Abstract

In my first essay, I study whether U.S.-Mexico dual residents borrow in pesos or dollars, using both theory and survey data. I attempt to understand the microeconomic and behavioral foundations of the actual and theory-predicted currency choice by households. I also attempt to reconcile the export- and import-related theory of currency choice proposed in the corporate finance literature with behavioral tests. Finally, I test the nominal uncovered interest parity condition established in McBrady & Schill [1] in a household setting by analyzing if there can be any cost savings from borrowing in dollars over pesos and vice versa. Theoretically, I show that agents can have loan cost savings and maximize their utility of consumption by allocating their borrowing to either dollars or pesos based on what currency they consume in, currency path expectations, interest rates, and relative risk aversion. I find that generally, survey participants prefer to borrow in dollars, which is likely driven by the income and spending matching motivations and the country of primary residence and employment. This behavior is similar to corporations matching the currency of their revenue from exports and invoicing to the currency of debt. However, contrary to the theoretical predictions, where more risk averse individuals should be borrowing in dollars, I observe that higher relative risk aversion decreases the propensity of borrowing in dollars. This result is likely driven by the fixed vs variable rate survey design. Additionally, I find evidence that survey participants who mostly earn in pesos and are highly risk averse have a 6 p.p. lower indifference rate than those who are majority dollar earners. In my second essay, I develop a model following the Jensen & Meckling [2] model of leverageinduced risk shifting. My theoretical model explains why sellers with higher LTV ratios list their properties at higher prices. In the residential real estate market, sellers with different unobservable characteristics experience different outcomes with regards to their time to sale. One such characteristic is the seller’s loan-to-value ratio. Empirical evidence shows that the higher the seller’s LTV ratio is, the longer it takes them to sell their property. However, there is a tradeoff between the time to sale and the final sale price that the seller receives for their house. This relationship can be interpreted as a risk-shifting problem where the homeowner increases the risk associated with the house sale by increasing the listing price, which results in a longer time to sale. As a result, the homeowner enjoys a larger upside (higher proceeds from the house sale) in case of success, and the bank bears a higher downside in case the house is not sold.

Committee Chair

Michaela Pagel

Committee Members

Deniz Aydin; Hong Liu; Maarten Meeuwis; Nishant Vats

Degree

Doctor of Philosophy (PhD)

Author's Department

Finance

Author's School

Olin Business School

Document Type

Dissertation

Date of Award

5-5-2026

Language

English (en)

Download

Included in

Finance Commons

Share

COinS