A0395
Title: Calm your portfolio: the importance of disciplining intelligent but fickle forecasts in portfolio optimization
Authors: Konark Saxena - UNSW Sydney (Australia) [presenting]
Abstract: How quickly should the portfolio choice be updated in response to new information? Traditional portfolio optimization methods often change weights frequently, resulting in high turnover, transaction costs, and unstable portfolio risk. To address this, a strategy is proposed that incorporates historical weight predictions and target volatility, penalizing changes in weights or portfolio risk to balance the benefits of new information against the costs of acting on noise. The effectiveness of these "calming constraints" is evaluated by comprehensively analyzing whether various forecasting methods can collectively enhance the out-of-sample performance of mean-variance efficient (MVE) portfolios. Choosing from the 500 largest stocks, it is found that MVE portfolios formed using intelligent forecasts do not outperform the passive strategy, even after considering transaction costs. However, calming constraints significantly improves their performance before and after costs. The investigation reveals that portfolios simultaneously targeting risk, managing transaction costs, correcting covariance matrix errors, and using simple linear Fama-MacBeth return forecasts achieve net Sharpe ratios greater than one, outperforming the passive portfolio by a significant margin. While no single idea alone surpasses the passive benchmark, portfolios that incorporate these multiple strategies demonstrate superior performance.
