Montier snaps at smart beta and risk parity.
(Maxwell) Smart Betas
James Montier of GMO's asset-allocation team (striking a blow against grade inflation, that appears to be his official title) is among the fund industry's top writers. He tackles important subjects, glides past the distracting details without oversimplifying, and knows his material. His letters are on my short list.
I quite enjoyed this month's paper, "No Silver Bullets in Investing (just old snake oil in new bottles)." In the article, Montier spanks two current investment trends, "smart beta" indexing and the "risk parity" approach of asset allocation. (As a bonus, he finishes with a brief but useful overview of how different assets tend to fare against inflation. I'll touch on that subject in tomorrow's column.)
His smart-beta argument is not new. As with several critics before him, including Morningstar's Paul Kaplan, Montier regards most smart-beta strategies--that is, indexing schemes that weight not by market capitalization but instead by some other process--as being value and/or small-company approaches in drag. Call itfundamental weighting, call it equal weighting, call it minimum variance, call it what you like. They all result in pushing a portfolio down and to the left in the equityMorningstar Style Box, away from large growth and toward small value.
This is clearly so, a fact that is openly acknowledged by one of smart beta's creators, Rob Arnott, who recently stated in a Morningstar interview that "what's smart about smart beta is breaking the link with price." He continues, "equal weighting does much the same thing [as does Arnott's approach of fundamental investing]."
Developing the idea further, breaking the link with prices does more than put a fund into smaller companies that sell at lower price/book multiples. It also moves the portfolio more heavily into stocks that have cheaper share prices, that have lower trading volume, and that are priced more cheaply relative to their companies' sales. That is, the portfolio gains not one, but five "smart betas": small company, value, low-priced stock, illiquidity, and fundamentals. (Actually, it gains more betas than that, but we'll stop at five for this discussion.)
These betas are related. Having more of one beta implies having more of all others. What remains to be determined is how to disentangle the effects. To the extent that smart beta works, what are the factors that drive that success? Do some of these factors count for more than others? Such research is in its early stage and hasn't yet yielded much insight, aside from the fact that the value factor seems to be more important than the small-company attribute. At this stage, the key lies not in the details but rather in the decision. Market weight to get traditional beta. Break from price to get "smart" beta.
There are two notable exceptions to the extended family--smart betas that appear to be unrelated. One is stock-price momentum, first documented in academic papers in the 1990s. Momentum is not related to small/value betas. The second is the newer idea of quality, which is variously defined, but which can be summarized as "company attributes that Warren Buffett would like." Examples would be high profitability, low debt, and stability of earnings. These are good things, rather than risks associated with the small/value spectrum. Whatever the payoff for quality--which is barely understood, the notion being so nascent--it doesn't come from the usual places.
Montier is silent about the prospects of momentum but not about small/value and quality. At today's prices, he writes, choose the latter. GMO finds both small and value stocks to be relatively expensive. It's not overly fond of quality U.S. stocks, either, but it regards them as the currently superior option.
Montier is considerably kinder toward smart beta than he is toward risk parity. The latter, he writes, "is the antithesis of everything at GMO we hold dear." That's putting the matter directly.
Risk parity refers to building a portfolio so that each asset in the portfolio has the same level of risk. This differs sharply from traditional asset allocation. Consider, for example, a 60% stock/40% bond portfolio, with stocks being judged as 6 times riskier than bonds (more on such judgments shortly). In such a case, 90% of the portfolio's overall risk comes from its 60% in stocks, and only 10% of risk comes from the 40% placed into bonds.
Risk parity therefore tends to lead to portfolios that are relatively heavy in bonds and light in stocks. It also can lead to leverage, as the resulting low-volatility portfolio also has low expected returns, unless the portfolio is leveraged so as to magnify potential gains (and losses).
I appreciate the concept. It's an academic construct, of a similar ilk as William Sharpe's tangency portfolio. From the academic perspective, why limit the search for efficiency by mandating that all assets be long? Instead, seek the most efficient portfolio that exists, regardless of constraints, then add or subtract leverage to reach the appropriate level of risk.
Incorporating leverage is logical; the idea of the tangency portfolio is one reason why Professor Sharpe won a Nobel Prize. It's not only an academic idea, either. In the marketplace, hedge funds have long used leverage. Nonetheless, leverage is a tricky thing, hampered not only by regulation but also by periodic credit squeezes. To those dangers, Montier adds the concern that a near-term market swing can squeeze a short into prematurely closing a position. Leverage, he writes, can transform "the temporary impairment of capital (price volatility) into the permanent impairment of capital." In contrast, steep losses to an unleveraged long holding need not force that asset's sale.
Montier also dislikes risk parity because its portfolios can be unstable, as changing measurements of risk can dictate large portfolio re-allocations. In addition, Montier disparages risk parity because it does not require that the user forecast asset-class returns. As forecasting asset-class returns is GMO's calling card, and a task it has conducted with some success, he is naturally scornful.
Those items don't bother me as much as they do Montier. What truly troubles me is something that Montier mentions in passing: the track record. Montier points out that per most risk-parity approaches, the stock weighting would have been increased in 2007 because stock prices were relatively stable (recall that in risk-parity schemes, the lower an asset's volatility, the more heavily it is purchased). In 2009, the stock weighting would have been reduced because stocks were gyrating.
A sophisticated, mathematical version of buying high and selling low is sophisticated and mathematical--but it loses money just the same.
In short, Montier chuckles at smart beta and winces at risk parity. In fact, he publishes two equations that summarize his feelings:
1) Smart Beta = Dumb Beta + Smart Marketing
2) Risk Parity = Wrong Measure of Risk + Leverage + Price Indifference = Bad Idea
Those strike me as about right.
John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.