Extended information ratio proposal for portfolio optimization based on market dynamics

It is well known that the information ratio is one of the most famous indices to evaluate the performances of portfolios. However, it is difficult to design a portfolio through the information ratio to keep the performance in its future period. In order to keep the portfolio´s performance that does not depend on the market dynamics in the future period, we propose the extended information ratios for the portfolio design that uses the beta and the beta prime in this paper. We apply a simulated annealing (SA) to optimize the portfolios employing the extended information ratios and then evaluate the efficiency of the portfolios. In the numerical experiments, we show that our portfolios keep the good performances when the trend of the market of the future period becomes different from that of the past period. Even though the portfolios employing the information ratio as an evaluation value display good performances, they have some computational flaws. We found that the evaluation values do not converge for some data periods. We discuss this problem in this paper as well.