Mutual fund performance: a new way to pick a persistent winner
A new academic paper introduces the MMT model, which improves upon standard methods by correcting for benchmark biases and incorporating peer performance, proving more effective at identifying persistent winner, writes Cesario Mateus, Irina B. Mateus and Natasa Todorovic
In the ever-evolving world of finance, selecting mutual funds that consistently outperform the market is a challenge that both investors and fund managers grapple with. Mutual funds are a cornerstone of modern investment portfolios, offering diversification and professional management.
However, the ability of funds to consistently deliver above-average returns, referred to as “persistence” in performance, has long been debated. Extensive research on actively managed equity funds has traditionally relied on the Fama-French and Carhart factor models to assess performance, generally finding that most funds fail to outperform their benchmarks after accounting for fees, with a few exclusions, and has provided scarce evidence of performance persistence. However, recent studies have criticised the Fama-French factor models for their arbitrary factor construction and, as a result, the disproportionate weighting of value and small-cap stocks, introducing bias. This results in non-zero alphas for passive benchmarks when evaluated against the Carhart four-factor model, undermining alpha as a reliable indicator of a manager’s skill. For example, research has shown a Carhart alpha of 0.82 per cent for the S&P 500 from 1980 to 2005, a high of 7.5 per cent for the Russell 2500 Value index from 1995 to 2004, and consistently negative alphas for the FTSE 100 from 1992 to 2013, illustrating how factor design distorts performance evaluations.
Active peer benchmarks
Emerging research also underscores the value of incorporating active peer benchmarks in performance analysis. Discrepancies in self-selected benchmarks, the limited availability of style-specific indices (e.g., FTSE or MSCI classifications), and the lack of focus on relative peer performance can misrepresent a fund’s success. Investors are particularly interested in how funds rank among peers, as a fund may surpass its benchmark but lag in its peer group, or it may underperform during market slumps yet still lead its peers.
A recent academic paper, Mutual Fund Performance: The Model for Selecting Persistent Winners by Mateus, C., Mateus, I.B., and Todorovic, N., published in The European Journal of Finance (2025, Vol. 31, Issue 5, pp. 647–669) suggests a new unified framework that eliminates the bias of passive benchmark alphas (as measures of excess performance) and enhances the Carhart model further by incorporating relative peer performance in the analysis. The new MMT model introduces a new, practical way to evaluate mutual funds that better aligns with how investment professionals actually assess performance. It offers fresh insights into identifying mutual funds with persistent superior performance, as it allows for identifying genuine, true winners —the funds with the best stock-picking ability in comparison to their peers, which drives future overperformance in the subsequent 36 months.
This research is of paramount interest to investors and practitioners because it provides a comprehensive, more reliable and easy-to-implement approach to identify mutual funds that genuinely add value and are likely to continue outperforming their peers and the stated benchmark. The findings highlight a clear economic impact: the MMT model consistently detects more positive and fewer negative alphas, while the Carhart model, during periods of both index outperformance and underperformance, identifies more fund periods with negative alphas and tends to classify more funds as underperformers. Winners selected by the MMT model consistently earned higher returns in excess of the benchmark and peer-group performance (e.g., 56.05 basis points per annum for Large Cap Value funds and 77.65 basis points per annum for Large Cap Growth funds) compared to winners identified by the standard Carhart model.
MMT model demonstrates its ability to identify “unique winners” (funds that are identified as outperformers by the new model but not by Carhart). These unique winners exhibit even stronger persistence and higher excess returns (above the benchmark and peer-group) in the subsequent 24 and 36 months (e.g., 210.87 basis points per annum in difference for Large Cap Value funds and 117.99 basis points per annum for Large Cap Growth funds), while winners identified only by the Carhart model often do not show persistent performance. The model is also effective in identifying unique and more extreme future losers. MMT findings are not limited to specific types of funds; the model’s effectiveness in selecting persistent winners and losers holds true across different style categories, including Large Cap Value, Large Cap Growth, Small Cap Value, and Small Cap Growth funds. The results are also robust to different holding periods and various statistical specifications for defining winners and losers, including analyses based on only statistically significant alphas, 4×4 contingency tables, and decile/quintile portfolio sorts.
In conclusion, the paper “Mutual Fund Performance: The Model for Selecting Persistent Winners” offers a sophisticated, yet practical, evolution in mutual fund performance evaluation. For finance professionals, this means a better chance of identifying funds that truly outperform, rather than those whose apparent success is merely a reflection of market or peer-group trends. It equips investors with a more reliable framework to make smarter investment choices, especially valuable in challenging market conditions where true skill becomes even more crucial.
Cesario Mateus is full professor of finance at Aalborg University, Denmark
Irina B. Mateus is associate professor of finance, Aalborg University, Denmark
Natasa Todorovic is associate professor of finance, Bayes Business School, UK