An empirical test on the determinants of households' stock market participation decision and portfolio composition.

Abstract This paper uses the U.S. Panel Study of Income Dynamics household level data to examine the factors that may affect stock market participation and portfolio composition. Three types of background risks sourced from labor income, business income and home equity, as well as some households' demographic characteristics are examined. It finds that all these background risks are statistically significant in determining households' stock market participation and portfolio composition. In addition, the effects of the correlation between these background risks and stock return, risk free rate, and the correlation between any two of the three background risks are also examined. It finds that, when the background risk is highly positively correlated with the stock return, or highly negatively correlated with the risk free rate, it discourages households' stock market participation and stock holdings; whereas when it is highly negatively correlated with stock return, or highly positively correlated with the risk free rate, it encourages stock market participation and stock holdings. However, the effects of these correlations vary in terms of statistical significance. In addition, households' stock market participation and stock holding is also affected by various households' demographic characteristics, such as age, education, wealth, labor income, business income, house ratio and mortgage ratio