Based On The Macroscopic Variables Of Industry Configuration

Sector is a group of corporations which have some common attributes. No matter in macroscopic or microscopic aspects, it is an important object for research. Sector rotation is an indispensable process in the from up to bottom asset allocation. It has significant meaning to final portfolio formation.According to the related research and experience abroad, selected sectors greatly relate to the yield of portfolio and a large proportion of stock yield can be explained by the sector selected. Investors really can make significant excess return by sector rotation according to the judgment of future macroeconomics. So there are strong links between sector rotation and macroeconomics, which is a key concern of scholars and market investors. On the one hand, how to grasp different sector growth in economic cycle to do sector rotation is a great concern for investors; on the other hand, how to use all kinds of data available in the market to predict future sector trend and get reasonable suggestion for investing by means of quantative methods is also a topic for scholars and market investors.This thesis tries to do sector rotation by introducing representative macroeconomic variables to the sector index yield regression and Logistic regress for the probability of rise. The final sector rotation method is based on the combined consideration of predicted index return and probability of rise. Furthermore, the thesis also uses the traditional mean-variance method to do sector rotation and compare the advantage and disadvantage of two methods.We find that the result using the method based on macroeconomics is stable, the sectors we long and the sectors we short accord with reality. While the mean variance method is not stable if the market does not have the short selling constraint.