Downside deviation as a measure of identifying underperforming assets

Abstract: Quantitative approaches to achieving excess return are becoming increasingly popular as computational capabilities increase. Today, the main issue at hand is the development of accurate and reliable models for predicting the return of individual instruments. Index tracking mutual funds are however seldom able to select specific instruments to invest in, but rather select individual instruments in an index not to invest in. This study seeks to investigate whether downside deviation is a suitable risk measure for identifying instruments that underperform among a collection of instruments, and thereby the possibility of using downside deviation in a stop loss algorithm. That is, an algorithm that analyzes a portfolio and indicates which instruments might underperform compared to the benchmark, helping the portfolio manager to avoid realizing negative returns. In order to simulate the suggested model, daily pricing data between 2009-12-30 and 2018-12-28 for the large cap list on OMX Stockholm per 2009-12-30 is used. A model generating excess return was indeed found using downside deviation as the performance measure. The model yielded an average excess return of 3.27% over 1000 simulations with p < .00001 for the unregulated case. When regulations such as UCITS IV was implemented, the statistically significant excess return turn to statistically significant deficit return.