Ebook Trade Adjustment in Large Crises
Episodes of large crises such as the Argentine crisis in 2002, the East Asian crisis in 1997-98, and the Mexican crisis in 1994 are associated with large reversals in the trade balance. The dollar value of imports in Argentina, for instance, dropped by 80 percent and the trade balance climbed from -1 to 18 percent of GDP between 2000 and 2002. A second striking feature of these episodes is the large measured productivity decline associated with the large declines in output.
Despite the dramatic nature of such events, little is known about the mechanics of trade adjustment at the firm and product level. Consequently, we lack a detailed understanding of exactly what are the costs associated with trade balance reversals. Consumer welfare might deteriorate due to the reduced number and quality of imported final good varieties. Additionally, as described in the influential works of Romer (1990) and Romer (1994), measured productivity can decline if if producers switch from foreign to domestic intermediates when they are not perfect substitutes.
A large and growing theoretical literature in trade and macro argues for the importance of the extensive margin, whether defined as changing firms or changing products at business cycle frequencies. Ghironi and Melitz (2005) build a model to highlight the impact of firm entry and exit decisions on business cycle moments. Corsetti, Martin, and Pesenti (2008) explore the role of the extensive margin in understanding current account adjustments. A relevant question then is how empirically important is the extensive margin at business cycle frequencies? A dramatic crisis event can help shed light on this question.
This paper aims to explore the importance of intensive and extensive margin trade adjustment during crises and the macroeconomic implications of their impact on domestic productivity. To do this, we evaluate detailed firm-level customs data covering the universe of import transactions for Argentina over the period 1996-2008. This period includes a dramatic nominal exchange rate depreciation and trade balance reversal. We combine this trade data with operating and financial information on the largest Argentine firms. One of our main findings is that the extensive margin, when defined as the entry and exit of firms into import status or when defined as the inclusion or exclusion of product categories (HS6 or HS10), plays a small role in understanding import adjustment during the crisis. However, at the firm level, the extensive margin plays an important role as firms drop products that constituted a significant share of their previous imports. The weighted average of Feenstra correction measures for the firm level imported input cost index ranges from about 1.04 to 1.09, implying a significant decline in measured productivity. We evaluate the trade adjustment patterns using a general equilibrium model with heterogenous firms that import differentiated varieties of inputs. Firms decide whether to import and, if so, which range of input varieties to import. We show that some of the predictions of the model are consistent with evidence in the data and imply that measured productivity declines for importing firms.
To elaborate, the extent of the external margin of adjustment is strikingly large when measured in terms of the number of firms or products. The number of importing firms dropped from over 15,000 to less than 7,000 in a span of only four quarters. The number of distinct 10-digit Harmonized Tariff Schedule (HTS) product codes imported dropped from about 13,000 to 10,000 over the same period. Measured by the volume of trade, however, country-level extensive margin adjustment is strikingly small. On average, the net contribution of new importers entering and old importers exiting trade accounts for only about 10 percent of the time-series variation in total import volume. This finding holds when looking at the quarterly or annual frequency, when defining the extensive margin by the changing set of firms or by the changing set of imported 10-digit HTS categories of goods, when looking at normal times as well as during the crisis and recovery, and when separately considering each end-use category.
Posted in :
