I’ve just returned from the Research Affiliates Advisor Symposium in Newport Beach, California last week. If the cloudy and sometimes rainy weather disappointed some of the attendees, the conference itself didn’t. As I reflect on it now, it stands as one of the finest investment conferences I’ve attended. This will be the first in a three-part series on the conference.
The argument against traditional indexing
Before I talk about the conference, a few words about Research Affiliates are in order for readers who aren’t familiar with the firm. Research Affiliates is the company of Robert Arnott who is the author of influential papers on investing and a former board member of the Jounral of Portfolio Management. Arnott arguably devised the first “smart beta” stock strategy which he calls the “fundamental index.” Instead of ranking stocks in an index based on market capitalization, which is how most index funds function, the fundamental index ranks them on four economic factors of the stocks’ underlying businesses – sales, cash flow, dividends and book value. In other words, the fundamental index or “RAFI” (Research Affiliates Fundamental Index) “breaks the link,” as Arnott puts it, between a stock’s rank in an index and its market capitalization or price. In fact, that’s what all smart beta strategies do; they are all based on some newly devised “beta” or index that doesn’t rank stocks based on their price or market capitalization. I like Arnott’s approach because it forces investors to think of stocks as ownership units of businesses and to rank stocks based on characteristics of their underlying businesses.
Products that use the fundamental index include the PowerShares FTSE RAFI US 1000 ETF (PRF) and the PIMCO RAE Fundamental PLUS Fund (PXTIX). Other funds apply fundamental indexing to foreign stock markets. Arnott’s insight that capitalization weighted indexes necessarily give an investor more exposure to more expensive stocks and his research into other forms of index construction started the smart beta revolution. Other “factors” or characteristics of groups of stocks that could help those stocks beat the market – namely small and value — were identified in a famous paper by Eugene Fama and Kenneth French. While Dimensional Fund Advisors took its bearings from the Fama/French research and built funds with capitalization weighted indexes that it tilted toward small-cap and value stocks, nobody created an alternative index or helped create funds that were invested in one until Arnott.
One award-winning academic paper has argued that the fundamental index has a “value bias” making the PowerShares fund’s S&P 500 Index-matching performance over the past decade, when value has underperformed growth, striking. (Arnott also classifies his index as value-oriented on the smart beta portion of the Research Affiliates website.) The PIMCO fund has outperformed the index by more than 300 basis points annualized, but it’s structured differently than the PowerShares fund. The PIMCO fund gains exposure to the fundamental index through a derivative, which it colateralizes with a bond portfolio. So it has two sources of return – the fundamental index and the bond portfolio that PIMCO manages in an effort to overcome the price of the derivative.
Since the development of products using the fundamental index, other funds with other “factors” have hit the market. Along with plain capitalization weighted index funds, smart beta funds have been attracting assets in droves while traditional active managers mostly bleed capital. But there aren’t as many factors that can consistently beat capitalization weighted indices as marketers might have you think. The “factor zoo,” as Research Affiliates calls it, may be well populated, but only a few species are worth much. Those are value, size, income, momentum, low beta, and quality. In some of its papers, Research Affiliates isn’t even sure “quality” is a legitimate factor, although the firm includes it in its factor appraisals on its website.
Not All Factors Are Equal — Or Always Well-Priced
The question about which factors might be overpriced comes from an argument about data mining. Most factors reflect data mining or observance of a one-time bump in valuation that isn’t sustainable. This has caused a public argument between Research Affiliates and Clifford Asness of AQR Management. If illegitimate factors are the result of a one-time or random bump in valuation, legitimate factors that have persistence can get expensive too. Furthermore, those factors can be embraced or shunned – in effect, timed — based on their relative historical valuations. By contrast, Asness thinks investors are better off sticking with a factor or two they like rather than trying to time them.
Parts of the conference addressed the valuations argument. Research Affiliates thinks momentum and low-beta or low-volatility strategies are expensive. That makes sense given that some investors have driven up the prices of Facebook, Google, and Amazon, and others have sought stock exposure with lower volatility in the aftermath of the financial crisis.
Most value strategies, on the other hand, are relatively cheap. They have suffered as Facebook, Amazon, Apple, Netflix, and Google have propelled growth and momentum indices higher in recent years. But investors might be wrong to choose the factors that have performed the best. Research Affiliates Head of Investment Management, FeiFei Li, gave a presentation showing among other things that choosing the worst performing smart beta factors can lead to better performance than choosing the best performing ones. Valuations are predictive of future returns. In other words, contrarian smart beta investors looking for a “reversion to the mean” might overweight value over momentum based on how each factor is priced relative to its history on the Research Affiliates website.
Li also argued that investors should look at how smart beta funds implement their strategies, because implementation can have a much greater effect on returns than the expense ratio. In particular, high turnover, low weighted-average market cap, and a low number of holdings can be warning signs to investors that a fund isn’t implementing its strategies in the most cost-effective ways.
Morningstar’s Director of Global ETF Research, Ben Johnson, followed Li’s talk with some similar themes. Johnson argued that re-framing factor investing as an evolution of active management was useful. Many traditional stock pickers use factors that characterize ETFs, just not as systematically and mechanically. Johnson also warned investors to be wary when funds with similar factors are introduced at once. That can be an indication that a factor has had a run that’s probably not repeatable.
In my next installment, I will report on the portions of the conference that were concerned with global asset allocation and asset class valuations.