are efficient, if all investors have identical information, and if investors maximize the expected return in their portfolios and minimize volatility, the expected excess return on the market portfolio is M-m = E oi W /,A (5.i; That is, the market portfolio's expected return over the riskless asset is the market portfolio's variance divided by the average across all market participants of the ratio of their wealth to their risk aversion. Unfortunately, to most of us this formula reveals no intuition whatsoever. However, we all agree on the concept of an equilibrium expected return to compensate investors for taking market risk. The difficult question is, how large is the market return premium? It's clearly not zero or negative, as investors extract a price in order to bear volatility in their wealth. On the other hand, it's probably not 10 percent per year above the riskless asset, because the market's volatility is of the same magnitude, which implies that holding the market causes a relatively small probability of negative return over one year, and even less than this at the end of five years. In this chapter, we attempt to arrive at a reasonable range for the market risk premium over the riskless asset. More specifically, we evaluate estimates of the equity risk premium (ERP), from which the market risk premium is easily derived using the CAPM. We consider two approaches to measure the ERP. Our first is purely empirical: We study the average returns of equity markets over long periods of time. In addition to looking at long-run averages, we also look at decompositions of these averages, in the hope that they will provide insights into the drivers of equity returns. Our second approach is more theoretical. In this approach, we look at the theoretical relationship in equilibrium between investor demand and asset supply. In particular, we are interested in exploring the role of investor preferences in shaping the equity premium.