A1754
Title: Quantile maximizer in action
Authors: Martin Hronec - UTIA AV CR vvi (Czech Republic) [presenting]
Jozef Barunik - UTIA AV CR vvi (Czech Republic)
Abstract: Out-of-sample performance of the portfolio is studied, selecting investors with $\tau$-quantile preferences. Investor's risk aversion is captured by $\tau$, where more risk-averse investor maximizes lower $\tau$-quantile. The quantile maximization is reformulated as a mixed integer programming problem leading to the computational complexity which allows finding the optimal portfolio weights under a reasonably large number of assets as well as reasonably long time periods. Using a number of empirical and simulated datasets, differences in optimal portfolios are documented across different levels of risk aversion. Optimal quantile portfolios are compared with benchmark portfolios from the out-of-sample perspective. It is documented that maximizing low $\tau$ -quantiles leads to more concentrated portfolios than global minimum variance portfolios achieving higher out-of-sample Sharpe ratios. Mean returns are typically larger for the low $\tau$ -quantile maximization portfolios compared to the global minimum variance portfolios.
