Are Star Fund Managers Actually Skilled? the Four-Factor Model

Every year, magazines crown the best-performing mutual funds and investors pour money into last year's champions, assuming a great year means a great manager. It's a natural bet: surely the funds at the top of the list got there through skill, and skill should continue.

In 1997, Mark Carhart tested that assumption with unusual rigor, and the answer was deflating. Most of what looks like "skill" in fund performance turns out to be a mix of luck, cheap fees, and accidentally riding known market patterns. To prove it, he built a tool that has since become the industry-standard scorecard for judging any fund or manager: the four-factor model. Here's the story.

The problem: does past performance predict future performance?

Mutual fund ads famously carry the disclaimer "past performance does not guarantee future results." Carhart asked the sharp version of that question: does past performance guarantee anything at all? If a fund crushed it last year, is it more likely to crush it next year?

To answer honestly, you first have to strip out all the boring, non-skill reasons a fund might have looked good. By the mid-1990s, researchers knew four such reasons, four broad patterns that could make a fund's returns look great without the manager being a genius. So Carhart's first move was to assemble a model that accounts for all four at once.

The key idea via analogy: grading on a proper curve

Imagine a student who scores 95% on a test. Impressive? It depends on the test. If it was an easy test where everyone scored high, 95% is nothing special. To judge the student fairly, you grade relative to how hard the test actually was.

Judging a fund manager is the same. A manager might post big returns simply because they happened to be riding forces that lift returns regardless of skill. Carhart's four factors are the "difficulty adjustments" you apply before you're allowed to call anything skill:

1. The market factor, did the fund just go up because the whole market went up? (The CAPM idea.)
2. The size factor, did it do well just by holding small companies, which tend to earn more? (From Fama-French.)
3. The value factor, did it do well just by holding cheap "value" stocks, which tend to earn more? (Also from Fama-French.)
4. The momentum factor, did it do well just by holding recent winning stocks, which tend to keep winning? (This is Carhart's addition, taken straight from the Jegadeesh-Titman momentum discovery.)

Carhart's real innovation was recognizing that a manager could look brilliant simply by buying recent winners, an easy, mechanical strategy, not genius, so momentum had to be one of the difficulty adjustments. Bolt momentum onto the three Fama-French factors and you get the four-factor model.

Only after subtracting all four of these "you-didn't-have-to-be-smart-for-this" sources of return do you get to the leftover, the alpha, the piece that might actually be skill. In one sentence: the four-factor model measures how much a fund earned that can't be explained by simply riding the market, small stocks, cheap stocks, or recent winners.

What he found: skill is rare, but bad funds stay bad

Carhart ran this scorecard across a huge sample of mutual funds, and the results were sobering:

• Almost none of the "hot" funds had real skill. When he graded last year's winners against the four factors, most of their glow disappeared. They hadn't been brilliant; they'd mostly been holding momentum stocks and other rewarded tilts. Strip those out and the genuine, skill-based alpha was tiny or nonexistent.
• The little persistence that existed came from momentum, not talent. Winning funds tended to stay winning for a short while, but mostly because they held momentum stocks that kept rising on their own, not because the manager kept making smart new calls. It was the stocks carrying momentum, not the manager carrying skill.
• Fees and costs quietly decided everything. The funds that persistently lagged did so largely because of high expenses and heavy trading. Costs are the one thing that's reliably persistent: an expensive fund tends to stay a drag, year after year, because the fees never stop.
• The clearest lesson: avoid the losers. The strongest, most reliable finding wasn't "pick winners", it was that the worst-performing, highest-cost funds tended to keep underperforming. You can't reliably spot future stars, but you can reliably spot and dodge future duds by watching fees.

Why it mattered so much

Carhart's paper is one of the most cited in all of finance, and its influence runs deep.

• The four-factor model became the default scorecard. For roughly two decades, if you wanted to claim a fund or strategy had genuine skill, you had to show positive alpha against Carhart's four factors. It's the referee academics and practitioners reach for by default. Beating "the market" stopped being good enough; you had to beat market-plus-size-plus-value-plus-momentum.
• It supercharged the case for index funds. If most active managers can't produce skill-based alpha once you adjust for these factors, then paying them high fees is a bad deal, and cheap index funds win by default. Carhart handed the low-cost-investing movement some of its hardest evidence.
• It cemented momentum's status. By making momentum one of the four essential yardsticks, Carhart signaled that the Jegadeesh-Titman effect wasn't a curiosity, it was a force so pervasive you couldn't fairly evaluate anyone without accounting for it.
• It reframed how we think about fund selection. The practical takeaway, "you probably can't pick winners, but you can avoid expensive losers", is now conventional wisdom, and it flows directly from this paper.

The honest limitations

The four-factor model is a workhorse, but it's not the final word.

• The factors are a choice, not a law. Why exactly four? Later work argued for more (profitability, quality, investment) or different ones. The four-factor model is a convention that worked well, not a complete or unique description of what drives returns. Fama and French themselves moved on to a five-factor model that, notably, doesn't even include momentum.
• Momentum is awkward to actually hold. The model treats momentum as a free-floating source of return, but as we saw, real momentum is costly to trade and occasionally crashes. Crediting or blaming a fund for its "momentum exposure" glosses over how hard that exposure is to capture in practice.
• Absence of measurable skill isn't proof of no skill. The model can only detect skill that shows up as steady, factor-adjusted outperformance. A genuinely talented manager with a lumpy or unusual style might have real skill the model can't see, though the burden of proof, rightly, is on them.
• Survivorship and data quirks lurk. Studying funds is tricky because failed funds vanish from databases, which can distort the picture. Carhart was careful about this, but it's a permanent hazard in this kind of research.

The one-line takeaway

Carhart added momentum to the Fama-French factors to build finance's standard skill-detector, and used it to deliver a humbling verdict: most "star" fund performance is riding known patterns, not genius, and the most reliable edge an investor has is simply to avoid high-cost funds.