20 years ago, Morningstar changed the methodology for their mutual-fund ratings, and since then strategies that employ momentum-stock investing have struggled to find success, contends an article in The Wall Street Journal. Momentum-stock investing uses the idea that winning stocks will continue to win and losing stocks will continue to lose—based on stocks’ general trajectory.

For more than 60 years, portfolios that were rebalanced to buy into 10% of the most winning stocks while shorting 10% of the stocks with the lowest returns at the same time, saw an annual appreciation rate of 17.4%. Then, in 2002, Morningstar changed the way it rates mutual funds. Since then, the same portfolio has dropped 3.6% annually, while the stock market in general stayed relatively stable throughout both before and after 2002, the article details.

Prior to 2002, Morningstar compared the past performances of all U.S. stock funds—no matter their focus—with the best ones garnering a 5-star rating and the losers getting assigned 1-star. Then in 2002, Morningstar changed their methodology to compare each fund with only those in the same style of investment. Large-cap growth was compared only to other large-cap growth funds, small-cap value to other small-cap value funds, etc. That allowed some funds that had gotten low ratings to move up in the ranks, because they were being compared to a much smaller subset, the article explains. After the change, investors still poured money into Morningstar’s top-rated funds, but that money was spread across a broad spectrum of investing strategies—including those that weren’t doing so great. Meanwhile, the best performers were getting less inflows, and therefore missing out on the boosts that made them so successful in the first place.

However, not everyone agrees with the research, which was published in a study by a team of researchers, including Itzhak Ben-David of Ohio State University. In an email to The Journal, Morningstar’s director of research John Rekenthaler said that mutual funds account for “a minority of the moneys that flow into stocks” and that the change in methodology isn’t the only thing contributing to momentum-investing decline in the last 20 years. But while Ben-David concedes that there are certainly other factors at play, he insists that the shift resulted “in a surge in mutual-fund flows, and that flow increases lead to contemporaneous stock-price appreciation.” In other words, the rating change did have a more significant impact on momentum-investing’s performance.

Momentum-investing strategies are still used, and can be effective, particularly ones that employ ETFs with a specific single-industry focus. Still, it is risky to bet on a stock based solely on its historical record, as its future success can’t be guaranteed. And, says Jiacui Li of the University of Utah, another author of the study, “financial systems are complex,” adding that “a small and innocuous change can have large unintended consequences on the entire system.”