One of the first things that we learn in life is the alphabet. Through it, we form words and are able to communicate with one another using a common language. In investing, it would be useful if one of the first things that we learned about was the performance of our alpha bets. Why is alpha important? If an investor can persistently generate positive alpha, this may well be linked to investment skill. So, what is alpha and how good has the real estate industry been at generating it?
Alpha comes from the heyday of work on quantifying risk-adjusted performance, the 1960s. Sharpe, Treynor and Jensen all made contributions to the debate at the time but it is from Jensen’s work that alpha derives. For Jensen, the first question is what return should I have received given the sensitivity (beta) of my portfolio to market movements? Once that is calculated, you can compare the answer to the actual return; the remainder is alpha.
alpha = Rp – Rf – (beta x (Rm – Rf))
In words, this equation says that alpha is equal to the return on a portfolio (Rp), less the return on the risk free rate (Rf), less the sensitivity (beta) of the portfolio to the market’s return (Rm) less the return on the risk free rate (Rf).
Let’s consider a set of performance data for two portfolios where:
Rf = 1%
Rm = 6%
Rp1 (the return on portfolio 1) = 5%
Rp2 (the return on portfolio 2) = 7%
Beta1 (the sensitivity of portfolio 1) = 0.7
Beta2 (the sensitivity of portfolio 2) = 1.3
What are the values of alpha for portfolio 1 and portfolio 2?
Portfolio 1 has underperformed the market (5% – 6% = -1%) and has a performance delta (difference) of -1%. The world is not without people who would say the portfolio has an alpha of -1% but that is not correct since this underperformance is not risk-adjusted. The alpha of the portfolio is calculated as 5% – 1% – (0.7 x (6% – 1%)) = +0.5%. So, this portfolio has underperformed the market but delivered more return than you would expect given its sensitivity to market movements.
Portfolio 2 has outperformed the market (7% – 6% = 1%) and has a delta of +1%. The alpha of the portfolio is 7% – 1% – (1.3 x (6% – 1%)) = -0.5%. This portfolio has outperformed the market but delivered less return than you would expect given its sensitivity to market movements. In this case, you could have carried less market-related risk (a beta of 1.2) and received the same return, assuming zero alpha. These two contrived examples go to show that:
Alpha and delta are not the same. When people say that they have delivered alpha of 1%, make sure it is actually alpha and not delta, just to make sure you are speaking the same language.
Alpha is not always positive. It is possible to deliver risk-adjusted underperformance.
It is possible to deliver positive alpha and negative delta; it is also possible to deliver positive delta and negative alpha. When people say they are looking for “risk-adjusted returns”, that implies that they would be happy to underperform the market (negative delta) but to have positive alpha, but they may mean that they want positive delta and alpha, or, conceivably, they might not know what they mean.
With the formula and maths out of the way, how have real estate investors’ alpha bets turned out? Well, real estate data analysis, inevitably, comes with measurement problems and the quantification of alpha comes with model specification problems so people will find fault with whatever the results are (unless they are enormously flattering of course). Nonetheless, here are a couple of examples of the research that has been carried out.
The first comes from way back in 2008 when Bond and Mitchell considered the question of “Alpha and persistence in UK property fund management”. This report is results dense and somewhat out of date but here are the findings that are most relevant to the question of alpha and persistence.
Looking at four, five-year periods (1982-1986, 1987-1991, 1992-1996, 1997-2001) they first ranked the Jensen alpha of each fund in each period. They then looked at the alpha of each fund in the following five-year period to assess whether any alpha (positive or negative) in the ranking period persisted into the evaluation period.
They found that the the probability of a top quartile alpha fund in the ranking period remaining in the top quartile in the following period was 25%, the same probability as the top-quartile funds being distributed randomly amongst the quartiles in the following period.
The authors concluded that the results show “no indication in persistence of quartile rankings” of alpha. This would be consistent with a lack of skill or skill that only worked in particular market environments.
They also found that over two ten-year periods (1987-1996, 1997-2006) only 8% of 170 funds and 26% of 85 funds respectively delivered an alpha greater than +1%. Since the performance data were gross of fees, the vast majority of funds were likely to have had negative alpha on a net of fees basis.
The second study comes from 2019, when Bollinger and Pagliari considered performance data for fund indices in the US by investment style, core, value-add and opportunistic, comparing risk-adjusted performance on non-core funds with core funds.
The good news is that, in absolute terms, value-add and opportunistic funds outperformed core funds, on average; positive delta, even net of fees. The not so good news is that the authors found, “over the 2000–2017 study period, value-added funds …, on average, generated a negative alpha of −3.26%; similarly, opportunistic funds … generated a negative alpha of −2.85%.”
Seeking to understand whether the GFC contributed to these overwhelmingly negative results, the authors subdivided the period into 91 different time periods of no less than five years each. For value add funds they found that, on average, alpha was negative for all 91 periods. For opportunistic funds they found that alpha was negative in all but 1 of the 91 periods.
The caveat here is that these results are “on average” and that it is possible, even likely, that, due to the dispersion of alpha at individual fund level, some value-add and opportunistic funds will have delivered positive alpha, but it probably isn’t many given the extent of the average negative alpha and evidence regarding dispersion from the Bond and Mitchell study.
It seems clear that it would be useful for real estate investors to understand the outcome of their alpha bets.
Article originally published here.
