Doctoral Dissertation Research in Economics: Behavioral Responses to Wealth Shocks: Evidence from Sweden

Harvard University, Cambridge MA

Investigators

Abstract

How do workers prepare for retirement after an unexpected change in their financial situation? How do sudden wealth changes affect earnings and health before and after retirement? Do people work more and become more stressed after unexpected losses in their portfolios? Do people spend down new-found wealth quickly, or do they save it? To what extent do people help each other out in the family? This study addresses these questions empirically with three novel research designs on Sweden's complete population registries. The project is the first to use this data for this purpose, a singular opportunity for any major developed economy in particular. The main design uses capital gains calculated from exceptionally detailed portfolio data in the universe of Swedish tax returns. Riskier assets should generate higher average returns, safer investments lower ones, but realized returns above or below these risk-specific levels constitute unpredictable shocks to one's wealth and savings. Auxiliary designs use sharp cutoffs in pension and inheritance tax policies as clear shocks to wealth. All designs lead to investigations of outcomes like the age of retirement, health and well-being, as well as what one earns, consumes or saves. Even more, the causal effects of wealth on labor supply or consumption can also be documented for family members -- a very important channel which has never been possible to study using the complete family network of an entire country before. This work advances our understanding of how human decisions change with wealth. Both the nature of the data and the shocks investigated surpass previous investigations using surveys: The effects are established in a much broader population, with more precise data (on earnings or consumption), and, in particular, on novel outcomes (like health). Most importantly, the rich data does not only offer precision, but also novel research designs with greater internal validity, i.e. clean comparisons of situations where the variation in wealth is close to random. The fact that this work can analyze negative shocks to wealth makes it stand out from prior (or concurrent) work on lottery winnings or inheritance. It is particularly important to establish whether people react to losses differently from commensurate gains. Also, people can think about changes to their savings unlike windfalls of equal size. Moreover, the study is among the first to properly take into account spillover effects where help can be important within the family, viz. "private insurance" of savings or health. A study with equally reliable data on the basically complete network of families in a country is an important advance in itself. Many results can have broader implications. For instance, if retirees are much healthier with some extra wealth, more powerful retirement saving schemes might look justified. On the other hand, if wealthier people leave work more easily, a means-tested retirement age might raise welfare. If family help still seem prevalent even in a generous welfare state, both policy proposals and other empirical studies can be expected to make bigger efforts to take such spillover effects into account. The wider research and policy community also cares about people responding to incentives depending on whether they or their family members are more or less wealthy.