A Study On The Trading Costs In China Stock Market

The stock markets in China have been growing up gradually for almost 20 years, and for the advantage of late-development, the market mechanisms and efficiency are being improved continuously. But, as an emerging market, China's stock markets inevitably have some unique characteristics, which will affect the efficiency and the further development of the market. Hence, it's necessary for us to study deep the microstructure in China stock market, understand the deep-seated reasons that have relations with the market fluctuation. And this will make active and far-reaching sense for the setting and healthiness of the mechanisms for price formation in China stock market, and improving the efficiency of the market.Based on the market microstructure theory, the dissertation combines the high and low frequent data, adopts the modern and practical econometrical and statistical analysis methods to empirically investigate the trading costs in China stock markets. According to the estimations of the trading costs in China stock market, the dissertation analyses the determinant character and the stochastic character about the trading costs, and provides some policy suggestions for different empirical results. The main research work and innovations are as follows:1. Confirming the measures of the trading costs applying to China stock market and their characters by comparing the various estimations of the trading costs.With the clear definition of the trading costs for China stock market, using the methods for estimating the trading costs in the international for reference, selecting the stock listed in ShangZheng 180 index as sample, the dissertation estimates the quoted spread, the efficient spread and the trading spread which can be used to measure the trading costs respectively, and simulates the trading costs with day data at the same time, then make a comparative and correlative analysis among these five measures. The empirical conclusions can be summarized as following:(1) The averaged quoted spread is about 0.15% on average, that is, the total trading cost in China stock market is about 0.15%. And the Roll estimation of the efficient spread that reflects the fixed cost is about 0.04% on average, which is far less than the level the stamp tax (0.1%), this means that the rate the stamp tax is too high in China stock market in some sense. The trading spread is close to the quoted spread very much, and its average is about 0.16%, but the latter is more volatile. From the general level, the trading spread and the quoted spread can both be used to measure the trading cost in China stock market. But since the trading spread is an average measure by estimating of the statistical model, the quoted spread is more likely to apply to analyze the trading cost.(2) By comparing the various estimations of the trading costs, we find there are other trading costs components in China stock market except for the order processing cost, and they will also increase over time. The trading cost expresses the common trends, and is also associated with the characteristics of the stocks.(3) By comparing the measures of the efficient cost for different frequency data, we find the Gibbs estimation of the trading costs based on the day data is not an applicable proxy for the trading cost in China stock market from both time and cross-section aspects.2. Analyzing the sources and characters of the trading costs in China stock market by decomposing the components of the bid-ask spread with panel data models.Based on the general approach of HS model which considered the effect of the trading is serial correlate, we decompose the bid-ask spread into two and three parts respectively by using intraday data for the stock from ShangZheng 50 index. The major results include:(1) If the trading is not correlate, the bid-ask spread is decomposed into order processing costs (0.2907) and the others (0.7093) including the inventory cost and the adverse selection cost. This implies that the adverse selection cost may be the main component in the bid-ask spread.(2) If the trading is correlated, the order processing costs become 0.1102, which is less than the corresponding level in two-part decomposition. The reason may be the correlation in trading. Surprisingly, the adverse costs is only 0.4514, and there is an unknown part. A rational explanation is that this part of spread reflects inventory costs. When the spread is decomposed into three-part, the order processing costs declines to 0.1116, the adverse selection costs is 0.4022 and the other component is 0.4862 on average, which is called order persistence costs in this study. That is, the effect on the price volatility in China stock market of the adverse selection costs and the order persistence costs is almost 90% of half bid-ask spread.(3) Moreover, we also find that stock price and its volatility, firm size and trading volume have some effect on the adverse selection costs and order persistence costs.3. Researching the determinants of the trading costs in China stock market based on the market microstructure theory.The dissertation studies the determinants of the intraday and inter-day trading costs by panel data model with fixed coefficient. We find the following results:(1) The trading costs have significant negative relation with stock price and positive relation with the risk of the stock. Although the trading volumn in money and market value have positive relations with the trading costs, they are weaker relatively. And the effects of the other variables on the trading costs are different with different frequency data.(2) To inter-day trading costs, the state fixed effect model explains best, and the time fixed effect is not significant for 30 minute and 60 minute data. Comparing the state fixed effect model and no fixed effect model, we can see that state fixed effect model can be almostly explained by the risk and market value.(3) To intraday trading costs, the determinants of the trading costs are slightly different. The two estimations of the trading costs both have significant state and time fixed effect, but the quoted spread, the estimation of total trading costs, can be explained better by the time fixed effect model, the efficient spread, the estimation of the fixed costs can be explained better by the state fixed effect model. The logarithmic market value has weaker effect on the quoted spread, and the trading volumn in money has also weaker effect on the efficient spread. Thus it can be seen that the quoted spread is highly correlated with variables which vary largely over time, but the case is contrary for the efficient spread.4. Analyzing the commonality in trading costs according to the'market model' for the trading cost.Based on the CRS model, the dissertation studies the commonality in trading costs by panel data model with varying coefficient in China stock market. The results show that the trading costs of the individual stock and portfolio both move together with market-and industry-wide trading costs, and the sources of this commonality is industrial trading costs. The trading costs have significant scale effect, that is, large firm spreads have greater response to the market-wide changes in spread.5. Researching the stochastic volatility of the trading costs in China stock market with ARMA and TARCH model.The dissertation describes the mean equation and variance equation which characterize the stochastic trading costs with ARMA and TARCH model, and analyses the intraday and inter-day volatilities of the trading costs in China stock market. The results show that the inter-day volatility is serial correlate and has the conditional heteroskedasticity, but only the trading costs portfolio has "leveraged effect" to the market shock. The intraday volatility not only is stochastic, but also has the deterministic trend, which seems to be "U" shaped. In addition, the intraday trading costs have not asymmetric response to the market shock.Finally, according to the empirical result, the dissertation gives some policy suggestions including improving the market transparence, controlling the private information-based trade and ameliorating the market trading system.For an emerging market like China stock market, domestic studies about trading cost mainly focused on the determinant characters as intraday pattern, components and determinants of the bid-ask spread, and rarely involved in the stochastic trading costs. The study in this dissertation can make up this deficiency, and deepens our understanding of the microstructure characteristics in China stock market, and enriches the market microstructure theory, and provides the theoretic supports and empirical evidences to further improve the trading mechanism in China stock market, make governing policy for government and improve the market efficiency. In the meantime, it will help the investors to understand fully the situation of the security market and make correct investment strategy, thereby protect the legal rights and interests of the investors. Hence, this dissertation has theoretic and practical meaning.