Date of Award

Spring 5-17-2010

Degree Type


Degree Name

Doctor of Philosophy (PhD)

Graduate Group


First Advisor

Joao F. Gomes

Second Advisor

Urban J. Jermann

Third Advisor

Andrew B. Abel


My dissertation aims at understanding the financing and investment decisions of firms. It contains two chapters.

Chapter One studies the currency composition of debt for firms in emerging economies. Using a dataset of traded Mexican firms, I document two stylized facts about firms in the non-tradable sector: (i) they take on large amounts of dollar-denominated debt and (ii) their earnings are not sensitive to the exchange rate. I propose an explanation based on imperfect competition in the domestic goods market that reconciles these seemingly contradictory empirical facts. First I develop a stylized model of production and financing for firms in an open economy. I show that non-exporting firms are exposed to exchange rate risk because of the presence of exporters in the economy, and that they hedge their currency exposure using dollar debt. An extended model is used to quantify how much of the dollar debt in the data can be explained through this channel. A calibrated version of the model can account for all of the dollar debt observed in the data.

Chapter Two investigates the relationship between the investment decisions of firms and their cost of financing. Recent empirical work using panel data documents that, while the correlation of investment and Tobin's Q is low, the correlation of investment and credit spreads is high. We propose an explanation for these empirical findings, based on time-varying risk, i.e. stochastic volatility. In our model, firms finance investments using defaultable debt as well as equity issuance, and they are subject to standard profitability shocks as well as shocks to volatility. An increase in volatility leads to an increase in the probability of default and hence the credit spread, while reducing investment and increasing equity value. This shock hence generates a negative correlation between investment and credit spreads, and between investment and Q, helping the model match the data.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."