Click to enlarge Image source: istockphoto.com. Used with permission. Quit Calling It Smart Beta Last week, an advisor forwarded me this Financial Times article, ” Smart Beta Not Quite as Clever as Marketed ,” asking for my comment, to which I immediately responded, “the only aspect of the article I agree with is the title.” Since arriving at 3D Asset Management, I’ve published two articles on ETF investing: ” ETF Product Development: Innovation Versus Over-Engineering ” and ” Why ETFs (and Why Strategic Beta ETFs) .” In both these articles, I have advocated for our use of factor-based ETFs, commonly known as alternative beta or strategic beta. Despite what appears to be an industry-wide adoption of the term, I refuse to use the label ‘smart’ beta for factor-based ETFs. Here is what I wrote in “ETF Product Development:” …’smart’ beta (factor) investing is not ‘smart’ at all but just a reformulation of the Dimensional Fund Advisors’ (DFA) strategy of investing in areas of the market which have afforded higher risk premia over the long run. ‘Small cap’ and ‘value’ factors outperform over the market because they come with higher risks which investors are compensated for over the long run – these factors are no ‘smarter’ than a traditional market-cap based approach such as the S&P 500. ‘Smart beta’ is a marketing label used by ETF sponsors to get advisors more comfortable with alternative methods of indexing. But it is no more nefarious than that, unlike what is implied by the Financial Times article critiquing factor ETFs. The rollout of strategic or alternative beta ETFs reflects innovation on the part of ETF and index providers to give retail investors access to market segments and themes historically only available to professional/institutional investors. What follows is a section-by-section critique of the FT article: Has the investment industry’s marketing push outsmarted itself? For several years, huge effort has gone in to selling “smart beta” funds. It has worked, creating great excitement. Now, not at all surprisingly, the backlash has begun. 3D’s Response: “Backlash” from whom? From those most threatened by the advent of factor-based ETFs? Strategic beta ETFs capture much of the systematic elements of many actively-managed strategies in cost-effective and tax-efficient vehicles. So who feels most threatened by this wave of product innovation? That is, of course, rhetorical. Smart beta comes up with a strategy to beat the index, which can itself be made into an index with simple rules. The advantage of doing this is that funds that track an index can be run far more cheaply than active funds, which face a far higher bill for research and managers’ salaries. So if a winning strategy can be reduced to an index, it should be possible to cut costs, and offer a superior return to investors . 3D’s Response: Agree with most of this statement, except this notion of “winning.” The goal of investing isn’t to ‘win’ but to achieve long-term financial goals, whether with indexing or with active management. If the focus is just on ‘winning’, then one misses the entire point of investing in the first place. For instance, a dividend-focused strategy is designed not so much to outperform the broad market-based benchmark (i.e. the S&P 500) but to provide investors with desired portfolio characteristics, namely 1) total returns sourced more from dividend income than capital appreciation and 2) lower equity market sensitivity. Does this strategy ‘lose’ if it delivers on this objective even though it may underperform the market-based benchmark? Smart beta strategies are now proliferating but most commonly stem from anomalies identified in the academic literature. Perhaps most importantly, there are Value (cheap stocks do better than expensive), Momentum (winners keep winning, and losers keep losing), and Low volatility (relatively stable stocks perform better). All will have periods when they do badly. All perform well in the long run. Other popular strategies involve weighting portfolios by companies’ sales, or revenues, or dividends . From these building blocks, investment managers have now built multifactor funds in different proportions, and come up with a dizzying array of new factors. And they have sold a lot of funds on the back of it. 3D’s Response: As a former quantitative portfolio manager, I can reasonably say that most of the historically effective factors are now captured in some form by ETF products. I don’t know what the true library of significant and investable factors consist of, but I am confident it is between the three originally proposed by Fama/French (market, size, value) and somewhere in the range of 15-20. But is it a “dizzying” array? First, one must distinguish between an ETF that captures an historical risk premium or anomaly such as value or momentum versus an ETF built on rules designed to capture a specific investment theme such as Goldman Sachs Hedge Fund VIP ETF which would screen for top hedge fund holdings of