What is alpha?
Tommi Johnsen, PhD | Advisory Board Trent Ambler, MSF | Portfolio Manager
Key Point: Alpha is realized return in excess of the risk taken. Outperforming a benchmark does not necessarily guarantee positive alpha. Delivering alpha means that the realized returns were achieved with less risk than otherwise would have been required by any investor.
Implications: Consistently providing alpha means realizing greater returns while exhibiting lower risk. That is, lower risk than should be expected. There are various sources of alpha which can be exploited by investors.
IS IT ALPHA OR IS IT RISK?
As a concept, “Alpha” is misunderstood. As a method of evaluating performance, it is complicated. One can only understand and use the term within the framework of the Efficient Markets Hypothesis (EMH).
We have written on the meaning of the EMH here:
Within the context of the EMH, alpha is best thought of as that portion of investment payoff unrelated to any source of risk. Alpha can be positive or negative, but it is not equivalent to the outperformance we constantly hear about in the financial press.
ALPHA and OUTPERFORMANCE ARE NOT INTERCHANGEABLE.
Alpha and outperformance are only equivalent if the risk of the portfolio is equivalent to the risk of the benchmark index. The difference between alpha and outperformance is not a simple story and here’s why: Outperformance does not take into account differences in risk and Alpha does.
When evaluating portfolio performance relative to any benchmark, it is always necessary to consider the riskiness of the investment relative to the riskiness of the benchmark. Managers should, on average, “outperform” an index like the S&P 500 if their portfolios have higher risks. Say the portfolio has a higher weighting in tech stocks than the S&P 500. Managers can likewise “underperform” the S&P 500 by investing in less risky portfolios. Say hold fewer tech stocks than the S&P 500 holds. It is possible to underperform an index and have positive alpha.
The misunderstanding here is a bit nuanced but extremely relevant. To keep it simple, let’s take the single index model as the relevant risk model, i.e. Beta, and consider possible performance outcomes and how they illustrate our point.
If an equity holding has a beta[1] of 1.2 then it is expected based on market exposure only that the given equity holding will deliver 1.2x the market. If the market delivers a 10% return and this particular holding delivers 12% then it could be said that the return realized by holding this stock was in line with its beta, in line with its risk exposure.
However, if instead of 12% the stock delivered 18% over the same period and if market exposure is the single risk exposure being used to describe such returns, then over that same period it could be said that this stock had positive excess return of 600 basis points (6%). The stock outperformed its beta and under a single index model (SI)[2] thus produced ‘alpha’ of 600 basis points.
Again, using the SI model when a stock with a beta of 1.2 delivers 12% return while the market delivers 10% it is not accurate to say that holding this stock delivered ‘alpha’ of 200bps (2%). The return realized by the stock is in excess of the market return, but such excess return is not alpha in this case. Instead, the excess return realized is better thought of as payoff that is commensurate with the excess ‘risk’ taken. It is the payoff for taking risk. It is not the payoff that is in excess of the risk taken. A stock with a Beta of 1.2 can be said to be 1.2x as ‘risky’ as the market and thus will on average perform according to that excess risk exposure.
PURSUING ALPHA
The ability to deliver alpha is quite often taken as evidence of superior skill or investment acumen. The sustained ability to produce alpha suggests a level of understanding about market dynamics that is not being fully priced by or is not available to the broader investment community.
How then is it done. To generate alpha consistently, where should an investment manager spend their time?
1. Generate a better set of information. Think Wall Street research here. Develop a method to produce better than average earnings forecasts, for example.
2. Develop a quantitative model that processes information superior to the competition and other investors. As more data is made available to investors, and in ways that are increasingly timely, this is even more relevant. Making sense of vast reams of information becomes critical as the information set balloons; paradoxically, more information delivered faster makes the valuation exercise harder and thus there is an edge to be had in the superior processing of data.
3. Exploit behavioral biases. The evidence is overwhelming. Investors make systematic errors that affect stock prices. Take advantage of them.
4. Achieve access to deals or deal structures that are not available to other players in the market. Insider trading is an illegal example of this but there are legal examples as well. If an investor by way of their size, sophistication, or ability to execute is afforded opportunities that are not available to the broader market this too can be a source of alpha. This also shows up when a controlling position is established in a yet to be executed larger play. If by way of networking ability or just pure sweat equity an investor is able to quietly and incrementally assemble what eventually turns into a larger more valuable position, there can be gains back to that investor which are in excess of the risks taken. Think establishing control of a corner in the boardgame monopoly.
5. Trade in markets where even a slight informational or procedural advantage can give substantial edge. If it is known that institutional level (‘smart money’) investors limit their exposure in smaller markets where there just isn’t sufficient scope or capacity for there to be an impact on their bottom line otherwise well capitalized non-institutional players may be able to garner an edge here.
The first three, generate a better set of information, develop a quantitative model that processes information better, and exploit behavioral bias are especially relevant for equity investing. If an equity investor is able to build a process that touches on all three of these there is potential for substantial and sustained alpha.
[1] Beta is defined as the covariance of returns for a particular equity holding with the market returns over some period (monthly returns over the past five years is often used here) divided by the variance of market returns over the same period. It is the slope of an Ordinary Least Squares regression of that equity holdings returns regressed on the market returns over that same period. Beta is an estimate of systematic risk.
[2] A Single Index Model is one where only one risk factor is used to explain returns. If market returns are used as the independent variable and equity returns for a single holding are used as the dependent variable then this is the Capital Asset Pricing Model or the CAPM.