A fractal approach to long-only portfolio optimization is proposed. The quantitative system is based on naive risk parity approach. The core of the optimization scheme is a fractal distribution of returns, applied to estimation of the volatility law. Out-of-sample performance data has been represented in ten period of observation with half year and one year horizons. Implementation of fractal estimator of volatility improves all performance metrics of portfolio in comparison to the standard estimator of volatility. The efficiency of fractal estimator plays a significant protective role for the periods of market abnormal volatility and drawdowns, which allows beating the market in the long term perspective. The provided results may be useful for a wide range of quantitative investors, including hedge funds, rob-advisors and retail investors.
Category: Economics and Finance