A Multiple Factor Model For Chinese A-Stocks

Multi-factor model describes the relationship between stocks return and risk factors in simple linear model, so the calculation of covariance about stocks return could be simplified by calculating risk factor returns covariance and specific returns covariance. We consider that there are n kinds of stocks in a market, we can write multi-factor model as follows:R_i =α_i +β_i1·F₁ +…+β_ik·F_k+ε_i i=1,…, n (1)Where R_i is return to security i ,α_i is a constant, which means the common effect of all risk factors,β_ij is sensitivity/exposure of security i to factor j, F₁,…, F_k are the kfactors,ε_i is specific return to security i. The covariance of portfolio could be calculated as follows:σ_p² = x_p^T·F·x_p + h_p^T·Δ·h_p = h_p^T·Σ·h_pthis paper present an empirical study focusing on the estimation of a fundamental multi-factor model for Chinese stock market. Follow the approach of BARRA E3 model, we consider some macro-economic factors, some fundamental factors and some technique factors, and select 9 factors that might have significant influence on stock returns and risks, then we have selected 8 factors that have significant influence. We build up multi-factor regression series, the explanatory powers in most cases illustrate that the result of regression are successful. So the stock return can be expressed with common risk factors and specific risk factors in a linear model, we can simplify the covariance of portfolio by using common factor returns and specific factor returns.