Essays on household portfolio and current account dynamics

This dissertation studies the household behavior of consumption and portfolio choice, and current account dynamics. After a summary in Chapter 1, Chapter 2 develops a model of household portfolio choice between liquid and illiquid assets under income uncertainty. The model demonstrates that households optimally hold the majority of precautionary wealth in the illiquid form, even though they have to pay transaction costs in changing illiquid asset positions upon adverse shocks. I also show that the marginal propensity to consume out of illiquid wealth is much smaller than the MPC out of liquid wealth in the short run.; Chapter 3 puts the implications of the model to the test in the Panel Study of Income Dynamics. I examine the steady-state relationship between income uncertainty and illiquid portfolios. My estimation results confirm that households facing more income uncertainty hold more illiquid portfolios in the steady state. However, consistent with the fact that illiquid wealth is held for the precautionary reasons, I find that households actually reduce their illiquid asset holdings after adverse shocks.; In Chapter 4, which is a joint work with Prof. Christopher Carroll, we propose a new method to estimate the wealth effect on consumption. We argue that the recently popularized cointegration method is misguided since it relies on the unrealistic assumption that consumption, income, and wealth have a unique, linear relationship in the long-run. Alternatively, we argue that a model of habit forming consumers is a good characterization of consumption dynamics. Using the habit model, we estimate that the immediate MPC out of wealth is 0.9 percent and the long-run MPC is 4 percent.; Finally, Chapter 5 applies the consumption theory to a model of current account dynamics. Simple intertemporal models of the current account commonly fail to match the volatility of actual movements. I extend the simple model to allow for the distinction between durable and non-durable goods and find that the extended model fits the actual data better for a small open economy such as Canada, but not for an large open economy such as the U.S.