Estimating the Impact of the Exchange Rate on the Trade Balance and Jobsby William R. Cline | November 1st, 2010 | 09:49 am
As pressure mounts on China to let the value of its currency appreciate, a debate has occurred among some economists over the potential effect of such a revaluation on the US current account deficit—or on the number of jobs in the United States. In recent congressional hearings, C. Fred Bergsten (2010), director of the Peterson Institute for International Economies, testified that a correction of China’s exchange rate undervaluation would produce approximately 500,000 American jobs. The purpose of this note is to set forth calculations that provided part of the basis for this estimate, in particular, and to further clarify estimations of the impact of the exchange rate on trade and jobs more generally.
Under circumstances of full employment, changes in the trade balance affect the composition of employment, but not the total number of jobs. An increase in exports and reduction in imports means a shift of workers from services and other “nontradables” into the tradable goods sectors. Any incipient increase in the total number jobs as a consequence of a rising trade surplus would tend to spur monetary correction aimed at avoiding overheating, curbing employment in nontradables.
However, the United States is not at full employment. Unemployment stands at 9.6 percent, and the IMF (2010a) projects that even by 2014 it will still be as high as 8.8 percent, far above the 2007 level of 4.6 percent. Moreover, job losses have been disproportionately concentrated in manufacturing. From 2007 to September 2010, manufacturing employment fell from 13.9 million to 11.7 million, or by 15.9 percent (BLS, 2010). Only construction employment fell proportionately more (by 26.6 percent, from 7.6 million to 5.6 million). For all other sectors, the decline was only 2.7 percent (from 116.1 million to 112.9 million). Because export jobs, and jobs in import substitutes, tend to be in the manufacturing sector, the high incidence of job loss in manufactures is an additional reason, beyond the high overall unemployment rate, to judge that under current circumstances a reduction in the trade deficit would translate into additional jobs.
Cline (2005) develops a current account model of the “Massachusetts Avenue” variety (so-named for the location of research organizations in Cambridge and Washington), in which exports depend on the exchange rate (with a lag) and foreign GDP growth, imports depend on the exchange rate and domestic GDP growth, and transfer payments as well as income receipts and payments on foreign assets and liabilities are taken into account.1 This model finds that a 10 percent effective depreciation of the dollar reduces the current account deficit by 1.4 percent of GDP after 3 years and about 1.6 percent after 5 years (p. 96). Specific simulations also indicate that about 80 percent of the change is from a reduction in the deficit on goods and services, with the remainder reflecting changes in payments on international assets and liabilities. With present US GDP at about $14.5 trillion, a change of 1.5 percent in four years would amount to $218 billion in the current account, and $175 billion in the trade balance (80 percent of the total change in current account). In other words, a 1 percent effective depreciation in the dollar translates to a reduction of about $22 billion in the current account deficit, with the great bulk occurring in the balance of trade in goods and services.
In terms of the observed changes in trade values, most of an improvement in the trade balance will occur on the export side rather than in a reduction in the dollar value of imports. The reason is that for imports the price and quantity effects work in opposite directions. Depreciation of the dollar boosts the import price and in response there is a decline in the quantity of imports demanded. These two effects approximately offset each other.2 In contrast, a more competitive dollar boosts the quantity of exports as foreign demand responds to the more attractive price, at a dollar price that is (largely) unchanged.3 So the nominal dollar values will tend to show little change on the import side and sizable increase on the export side as the response to a significant depreciation in the dollar. Thus, from 1985 to 1987, the effective real value of the dollar fell 19.6 percent (Federal Reserve, 2010). Allowing for a two-year lag from the exchange rate signal to the trade outcome, from 1987 to 1989 the dollar value of exports rose by a remarkable 40 percent in response, whereas the dollar value of nonoil imports of goods and services rose by only 15.5 percent.4 The strong adjustment thus showed up mainly on the export side.
When the focus is on jobs, however, the adjustment occurs substantially on the import side as well because it is the real value of imports rather than the nominal dollar value that matters for production of import substitutes and employment needed for that production. Cline (1989, p. 360) estimates that under the most likely combinations of import and export “elasticities” (response of volume to price) and “pass-throughs” (proportion of the exchange rate change that is allowed by producers to reach markets rather than being offset by strategic behavior of “pricing to market”), the ratio of the “real” to the “nominal” change in trade balance from exchange rate depreciation amounts to about 1.6 to 1. So the real GDP change associated with a 10 percent dollar depreciation would reach about $280 billion ($175 billion x 1.6), or almost $30 billion change in the real trade balance per percentage point change in the effective exchange rate.
