Participation Effects in Household Financial Decisions

This dissertation consists of two essays investigating participation effects in household financial decisions. In the first essay, entitled "Household Mortgage Choice and Mortgage Market Participation," I empirically study a household's choice of an adjustable rate mortgage (ARM) over a fixed rate mortgage (FRM) across time. This decision has been investigated in the cross-section previously, but to date, no one has studied how a household's choice of mortgage contract type changes as they gain experience in the mortgage market. This study investigates whether mortgage market participation has a systematic effect on the choice of an ARM vs. an FRM within a household. Using the Panel Study of Income Dynamics (PSID) and the Survey of Consumer Finances (SCF), I document a novel stylized fact: a household's propensity to choose an adjustable rate ARM over an FRM increases with the number of previous mortgages the household has used. Households do not choose an ARM due to budget or liquidity constraints when increasing housing consumption; nor is the observed pattern of increased propensity to choose an ARM with mortgage market participation explained by the simultaneous relaxation of budget constraints as homeowners participate in the mortgage market. Stabilization of a household's income stream and rising home prices are also ruled out as the source of increasing ARM choice propensity with greater utilization of mortgages, as is expected length of tenure. Evidence is presented supporting the hypothesis that households learn about mortgage products by participating in the market.

In the second essay, entitled "Participation Effects in Refinancing Decisions", I investigate household refinancing decisions in the context of market participation. Using optimal refinance interest rate differentials as derived in Agarwal, Driscoll and Laibson (2013), I document an important participation effect in the Panel Study of Income Dynamics, whereby households with greater mortgage market participation, as measured by previous mortgages used, are more likely to refinance optimally. This result is robust to potential liquidity constraints, where the household fails to refinance due to an inability to pay any fixed costs associated with the transaction. Participation effects persist even when controlling for the potential of equity extraction as the primary motivation for refinancing. These results are consistent with an information acquisition model, whereby households gain knowledge and understanding of financial transactions by participating in financial markets.