Investing in equities hopefully provides investors a return that exceeds the risk free rate of return. The difference between the two is also known as the equity risk premium, or ERP. The average long-term ERP exceeds the level that classical equilibrium theory predicts. Mehra and Prescott [1985] showed that for the US in the period 1889-1978 the ERP has been in excess of 6% per annum. In a consumer CAPM framework this corresponds with a level of relative risk aversion of 26. This is far higher than experimental theory predicts; values between 1 and 10 can be seen as normal within this framework. Being incapable of associating the measured risk with the observed return, Mehra and Prescott dubbed the phenomenon the equity premium puzzle. Ever since, the issue caught the attention of academics as well as practitioners and spawned a whole new literature based on two classes of explanations for the existence of the ERP puzzle: theoretical and empirical. This paper is a contribution to the empirical research.
From the empirical side Siegel [1992] extends the Mehra-Prescott sample to 1802-1990 and observes an ERP of 5.3% per annum over the entire period. Moreover, early stock returns did not exceed the risk free rate of return by nearly the same magnitude as they did in recent data. Brown, Goetzmann and Ross [1995] and Goetzmann and Jorion [1999] suggest the high equity premium in US equities to be the exception rather than the rule. They hint at the issue of survivorship bias by only observing the US, a clear survivor in a turbulent century. Blanchard [1993], Fase [1997] and Dimson, Marsh and Staunton [2001] show the robustness of the puzzle by studying a number of developed countries. Fase finds the theory to be even more at odds with reality for Belgium, France, Germany, The Netherlands, and the UK in the post-war period compared to those in the US. This, however, still does not completely resolve the issue of survivorship bias.