The Commerce Department (Commerce 2010) estimates that in 2008 US exports of goods and services of about $1.7 trillion accounted for 10.3 million US jobs, a rate of about 6,000 jobs per billion dollars. At this rate, a change of $280 billion in the real trade balance as a consequence of a 10 percent depreciation would translate to 1.68 million additional jobs.5 The first central estimate of this note, then, is that a 10 percent real effective depreciation of the US dollar generates about 1.7 million trade-related jobs. These jobs are gross gains under more usual circumstances, and they are also likely to be net job gains under conditions of high unemployment.
Because the most important currency misalignment in the global economy today is that of the undervalued Chinese renminbi, it is important to consider the corresponding magnitude of US job effects from an appreciation of the renminbi. Cline (2010) conducts statistical tests showing that China’s current account responds to its exchange rate, and that the US-China bilateral trade balance responds to China’s exchange rate as well. A 10 percent rise in the value of the renminbi leads to an improvement in the US trade balance by an amount ranging from $22 billion to $63 billion, under a variety of alternative calculations. The central value of $40 billion represents about one-fifth of the trade balance change from a 10 percent overall depreciation of the dollar (the $175 billion identified above).6 So the impact of a 10 percent rise in the renminbi on US jobs would amount to about one-fifth of the 1.7 million jobs created by a 10 percent overall depreciation of the dollar. About 350,000 US jobs ( = 0.25 x 1.7 million) would thus be created as a consequence of an appreciation of the renminbi by 10 percent.
Cline and Williamson (2010) calculate that the effective exchange rate of the renminbi needs to rise by about 15 percent in order for China to reduce its current account surplus from the 2015 level of 8 percent of GDP projected by the IMF to a benchmark of only 3 percent of GDP in an environment in which there are concerted international efforts to reduce global imbalances. This estimate is within the approach of Fundamental Equilibrium Exchange Rates (FEERs) developed by Williamson and implemented with current account targets of ±3 percent of GDP. On this basis, the number of US jobs that would be expected to be created by correction of the undervalued renminbi would amount to 350,000 x 1.5 = 525,000.
Two myths have stood in the way of recognition that an appreciation of the renminbi could improve the US trade balance and create US jobs. The first is that the United States no longer manufactures anything imported from China, so if China’s exports did decline they would simply be replaced by imports from other countries. This characterization reasonably applies to some products such as footwear, television sets, toys, and sporting goods.7 However, more generally it is inaccurate even in some industries typically perceived as the domain of China and other developing countries. In the important case of apparel, in 2008 US imports from China at $36 billion in 2008 only modestly exceeded domestic production at $30 billion. For many other sectors with sizable imports from China, it turns out that US production is much larger than these imports, debunking the myth of the lack of domestic productive capacity to replace them. Domestic production dwarfs imports from China in semiconductors ($60 billion versus $2 billion), textiles ($58 billion versus $2 billion), automobile tires ($17 billion versus $2.0 billion), furniture ($49 billion versus $13 billion), fertilizer and pesticides ($40 billion versus $0.8 billion), basic chemicals ($98 billion versus $4 billion), iron and steel products ($124 billion versus $4 billion), and many other sectors.8
The second myth is that China’s exchange rate does not affect its trade with the United States because in the period when it strengthened, from 2005 to early 2008, the imbalance with the United States widened instead of narrowing. As shown in Cline (2010), this interpretation fails to recognize the lagged effect of the exchange rate in driving trade. Cline (2005, pp. 77–78) provides statistical evidence that US trade responds with approximately equal weights to the exchange rate in each of the two prior years, rather than the current year. The empirical literature typically shows exchange rate lags centered around two years for many countries (see, for example, Bayoumi 1999). For US-China trade, the tests conducted in Cline (2010) specifically found greater statistical explanatory power when including not only the previous year’s exchange rate but also that of the year before. Once the lags (and the time trend toward a widening US-China bilateral deficit) are taken into account, the exchange rate is indeed shown to influence US-China trade in the expected direction. The rise of the renminbi during 2006–08 thus played an important role in the decline of the imbalance in 2009–10.
