Financial theory has taken a lot of abuse recently, specifically some of the basic tenets of modern portfolio theory. Some of it we at AQR aid and abet (just two examples: markets are good but not perfect, and risk balancing beats capitalization weighting). But a fair chunk of the abuse comes from our industry’s collective tendency to judge ideas over relatively short periods, even if — or precisely because — short periods often feel like long periods. Thus, it’s important to occasionally step back and note that when examined properly, the very basics hold up better than many think or sometimes casually assert.
To explore this, I’m going to do some analysis very similar to something I saw Gary Brinson do many years ago (at least I think it was Gary; of course, while I’ve stolen the basic idea, any errors of concept or execution are my own). While the whole concept is borrowed, a fair amount of this will be out of sample as I’m getting old and it was a while ago! To start off, if you wanted to create an efficient frontier of stocks, bonds and commodities (here, I just look at the U.S., though I think, from distant memory, that Gary’s version included international and possibly small stocks, too) you’d probably expect:
All of the analysis below is done using excess returns over cash, and represent annualized realized monthly arithmetic average returns and annualized monthly standard deviations.
- Bonds (U.S. 10-year Treasury bonds) to be lower volatility and lower expected return;
- Stocks (S&P 500) and commodities (S&P GSCI) to be higher volatility and higher expected return. Having said this, commodity expected returns are more difficult to pin down. Why? Certainly they are harder because their weight within the market portfolio is quite arguable, but also because they seem to have low correlation with the market portfolio of wealth. So, unlike equities you probably expect lower returns compared with their volatility. In fact, whether or not you expect positive excess returns over cash for commodities is a controversial topic. We would (just our opinion) expect positive — but not as high as stocks (again, because their risk is more diversifiable) — even though commodities have been, and probably will remain, quite volatile. Of course, we also have the benefit of knowing what the final figure at the end of this piece looks like!
We begin with 1970-1974 and proceed in five-year increments (for all of the charts I keep the range of the X axes constant, as is my preference for comparing exhibits, but I can’t indulge this preference for the Y axes because they just vary too much).
Well the first five-year period certainly doesn’t look right! Stocks shouldn’t have a lower expected return than bonds (I’m foreshadowing the ultimate point here of course as these are not “expected” returns but realized returns). And negative to boot! And commodities can’t be that good. What gives? This theory stuff is not starting out well… Maybe the next five years will restore some sanity?
Darn, very similar, but even worse. Nobody expects to lose money on stocks and bonds for five years, or make that much on commodities. This finance stuff makes no sense!......MUCH MORE
HT: Levine@Bloomberg
