UC Berkeley

In this dissertation, I explore the interactions between financial markets and real economy activities. In the first chapter, I use the evidence from an emerging market to study how the development of its financial system could affect activities in its real economy. In the second chapter, I look at excess returns in the US treasury bond market and try to understand the economic fundamentals driving the risk premia. In the final chapter, I examine corporate financing decisions using publicly traded firms in the US. The patterns in their financing decision can be partially explained by the information embedded in the financial market.

To what extent the development of sophisticated financial markets benefits emerging economies is an open question. In the first chapter, I use a unique data set on all currency derivative transactions by non-financial firms in 2006 and 2007 in Colombia to provide new evidence on one aspect of this question: the effect of participation in derivatives markets on firm capital formation. I use a difference-in-difference propensity score matching approach in order to control for self selection and common trends. I find a large positive effect: firms using currency derivatives invest on average 5.7 percent more, which is about 40 percent of their average investment rate. This investment-enhancing effect is entirely driven by firms taking long positions (i.e. dollar buying) in the derivatives market. For firms taking short positions, typically exporters, the use of derivatives does not have any discernible impact on investment. One possible explanation is the asymmetry in the impact of the exchange rate movement on exporting and importing firms.

In the second chapter, I propose a latent variable approach within a present value model to estimate the expected short rate changes and bond risk premia. This approach aggregates information contained in the history of yield spreads and short rate changes to predict future bond excess returns and short rate changes. I find that the factor from Cochrane and Piazzesi (2005) fails to predict bond excess returns when I consider different maturities of the underlying short rate. From the proposed present value model, I find a significant predictable component in short rate changes with R-square ranging from 29 precent to 80 percent, and a moderate R-square about 12 percent for predicting bond excess returns. Both expected short rate changes and bond risk premia have a persistent component, but bond risk premia are more persistent than expected short rate changes. In addition, the bond risk premia become more persistent as I increase the maturity of the underlying short rate. Finally, I explore the source of the time variation in bond risk premia, and find that monetary policy plays an important role.

In the third chapter, I document a strongly decreasing time trend in firms' leverage ratio at their IPO years over the period from 1975 to 2006. This trend survives when typical factors are controlled for, including industry fixed effect. Furthermore, I find that firms listed more recently are more adverse to debt financing. A deeper examination shows that the risk associated with firm's operation provides a limited explanation for this finding. However, the underpinnings of the observed pattern of firms' leverage ratios at IPO are still largely unresolved.