Research On Chinese Stock Market Anomalies And Quantitative Trading Strategies

After more than 30 years of development and reform,the depth and breadth of China's capital market has continued to expand,and the market vitality and potential have been gradually released.On the one hand,the reform of the capital market system has been vigorously promoted through strong measures such as the reform of the split share structure,the introduction of margin trading,and the development of a registration system,and the market transaction mechanism has been continuously improved.On the other hand,the scale of funds has increased significantly,and institutional funds such as insurance funds,social security funds and corporate annuities have entered the capital market,and the structure of market investors has improved.So,how is the overall efficiency of the current Chinese stock market? Does the overall market anomaly change over time? What factors are driving this change? Can investors benefit from trading strategy based on anomalies? These issues prompted us to write this article.This article takes the overall anomalies of the Chinese stock market as a research object,analyzes its cross-sectional effectiveness in the Chinese market and its changes over time,and explores the driving factors of evolution of anomalies from the perspective of the margin trading,institutional investor participation and analyst investment advice.We consider using a variety of anomalies that are significantly present in the Chinese stock market to predict returns,construct quantitative trading strategies,and provide rational portfolio recommendations.Specifically,the research conclusions in this paper are:First,through research covering 6 categories anomaly factors,it is found that 30 of these single factor combinations can provide significant non-zero portfolio returns,thus the overall market efficiency is not high.From the perspective of factor categories,the results of this paper verify the existence of momentum anomalies in the Chinese market,and at the same time find that corporate investment,intangible assets and transaction friction category factors are the main driving forces for portfolios returns.Second,only part of the anomaly factors can continue to provide excess returns,indicating that market efficiency continues to improve.Among them,the factors of the momentum effect and transaction friction categories are basically declining,indicating that the market continuously digests the information of these two categories of factors over time.Third,margin trading,the participation of institutional investors and the investment advice of analysts have played a driving role in the evolution of China's market anomalies.Specifically,margin trading traders obtain profits by trading stocks that are short in anomalies,while financing traders cannot.Institutional investors also make use of the anomalies,thus playing a role in correcting market mispricing.When the analyst released the research report,his recommendations were in the same direction as the forecast direction of the four anomaly factors of value and growth,corporate investment,profitability,and intangible assets,conveying information that can correct mispricing.Fourth,a portfolio constructed based on effective anomaly factors can provide excess returns,and the linear machine learning algorithm provides a higher predicted return rate than traditional OLS regression models.Through the use of traditional and linear machine learning methods to predict the rate of return,we found that linear machine learning algorithms can effectively integrate factor information and provide a higher rate of return than traditional models.We also find that strategies that using linear machine learning models have better performance in terms of overall risk-adjusted returns and returns that consider the maximum drawdown risk,and their high-order moment risk is at a lower level.Fifth,when considering reasonable and appropriate transaction costs,or changing the length of the model 's rolling forecast window period,a portfolio constructed based on effective anomaly factors can provide excess returns,and the performance of linear machine learning algorithms is still superior to traditional OLS methods.The conclusions of this article have certain practical significance.First,the designer of the market mechanism should consider appropriate relaxation of shortselling restrictions.Second,investors should pay attention to the evolution of capital markets and adjust investment strategies accordingly.Third,analysts should pay full attention to both market information and fundamental information.