The Calendar Effects Based On EGARCH-M Model

The Calendar effects in the stock market prove that the market is non-effective, so there will be a great risk by use the theory and model under the assumptions of the efficient market. And with the extremely rapid development of China’s stock market, investors are also growing, so it is very significance for depth studying.calendar effects. But the research on calendar effects main emphasis on the mean and variance, this article will study calendar effects from a more comprehensive perspective, while taking advantage of the calendar effect to guide investors for the choice of market timing. Research methods main include descriptive statistics analysis and econometric model analysis. Description statistical analysis includes:mean, variance, skewness, and peak-degree. Econometric model analyses mainly utilize the EGARCH-M model under t distribution; the analysis mainly includes the fluctuations asymmetry, the impact of expected risk, the impact of lag revenue, the impact of lag phase fluctuations and leverage effect.This study ideas and content as follows:This study ideas as follows:The topic backgroundâ†'Literature reviewâ†'The concept of the theory and model elaboratedâ†'Data processing and model selectionâ†'Use descriptive statistics and econometric model empirical to analysis and summaryâ†'Explain the reason of calendar effectsâ†'Use the conclusion guiding investors for time choices and proposed policy recommendations to management market. The article is divided into seven chapters, as follows:The first chapter has a description in the background of topics. China’s stock market develops rapidly while investors lack professional knowledge, investors’ irrational transaction make market become more non-effective:calendar effect. The research about calendar effects were mainly concentrated in the mean and variance, but leverage effects, expected risk, skewness, kurtosis all will affect the distribution of investment earnings. So if they are not taking account into the relevant theory and model, it is bound to generate a great of uncertainty risk. At the same time, a lot of research theories assume that the market was efficiency, and some modes assumed that the data followed the normal distribution, but the actual proof is opposite, so it is important to study the calendar effect for risk control and update the theory and model.The second chapter content is literature review. It will be elaborate the research results of calendar effects in domestic and foreign from the week effect, the January effect and the seasonal effect. The Literature review will expound calendar effects evolving process.The third chapter is related to the concept of the theory and model. It mainly elaborated the concept of calendar effects, skewness and kurtosis, Overreaction, loss aversion, herding behavior in behavioral finance, asymmetric wave theory and graphics, the origin and development of GARCH model. Theoretical concepts and models of this chapter will mainly services for empirical analysis and analysis cause about calendar effects in the next chapter.The fourth chapter main focus on the data processing and model selection, the article will select two representative indices of Shanghai market and Shenzhen market:the Shanghai Composite Index and Shenzhen Component Index. A daily limit of normality and volatility will be taking into account the sample selection. Data processing includes sequence normality test, the stationary test and ARCH Tests. Normality test use JB and KS test, the stationary test use unit root test, the ARCH test check autocorrelation on residual sum of squares of sequence. In the model selection, TARCH-M and EGARCH-M model fit two indices respectively in the GED distribution and t distribution; it is found that EGARCH-M model has the best result on the t-distribution, so select the EGARCH-M model for the latter part of the empirical analysis.The fifth chapter is the empirical analysis section of the article. This chapter focuses on the use of descriptive statistics and measurement model for calendar effects, including analysis of the week effect, the January effect and seasonal effects, research from these aspects:the mean, fluctuation variance, skewness, kurtosis, the expected risk, lag earnings, asymmetric volatility, leverage effect and volatility clustering. Then there will be comprehensive description of the characteristics of calendar effects and the last is the conclusion.The sixth chapter will explain the reason about the calendar effect and analysis the impact on investors. Overreaction, herding, loss aversion, the tightness of the capital flow, the effect of the Spring Festival, the financial statements announced investment behavior all are the reason for calendar effects. Calendar effects exist, prove that the market is non-effective, and therefore it inevitably produces a very risky use market theories and models that assumptive the market is effective to guide investment.The seven chapter is proposed timing strategy and policy recommendations, the main idea of this chapter is to use the calendar effect conclusions from empirical analysis to guide practice guide investors to identify the timing into the market, and advises managers to formulate laws and regulations management market; make the market more sound and effective。For calendar effects, the main conclusions of this study are as follows:Week effects:(a) the mean revenue:the highest mean revenue is on Wednesday, the lowest is on Thursday (b) fluctuations variance:Monday has the maximum fluctuation variance, Thursday has minimum fluctuations variance (c) skewness:Tuesday has the most serious left side of skewness, Wednesday has the most serious right side of skewness (d) kurtosis:Monday is weakest on the spike thick tail, Tuesday is most serious on leptokurtic (e) lag revenue:Mon revenue is significantly impacted by lag revenue with a positive impact, Friday revenue is significantly impacted by lag revenue with a negative impact (f):asymmetric fluctuations and lag-period fluctuations did not show significant effect of weeks.Month effect:(a) the mean revenue:there are higher gains in February, March and April, and it is highest in February, the lowest mean revenue on August (b) fluctuation variance:February has the most volatile and December fluctuations is minimum (c) skewness:May has the most left-biased, and September skewed to the right (d) the kurtosis:February has strongest fat tail of the peak and December is the weakest (e) expected risks:March, May and July gains are substantially impacted by the expected risk (f) lag revenue:March and May were significant impacted with by the lag (g) there is obvious presence of the leverage effect on March and August, and December obviously does not exist leverage effect (h) lag fluctuations:January, April, June and August all are not significant impacted by lag fluctuations, but other months all are affected significantly.The seasonal effect:(a) the mean revenue:the first quarter has the highest gains, and third-quarter has lowest revenue (b) fluctuations variance:a minimum of fluctuations in the second quarter, the largest fluctuations in the fourth quarter,(c) skewed:the fourth quarter skewed to the right, and the other quarter all are left partial (d) kurtosis:the first quarter has the most fat tail, the fourth quarter is the weakest (e) the expected risk:the first quarter, second quarter, third quarter earnings all were significantly impact by the expected risk (f) lag earnings:lag revenue had no significant effect any quarter,(g) it is obvious that there is a leverage effect on the third quarter and the fourth quarter (h) lag fluctuations:All quarter of the current fluctuations are significantly affected by the lag phase fluctuations.The main innovation of this paper is as follows:a. Select the certificate Composite Index and Shenzhen composition index as a sample, consider the daily limit restrictions for volatility, and remove the data before the daily limit, the data conducted subdivision by week, month and season for the sub-sample study. Combine the description of statistics and EGARCH-M model fitting analysis on the week effect, January effect and seasonal effects.b. More comprehensive view analysis on the calendar effect. Use the non-normality EGARCH-M model fit the data of calendar effects, and put the expected risk variables in the model so as to forecast the relationship between the risks and revenue under conditional variance, make more convenient to calculate the expected return, and also allow investors to manage risk effectively.c. The paper argues the causes of calendar effects, mainly due to the following aspects:(a) investors Overreact to information, Herding, and Loss Aversion.(b) Spring Festival effect increase good news on retail market and financial sectors.(c) The financial statements disclosed the bad and good news to investors and reflect is inconsistent.(d) The policy announcement is centralized and regularity on the time,(e) the investors’ funds flow tightness (f) investor’s investment habits.d. This paper use calendar effects guiding investors for the market time choice. Proposed week effect is suit for short-term investors, January effect is suit for the medium-term investors and the seasonal effect for long-term investors, and proposed government intervention in management, suggested that the Government avoid concentrated and regular publication of information, to regulate market information disclosure system, develop expertise and risk awareness of investors.