Fractional Momentum Strategy I

Introduction

The Fractional Momentum Strategy leverages empirical finance research and advanced computational methods to optimize investment portfolios. It applies a fractional-difference filter to historical price data, preserving long-term memory while mitigating large drawdowns common in traditional momentum and reversal strategies. By incorporating a comprehensive range of historical data, the strategy enhances risk-adjusted returns and remains resilient against transaction costs and market frictions.

Empirical evaluations demonstrate its consistent performance across various asset classes and international markets, with portfolios rebalanced on a weekly basis. The evaluation process accounts for dividends and stock splits.

Performance Teaser

The following plot illustrates the evolution of daily prices of the 30 DJIA US stocks. These 30 stocks represent a subset of the 1,000 to 2,000 stocks analyzed. Viewing them this way does not reveal how to construct an efficient portfolio.

The next plot below shows performance over time, out-of-sample, when using our Fractional momentum methodology. It is based on a selection methodology from the 1000 to 2000 stocks in our universe. When comparing our method to other strategies such as Markowitz (with GARCH), but notably, as shown in the graphic below, the more advanced strategies of momentum, short-term reversal and momentum and short-term reversal, ours is superior with respect to all measures of interest, such as total wealth accumulation and all common risk measures.

Key Performance Results

The following bullet points indicate relevant information on our methodology and choice of data used to generate the numbers in the table above:

• The table presents the economic performance, net of a 10-basis point transaction cost, of our 'Fractional Momentum' strategy versus benchmark strategies.
• It details our dollar-neutral strategy that invests in the top 10% and shorts the bottom 10% of stocks ranked by fractional momentum.
• The investment universe includes a dynamic range of 1,000 to 2,000 U.S. stocks, following Fama and French (1993), avoiding selection bias by accounting for both listings and delistings.
• To minimize the impact of microcap stocks, those below the 20th percentile of NYSE market equity are excluded, with selection weighted by market capitalization.
• Our Momentum (MOM) strategy is based on past twelve-month performance, excluding the latest month, while the Short-Term Reversal (STR) strategy focuses on the past month's negative performers.
• The STRMOM strategy employs a double-sort method, sorting stocks by both MOM and STR criteria for a strategic long/short portfolio based on short-term underperformance and medium-term performance.
• The fractional momentum approach combines momentum and reversal signals through a fractional-difference filter, significantly improving risk-adjusted returns and reducing drawdowns compared to standard strategies.

Showcase of our strategy

Explore this strategy

            Chitsiripanich, Soros and Paolella, Marc S. and Polak, Pawel and Walker, Patrick S., Momentum Without Crashes (November 14, 2022). Swiss Finance Institute Research Paper No. 22-87, Available at SSRN: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4280465
            

            @article{momentum-without-crashes,
                author = {Soros Chitsiripanich and Marc S. Paolella and Pawel Polak and Patrick S. Walker},
                title = {Momentum Without Crashes},
                journal = {Swiss Finance Institute Research Paper No. 22-87},
                year = {2022},
                note = {https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4280465}
            }