To conclude, this note has sought to provide some summary quantitative relationships between the exchange rate, the trade balance, and impact on jobs under conditions of excess capacity. In particular, it reports the basis on which one could expect correction of undervaluation of the Chinese renminbi to generate some 500,000 US jobs. At the same time, the discussion seeks to dispel myths about the unavailability of US domestic production to replace imports from China, and about the supposed past failure of the renminbi to influence US-China trade.
Bayoumi, Tamim. 1999. Estimating Trade Equations from Aggregate Bilateral Data. IMF Working Paper 99-74. Washington: International Monetary Fund (May).
BEA (Bureau of Economic Analysis). 2010a. US International Transactions Accounts Data: US Trade in Goods. Washington: Department of Commerce. Available at www.bea.gov/international.
BEA. 2010b. Gross Domestic Product (GDP) by Industry Data. Washington: Department of Commerce. Available at www.bea.gov.
Bergsten, C. Fred. 2010. Correcting the Chinese Exchange Rate. Testimony before the House Committee on Ways and Means (September 15).
BLS (Bureau of Labor Statistics). 2010. Employees on Nonfarm Payrolls by Major Industry Sector, 1960 to Date. Washington: Bureau of Labor Statistics (October). Available at ftp://ftp.bls.gov/pub/suppl/empsit.ceseeb1.txt
Census (US Census Bureau). 2010. US Imports from China by 5-digit End-Use Code, 2005–2009. Washington: Department of Commerce. Available at www.census.gov.
Cline, William R. 1989. United States External Adjustment and the World Economy Washington: Institute for International Economics.
Cline, William R., 2005. The United States as a Debtor Nation. Washington: Institute for International Economics.
Cline, William R. 2010. Renminbi Undervaluation, China’s Surplus, and the US Trade Deficit. Policy Brief 10-20. Washington: Peterson Institute for International Economics (August).
Cline, William R., and John Williamson. 2010. Estimates of Fundamental Equilibrium Exchange Rates, May 2010. Policy Brief 10-15. Washington: Peterson Institute for International Economics (June).
Commerce (International Trade Administration). 2010. Exports Support American Jobs. International Trade Research Report no. 1. Washington: US Department of Commerce.
Federal Reserve. 2010. Price-Adjusted Broad Dollar Index. Washington: Federal Reserve. Available at www.federalreserve.gov.
IMF (International Monetary Fund). 2010a. World Economic Outlook Database (October 2010).Washington: IMF.
IMF. 2010b. International Financial Statistics. Washington: IMF. CD Rom.
1. The role of the time lag for the influence of the exchange rate is particularly important in understanding trends in US-China trade, as discussed below.
2. Technically, the price elasticity of imports is approximately unity so that a 1 percent rise in price induces a 1 percent decline in volume, leaving no change in the dollar value of imports.
3. Studies have shown that on the export side, firms “pass through” most of the depreciation by refraining from boosting their dollar prices even though the result is a cheaper foreign currency price in the foreign market.
4. Calculated from IMF (2010b) and BEA (2010a).
5. Assuming that the job intensity is the same for both imports and exports. Because US imports from developing countries are likely to be in sectors that are more labor intensive than US exports, the use of the same employment coefficient for both imports and exports will tend to make the job impact estimate conservative.
6. This share is about twice the direct share of China in US trade, with the increase representing various combinations of higher trade elasticity estimates in Cline (2010) than in Cline (2005) and/or assumption that other “satellite” economies would also appreciate against the dollar if China were to do so (Malaysia, Singapore, Taiwan, Hong Kong). Note also that simply multiplying the $40 billion estimate by 6,000 jobs per billion dollars would understate the job impact, because of the need to expand from nominal to real trade impact, as discussed above.
7. In 2008, US imports from China substantially exceeded US domestic production in footwear ($11.6 billion versus $1.9 billion); television sets, radios, and other video and audio equipment ($19.3 billion versus $6.1 billion); and toys and sporting goods ($29.2 billion versus $15.6 billion). Calculated from Census (2010) import values and BEA (2010b) gross domestic product by sector.
8. In computers, domestic production also substantially exceeds imports from China ($72 billion versus $52 billion), but imports are dominant in the subsector of personal computers.