individual stocks. Second, the ‘dizzying’ array of factors is partly driven by legitimate differences on what is the best rule-based design to capture a factor. Strategic betas such as value and quality can be defined differently, but many of them achieve similar results. ETF strategists such as ourselves conduct due diligence to get underneath the hood on what the factor is trying to achieve and then provide research opinions on whether a particular ETF provides the best rules-based design given the cost. But there is a problem. In theory, and in practice, once a market anomaly has been observed, it cannot continue. There are two reasons why future performance may be worse than the historical backtest suggests, outlined by Pete Hecht, chief market strategist for Evanston Capital Management, in a recent paper. First, the back-test may have been “data-mined.” In other words, the researchers fiddled to find a formula that delivered the very best result for the period they were looking at. This may be due to dishonesty, or may happen unconsciously. 3D’s Response: First, a straw man warning. Pete Hecht from Evanston may be correct in his observation of Fama/French, but some disclosure should have been made that Evanston has a vested interest in dismissing smart beta as a flash in the pan fad since Evanston serves as a multi-manager hedge fund-of-fund, whose value proposition (along with much of traditional active management) is under competitive threat from strategic beta investing. That said, Hecht’s first argument is facetious at best and indicts any data-driven approach to investing. There are clear standards for what constitutes robust backtesting, namely is the observed factor anomaly consistent, robust, and reliable across time and across different markets? There is ‘true’ data mining which is to throw as much stuff on the wall and see what sticks and there is robust backtesting where there is a clear and intuitive rationale for the observed behavior. A second problem is arbitrage, and the very existence of smart beta funds feeds this problem. Once you know that cheap stocks outperform, the logical response is to buy cheap stocks. If many do this, cheap stocks’ price will rise until they no longer outperform. 3D’s Response: This ignores the fact that underlying factor behavior is a risk-premium or behavioral explanation. ‘Arbitrage’ implies there exists some systematic mispricing between two or multiple assets and that a well-functioning market should reduce this mispricing such that it cannot be exploited over a sustained period of time. However, those who hold to a risk premia view of factor investing believe such premia exist precisely due to rational pricing on the part of investors. Take the value risk premium as an example. As Fama/French originally argued in their 1992 paper, ” The Cross-Section of Expected Stock Returns ,” the value effect (as proxied by book value/market value ratio) is “the market’s assessment of its value [which] should be a direct indicator of the relative prospects…” In other words, ‘cheap’ stocks are cheap for a reason, because they are riskier than the overall market. Cheap stocks tend to be financial companies that trade close to book value due to regulatory or financial leverage reasons or companies with highly cyclical (and uncertain) earnings that the market is not willing to assign a premium multiple to. I won’t delve into the rationales driving the other main factor categories, but the historical size and value risk premia as documented by Fama/French are rooted in highly intuitive economic rationales and not the result of some creative backtests with no fundamental rooting as implied by the FT article. Mr. Hecht took Mr. Fama’s formulas for determining which stocks were cheap, and saw how the strategy would have performed starting in 1992 and carrying on to the present. In all cases, whether measured by straight performance or adjusted for risk, they did much worse after the paper’s publication than they had before it. The reduction in performance ranged from 30 to 71 percent. The value effect had diminished. 3D’s Response: There is some evidence that the long-term risk premium has shrunk as market participation has expanded and has become more electronic and global. However, Mr. Hecht’s observation wouldn’t just apply to the value premium but to the entire market risk premium itself. If equity market investing has become less ‘risky’ then investors should expect to be compensated with less return over time. But common sense intuition would hold that equities are riskier than bonds over the long run and that risk can fluctuate during different economic and inflationary cycles. Otherwise the entire notion of rational capital pricing would collapse into absurdity. No one would suggest that the equity risk premium should go to zero (except these guys ) just because the markets have gotten more efficient. Likewise, few would argue that small cap stocks should not trade at a premium versus large cap