The equity-bond risk premium — the long-run expected return advantage of stocks over government bonds — is one of the biggest questions in financial markets. The extent of the premium is widely debated, but it is reasonably clear that it declined in the last quarter of the 20th century, to partly rebound in the first years of the 21st century.
Our review provides a road map to the complex literature on the topic. We explain the key drivers of the risk premium and varying assumptions about them, letting investors themselves assess the long-run prospects for stocks versus bonds. Long-term government bond yields are known, while prospective equity returns are inherently less transparent and thus more open to question.
There is an ongoing shift in opinion about expected returns. Long-term equity premiums have traditionally been predicted from historical average asset performance assuming a constant risk premium, but today they are increasingly predicted with the help of dividend discount models, assuming time-varying expected returns.
We first review the historical average returns of major asset classes and explain why we believe these are misleading guides for the future. Essentially, the double-digit returns of the 20th century were due to equities starting cheap and getting richer over time. Many investors extrapolated this past performance and expected (at least) as high future returns. Investors thus missed, first, the fact that a part of realized returns was unexpected windfalls from rising equity valuation multiples, and, second, that when starting from high valuation levels it is not reasonable to expect as high returns as in the past.