How Is This Research Useful to Practitioners?
The author addresses a well-known tenet of asset-pricing theory that often informs the growth-versus-value stock choice in active management. The value premium, or the comparatively higher future return known to accrue to stocks with high book-to-market ratios (B/Ms), has been linked in recent literature to timing differences between the cash flows of value and growth stocks. Because growth stocks are believed to have substantially higher future cash flow growth rates and longer cash flow durations, they draw increased investor interest, which lowers their relative returns. The author finds evidence to persuasively refute this belief (while confirming the value premium’s existence) and shows this misperception is attributable, in part, to valuation model biases.
Summarizing his results in the context of both empirical asset pricing and firm-level behavior, the author finds that in rebalanced portfolios, value stocks have faster-growing cash flows. In buy-and-hold portfolios, growth stock cash flows do not grow substantially faster than those of value stocks; they often grow more slowly. In the more recent historical period (1963–2001), growth stock cash flows increase only 2% faster annually, or about one-tenth of what prior research has found. Dividends of growth and value stocks rise at similar rates, whereas earnings for the latter actually grow faster. Finally, after accounting for survivorship bias, regressions of future dividend growth rates on the B/M favor value stocks.
Acknowledging that the supposed growth stock advantage depends on empirical evidence, the author shows that conceptual biases related to the permanence of return on equity (efficiency growth), dividend growth (survivorship bias), and persistence of cash flow growth rates early in company history (lookback bias) are likely responsible for the false conclusions.
How Did the Author Conduct This Research?
Accounting information is from Compustat, data on stockholder equity are from Moody’s Investors Service, and stock return and dividend data for NYSE, NASDAQ, and AMEX stocks (excluding financials and utilities) are from CRSP. The total sample period is 1926–2001, partitioned into roughly equal intervals (1926–1962 and 1963–2001) to enable comparisons between “early” and “modern” eras and over the complete period.
Following standard research methodology, the author creates annual portfolios and then subdivides them into quintiles sorted on the B/M, ranging from growth (low) to value (high). Performance based on a $100 investment in selected portfolios is first tracked annually over 10-year intervals in terms of average real dividends per portfolio and their average percentage growth rates for groups with either fixed constituents (buy and hold) or rebalanced to maintain a consistent growth-to-value stock ranking. The author estimates long-run growth rates of up to 35 years and performs time-series analyses of average dividend growth rates by portfolio.
The author posits that the conventional wisdom is based principally on four pillars: 1) the structure of dividend valuation models, 2) a belief that growth stocks have higher ROEs, 3) sharply negative coefficients in standard regressions of dividend growth rates to B/M, and 4) recent findings that growth stocks have longer cash flow durations. He then proceeds to undermine each pillar by offering new evidence or by identifying biases in the existing research.
Abstractor’s Viewpoint
In investing, as in many process-driven activities, major surprises can lurk within the unexamined certainties of received wisdom. The author provides a cautionary tale about the foundations of the value premium, which ostensibly have sound support in the academic literature. The research is a thoroughly investigated and multidimensional critique against drawing convenient (yet plausible) conclusions (e.g., growth stocks have more dynamic and extended cash flows/dividend streams).
Although the limited number of observations and time period under investigation represent serious flaws in this research, the work serves as a prudent reminder to practitioners that shortcuts to outperformance more than likely lack staying power. For example, the discussion of the lookback bias illustrates the fundamental foolishness of projecting a new star company’s early performance as the expectation for its future. Similarly, considering survivorship bias, while ventures such as Google/Alphabet and Amazon stand as growth paragons, the reality is that most start-ups ultimately vanish. Delisting possibilities are more salient for one’s portfolio calculations.
The author seems to come up short by not providing an alternative explanation for the value premium. Of course, research perfection is most likely unattainable, with this further quest a suitable follow-on investigation that bears watching. And explaining how the growth rates of buy-and-hold and rebalanced portfolios may be related could be of greater interest to wealth managers, anyway. The related finding that value stocks have higher dividend growth rates in rebalanced portfolios and growth stocks have higher dividend growth rates in buy-and-hold portfolios may serve as useful tactical guidance for profitable asset management.