Beliefs about the future self's financial conditions and preferences are important when deciding how much to consume and leave for tomorrow. This thesis seeks to answer how the perception regarding future behaviors affects consumption and savings decisions and how the heterogeneity in the perception can explain wealth inequality and excess sensitivity of consumption to temporary income shocks. In Chapter 1, I consider households with an imperfect perception of expenditure shocks. Households may underestimate the expenditure shocks in the future and then undersave than the case they had a belief consistent with truth. The model requires many households who underestimate future shocks to match the distribution of liquid wealth found in the data. Next, Chapter 2 (coauthored with Anujit Chakraborty, Claudia Cerrone, and Leonhard Lades) estimates degrees of present bias and sophistication in the quasi-hyperbolic discounting framework. We find a significant degree of present bias in the effort domain using an online experiment. Also, the results suggest a large dispersion of present bias and sophistication exists among the experiment participants. Finally, Chapter 3 incorporates distributions of present bias and sophistication elicited in Chapter 2 into a life-cycle framework to explain the dispersion of wealth.
Chapter 1 introduces a life cycle model of consumption and savings where households face exogenous expenditure shocks. Households are heterogeneous as they have different levels of perceptions of expenditure shocks. My model predicts that households who underestimate the expenditure shocks tend to spend more now and save less for the future. Using this model, I calibrate the distribution of the perception of future shocks to match the dispersion of liquid wealth in data. Based on a realistic level of liquid wealth over the life cycle, the model features many households underestimating the future expenditure shocks, which generates a high marginal propensity to consume overall. I also provide a policy recommendation that can enhance overall welfare.
Chapter 2, joint work with Anujit Chakraborty, Claudia Cerrone, and Leonhard Lades, estimates the degrees of present bias and sophistication over effort and money using a simple yet novel experimental design. We find a significant degree of present bias in the effort domain but not in the money domain. However, we find a significant correlation between the estimates of present bias across effort and money domains. Furthermore, we find that subjects are partially sophisticated in the effort domain, though it is insignificant at the aggregate level. Lastly, we find a severe dispersion of present bias and sophistication in both effort and money domains.
Chapter 3 investigates the role of heterogeneous present bias and sophistication in a life-cycle model with both liquid and illiquid wealth. Using numerical simulations, I confirm the findings of previous literature that it is hard to draw monotonic relationships between consumption and the level of sophistication. However, under some conditions, such as binding borrowing constraints, I show that agents with a higher degree of sophistication tend to prefer commitment and save more illiquid assets in a stylized setting. Finally, I incorporate the distribution of present bias and sophistication measured in Chapter 2 into a more realistic life-cycle setting and show that heterogeneity of present bias and sophistication can lead to a severe dispersion of wealth and high dependence on borrowing.