Essays in Behavioral Household Finance

Abstract

This dissertation investigates some of the factors affecting modern household finance decisions in the United States using natural experimental variation and administrative data. In Chapter 1 I estimate the effects of financial education on retirement savings decisions. Between 2007 and 2008 the U.S. Army implemented a mandatory 8 hour Personal Financial Management Course (PFMC) for new soldiers. Staggered implementation across locations and time provides quasi-experimental variation in whether an individual received the training. I find that the course has large and lasting effects on individual retirement savings in the Thrift Savings Plan, a tax-deferred account similar to a 401(k). The course doubles savings, has significant effects throughout the distribution of savings and the effects persist out to two years. The mechanism for the effects is likely a combination of both human capital and behavioral assistance. In Chapter 2 I estimate the effects of financial education on a variety of other economic behaviors. I rely on the same natural experiment as in Chapter 1 but I use individually matched credit data to estimate the effects of financial education on credit scores, credit balances for several types of accounts, monthly payments and adverse legal actions. In some areas I find that the PFMC has beneficial effects, reducing cumulative account balances (especially for automobile accounts) and aggregate monthly payments. In other areas, including credit scores, the probability of being active in the credit market and the number of adverse legal actions, the PFMC has no statistically significant effects on financial behavior. In Chapter 3 I estimate the effects of stress on financial decision-making. I use the natural variation in the casualty rates faced by individuals deploying overseas an exogenous source of stress and I measure the effects of this stress on individuals' participation in the Savings Deposit Program (SDP), a risk-free 10% annual percentage rate savings account. I find a modest and statistically significant negative relationship between the stress of casualties and SDP participation on the order of 5%. Some failures of the randomization test and the confounding effects of overall activity levels and rural locations cannot be eliminated as a source of the observed savings differences and as a result, these results should be considered suggestive evidence of the adverse effects of stress on financial decision-making.