Natural experiments in corporate investments and financing

Natural experiments allow researchers to more closely approximate randomized experiments. As corporate investment and financing decisions are made jointly, I study three natural experiments and show that in these quasi-experimental settings I can make more precise statements about causality.;The first chapter examines the link between financial development and economic growth using a natural experiment from an emerging economy. Administrative foul-ups in the implementation of exchange controls in Venezuela in 1983 lead to a large unexpected credit freeze. I exploit this exogenous variation in financial development to identify the effect of financial development on plant performance. My results show that the negative shock to financial development induced by exchange controls reduced plant output and productivity, but did not affect capital stock or other inputs. This suggests that causality runs from finance to growth and that finance affects growth mostly through a productivity channel.;The second chapter investigates how the subprime credit crisis in 2007 affected corporate investment. I argue that the credit crisis is unlikely to be a good natural experiment as it was followed by a more severe economic and financial meltdown in late 2008. To better assess how the credit crisis affected corporate investment, I investigate if the subprime crisis affected discount rates by estimating a structural model. Consistent with macroeconomic trends, I find little evidence of significant changes in discount rates after the crisis. Using difference-in-differences techniques I fail to find convincing evidence that the shocks to credit affected the investment of financially vulnerable firms.;The third chapter examines whether firms react to cash shortfalls by cutting investment. We reexamine the puzzling evidence in Rauh (2006) that mandatory pension contributions cause sharp investment declines, finding that these results are likely due to the endogeneity that this study is trying to avoid. We also compare firm-year observations in which the firm's pension assets are barely less than its pension liabilities to observations in which assets are just greater than liabilities. In this quasi-experimental setting, we find little evidence that firms cut back on investment. Instead, they mostly use a variety of financial tools to fund their pension liabilities.