What’s Alpha?
These days, the statistic’s interpretation is taken more figuratively than literally.
These days, the statistic’s interpretation is taken more figuratively than literally.
Alpha is among the most common of investment terms. It is also among the most confusing, given its multiple meanings. Today’s column will disentangle the threads.
The expression originated in 1967, appearing in “The Performance of Mutual Funds in the Period 1945-1964,” by Michael Jensen. (Consequently, alpha is sometimes called by traditionalists “Jensen’s Alpha.”) It was a mathematical result, providing the intercept on the y-axis for a best-fit line—defined as alpha’s counterpart, beta, which describes a fund’s level of risk—on a scatterplot that places stock market performance on the x-axis and the fund’s totals on the y-axis.
Which is a mouthful. Had there been no illustration following the invention, alpha would likely have reached a limited audience. However, the picture extends the concept. The general direction of the beta and alpha immediately tell a fund’s story. In the example below, the gently rising line indicates a relatively conservative fund, with a beta of less than 1. Meanwhile, an intercept above zero shows that the alpha is positive, which means ... well, therein lies the problem.
Source: Campbell Harvey, Duke University
Jensen ascribed the greatest possible power to alpha. In his view, fund performance came from two sources: general and specific. The general source was the level of the fund’s stock market exposure, as measured by beta. The specific source was the portfolio manager’s decisions, as captured by alpha. Consequently, one could rely fully and completely on the alpha statistic to judge the ability of portfolio managers.
The results left him unimpressed.
Nobody serious about investment research would write those words today, based solely on Jensen’s computation. They would recognize that funds could have ongoing exposures that make them behave very differently from the overall stock market. To cite a particularly dramatic example, Vanguard Small Cap Value Index VSIIX opened this millennium by gaining 22%, in a year when the S&P 500 lost 9%.
(Note: This column’s alpha examples are unrelievedly positive, because why not? It’s fun to imagine success. But it should be remembered that alphas can as easily land below sea level as above it.)
That fund’s management, of course, didn’t succeed by predicting “security prices.” Rather, the fund thrived because its portfolio deviates sharply from the mainstream, and 2000 happened to favor that dissimilarity, just as the previous years had punished them. The alpha scores merely restated the obvious. We already knew that Vanguard Small Cap had enjoyed a great year; what we sought to know was why. On that question, alphas were silent.
As recognition grew that investment managers should neither be rewarded nor punished for their portfolios’ accidents, the alpha calculation became increasingly complicated. Initially, it expanded from comparing funds against the single benchmark of the U.S. stock market, to charting them against three factors: stocks overall (the same as before), company size, and value/growth. This became codified as the Fama-French three-factor model.
In the 1990s, academic researchers showed that portfolios that held average exposures to each of those three factors would have posted above-zero alphas, if they held only securities that had enjoyed high recent performances. This finding not only torpedoed the strict definition of market efficiency, which stated that price movements convey no clues, but also undermined the three-factor model. If funds indexed to this “momentum” factor recorded positive alphas, then the measure could no longer be interpreted as revealing “manager contribution.”
You know what comes next, even if you don’t know what comes next. The three-factor model became the four-factor model. That, of course, was not enough. Dividend-yield funds frequently recorded positive alphas, even if they lacked exposures to the other factors. Owning illiquid stocks could prove profitable. Others found that, on a risk-adjusted basis, stocks that had relatively low standard deviations outperformed their more volatile rivals.
The models kept expanding. They will always expand, because there is no practical limit for the number of factors that are required to estimate the effect of the manager’s decisions. Specify 15 factors, the fund might benefit from a 16th. For example, no matter how many items that beta calculation incorporates, its alpha intercept will be skewed for a fund that always favors financials, if industry exposure is not one of the specified factors.
In practice, then, while researchers (including Morningstar) continue to publish alphas, the statistic resists easy interpretation. Sometimes, a fund’s alpha is a reasonably good estimate of the managers’ contributions. Other times, it fails. Unfortunately, it’s difficult to distinguish between the two cases. Whether calculated by one factor or many, a fund’s alpha may provide useful insight into management’s abilities. Or it may not.
That’s not very helpful. For that reason, the investment industry now tends to use the term alpha loosely, rather than technically. We cannot know a fund’s true alpha; the best we can do is study the shadows on the cave walls and guess at the shapes of the puppets that dance behind our heads. But we can define what alpha would be, if we possess the ability to identify it.
Some have described alpha as “knowing something others don’t know.” That is too strict. If two people on the planet profit from the same investment approach, surely they each create alpha, even if they are not alone. On the other hand, benefiting from widely publicized and imitated trades, as hedge funds once did (with, for example, convertible-arbitrage strategies), is too public to qualify as true alpha. Such opportunities disappear once money piles into the trades.
Thus, my definition: Alpha is the result of decisions that cannot be captured by any factor model, no matter how intricate the model, because the insight that underlies those decisions has not yet become public knowledge. When and if the investment tactic does become known, then it no longer is alpha. It is instead one of the many components of beta.
Note: A version of this column was originally published on Jan. 28, 2020.
John Rekenthaler (john.rekenthaler@morningstar.com) has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.
John Rekenthaler does not own (actual or beneficial) shares in any of the securities mentioned above. Find out about Morningstar’s editorial policies.
Transparency is how we protect the integrity of our work and keep empowering investors to achieve their goals and dreams. And we have unwavering standards for how we keep that integrity intact, from our research and data to our policies on content and your personal data.
We’d like to share more about how we work and what drives our day-to-day business.
We sell different types of products and services to both investment professionals and individual investors. These products and services are usually sold through license agreements or subscriptions. Our investment management business generates asset-based fees, which are calculated as a percentage of assets under management. We also sell both admissions and sponsorship packages for our investment conferences and advertising on our websites and newsletters.
How we use your information depends on the product and service that you use and your relationship with us. We may use it to:
To learn more about how we handle and protect your data, visit our privacy center.
Maintaining independence and editorial freedom is essential to our mission of empowering investor success. We provide a platform for our authors to report on investments fairly, accurately, and from the investor’s point of view. We also respect individual opinions––they represent the unvarnished thinking of our people and exacting analysis of our research processes. Our authors can publish views that we may or may not agree with, but they show their work, distinguish facts from opinions, and make sure their analysis is clear and in no way misleading or deceptive.
To further protect the integrity of our editorial content, we keep a strict separation between our sales teams and authors to remove any pressure or influence on our analyses and research.
Read our editorial policy to learn more about our process.