stocks (i.e. would you be willing to accept the same premium for holding Pfizer (NYSE: PFE ) than you would a small, speculative biotech?). And the same holds for value stocks as mentioned above. However, Hecht’s second argument brings up broader implications for active management. If there is a diminishing return to value and small cap investing (and it’s debatable since size and value have shown long-term persistency across multiple markets), then this would have profound implications for all of active management, not just strategic beta. Hecht’s arguments of diminishing returns to factor investing has even worse implications for traditional active management and hedge funds where much of the alpha they provide can be sourced from such systematic factors. As more active managers and hedge funds become ‘discovered’ and ‘popular’, then presumably their edge in delivering alpha should also be reduced. That leads to another problem, identified by Rob Arnott in a paper for Research Affiliates, a pioneer of smart beta. A strong backtest at any point in time, he reasons, may be because the factor tested has become expensive. Very perversely therefore, a strong backtest almost becomes a reason not to buy into a strategy. And if a strategy looks good now simply because it is expensive, that may be an active reason to fear that it will now perform badly. Conversely, it might imply that factors that have done poorly of late – and as the chart shows, value has badly lagged behind the market ever since the financial crisis – are now cheap and worth buying, for those with the intestinal fortitude to do so. Meanwhile it is worth checking whether low-volatility and high-momentum stocks, both still performing well, look over-expensive and due to revert to the mean. 3D’s Response: Are there more opportunistic times to buy factor portfolios like value? Sure, that’s almost axiomatic. Buying value seemed to be the only strategy that worked in 2009 following the steep sell-off during the global financial crisis, but when it looked like large banks like Citigroup (NYSE: C ) and Bank of America (NYSE: BAC ) would go under, it would have required heroic “intestinal fortitude” to bet on value at the height of uncertainty. When a long-term strategy looks especially cheap versus its history, then it will likely have a higher payoff if you believe in the long-term rationale for owning that strategy in the first place. I find it interesting that Research Affiliates is now just commenting on this after an especially tough period for the RAFI indices, notably in emerging markets. Despite de-emphasizing the role of ‘price’ in its construction, RAFI has historically correlated with value-style indices. One rarely hears from value managers to avoid their strategies because the underlying valuation spreads are narrow (and less opportunistic to buy into the strategy). During these moments when valuation spreads were narrow, chances are the prospective investor had been regaled with Morningstar ratings on past performance, sort of like the backtests Research Affiliates is critiquing. I do believe factor (and strategy) timing is difficult, particularly if one uses valuation spreads as part of the allocation process. Please see this article published by Larry Swedroe on ETF.com where he summarizes research questioning the use of valuation as a timing tool for factor selection. However, I would not rule out valuation as a reason to avoid a factor or to opportunistically invest in it; it is a matter of perspective and what other aspects are incorporated into the decision-making. The bottom line is that investment processes incorporating strategic beta ETFs should not bet on one or two factors but should be diversified across a variety of factors so as to minimize the disruption due to valuation conditions. A final issue: risk. Piling into one particular factor is inherently more risky. For Andrew Lo of Massachusetts Institute of Technology, one of the world’s most respected financial theorists, the problem with “smart beta” is that it can easily morph into “dumb sigma” – the Greek letter used for volatility. 3D’s Response: This is awkwardly written. Yes, investing in just one factor is riskier than being diversified across multiple factors. Yet, how does this morph into “dumb sigma?” Many things can lead to “dumb sigma.” Buying equity on margin can be considered “dumb sigma.” “Dumb sigma” just reflects poor portfolio construction and risk control and is not necessarily indicative of strategic beta investing. The main takeaway is that alternative / strategic beta ETFs have given retail investors more options than what has been historically available. These are not a panacea for beating the markets but can serve as cost-effective to gain targeted exposures not directly accessible via traditional market indices. The goal is not to “win” but to build more robust portfolios to achieve long-term investment goals and objectives. Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.