I’m just going get this out of the way and say it: There’s no alpha in ETFs.
That may be a given for many, but with all the new products being launched based on some sort of active management (including some rather “black box” mystique within the ETF space), I thought a discussion of the concept and perhaps definition of alpha was required. Furthermore, as a market participant very much interested in the management of betas (different market exposures), this topic is really about what I do. Consider that some disclosure on my bias in this matter.
So you can go from market cap weighted indexation to fundamental weighted indexation. You can call a product an Intellidex or Alphadex or some other term that seems to meld the idea of active management implemented within some form of index fund or ETF. But I can’t get around the fact that it’s just not alpha.
If what is being proposed is a US large cap equity strategy (or whatever) based on some set of rules that, in the long term, beats the S&P 500 Index (or the relevant benchmark index), that may be deemed as alpha. It certainly could be evidence of a strategy that over a certain period of time beats a market cap weighted index. This is “alpha” in a statistical context based on portfolio theory such as CAPM. [As an aside, there was considerable debate about the merits of CAPM but generally most market participants now agree that the flaws (due to generalizations and required assumptions) of CAPM make it simply good theory but weak on application. Do a Google search for “CAPM is dead” to find out more on this discussion.]
Thus far I’ve discussed what can be defined as statistical alpha. But is it really alpha?
To me alpha is a very rare thing. Stepping away from a mathematical definition, I look at alpha as returns that can not easily be duplicated. It would be a rare finding in which a market participant can truly exploit an inefficiency in the market. I would agree with the observation that there’s less alpha now than in the past. There’s certainly more market participants out there trying to uncover every stone to find their elusive alpha. There are more hedge funds whose purpose is to provide alpha … we would hope it’s “pure alpha” if they specifically aim to perform strongly in both good and bad markets. Perhaps there’s no change in the amount of alpha out there, but because of so many participants, especially hedge funds, there’s simply less to go around “per capita”.
What I’m getting at is that there are really two definitions of alpha. One is the statistical definition. If a manager or investor can beat their benchmark index by X% on average every period, that’s their “statistical” alpha. So for every XYZ Capital Management out there who has shown that kind of result, they can calculate their alpha or outperformance over the index.
However, the other definition of alpha that I am trying to focus on here is that which relies more on a qualitative rather than quantitative due diligence on a manager or instrument including certain ETFs that are loosely classified as “quasi-indexing” which I list below. By the way, I’m not sure what the right term for these is, but I don’t see anything wrong with quasi-indexing.
Isn’t alpha supposed to be the returns from a strategy that is based on market inefficiencies? If so, then shouldn’t these market inefficiencies disappear as other players enter the field? This line of thinking is discussed in this paper available on SSRN:
By the way, this paper was written by market participants. At the time of its publishing, the authors worked for ABP Investments which, according to its website, is the 2nd largest pension fund in the world. This paper discusses how investment processes can be divided into two groups:
Traditional beta which they call “commoditized beta”. This is based on exposures to various markets and is the classic definition of market risk. Traditional alpha which they call “non-commoditized beta”. This is based on other risk factors not associated with market exposure.
Basically, the writers of this paper sum up things well in four points (bottom of page 4) provided here ad verbatim:
1. Any investment process which today generates return by taking exposures, which are not well known, will become obsolete progressively, as the exposure premium reduces or the exposure is commoditized.
2. As exposures become commoditized, the space to generate additional alpha return (from the residual) decreases, and the space for beta return increases.
3. There is practically no alpha based exposure, which cannot be commoditized.
4. The investment problem which originated in finding exposures to generate alpha will gravitate towards becoming the process of analyzing when to take a specific commoditized beta exposure.
They go on to say that the alpha versus beta debate is irrelevant. “The eventual investment problem is therefore not to now generate exposure combinations, which would generate alpha, but to be able to time the betas of the existing exposures. We therefore believe that active management will devolve to an exposure based allocation process, where the objective is largely to allocate to different forms of beta, and where alpha does not actually exist. Portfolio diversification is then just the diversification obtained by applying the forecasting process to more than one beta.” This is basically a rewording of point #4 above.
Yikes. Does this mean that hedge funds and the search for alpha are synonymous to money down the drain? No, I would say this is like any analysis of market efficiency. It’s definitely not black and white. The US market is highly efficient with thousands of participants providing tons of analysis and many more putting their money where their mouth is. You have less analysis and less market participation in certain developing markets (in addition to other factors like government regulation) which creates a less efficient market environment. So in the US, there’s a far greater argument to use ETFs than for any given emerging market exposure. At some point in time, with greater “commoditization” you move from a market that looks more like India (for example) to one that looks more like the US. This would seem to agree with points 1 and 2 above.
Similarly, with hedge funds, there simply has to be some evolving where one form of alpha has to be replaced with another. Hedge funds (and especially “fund of funds”) and multi-strategy funds are in the business of finding the next alpha. Whether this is something that can be done successfully and on a repetitive basis is a very good question. Like anything else, the market will decide.
But this leads to point #3 above. As alpha is exploited, it will be commoditized and become some alternative form of beta. In my opinion, this is the real key point of this paper. Alpha will eventually become beta. If it’s a good idea, a set of rules can be force fed into a computer model that can run it on autopilot. (I will admit that not every strategy can be transformed into a quantitative model and thank goodness because there’s got to be some active management left for the industry! Many pros rely on performance fees. This makes me wonder … performance fees are a good reward for true alpha. Should quasi-index ETFs charge performance fees if they truly provide alpha?)
Strategies on autopilot. Is this not what these ETFs listed below are all about? All of these ETF’s have in some way turned a systemic market “inefficiency” into a commodity:
PowerShares Dynamic Market Portfolio (NYSEARCA:PWC), PowerShares Dynamic OTC Portfolio (PWO) and the various other Intellidex based ETFs from PowerShares The various FTSE/RAFI ETFs from PowerShares and Claymore WisdomTree ETFs First Trust DB Strategic Value Fund (FDV)
I am not negative on ETFs such as these which I would broadly classify as quasi-index ETFs. I simply don’t think that they provide alpha. They provide what may loosely be considered as alternative forms of beta or broad market exposure. Traditional ETFs have been based on traditional market cap weighted indexation. These new (some not so new) offerings provide a change, and in this case I think change is good because we are getting new forms of broad market exposure. However, some clearly, or at least in my opinion, provide better value to the investor than others. That’s another discussion.