In their May 2005 paper entitled “The Market Equity Risk Premium”, Brian McCulloch and Dasha Leonova present a comprehensive review of equity risk premium research to support decision-making regarding the annual capital contribution to New Zealand Superannuation Fund, a government-managed pension fund. They seek the best estimate of the future annual premium of nominal long-term equity returns over nominal long-term bond returns. Based on international experiences and forecasts over many decades, they conclude that:
“The traditional view has been that realized US market returns over the past seventy-five years (giving an average equity risk premium over long-term bonds in the region of 7%) provide an unbiased estimate of the expected future long-term equity risk premium. This view has given way over the past few years to a consensus that the future expected risk premium is actually somewhat lower.
“This view of a declining expected long-term equity risk premium reflects a range of recent empirical evidence and theoretical analysis. Richer evidence on historical returns has been presented by analysing longer time series of US capital markets data and by looking over multiple international capital markets. Additional historical information on dividends, earnings and the variability of returns has also been harnessed. Various methodologies have been adopted to analyse this historical record. The results from these studies now consistently suggest that the expected long-term equity risk premium sits in the range of 0% to 7%. The theoretical result that the traditional view was not consistent with that implied by standard economic models of consumption, utility and risk aversion (the “equity premium puzzle”) has helped bolster the emerging view of a lower expected equity risk premium. However, the very low (less than 1%) expected equity premium implied by the theoretical results is not now seen as having strong normative implications.
“Two potential limitations need to be taken into account in interpreting the historical analyses. First, although seventy-five or more years of data across several capital markets may seem like an abundance of data, it is not actually enough to allow very statistically precise estimates to be made. This is because of the high volatility of capital markets. Second, the external validity of the historical record for predicting future expected long-term capital market behaviour is of question. Recent all-time high equity prices have generated broad-ranging speculation about future capital market behaviour. In addition, trends in capital markets over time (including new institutions, declining transaction costs, and a widening pool of investors) along with macroeconomic developments (including changing inflation and GDP growth expectations, and globalisation) and impending demographic changes provide the possibility that future capital market behaviour may well deviate substantially from past experience, resulting in a general decrease.
“Although the long-term expected future equity risk premium is not directly observable (and hence the past emphasis on the historical information that is available), there is other forward-looking information that is informative about future expected capital market behaviour. These include survey approaches, in which expectations are elicited directly from market experts, and approaches that infer the expected equity risk premium implied by analysts forecasts of dividends and earnings. The most recent survey results are providing median estimates of the long-term equity risk premium in the 2% to 5% range. This represents quite a decline over the past few years. A year or so ago, a median in the 5% to 7% range was indicated. However, by their nature, survey results are not necessarily very diagnostic and should be treated with caution.
“Market analysts routinely put considerable effort into making forecasts of firms’ expected future earnings performance. Estimates of the market equity risk premium implied by these forecasts are therefore a potentially strong source of information about market expectations. Research using analysts’ earnings forecasts has gone through an evolution starting with straightforward applications of the dividend growth model. These initially gave an expected equity risk premium similar to the 7% historical result. However, the assumed long-term earnings growth rate has a crucial impact on the result in this model and there was some doubt about the realism of analysts’ growth forecasts when seen in the context of expected growth in the economy as a whole. Later applications using analysts’ near-term forecasts but lining long-term growth up with expected growth in the economy obtained estimates that were more in the 3% to 5% range. A more recent innovative development has been the application of the residual income model in place of the dividend growth model. Although isomorphic to the dividend growth model, the residual income model has the advantage of being based on the underlying earnings that analysts forecast, rather than using those forecasts as a proxy for dividends. A few approaches have been adopted applying the abnormal earnings methodology, with the results generally coming out lower than the earlier historical-based approaches, with range for the expected equity risk premium of 2.7% to 5.3%.
“… In summary, the traditional view based on historical returns of an expected equity risk premium of 7% has given way over the past few years to a consensus that the long-term future expected equity risk premium is somewhat lower. We believe that the long-term annual (arithmetic) expected future equity risk premium now sits in the range of 3% to 5%. In forming this view, we found the research based on analysts’ earnings forecasts with a residual income model to be the most persuasive, and those results are broadly consistent with the latest research using historical data. While the survey results also happen to be broadly consistent with this as well, we put less weight on them. The risk premium we have considered is a premium for a world portfolio, which is largely dominated by the US market.
“For the purpose of calculating the required capital contribution to the New Zealand Superannuation Fund, the Treasury is adopting the assumption of a long-term annual (arithmetic) expected future equity risk premium of 4%. In the normal course of events, we would expect this assumption to stay stable over long periods of time.”
If the authors are right and if U.S. Treasury note yields remain the the 4%-5% range, expect single-digit long-term returns from stocks.
This paper is rich in citations to relevant academic research on the equity risk premium.