HOW DID EXCHANGE RATES AFFECT EMPLOYMENT IN U.S. CITIES?

Abstract

I. INTRODUCTION Substantial U.S. dollar depreciation is sometimes considered a solution to the problems of the large U.S. current account deficit and declining manufacturing employment. However, existing studies report mixed evidence about the size of the effects of exchange rate movements on employment in advanced economies. Two early studies based on data from the United States, Branson and Love (1988) and Revenga (1992), did report large estimated effects. Revenga (1992), for example, finds that the U.S. dollar's appreciation in the early 1980s reduced employment by about 6%. Some later studies based on larger samples, however, report much smaller estimates of the employment effects in the United States (Campa and Goldberg 2001; Goldberg and Tracy 2000; Klein, Schuh, and Triest 2003). Klein, Schuh, and Triest (2003), for example, find that two consecutive annual 5.4% (one-standard deviation) appreciations of the cyclical component of the exchange rate reduce net employment growth in the manufacturing industries by 0.7%. Meanwhile, other recent papers report a large effect of exchange rate on employment in advanced economies, including Canada (Huang, Pang and Tang 2014; Leung and Yuen 2007), France (Gourinchas 1999; Hatemi-J and Irandoust 2006), Italy (Nucci and Pozzolo 2010), and the United States (Faria and Leon-Ledesma 2005). In these papers, a 1% appreciation is typically associated with a 0.5% to 1% decrease in employment. Almost all these previous studies focus on manufacturing industries, which are traditionally regarded as the tradable industries in an economy. Although the manufacturing sector plays an important role in the economy, its share in total employment is typically below 15% in developed countries. Consequently, if the exchange rate affects employment only in the manufacturing industries, its effect on national employment would likely be small. Meanwhile, there are a number of ways in which the exchange rate can affect nonmanufacturing industries, even if those industries have little or no exposure to international trade. For example, one such channel is a potential demand knock-on or spillover effect. If a depreciation strengthens the demand facing the domestic manufacturing industries, these industries, and their workers, will in turn demand more products and services from the domestic nonmanufacturing sector, potentially boosting its employment. In this article, we update the research on the employment effects of exchange rates with more recent data and using a novel source of variation, those arising from differences in local industrial composition. In addition, we broaden the analysis to include the exchange rate movement's indirect employment effect on the non manufacturing sector, which hires far more labor than the manufacturing industries. Specifically, we analyze the data from major U.S. Metropolitan Statistical Areas (MSAs), which we will refer to as cities. These cities have different mixes of manufacturing industries that have potentially different trading partners. In a particular year, a specific city can be subject to larger or smaller exchange-rate changes than other cities because of different industrial compositions. We will exploit this variation of the exchange rate across cities to estimate their effects on manufacturing jobs. To estimate the exchange rates' spillover effects on nonmanufacturing industries, we exploit nonmanufacturing products' relative lack of tradability. The key assumption is that the demand for nonmanufacturing products is, in part, local. Hence the economic misfortune of a city's manufacturing industries has negative impacts on local nonmanufacturing industries. The local connections could arise because the manufacturers source some of their inputs locally (think construction, maintenance, or other business services that cannot be provided remotely), or because those manufacturers' employees must consume locally produced goods and services like housing, retail, and restaurant services. …