Empirical studies of portfolio choice and asset prices

Sammanfattning: This thesis contains empirical studies of portfolio choice and asset prices. The first two chapters deal with incorporating labor supply into models traditionally only focusing on consumption. Can the risk premium on stocks be better understood when taking labor supply into account? This is the topic of the first chapter. Do possibilities of varying labor supply, and thus hedging stock market risk, help explain the stock ownership patterns of households? This question is what the second chapter tries to answer. If labor income moves with the stock market, an attempt should be made to hedge this with a lower share of stocks in the portfolio and, but do households act according to this rule? This is what the third chapter investigates. Chapter one, Labor Supply Flexibility and Portfolio Choice: Evidence from the PSID, examines the relationship between labor supply flexibility and portfolio choice. Theoretical articles have shown that, ceteris paribus, the optimal portfolio share of risky assets (stocks) increases with labor supply flexibility, due to increased possibilities of hedging financial risk by adjusting the labor supply. Using PSID household data, this hypothesis is tested using a direct measure of labor supply flexibility from survey questions. The results indicate that the total portfolio share is increased by labor supply flexibility. When separated, most of this effect seems to come from the increased probability of stock ownership due to flexible labor, rather than an increased portfolio share among stockholders. Chapter two, Can Leisure Explain the Equity Premium Puzzle? An Empirical Investigation, investigates the asset pricing properties of non-separable utility functions with consumption and leisure. The parameter restrictions needed to match the historical equity premium are explored using US data on consumption, hours and returns. Empirically, it is shown that to match the equity premium with a low level of risk aversion, consumption and leisure need to be strong complements, i.e. have a very low substitution elasticity. Chapter three, Income Risk and Stockholdings: Evidence from Swedish Microdata, examines the relationship between income risk and portfolio choice. It empirically investigates whether the stock market risk (the covariation with the stock market) in labor income is reflected by an offsetting lower share of stocks in financial portfolios, an effect that has been shown to exist in theoretical articles. Swedish microdata from HINK on households’ income and wealth are used for this purpose. In repeated cross-sections, households are divided into "portfolio cohorts" corresponding to percentiles of the share of stocks in financial assets. Income risk, i.e. the regression beta of (log) income growth on aggregate stock returns, is compared for the different groups. As predicted by theory, the results provide some support for a negative relationship between income risk and the share of stocks.