Anchoring and loss aversion in momentum portfolios: a backtesting study of the S&P 500 (2000–2025)

Purpose This research aims to investigate the impact of two behavioral biases, loss aversion and anchoring, on momentum strategies by combining behavioral finance theory with an empirical analysis of monthly S&P 500 total returns data over 25 years. Its objectives are to estimate the relative importance of these biases on the momentum portfolio and to provide actionable investment recommendations. Design/methodology/approach This research employs a quantitative approach to examine the effects of anchoring and loss aversion on momentum strategies, using a historical universe of 1104 S&P 500 constituents from January 2000 to March 2025. The approach combines behavioral finance theory with strong backtesting. This research has constructed three portfolios: (1) a baseline momentum portfolio based on six-month historical returns; (2) an anchoring-adjusted portfolio that excludes stocks close to their 12-month high; and (3) a loss aversion momentum portfolio that filters high downside volatility. Findings All portfolios generated positive mean monthly returns ranging from 0.54% to 0.61% under turnover-proportional transaction costs but behavioral adjustments did not yield statistically significant alpha. The baseline strategy achieved the highest annualized return (4.15%) and Sharpe ratio (0.2346). The anchoring-adjusted portfolio recorded a mean monthly return of 0.5955% and an annualized return of 3.79%. In comparison, the loss-aversion-adjusted portfolio achieved a mean monthly return of 0.5363% and an annualized return of 2.53%. Performance differences remained statistically indistinguishable from the baseline across all comparisons. This suggests that behavioral filtering modifies the risk-return profile without consistently improving risk-adjusted returns in efficient large-cap indices. Originality/value This study contributes to the behavioral finance literature by integrating theory with empirical evidence and offers practical implications by demonstrating that mechanical adjustments to mitigate biases may increase complexity without consistently improving risk-adjusted abnormal returns. It highlights systematic momentum resilience while noting the limitations of static behavioral filters in efficient market environments.