Using modified run analysis to predict active investments outperformance: a case of US equity assets

What the paper says

Purpose This paper deals with the forecasting application of runs analysis based methods in the field of active investments’ alpha. This study aims to use different models exploiting possible momentum and persistence in the active investments’ returns, specifically in relation to the prominent field of their (possible) outperformance. Design/methodology/approach Gathering a large set of managed assets from the US large-cap equity segment, this study examined a variety of simple and more complicated models, from unpretentious probability, to the full-scale runs analysis approach. These models were evaluated on monthly and yearly data, with simulation of their performance on 30 years period also carried out. The study then tested the results for significant differences, as well as separately evaluated them for individual types of investments and equity strategies (styles). Findings Simpler forecasting methods provided generally more efficient outcomes. Choosing simply investments with the highest past alpha (simple strategy), highest ratio of positive alpha in the past (opportunistic strategy) or in the given stadium of the market cycle (market strategy) would result in a low relative error, while having limited data and computational requirements. The runs analysis itself performed less favorably, being on one hand quite successful in the alpha direction forecast, but subpar when it comes to average return. Originality/value The results also point to short-term persistence in surveyed investments performance, indicating a yearly interval as optimum for portfolio reallocation. Overall, this outcome implies the rehabilitation of technical analysis as a principal approach of estimating future performance, in the field of managed equity investments.