هزینه های معامله و کارایی بازار: شواهدی از مقررات زدایی کمیسیون

This study analyzes the impacts of explicit transaction costs on weak-form market efficiency within the context of the brokerage commission deregulation in Japan in October 1999, which led to lower commission rates across the market. Applying two alternative statistical tests to both daily and weekly data, we find that return randomness (unpredictability) increases significantly for stocks listed in Japan, but not for the Japanese stocks dually listed in the United States, which are immune to the deregulation. These results suggest an inefficiency loss or an efficiency gain in the Japanese equity market following the deregulation, insofar as randomness proxies for efficiency.

Previous studies of explicit transaction costs, such as brokerage commissions, have largely focused on the liquidity and volatility impacts, as prompted by concerns from regulators about the potential destabilizing effect. Overall, existing evidence points to a negative (positive) relationship between transactions costs and liquidity (price volatility).2 However, transaction costs may also affect other important dimensions of market quality, such as market informational efficiency. Indeed, transaction costs should be important determinants of informational efficiency, according to the definition of an efficient market, which fully reflects all available information promptly when there are no information and transaction costs and when investors hold homogeneous beliefs ( Fama, 1970 and Fama, 1991). 3 To the extent that market transactions are not frictionless, market informational efficiency may not hold ( Grossman and Stiglitz, 1980 and Jensen, 1978). The reason is that high transaction costs discourage information trading and arbitrage trading, resulting in less information being revealed and impounded into stock prices. In a similar vein, the burgeoning behavioral finance literature cites transaction costs as a main source of market inefficiency ( Barberis & Thaler, 2001). As such, the major goal of research in this area is to measure the impacts of these market imperfections on efficiency ( Fama, 1970). Furthermore, a study of the efficiency impacts of transaction costs may also offer valuable empirical evidence to policy makers in their debates about important issues such as securities transaction tax (STT). There has been a debate on the efficiency impact of introducing an STT, the opposite of brokerage commission deregulation. On the one hand, an STT could discourage information and arbitrage trading by increasing transaction costs, leading to lower market informational efficiency (see, e.g., Schwert & Seguin, 1993). On the other hand, an STT should enhance market informational efficiency by reducing noise trader risk and increasing arbitrage efficiency (see, e.g., De Long et al., 1990, Shleifer and Vishny, 1997 and Summers and Summers, 1989).4 Together, these opposing arguments leave the efficiency effect ambiguous, ex ante. An empirical investigation of the efficiency effect from the opposite should help resolve the controversy. Examining the changes in STT and capital gains tax in Japan on April 1, 1989, Liu (2007) provides direct evidence for the first time in the literature that lower exogenous transaction costs help to enhance market informational efficiency, as measured by lower first-order autocorrelation in stock returns. The purpose of this study is to further investigate the effects of explicit transaction costs on the weak-form market efficiency, making use of the brokerage commission deregulation in Japan in October 1999. Such an effort is worthwhile for at least three reasons. First of all, in view of the theoretical importance of the determinants of market efficiency in the finance literature, further evidence should help to generalize the conclusions drawn from a single clinical data point. Moreover, as Jones and Seguin (1997, p. 729) point out, avenues other than transaction taxes have been proposed to curb the undesirable effects of noise trading. Results from an investigation of the commission deregulation should complement those from the STT changes, helping to generalize the policy implications of different types of transaction costs. Finally, the current study employs more statistical tests and additional data to ensure the robustness of results. Specifically, besides the first-order autocorrelation test and daily data employed by Liu (2007), we also use the nonparametric runs test and weekly data in the current study to measure the efficiency impact of brokerage commission rates. With the brokerage commission deregulated, competition among brokerage houses should reduce the commission rates paid by investors, large or small, as reported in the following section, leading to more active trading and speedier incorporation of the same relevant information, private or public, and enhancing informational efficiency. That is, according to the definition of market efficiency, lower transaction costs should boost market efficiency by incorporating relevant information into prices more promptly through more active trading, if not more fully in the case where there is no change in the information content of trading. Actually, more informative trading could result from more active institutional trading, which is likely to be informational. 5 Together, market informational efficiency should increase owing to faster incorporation of the same, or possibly more, information into prices following the commission deregulation. As noted by Lo and MacKinlay (1988), while a precise test of the weak-from market efficiency necessitates an explicit economic model and, thus, is subject to the joint-hypothesis problem, the prior literature has conventionally couched the test within the framework of predictability of stock returns, or, more specifically, the random walk hypothesis. We follow this convention in our exploration of the impacts of the commission deregulation in Japan on the weak-form efficiency in the Japanese equity market, and document the following major findings. Using the runs test and the alternative autocorrelation test applied to both daily and weekly data, we find that equity return in Japan becomes significantly less non-random and, thus, less predictable following the commission deregulation. Meanwhile, no such change accrues to a control portfolio of Japanese stocks dually listed in the U.S. as American Depository Receipts (ADRs), which should be subject to the same marketwide economic forces in Japan except the deregulation under study. These findings indicate that, as anticipated, inefficiency (efficiency) decreases (increases) in the Japanese equity market following the commission deregulation, to the extent that return randomness suggests informational efficiency, as conventionally assumed in the prior literature. These results together with those from Liu (2007) imply that policies that help to reduce transaction costs should enhance market efficiency. The remainder of the paper proceeds as follows. The next section provides a brief background for the brokerage commission deregulation in Japan and formulates the hypothesis. Section 3 discusses the data and sample construction. Section 4 describes the empirical methods and reports the main efficiency impacts from daily data. Section 5 presents the findings from some robustness tests, including the efficiency effects from weekly data and the trading volume effects from daily data. The final section summarizes and concludes.

Brokerage commission rates were fully deregulated in Japan on October 1, 1999. As expected, commission rates fell across the board following the deregulation. With lower transaction costs, stock trading should be more active and, thus, more information should be impounded into stock prices at an enhanced rate, leading to higher informational efficiency in the equity market. We examine the efficiency effects of the commission deregulation in Japan in this study using two alternative methods, and document the following major findings. Applying two alternative empirical methods to both daily and weekly data, we find that, following the commission deregulation, equity return in Japan becomes significantly more random and, thus, less predicable. In the concurrent period, however, no such change occurs to a control portfolio of Japanese ADRs, whose returns should be subject to the same economics forces in Japan but the commission deregulation. Insofar as randomness proxies for market informational efficiency, as assumed conventionally in the literature, these findings suggest that the commission deregulation has enhanced the efficiency in the Japanese equity market. Moreover, these results suggest that any policy that helps to reduce transaction costs should improve market efficiency, lending another clinical data point to the debate on the efficiency effect of exogenous transaction costs. Given the theoretical importance of the efficiency issue in the finance literature, as demonstrated in the surveys by Fama, 1970 and Fama, 1991 and Barberis and Thaler (2001), and its implications for important public polices (such as brokerage commissions, margin requirements, and transaction taxes), further explorations of different types of transaction costs in different institutional settings should be worthwhile.