with data in developed economies lies in identifying windows of low information. In this connection, Botosan (1997) finds that greater disclosure through the annual report is associated with a lower cost of equity for firms that attract a low analyst following. On the other hand, she finds no evidence of an association between her measure of disclosure level and cost of equity for firms that have a high analyst following. In this study, we evaluate information asymmetry in a low information environment where the problem is likely to be significant.4 We identify an empirical setting that allows us to evaluate this research problem. China offers an ideal low information environment. We evaluate the relationship of private information production and voluntary disclosure with certification and the variation of asset returns in both the primary and secondary equity markets in China.5 China is an interesting test case because there is a dearth of security research relative to what is available in developed economies like the U.S., Switzerland and Germany. Moreover, the Chinese retail investors are less sophisticated than those in the developed economies. Therefore, the quality of information and the level of disclosure that is relevant and useful for the Chinese investors are more basic than what investors in the developed economies are used to. We construct measures for the quantity and quality of private information produced, as well as the quantity and quality of the voluntary disclosure for each firm in both samples for the primary and secondary markets. Following Lewellen and Shanken (2002) and Zhang (2001), this study evaluates the relationship between private information production and disclosure about the firm's underlying cash flow and the variation of asset returns. We hypothesize that there is a negative relationship between cross-sectional beta and variance estimates and the information measures. Consistent with our prediction, we find that the variance of risk-adjusted asset returns in the secondary markets is related negatively with voluntary disclosure according to internationally accepted standards like the U.S. Generally Accepted Accounting Principles (GAAP) as certified by credible audit firms. Setting our study in China also allows for the isolation of information asymmetry effects on beta and variance estimates. Zhang's (2001) model shows that the relationship between disclosure and the variation of asset returns depends on whether the variability of asset returns is driven by information traders' activities or information asymmetry. To illustrate the point, Bushee and Noe (2000) evaluate institutional investors in a highly developed capital market (U.S.) with strongly enforced disclosure standards. Our study evaluates a new and developing capital market (China) that has lower quality disclosure and enforcement regimes at the time of the study. Bushee and Noe (2000) find that yearly improvements in disclosure rankings actually increase return volatility. They suggest that U.S. institutional investors who are attracted to greater disclosure actually increase return volatility through their aggressive trading strategies. In China's case, institutional investors account for less than 20% of total shareholders in the A share markets.6 Moreover, Chinese equity markets are closed to foreign investors except for the B shares of listed firms.7 Our samples are drawn from listed firms that have issued only A shares which are theoretically accessible to local investors only. Our finding a negative relationship between certified disclosure and the variation of asset returns is therefore consistent with the prediction of Zhang's (2001) model when the variability of asset returns is driven by information asymmetry rather than information traders' activities. In examining firms that choose to disclose according to U.S. GAAP, we are also contributing to the international accounting literature. While voluntary disclosure in itself does not necessarily resolve information asymmetry, it can become an effective signal for quality information when certified by a credible audit firm. Our findings throw light on the dynamic resolution of information asymmetry in the development of capital markets. While voluntary disclosure about the firm's underlying cash flow as certified by credible audit firms is associated with lower risk in the Chinese secondary markets, we find that the quantity of private information produced is not. Yet, we find that Chinese firms that release quality information also have a better coverage. These findings highlight the dominant role of credible certification agents and standards in the early stages of development of the Chinese capital markets. Lastly, we report that neither private information production nor voluntary disclosure with certification is associated with the variation of asset returns in the Chinese primary markets. This study is organized as follows. Section 2 gives the background on the development of Chinese markets to explain why it is an interesting case of a low information environment for our study. Section 3 discusses the hypothesis development, data and method. Section 4 reports the results and the analyses. Section 5 concludes and highlights the limitations to this study.

This study evaluates investor uncertainty over the firm's underlying cash flow or return generating process by measuring its direct impact on the observed beta and variance of returns. If private information production or disclosure is any effective in resolving this information asymmetry problem, it would be associated with lower beta and variance estimates. We find preliminary evidence that quality disclosure according to U.S. GAAP with certification of a credible audit firm is associated with lower variance of risk-adjusted returns. In the Chinese IPO markets where the information asymmetry problem is heightened, we find no association between information production or disclosure with the variation of asset returns. One explanation is that when the probability of loss to a successful IPO subscriber is perceived to be insignificant, the expected benefits of producing information or signaling in such a market would be limited. We wish to qualify the results in several ways. This study is limited to the early years of development of the Chinese equity markets. There were intermittent speculative runs followed by policy reversals during the test period. It means that the data are expected to be very noisy. Moreover, Chinese data are not as clean as U.S. data, for example, especially as data for the information measures in this study are hand-collected and are less rigorously screened for errors. The results should therefore be interpreted against the constraint of working with a dearth of good data. We also highlight that the Chinese experience in the development of their capital markets is unique. These results are hardly generalizable to other markets, nor do we expect them to hold with changes in regulatory and institutional framework over time. Notwithstanding its limitations, this study has implications for management, investors, and regulators. Regulators set the legal, institutional and disclosure framework that affects the supply of and investors' demand for information. Management can help to resolve the information asymmetry problem between the firm and investors through compliance with mandatory disclosure or voluntary disclosure with signaling. What constitutes a credible signal varies from firm to firm and market to market. Healy and Palepu (1995) suggest that investor communications costs must surely be justified by avoidance of a greater potential loss in firm value. Finally, utility-maximizing investors need to evaluate how imperfect information about firms' cash flows would affect their investment decisions, and whether they can resolve part of the uncertainty by considering available information production and disclosure processes.