Portfolio management : adjusting optimal portfolios with momentum and sentiment factors

Abstract

This thesis aims to find whether a stock portfolio using Markowitz (1952) optimization method can achieve improved performance with addition of momentum and sentiment factors. The optimization method does by itself have its drawbacks, but the aim is to find whether the addition of momentum or sentiment factor can improve optimized portfolio’s performance while still keeping the simplified methods for optimization. The results suggest that momentum factor, which is based on time series momentum, provides consistent improvement on optimized portfolio with increased returns and increased Sharpe ratios over a long investment period. Sentiment factor was used as a contrarian indicator, and it provided more mixed results while having less significant effect on the portfolio performance than momentum factor. To measure sentiment a sentiment index was built out of indicators prone to capture investor sentiment.