Event Study Based On The Empirical Likelihood Method

As we know, China’s financial markets have great achievements in various aspects after decades of development. The function of the financial market is becoming more and more perfect; the coverage is more and more wider in the financial economy. Under this background, how to evaluate the influence on enterprises of economic events is very important. As a kind of empirical analysis method, the event study method has been widely used in the study of financial market. But we noticed, selection of hypothesis test method, for using of event study method to study economic events or financial events, is very important.The method is often used to test the hypothesis test in event study such as parametric test, the sign test and rank test, but these methods have their respective constraints. For example, parametric test has actually explicit assumption of the distribution of abnormal returns, testing the feasibility by the limits of distribution to a large extent; The sign test requirements of income distribution is symmetrical, but in the use of daily return data in case it is difficult to ensure the distribution of cumulative abnormal returns are symmetrical; But the rank test drawback is the estimation variance, when the variance estimation not punctual, rank test effect will be greatly reduced.In contrast, condition data from a known distribution family is not a necessary assumption of the empirical likelihood method, which does not require distributional assumptions on the data strongly; When data from the non-normal or variance estimation is not stable, the empirical likelihood method than other normal approximation method is usually to be precise; The empirical likelihood method has robustness results in many cases, is not easily affected by outliers in the sample; Secondly, the empirical likelihood method does not need to build the shaft statistics, avoiding the estimation variance. In this case, we use the empirical likelihood method which with many advantages in the event study, enough to make up for the test. The empirical part, we selected two events in2012the stock market of our country for study. One is for illegal operation by processing the securities regulatory departments of the Listing Corporation, another kind is the implementation of pure cash dividend policy of dividend distribution of listing Corporation. For these two classes of events we used two models (constant mean model and the market model) to calculate abnormal returns during the event window. Then using parametric and non-parametric test method to test the abnormal returns. Parameter test part gives two kinds of test statistics (J1and J2) results. Nonparametric test part gives the sign test and empirical likelihood test results. Finally, a comparative analysis of parameter test and the nonparametric test.The conclusion shows:we found in the handling of violation events listing Corporation, different nonparametric tests and parameters. It is difficult to determine the future volatility of stock price announcement and the handling of violation notice related. For the listing Corporation to take pure cash dividend distribution policy events, nonparametric test results and test results have little difference. We have reason to believe that the listing Corporation pure dividend policies on the volatility of stock prices have a greater impact.