Momentum Crash And Risk Management

Momentum is one of the important anomalies in traditional finance.Since foreign scholars made system test in 1990s and found that there was monthly momentum effect in the stock market,a growing number of scholars began to study the momentum effect.Firstly,from the perspective of time series and cross section,some scholars tested the differences between different momentum effects,drawing a conclusion that both time series momentum effect and cross section momentum effect existed.But the mechanism for forming momentum effects were different.Next,some scholars tested the momentum effects from the perspective of different asset classes.It was found by empirical test that various degree of momentum effects existed in the market of most assets such as stocks,bounds,futures,commodities and foreign exchanges.Thus,momentum effects had universality.Finally,some scholars tested the existence of momentum effects from different market angles.It was found by empirical test that momentum effects existed in not only the markets of developed countries,but also the emerging markets of developing countries.It illustrated that momentum portfolio had great value for investors in most countries around the world.In recent years,financial risk accumulates,and asset prices are of wide fluctuation especially the stock markets,because of the global economic recession.Stock markets boom and bust frequently.Because in developed countries,bull market lasts long and bear market is short,as well as the stock market is volatile,some portfolios will be withdrawn substantially when the market continues to fall and bounces back suddenly.Momentum portfolio is the most obvious.According to the empirical study of foreign scholars,American stock market momentum returns had severe pullbacks from July to August in 1932,and from March to May in 2009.Moreover,substantial retracements of momentum returns were observed in other countries' markets and other asset classes.Such phenomenon is called momentum crash.Developed markets continue to fall and rebound suddenly,resulting in momentum crash.How to manage the crash risk is practically valuable for the use of momentum portfolio.Thus,many foreign scholars study the mechanism of momentum crash,drawing the conclusion that momentum crash results from that past losers have larger benefits than past winners,and the losers have larger beta value than the winners when the market continues to fall and rebounds suddenly.Foreign scholars build variables that describe market conditions in order to predict the occurrence of momentum crash,and further build momentum portfolio of dynamic adjustment in order to improve the primary momentum portfolio.It's controversial that whether momentum effect exists in Chinese stock market.And studies of the risk and management of momentum crash are very few.There were some scholars once tested the momentum effect in China's A-stock market.Most conclusions were that there was weekly momentum effect and no monthly momentum effect.There was reversal within a month.Therefore,this thesis will firstly verify the existence of weekly momentum effect,then test the existence of momentum crash,and finally improve the existing momentum portfolio.This thesis selects all the weekly rates of return data of A-stock in China from Jan.2007 to Jan.2017,to test the short-term weekly momentum effect and the existence of momentum crash.Moreover,with the method of factor construction,this thesis forecasts the momentum crash from perspectives of market factors,market volatility factors and momentum return volatility factors.The thesis further tests the predictive ability of risk factors by distinguishing between the normal period and the collapse period of the momentum portfolio and selects the risk factor with strong predictive ability to construct dynamic stopping time momentum portfolio based on target volatility.The study found that there was weekly momentum effect in Chinese stock market from Jan.2007 to Jan.2017.Momentum crashes were observed in Jul.2015 and Mar.2016.There was negative correlation between market volatility and the returns of momentum portfolio,which could well predict momentum crashes,whereas other factors in Chinese stock market couldn't predict momentum crashes.Dynamic stopping time momentum portfolio based on target volatility can avoid the momentum crashes and substantially improve primary momentum portfolio's returns and Sharp Ratio.The finding not only proves the existence of momentum and momentum crashes,but also shows the function of dynamic risk management,which is instructive for building investment portfolio and risk management.This thesis has three innovations:the first is using weekly return data of China's A-stock from Jan.2007 to Jan.2017 to reexamine the existence of momentum effect after the reform on share structure,and to check the phenomenon of momentum crash based on the weekly momentum effect.The second is systematically analyzing the risk exposure of momentum portfolio and the cause of momentum crash.In order to predict the occurrence of momentum crash,this thesis builds two new risk factors that are the volatility as well as realized volatility of market returns based on the domestic and foreign literature.Thus,there are six risk factors in all including the other four risk factors built by foreign literature,to forecast the momentum crash.The risk factor that can best predict the momentum crash is selected to make subsequent adjustment for momentum portfolio.The third is constructing the dynamic stopping time momentum portfolio based on the target volatility.Whereas market volatility and momentum return volatility can forecast the momentum crash partly,the dynamic stopping time momentum portfolio can be constructed based on the target volatility considering the stopping time portfolio's advantage of saving cost,when the volatilities of market and momentum return increase but the benefits decrease.