The term “smart beta” has recently become a favorite buzzword among ETF pundits. Embedded in this term is a tacit recognition that alternative approaches to portfolio construction, beyond market cap-weighted benchmarks, may provide better, more intuitive ways to invest. One of the difficulties with a term like “smart beta,” however, is that the category has been broadly defined by some to include any index that does not follow a market cap-weighting methodology. This is problematic because it implies similarity among a variety of approaches that may be markedly different.
Nevertheless, rather than debating the merits of “smart beta” as a category, the purpose of this newsletter is to help investors understand what makes the AlphaDEX methodology (upon which the underlying indices of 39 First Trust AlphaDEX ETFs are based) distinct from market cap-weighted benchmarks, as well as other approaches that have been deemed “smart beta.” To that end, the following takes a closer look at the AlphaDEX methodology, highlighting the underlying logic upon which this model is based, as well as drawing performance and holdings comparisons with traditional market cap-weighted indices.
Overview of the AlphaDEX Methodology
In order to stay true to the mandate of each category to which the AlphaDEX methodology is applied, and to avoid style-drift, the universe from which potential holdings are selected is established from well-known benchmark indices. For example, the First Trust Large Cap Core AlphaDEX Fund (FEX) selects approximately 375 stocks from the S&P 500 Index. The candidates are scored based on a value model and a growth model (which are described in more detail below), in order to select portfolio holdings. “Value” and “growth” are scored separately based on the premise that there are different attributes that have tended to drive performance for value stocks versus growth stocks. Portfolio weightings are also determined by these scores so that stocks with more desirable attributes, as determined by the scoring methodology, may have a greater impact on performance than stocks with less desirable attributes. This model is reapplied quarterly for domestic AlphaDEX ETFs, and semi-annually for international AlphaDEX ETFs, at which time portfolio holdings are rebalanced (see chart for domestic AlphaDEX ETFs).
The AlphaDEX “Value” Model
The AlphaDEX “value” model scores stocks on three separate factors: price-to-book ratio, price-to-cash flow ratio, and return on assets. The first two factors seek to identify whether a stock is cheap or expensive, relative to certain underlying fundamental measures (book value and cash flow), while the third factor provides an indication of balance sheet quality. Essentially, this model seeks to capitalize on the so-called “value” anomaly that has been well established by several academic studies, wherein a pattern of excess returns is associated with cheaper stocks versus more expensive stocks.
In seeking to understand the existence of a factor that predicts the relative outperformance of cheap stocks in the context of the “efficient market hypothesis,”2 many have concluded that these stocks must be riskier than more expensive stocks. After all, they reason, there is no free lunch. However, a number of compelling alternative explanations for the value anomaly argue that excess returns associated with cheap stocks are primarily rooted in a pattern of behavioral mistakes made by investors.
For example, because investors are emotional beings, rational analysis is often supplemented with investor optimism and/or pessimism. On the one hand, expensive stocks tend to have high, optimistic expectations about future earnings growth, often extrapolated from previous periods of strength. However, investors may overlook potential challenges to future growth, which may arise from changes in the competitive environment, shifts in consumer preferences, or a variety of other unforeseen circumstances. As a result, when an expensive stock announces earnings results that fall short of analysts’ expectations, it may be punished more severely than a cheap stock that misses earnings estimates.
On the other hand, cheap stocks tend to have low, pessimistic expectations about future growth, which may be inferred from past weakness. In such instances, investors may overlook potential improvements, which may result from possible changes in leadership, restructuring, innovation, or even the inherent cyclicality of certain businesses. Hence, when a cheap stock announces earnings that exceed expectations, it may be rewarded more than an expensive stock.
The influence that these misjudgments may have on investor behavior helps to explain the persistence of the value anomaly, in our opinion.
The AlphaDEX “Growth” Model
The AlphaDEX growth model scores stocks based on five separate factors: three-month price appreciation, six-month price appreciation, twelve-month price appreciation, one-year sales growth, and price-to-sales ratio. The first three factors measure momentum over various time periods. The fourth factor provides a measure of fundamental growth. The fifth factor provides a measure of value. This model seeks to systematically profit from the so-called “momentum” anomaly, which is the tendency for stocks that have performed best in the recent past to continue outperforming over the next three to twelve months.
Although the success of momentum as a predictor of excess returns is well-supported by empirical evidence, risk-based explanations for this factor generally fall short in explaining why it actually works.
However, one compelling behavioral explanation is related to investors’ hesitance to adjust their beliefs about the intrinsic value a stock to reflect recent fundamental improvements. As a result, if these improvements persist, investors end up chasing stock prices higher, resulting in a pattern of excess returns in the near future.
The AlphaDEX growth model includes “one-year sales growth” in order to favor stocks whose superior market performance over the past 3, 6, and 12 months is supported by underlying fundamental growth. Finally, this model utilizes a value factor, price-to-sales, for many of the same reasons described in the above discussion of the value model.
AlphaDEX Versus Traditional Index ETFs
The application of the AlphaDEX methodology has resulted in significant differences between AlphaDEX ETFs and traditional market cap-weighted ETFs, in terms of both portfolio allocations as well as performance.
As for portfolio allocations, one of the most notable differences has been the tendency for AlphaDEX ETFs to be less concentrated than market cap-weighted ETFs among top portfolio holdings. This is especially apparent among sectors and individual country ETFs. For example, as of 2/28/14, the top 10 holdings for the nine domestic sector AlphaDEX ETFs represented an average weight of 27%, versus an average weight of 51% for the three largest comparable market cap-weighted ETFs within each sector. While such allocations may arguably “bias” returns in favor of the smaller stocks within each fund, this is both a necessary and desirable trade-off for reducing the degree of stock-specific risk inherent to many top-heavy portfolios, in our opinion.
As for performance, there are currently 16 AlphaDEX ETFs (out of 39) that have accumulated track records longer than 5 years, including 7 domestic size/style ETFs and 9 domestic sector ETFs. During the five year period ending on 2/28/14, 15 of 16 of these funds produced positive excess returns relative to their benchmarks. Risk-adjusted returns were similarly impressive, as 11 of 16 produced positive alpha over the past 5 years. Moreover, as a group, these funds captured an average of 123% of the positive performance (upside capture) of their respective benchmarks, while capturing an average of 100% of the negative performance (downside capture) of their respective benchmarks. (Additional 5-year performance details for each ETF can be seen in the table on the following page and the standardized performance for each ETF can be seen in Appendix A). Of course, past performance is not a guarantee of future results.
The Growth of AlphaDEX ETFs
Since the launch of the original 16 AlphaDEX ETFs in the spring of 2007, this family of funds has expanded to 39 ETFs, including both domestic and international equity categories. As of 2/28/14, AlphaDEX ETF assets under management (AUM) totaled $12.3 billion, compared to $4.8 billion a year earlier. This growth was broadly distributed, with 13 ETFs increasing by more than $250 million.
We believe this growth provides evidence that ETF strategists and portfolio managers have begun to look beyond the first generation of market cap-weighted ETFs. While various ETF sponsors have their own take on what constitutes a “better” way to construct an investment portfolio, the AlphaDEX methodology relies upon the insights of academic finance in seeking to provide a set of low-cost, tax-efficient tools, with which investors may implement a variety of investment strategies, ranging from sector rotation models to traditional strategic asset allocation portfolios.