Jensen's Alpha: The Number That Told Fund Managers They Were Not Special

The word "alpha" is now so embedded in finance that people use it as a noun for skill itself: seeking alpha, alpha generation, he has no alpha. That entire vocabulary comes from one 1968 paper, in which Michael Jensen defined the number, computed it for real mutual funds, and delivered a verdict the industry has been trying to live down ever since.

The problem: how do you tell skill from a beta bet?

Suppose a fund returns 15% in a year when the market returns 10%. Is the manager brilliant?

Not necessarily. Maybe the fund simply held riskier stocks. If the portfolio was 1.5 times as sensitive to the market as the index, then in a 10% market it should have made roughly 15%. The manager did not outperform. They just turned the dial up and let the market do the work.

That is the whole problem, and before Jensen there was no clean way to separate the two. Managers who had merely taken more risk were being celebrated as stock pickers, and it was very hard to argue otherwise, because the benchmark for "what they should have earned" did not exist.

The CAPM, freshly minted, supplied it.

The key idea, via analogy

Imagine grading students who all sat exams of different difficulty. Comparing raw scores is meaningless. What you want is: given the difficulty of the exam this student took, what score would we expect from an average student, and how far above or below that did they actually land?

The residual, the amount by which you beat your own expected score, is the only fair measure of ability.

Jensen's construction is exactly this. Two steps:

Step one: work out what the fund should have earned. Measure how sensitive the fund's returns were to the market (its beta). The CAPM then tells you the return a passive investor could have gotten simply by holding that same amount of market exposure, with no skill involved at all. That is the fund's "expected exam score," its fair benchmark given the risk it chose to take.

Step two: subtract. Whatever return is left over, above or below that benchmark, is alpha. Positive alpha means the manager delivered something the market alone could not have. Negative alpha means the investor would have been better off just buying the market with matching leverage.

Mechanically, alpha is the intercept when you regress the fund's excess returns on the market's excess returns. If the CAPM held perfectly and no manager had skill, every intercept would be zero. Jensen's insight was that the intercept was not a nuisance term; the intercept was the whole point.

Then he did the thing that made the paper famous. He collected data on 115 mutual funds over a twenty-year stretch and computed alpha for every single one.

What he found

The results were bleak for the industry.

On average, after fees, alpha was negative. The typical fund did not beat the passive benchmark that matched its risk. It lagged it, by roughly the amount you would expect given what it charged.

Even measured before deducting expenses, so measured on gross performance, the evidence that managers could pick stocks well enough to cover their own costs was weak. The average fund's gross alpha was not convincingly positive.

And crucially, the funds that did well did not keep doing well. The distribution of alphas across funds looked close to what you would get from pure chance. A handful of star performers existed, but no more than randomness would produce, and their success did not persist in a way that suggested skill rather than luck.

Why it mattered

• It gave the world the word "alpha." The entire modern language of active management, alpha versus beta, skill versus exposure, comes from this paper's decomposition.
• It made the case for index funds before index funds existed. If the average active manager subtracts value after fees, the rational response is to stop paying for active management. Jensen's paper is one of the intellectual pillars under Vanguard, which launched its first index fund less than a decade later.
• It was strong evidence for market efficiency. If professionals with research budgets and full-time attention cannot beat a passive benchmark, that is exactly what you would expect in a market where prices already reflect available information. Fama's efficient markets hypothesis got a powerful empirical ally.
• It raised the bar for everyone. After Jensen, "I beat the market" stopped being a defensible claim on its own. You had to show you beat the market given the risk you took, and that bar has only kept rising as more factors were discovered.

The honest limitations

• Alpha is only as good as your benchmark. Jensen used the CAPM, meaning market beta was the only risk being adjusted for. When Fama and French later added size and value factors, many managers who had shown "positive alpha" turned out to be simply tilting toward small and cheap stocks, earning premia that were already known and freely available. Carhart's addition of momentum did more of the same. Each new factor retroactively destroys alpha that earlier researchers thought was real. Today's alpha may be tomorrow's known factor.
• Roll's critique applies. Since the true market portfolio is unobservable, and alpha is measured relative to a proxy, every alpha estimate is contaminated by the choice of index. Roll made this point explicitly about performance measurement.
• Twenty years is not enough data. Alpha is estimated with wide standard errors. Detecting genuine skill against the noise of markets requires far more data than any individual fund's track record contains, so "we found no skill" partly means "we could not detect skill even if it existed."
• Survivorship bias cuts the other way. Jensen's sample was of funds that existed and reported. Failed funds that closed are typically underrepresented in such datasets, which if anything makes the industry look better than it was. The true picture may be worse than his numbers.
• A fund's beta is not constant. Managers change their market exposure over time, sometimes deliberately. A single fixed beta estimated over a long window can mismeasure the benchmark and therefore mismeasure alpha, which is precisely the gap Treynor and Mazuy set out to probe.

The one-line takeaway

Jensen defined alpha as the return a manager earns beyond what their risk level alone would have delivered, then showed that across 115 mutual funds it was, on average, negative after fees, which is to say the industry was charging investors for the privilege of underperforming a passive benchmark.