Essays on debt, financial crisis, and impulse response functions

The first two chapters of my dissertation study the impact of private external debt on the economy, both theoretically and empirically. The third chapter is focused on an econometric method that is of central interest in empirical macroeconomics---impulse response functions.;The first chapter analyzes the effects of a shift from sovereign to private sector borrowing on credit costs, default frequencies, aggregate debt levels and country welfare, using a theoretical model with decentralized borrowing and centralized default. Relative to a model with centralized borrowing, decentralized borrowing drives up aggregate credit costs and sovereign default risk, and reduces aggregate welfare. Private agents do not internalize the effect of their borrowing on economy-wide credit costs and tend to overborrow. Depending on the severity of default penalties, decentralized borrowing may lead to either too much or too little debt in equilibrium.;The second chapter empirically studies the effects of firms' foreign debt and other characteristics on firm performance in financial crisis. Using Korean firm-level data on publicly-listed and privately-held firms together with firm exit data, we find strong evidence of the balance-sheet effect for small firms at both the intensive and extensive margins. During the financial crisis, small firms with more short-term foreign debt are more likely to go bankrupt, and experience larger declines in sales conditional on survival. The extensive margin accounts for a large fraction of small firms' adjustment during the crisis.;The third chapter studies whether the choice of different methods of constructing impulse response functions leads to different conclusions about the responses of monetary policy. The impulse response confidence intervals based on local projections (LPs) tend to be less accurate and wider than those based on standard vector autoregressive (VAR) models. The accuracy of joint intervals can be more erratic in practice than typical pointwise confidence intervals. For a standard model of monetary policy, the use of LP estimates or joint intervals does not overturn the substantive findings of VAR studies based on pointwise intervals.