The Importance Of Size Factor In Quantitative Models

The multi-factor model divides the factors into three categories,market factors,industry factors,and style factors.Quantitative fund investment managers generally adopt a method of controlling the influence of market factors and industry factors on stock returns,and then selecting one or more style factors to construct a quantitative model to obtain returns.From2017 to 2018,many of the quantitative factor managers used by the quantitative investment managers failed,and the net value plunged sharply.These include the most effective style factor in the A-share market in the past – the size factor.Then will the size factor and other style factors affect each other? Does the failure of the size factor affect the benefits of other style factors? Does the investment manager add weight to the stocks when trading other style factors to get higher returns? How to understand the relationship between size factor and other style factors,and explore the effect of the connection between the two on actual investment?Firstly,this paper selects nine style factors for research by analyzing the Barra model,and then uses Fama-Macbeth regression method to analyze the influence of each style factor on stock returns,and then explores the size by calculating the correlation of income and double sorting.The interaction between factors and other style factors.Finally,through factor strategy backtesting,the performance of each style factor in the case of scale factor neutral(market neutral)and scale factor non-neutral is explored.The results show that under the market selection,the book market value ratio,price-earnings ratio,turnover rate,size,momentum,volatility six factors by factor After sorting,the ten groups of incomes are sufficiently monotonous,and the stocks have a large multi-end gain.However,the cumulative income of the size factor is greatly reduced from2017 to 2018.If only the size factor is used as the quantitative model,that is,the stock with the lowest market value per industry per month is bought,and the size factor from 2017 to 2018.Compared with the CSI500 Index,the annualized excess return is-3.4%,while the annualized excess return of the size factor from 2008 to 2016 is 28%.At the same time,all other style factor excess returns are reduced by the factor size factor gains.After removing the size factor effect of all factors,only the book market value ratio,the price-earnings ratio reciprocal,the turnover rate factor and the volatility factor income are still positive,but the four returns are also smaller than the factors of 2008-2016.Therefore,when using the style factor to do the quantitative model,we must consider the impact of the size factor on the return.In the end,from the four factors that are still positive in 2008 to 2016,the three factors of the price-earnings ratio reciprocal,book-to-market ratio and turnover rate are selected,and the three-factor model is constructed by using the three equal rights.Since the annualized gains of the volatility factor are too small,close to 1%,they are excluded.The three-factor model takes the same backtesting cycle and backtesting samples with a single style factor,and controls the influence of the size factor on the three-factor model through a single equity re-optimization.The results show that the three-factor model ten-row sorting returns are very monotonous,and the stocks do multi-end excess returns are high enough.Compared with other factors,the annualized volatility of the income is small,and the annualized Sharp is high.The three-factor model can avoid a sharp retreat from 2017 to 2018,and the cumulative income will hit a new high.