Empirical Study Of Stock Forecast Based On GARCH Models

The stock market is an important part of China’s capital market,and individual stocks are the core trading object of the stock market.On the one hand,it has direct practical guiding significance for investors and non investors,on the other hand,it can provide theoretical basis and technical guidance for market supervision.GARCH model family is a classical model to study stock forecasting.Compared with other mainstream forecasting theories and models,GARCH model family can not only predict the stock price or return,but also predict the fluctuation of stock price or return.The domestic research on stock market based on GARCH model family mainly focuses on stock index,and there is less specific research on individual stocks;At the same time,the research mainly focuses on the effect verification and mechanism analysis,and there is less research on the prediction,especially the prediction of fluctuations outside the sample.In this context,based on the daily data of China’s A-share market from June 25,2019 to February 11,2022,this paper calculates the correlation between Shanghai stock index and A-share stock,classifies it according to the degree of correlation,comprehensively considers various conditions,especially whether the stock series supports the establishment of constant term mean equation,selects representative samples,and uses empirical analysis and comparison methods to study the actual effect of standard GARCH model and asymmetric TGARCH model in GARCH model family on the prediction of stock fluctuation outside the sample.Through the analysis and comparison of the prediction results,the following relevant conclusions can be drawn.First,GARCH model family has a good prediction effect on all kinds of stock fluctuations,and it is still effective in the case of sudden and significant changes in the macro environment.Second,for low-order GARCH models,the shorter the prediction cycle,the higher the prediction accuracy.Third,the prediction effect of TGARCH model is not necessarily better than GARCH model,but depends on the leverage effect of the analysis object itself.Fourth,as of February 2022,there is still obvious arch effect in China’s stock market.According to the above conclusions,this paper gives relevant suggestions on information disclosure,market supervision and investor decision-making.