As we have noted over the past several years, alpha is notoriously difficult to define and isolate since its existence depends not only on the target being observed, but on the perspective of its observer. The more we study alpha, it seems, the less we know about it.
We are reminded of this irony by today’s CAIA guest post. Erik Einertson is with Los Angeles quant manager First Quadrant. (Regular readers may remember this guest post by Einertson’s colleague Jia Ye.) He takes a very comprehensive look at what is and what isn’t true alpha – so comprehensive that we decided to deliver it in two parts with part one below.
Special to AllAboutAlpha.com by: Erik Einertson, CFA, CAIA, First Quadrant, L.P.
Following the asset crash of 2008 and the strong rebound of 2009, much has been made about the failure of alpha to remain uncorrelated to beta. We saw the collapse of some strategies of “alpha” that had produced fantastic risk-adjusted returns throughout the previous decade.
How did this happen? Wasn’t the original goal of an alpha program to deliver excess return without subjecting the portfolio to additional market exposure?
How does an investor know when they are getting market exposure instead of alpha? Often correlations are used, but strategies like Private Equity and Real Estate can smooth their returns, making direct correlations look smaller than they really are. Even passive-options strategies on equity markets with deep out-of-the-money strike prices can also look uncorrelated to market indices until the strike prices are hit in unexpectedly volatile markets. In fact, much of what we consider alpha actually has beta exposure at exactly the times you don’t want it, especially during periods of market volatility like we saw in 2008. In the end, not all types of “alpha” are equal.
I categorize alpha into five types based on statistical characteristics of the skew distribution. The order of the categories are from the least valuable to the most valuable:
- Insurance Beta
- Exotic Beta
- Skill based Alpha and
- Protection Alpha
Each type of alpha has its benefits and risk characteristics. So investors need to consider each type in regards to a typical beta heavy portfolio. (click to enlarge)
The definition of alpha is about as controversial as the question about its existence. Investopedia describes Alpha as “The abnormal rate of return on a security or portfolio in excess of what would be predicted by an equilibrium model like the capital asset pricing model (CAPM).” Another common definition is any value added over an appropriate benchmark.
But what if I am just selling out of the money at a distant strike price? Have I really created alpha? This strategy could go long periods without generating any real equity beta exposure and can probably beat a cash benchmark. Unfortunately, a levered strategy like this will probably eventually generate huge losses because of the risk of a large negative tail event.
Recent discussion has centered on the suggestion that alpha is something more than just beating a benchmark. You can beat a benchmark by introducing an array of systematic risks ranging from what we’d more traditionally refer to as beta risk, to risks that are less commonly considered beta but are no less systematic than beta. A more appropriate definition of alpha should be “the time-varying selection of systematic risks, delivered without creating any average – over time – exposure to systematic risks.”
Some way to segregate and define alpha is something important to review more closely because obviously, all alpha isn’t equal.
If there are different types of alpha, there needs to be some way to identify where our alpha comes from and the risks associated with each type. Using skew as a proxy, for example, investors can gain some transparency about the source of any excess return and the risks associated with that return.
Alpha #1: Beta
Unfortunately, much of what we call “alpha” is simply hidden market exposure. This shouldn’t be considered alpha, but often is. Common reasons for beta being considered alpha are:
- A benchmark mismatch – For example, using a cash benchmark when an investment manager is buying actual market exposure through physicals or futures. While the client may be aware, they do not account for the mismatch in their performance evaluation.
- Lack of transparency – The end client does not know what instruments the investment manager is using and is unaware that they are simply loading up on simple market exposure to take advantage of market premiums. An example of this could potentially be many “Black Box” hedge fund strategies where investors are not given the holdings of the strategy. The investment manager simply takes the cash and invests it in equities or bonds through physicals or the futures market.
- Systematic Tilts – Systematic tilts towards higher risk instruments (stocks over bonds or small cap stocks over large cap stocks) that may or may not be transparent to the end client. These tilts may be termed tactical to investors, but a manager without skill uses a systematic tilt to generate returns based on market premia, not for alpha generation due to skill.
The real risk to the end user of hidden market beta is two-fold. First, many investors purchase alternative investments for portfolio diversification. The thought is that by investing in uncorrelated assets they can reduce overall portfolio risk and improve returns. If an investor is simply buying market exposure, they are not getting the diversification they are expecting (and if there is leverage they may actually be increasing overall portfolio risk). The recent performance of many hedge funds that coincided with the downturn of the S&P 500 is one obvious example
The second primary risk for investors is in regards to fees. In many cases, especially where there is a lack of transparency, a client may be charged hedge fund like fees (typically “2 and 20”) for something the client could buy for just a couple of basis points in the futures market. To compare strategies with embedded betas to those with no beta exposure, one should strip out the portion of the performance that is caused by this beta exposure, including fee differentials. An equation to account for this Beta exposure is shown below:
Where a is alpha, P is portfolio performance, B is the beta of the strategy with embedded beta exposure, mp is the performance of the market, HF is the fees for the strategy (should include performance fees) and FF is the cost of getting that exposure through the futures market.
For example, a hedge fund that has shown a long-term beta of 0.5 should have that portion of performance stripped out to arrive at an “alpha” portion of the portfolio before comparisons begin.
A simple question could be, “how much beta is there in my hedge fund”? According to Fung, Hsieh, Naik and Ramadorai (2008), between 1995 – 2004, when stripping out a variety of betas using their Fung-Hsieh model, Funds of Funds only delivered alpha during the October 1998 to March 2000 bull market, and only 22% of Funds of Funds actually created alpha over the entire period. That means a lot of beta is sitting in our alpha pool.
Mistaking beta as alpha will tend to make your portfolio less diversified, and, when used in portable alpha schemes, can cause you to inadvertently “double down” on your existing exposure. This reinforces the need to give appropriate benchmarks to strategies in a portfolio.
(Ed: More on the “other four faces” of alpha tomorrow in Part 2 of Einertson’s post)