This is a study using a unique body of expectations data collected over the decade of the 1960s. After describing the data, this paper first looks at the extent of consensus among those financial institutions providing the forecasts and measures the accuracy of the forecasts. We then ask if the forecasts are consistent with the hypothesis that tile expectations are "rational". We then turn to the relationship of the forecasts to security valuation. We develop our own variant of the popular capital asset pricing model using a framework suggested by Ross for this arbitrage model. Alternative specifications are developed relating expected returns to risk variables and relating securities prices to expectations and risk variables. We find that the expectations data of the sort we have collected do appear to influence security prices in the manner suggested by the theory. We also find that the expected security returns implied by the expectations data are related to "systematic" risk measures appropriately defined. Nevertheless, we find that, even when a variety of systematic influences are used, other risk measures, possibly related to their own variance of the securities, appear to play some role in security valuation.
