Working Papers


Abstract: I explore the incidence and duration of unemployment spells induced by foreign competition. Using German administrative data, I begin by demonstrating two basic facts. First, as German imports from China and Eastern Europe surged during the 1990s and 2000s, longer unemployment spells accounted for about a half of the import-induced earnings losses for an average worker. Second, controlling for 3-digit industry, the duration of unemployment spells became relatively higher for workers in low-wage occupations. This difference accounted for about a quarter of the heterogeneity in earnings losses between low- and high-wage occupations. I next build a dynamic search model with heterogeneous workers that explains these patterns. The model allows unemployment outcomes to be influenced by general equilibrium effects of costly worker reallocation between sectors and occupations. Calibrated to pre-shock quarterly data, it successfully predicts trade-induced relative changes in unemployment spells of workers in different occupations. Counterfactual experiments suggest that unemployed workers in sectors not directly affected by foreign competition have also experienced big changes in duration of unemployment spells.



Abstract: We investigate open economy dimensions of optimal monetary and fiscal policy at the zero lower bound (ZLB) in a small open economy model. At positive interest rates, the trade elasticity has negligible effects on optimal policy. In contrast, at the ZLB, the trade elasticity plays a key role in optimal policy prescriptions. The way in which the trade elasticity shapes policy depends on the government’s ability to commit. Under discretion, the increase in government spending at the ZLB depends critically on the trade elasticity. Under commitment, the difference between future and current policies, both for domestic inflation and government spending, is smaller when the trade elasticity is higher. 


"Reputational Effects in Sovereign Default" (2016), with Michal Fabinger

Abstract: We present a tractable, quantitative model of sovereign borrowing that delivers empirically relevant regularities, such as graduation from default, sovereign debt spreads that may be high for an extended period of time, high debt-to-GDP ratios, and high default rates. The model is an asymmetric-information extension of otherwise standard models of endogenous default on sovereign debt, with borrowing levels determined in equilibrium. Governments could be of different types based on their level of responsibility (cost of default as perceived by the politicians). Only the governments observe their level of responsibility. International investors try to infer the unobserved types based on the history of all observable actions, which gives irresponsible politicians an incentive to choose the same actions as responsible ones would. Governments could tolerate periods of high interest rates without defaulting to signal that they are of better type and to gain good reputation. This leads to lower interest rates during future recessions. For the same reason, even responsible governments should pay at first high interest rates in order to signal their type and thus “graduate from default” afterwards. A calibrated version of the model features these regularities, matches standard business cycle moments, and leads to a more realistic default rate in equilibrium, with parameter values same as in the existing literature.