Firms face uncertainty on many different dimensions: demand level, productivity and input prices, taxes and regulations. Furthermore, some argue that uncertainty is higher in recessions (cf. Bloom et al. (2012)) and one of the causes of the slow recovery during the recent Great Recession (cf. Stock and Watson (2012) and Baker et al. (2012)). However, most trade models assume uncertainty away by considering a deterministic framework or introduce uncertainty in a very limited way.

In this dissertation, I argue that uncertainty can be particularly important for two topics in international trade: (i) firms’ global sourcing decisions and (ii) firms’ exports decision when facing multiple sources of uncertainty. Firms’ decisions to enter new foreign markets, exit from foreign markets that they are currently serving and whether to vertically integrate or outsource with foreign firms (i.e. their global sourcing decisions). Not only do these decisions require high sunk costs (cf. Roberts and Tybout (1997) and Antras and Helpman (2004)) but they are also subject to an additional set of uncertain conditions, e.g. exchange rates, foreign market conditions, and foreign policies. In particular, these potential multiple sources of uncertainty can work as an amplification mechanism, specially during recessions.

The first chapter discusses the key insights that motivates my dissertation. The second chapter develops a dynamic model of international trade with heterogeneous firms who endogenously decide when to start exporting to foreign markets, under which sourcing scheme, and when to exit foreign markets in a framework with foreign demand uncertainty. The third chapter focuses on empirically evaluating the theoretical model of the previous chapter using U.S. firm-level data. I find that integration reduces the probability that a firm exits by as much as 8%, while uncertainty increases this probability by 23%. The fourth chapter looks into the interaction between demand and policy uncertainty during the Great Trade Collapse and is joint work with Kyle Handley and Nuno Limao. We examine if the resulting change in policy uncertainty initially deepened the collapse and then helped reverse it, when the worst fears of protection were not realized.