At the start of November of 2007, I published an analysis in which I provided outlooks for a series of sector ETFs . The projections were generated using Quantext Portfolio Planner (QPP). That analysis suggested that most major asset classes were due for a substantial reversion to the mean—downwards. In particular, the worst looking asset classes were emerging markets (NYSEARCA:EEM), developed international markets (NYSEARCA:EFA), and telecom (NYSEARCA:IXP).
One of the strongest results from that analysis was a back-of-the-envelope calculation that suggested that most major asset classes would need to under-perform TIPS (NYSEARCA:TIP) for the next three years in order to bring long-term returns into equilibrium. The asset classes that were projected to under-perform TIPS in this approach included the S&P500 (both market cap weighted (NYSEARCA:IVV) and equal weighted RSP, EFA, EEM, IDU (utilities), RWR (REITs), and a number of other major sectors. Given that TIPS have a much lower risk level than any of these broad asset classes/sectors, an investor might logically have chosen TIPS in preference to any of these asset classes.
In real markets, values do not simply settle back to equilibrium. On the contrary, asset values tend to overshoot—and they have. From November 1, 2007 (when my sector outlook article was published) through the end of November 2008, we have seen a huge decline in equities and a range of other asset classes:
click to enlarge
The table above shows the Projected Difference in returns between the trailing three years through October 2007 and what QPP projected the expected returns to be. A positive value means that an asset class had been under-performing and would be expected to generate additional returns (i.e. it was under-valued). A negative number means that the asset class has been out-performing and is due for reversion to the mean—downwards. EEM, for example, had a trailing average annual three-year return of 35.2%. QPP projected that the expected return of EEM was 13.4%, so that EEM had been generating much higher than “fair” returns---to the tune of excess return of almost 22% per year. This is not sustainable, of course. By contrast, QPP projected that TIPS (TIP) were substantially under-valued. The details of these calculations are provided in the original article.
In the thirteen months since this article was published, every major asset class has declined dramatically, ranging from a 58% drop for EEM to a 4.7% drop for TIPS. Reversion to The Mean (NYSEARCA:RTM) has hit with a vengeance. When we look at the same calculations today (using data through the end of November of 2008), we get the following:
QPP projects that every asset class has dropped so dramatically that a major reversion to the mean is in order—upwards. The most dramatically under-valued sectors are REITs (NYSEARCA:RWR) and high-yield bonds (NYSE:COY). Emerging markets (EEM), while substantially under-valued, are not even in the top ten in terms of relative under-valuation. Quite by coincidence, the S&P500 has rank 10 in terms of relative under-valuation—so ranks smaller than 10 are relatively more attractive than the S&P500 and vice versa.
The challenge with this kind of analysis is knowing when markets will revert to the mean and how fast they will do it. Momentum adds persistence to runs either up or down and the broader economic conditions that correspond to a bull or bear market (such as a recession) can persist for long periods.
It should be noted that more aggressive asset classes will tend to look more attractive in the measures in the table above, because the projected ‘fair’ return will be higher for these asset classes—we expect that TIPS will generate much lower returns than broad equity indices over the long haul, so TIPS will never look all that great when the market is under-valued.
There are several main points to take away from all this. First, mean reversion has once again been validated in a big way. Second, while QPP did not capture the extent of the decline, QPP was signaling that almost all major asset classes were substantially over-valued. Third, QPP is suggesting that the markets have way overshot the equilibrium level of returns, and thus are setting the stage for some very high future returns.
The greatest value of mean reversion models is to remind investors that markets tend to out-perform and under-perform, but that the most reasonable basis for planning is to use a long view rather than being reactive to recent conditions. There is a natural human tendency to believe that the recent past is a good basis for making decisions—but it's not. When markets are way down, investors get gloomy and sell. When markets are way up, investors get too optimistic and pile in. A mean reversion world view has consistently been shown to provide a better estimate of future performance.
There are many notable examples of the dominance of mean reversion for planning, but one stands out: the NASDAQ during the last crash. I performed a long-term analysis of QQQQ [pdf] to see how QPP’s projected performance compared to using trailing performance as a guide. From 2000 through 2002, QQQQ had average annual return of -28.5%. Using data from 2000-2002, QPP projected a future average return of 16.7%. Standing at the end of 2002, you would have had to be a committed believer in mean reversion to take such an estimate seriously. The average annual return for the period from 2003-2005 was 18.7%. There are numerous other examples in this paper and others cited in Footnote 3.
Investors have a much easier time “believing” in momentum than mean reversion—and this is explained by studies of heuristics. Momentum means believing that what happens next month or year will be very similar to what happened last month or year. Mean reversion is based on the idea that events will vary, but tend to vary about some unknown equilibrium condition. RTM means that you cannot expect momentum to last forever—in either direction. A range of research has shown that momentum effects tend to dominate on the short-term but that mean reversion is dominant over the long-term. In the current market conditions, with lots of bad news, many investors have concluded that the wheels are coming off the system.
Perhaps they are right this time. History suggests that RTM is the dominant long-term driver, however, and that the markets will recover. It is important to understand, however, that this does not mean that markets will rebound in the near term (i.e. within a year). A recent survey of the worst years for the S&P500 in history (2008 now ranks second only to 1931) shows that the five out of the other nine worst years were followed by a year of positive returns, and four were followed by another year of negative returns. If those odds don’t look sufficiently promising, there are ways to invest while still hedging against additional potential losses—and one of the more attractive approaches is to exploit the current record-high volatility by selling out-of-the money covered calls against holdings.