Three Essays in Microeconomic Decision Making: Experimental and Empirical Studies
AuthorSafford, Amanda Harker
StatisticsView Usage Statistics
Chapter 1 - Does experiencing a Crash make all the difference? An Experiment on the Depression Babies HypothesisThe depression babies hypothesis holds that people who grew up in the depression take fewer financial risks because of the negative returns experienced. This experiment tests this hypothesis and finds that subjects who experience a great crash hold, on average, 6% less of their assets in stocks than subjects who did not experience the crash, after controlling for gender, employment status, and financial literacy. Our results suggest that subjects who experience a significant market crash have lower and more volatile beliefs of future stock returns, and it is the change in beliefs, rather than an increase in risk aversion, that is driving behavior. Further, we find that experiencing a crash causes a significant difference in the overall belief distributions between the two groups, with the depression babies cohort holding more realistic beliefs about future stock market returns.Chapter 2 - The Influence of Social Comparisons on Risk Taking: An Asset Allocation ExperimentSocial comparison research has shown that individuals compare themselves to others in a number of domains, which can alter risk appetite and risk aversion. This set of experiments aims to test how social comparisons influence risk taking in a financial domain. We study the effect of providing three types of performance information to subjects: 1) their ranking and performance compared to a group of peers; 2) the decisions and performance of a group expert; and 3) the performance of a portfolio with a 100% allocation to stocks. We find that as the comparison standard increased, subjects' systematically allocated a higher percentage of the portfolio to a risky (stock) investment compared to a control group. Chapter 3 - Big Data for Small Businesses: Using Econometric and Machine Learning Techniques to Predict Restaurant RevenueAs the ability of businesses to capture, store and analyze data has increased, there has been a large increase in the use of data to inform decision making. Recent research has advocated an interdisciplinary approach to data modeling and analysis. We use both econometric and machine learning data analysis techniques to build a model of revenue for a small chain of restaurants in the Reno area. Benefits and drawbacks of both methods are discussed. This paper then looks at how using small business data may improve current revenue forecasting methods at the local government level.