This In Practice piece gives a practitioner’s perspective on the article “The Long-Run Drivers of Stock Returns: Total Payouts and the Real Economy,” by Philip U. Straehl and Roger G. Ibbotson, published in the Third Quarter 2017 issue of the Financial Analysts Journal.
What’s the Investment Issue?
In identifying the drivers of past stock performance to better forecast future returns, investment analysts have predominantly focused on fundamentals, such as historical dividends, earnings, and book value. But since the 1980s, quoted firms have increasingly used share buybacks instead of dividends to reward shareholders. In fact, across the US stock market, buybacks have surpassed dividends in 8 of the last 10 years.
Yet, the impact of buybacks on stock returns has been overlooked or misinterpreted by investors. Measuring corporate performance based on payouts, the authors believe, could thus be more accurate than using accounting-based measures, such as earnings.
This study presents a new model of stock returns based on total payouts—dividends and buybacks—which the authors believe will lead to more accurate forecasts.
How Do the Authors Tackle the Issue?
The authors create total payout models for US stock returns by using data from 1871 to 2014.
They develop three models, each of which mimics the behaviour of a hypothetical investor type in the case of a stock buyback. These models are designed to show the relative importance of dividends and buybacks as drivers of stock performance.
The three investor types are as follows:
• “Buy-and-hold investors,” who exercise their right to hold on to their shares during a buyback offer and, given that the total shares in issue are reduced after the buyback, wind up owning a bigger proportion of a company
• “Pro rata buyback investors,” who sell a proportional number of their shares back to the company and take cash
• “Cap-weighted index investors,” who sell some shares and hold on to the rest
The authors then compare growth in total payouts against growth in the real economy to quantify the relationship. They assess whether the total payout models can generate better forecasts of future equity returns than dividend-based models. Finally, the authors examine to what extent returns from various valuation measures are stable and consistent over market cycles. The authors believe this extra test is important because buybacks tend to be more sensitive to market cycles than dividends.
What Are the Findings?
In terms of forecasting future returns, the cash flows received by pro rata buyback investors and cap-weighted index investors are the same, despite the first group receiving their returns through buybacks and the second group relying on dividends.
All three models show that long-term equity returns are almost completely driven by total payouts—dividends and buybacks. The total payout model is found to more accurately attribute long-term returns than models based on earnings.
The study finds that total payouts per share grow in tandem with economic productivity, implying that total payouts contribute to the long-term growth of the real economy. Indeed, a significant proportion, 32% between 1980 and 2014, of the economic growth measured is attributable to buybacks, not to the growth of the underlying cash flows of the businesses.
Significantly, the authors show that the widely used DDM substantially underestimates expected returns when compared with the total payout model. Historical returns for the total payout model are 6.79%, but these fall to 6.21% for the dividend model. The discrepancy arises because the dividend model excludes buybacks and also underestimates growth that results from buy-and-hold investors rejecting cash and increasing their shares in companies.
Finally, the study finds that cyclically adjusted total yield is at least as predictive of future stock market returns as the cyclically adjusted P/E, a version of which is used in the well-known Shiller P/E. So, the volatility of buyback payouts does not, as it turns out, affect the ability to forecast stock performance over the longer term.
What Are the Implications for Investors and Investment Professionals?
Analyst forecasting accuracy is a long-standing area of controversy. This study proposes an easily implementable and helpful way to improve forecast accuracy at the stock and sector levels. The method could become another tool with which investors can attempt to predict future equity returns.
The total payout model developed and tested in this study seems to represent a viable alternative to such traditional models of stock returns as the DDM. It not only provides a more consistent framework for the attribution of returns, but it also appears to provide more accurate forecasts of future long- and short-term expected returns, which could be applied at a macro level.