Smart beta strategies interpret the value anomaly as the effect of contrarian rebalancing against long-horizon mean reversion in equity prices. This approach to extracting the value premium differs sharply from the traditional style of value investing predicated on buying stocks with low P/B ratios (or comparable measures) in proportion to their capitalization weights.
Building on this insight, the author first describes how conventional value-style index construction leads to industry concentrations and explains how price-based weighting adversely affects returns. He then introduces mean reversion in stock prices and presents rebalancing as a mechanism for implementing a more-diversified value strategy. The author proceeds to evaluate evidence that, like the equity premium, the value premium is also mean reverting, and demonstrates that fundamentally weighted smart beta indexes implicitly engage in dollar-cost averaging to capture the full effect of this second-order mean reversion. He concludes with remarks on the relative strengths of the traditional and smart beta approaches to index investing.
According to financial lore, the value style outperforms the market in the long run. The author observes, however, that the major value-style indexes have had mixed results relative to broad cap-weighted indexes. Over the 10 years ended December 31, 2013, for example, the annualized return of the S&P 500 Value Index was 36 basis points (bps) below that of the S&P 500 Index, while the Russell 1000 Value Index underperformed the Russell 1000 Index by 20 bps. The author suggests that the strategies exemplified by the S&P and Russell value indexes might not capture the full value premium.
The conventional style indexes have two built-in attributes that raise further questions. First, growth industries are represented only to the extent that stocks categorized as predominately value also incidentally have growth characteristics. The resulting indexes are unrepresentative of the broad market’s economic exposure because they are dominated by aggressive industry bets. As of February 28, 2014, for instance, the S&P 500 Value Index held 23.7% financial stocks, 14.7% energy stocks, and 9.8% technology stocks. The corresponding sector weights in the S&P 500 Index were 15.9%, 10.0%, and 18.8%, respectively.
Second, traditional value-style indexes assign their constituent stocks’ weights on the basis of market capitalization. Consequently, the weights of the stocks fluctuate with their prices. Prior to the global financial crisis, for example, bank stocks became more expensive and correspondingly took on a heftier weight in the index, representing 8.5% of the Russell 1000 Value Index as of May 31, 2007. (By comparison, the broadly diversified Russell 1000 Index held 4.4% of its total capitalization in bank stocks as of that date.) During and after the crisis, prices fell, until bank stocks represented only 4.6% of the Russell 1000 Value Index (2.1% of the Russell 1000 Index) as of February 27, 2009. Consequently, bank stocks contributed more substantially to the cap-weighted value-style indexes’ losses during the crisis than they added to gains in the recovery.
Cap-weighted indexes automatically adjust positions as constituents’ market prices change; annual reconstitutions add newly qualified stocks and remove those that no longer meet the value screening criteria, but there is no rebalancing to exogenous weights. A fundamentally weighted index, the prototype of smart beta investing, adopts another approach to capturing the value premium. It weights stocks in accordance with accounting measures, such as cash flow and book value, that track capitalization over time (producing sector exposures reasonably similar to the broad market index), but are not directly related to stock prices. A fundamentally weighted index also rebalances annually against valuation ratio movements over the entire cross-section of stocks, without regard to their industry assignment. This contra-trading captures the price mean-reversion effect—the underlying source of value returns—through the law of large numbers in the equity universe.
Over the 30-year period ending December 31, 2013, the simulated return of the fundamentally weighted index was 13.14% a year, more than 200 bps higher than the annualized returns of the S&P 500 Index (11.09%) and the Russell 1000 Index (11.12%). In the same period, the performance of the S&P 500 Value Index fell behind the S&P 500 Index by 18 bps, while the Russell 1000 Value Index exceeded the Russell 1000 Index by 41 bps. The smart beta strategy’s outperformance was not attributable to added risk: the fundamentally weighted index had a Sharpe ratio of 0.49 compared to 0.36 for the S&P 500 Value Index and to 0.41 for the Russell 1000 Value Index.
The author cites empirical studies demonstrating the strategically significant fact that the value premium itself is mean reverting, and he reviews past market cycles to illustrate the return impact of reversals in the ratio of growth P/B to value P/B multiples. For instance, in January 2006, when the housing and subprime mortgage bubble had driven up prices for the value-laden banking and consumer staples sectors, the growth P/B was only 4.36 times the value P/B, and value cumulatively underperformed growth by 33.1% in the subsequent three years. The ratio expanded as the economy recovered from the global financial crisis, reaching 11.5 times in March 2009, and value cumulatively outperformed growth by 44.4% in the following three years.
Market participants can capture the value premium either by investing in low P/B stocks (the traditional cap-weighted approach) or by rebalancing from the last few years’ winner stocks into the losers (the smart beta approach). But when momentum prevails for an extended period, rebalancing can cause value stocks to underperform, perhaps substantially. It makes sense to dollar-cost average contrarian bets. The author brings this point home by comparing two hypothetical portfolios, one that allocates a constant tracking error to low P/B stocks and one that dynamically allocates more tracking error when the gap between growth and value P/B ratios widens.
In a simulation over the period 1963–2013, the value portfolio with the dynamically adjusted tracking error (that is, the value portfolio that automatically engaged in dollar-cost averaging) outperformed the value portfolio whose tracking error was held static (the traditional portfolio with a value bias) by 49 bps a year with no incremental risk. Compared to cap-weighted value-style indexes, smart beta strategies whose rebalancing rule effectively carries out dollar-cost averaging may thus capture considerably more of the value premium.
The author concludes that the style-based investing approach implemented by traditional value indexes neither wholly captures the value premium in individual securities nor fully exploits it across all industries and all stocks. In addition, the dollar-cost averaging inherent in fundamentally weighted strategies takes advantage of mean reversion in the value premium as well as in stock prices. Nonetheless, the author cautions that smart beta strategies’ efficiency, as measured by transaction costs, investability, and underlying economic exposure, can vary appreciably. Investors need to be smart when it comes to analyzing smart betas.
Summarized by Philip Lawton, CFA