The burdens of global rebalancing

Robert Dekle, Jonathan Eaton, Samuel S. Kortum 27 August 2008



The US runs the largest current account deficit in the world. In 2006 the US deficit reached $788.1 billion, nearly 6% of US GDP and more than the surpluses of Japan, Germany, and China combined. The US deficit has been financed largely through the accumulation of US liabilities by foreign central banks. As their appetite for US IOUs can only be finite, a predominant view is that a reversal in the US trade deficit is inevitable. As Herbert Stein diagnosed a similar situation long ago, "If something can't go on forever, it won't."

While the dynamic forces driving trade imbalances remain too poorly understood to allow us to say with much precision when a correction will occur, our understanding of what drives international trade can tell us a lot about what a correction will look like when it does happen.

What must happen to fully rebalance the US current account?

What would a full "correction" of current account imbalances mean for the value of the dollar, the relative size of the US economy, and US living standards? What sort of adjustments inside the US economy will be needed? What will happen to the major surplus countries as well as to smaller players whose economies are tightly linked to the US, such as Canada and Mexico? To answer this question, we must have a way of linking trade flows – in particular US exports and imports – to economic factors such as the real exchange rate, US GDP, etc. Here the gravity model comes in.

A successful venture in international trade has been the gravity model of bilateral trade. Pioneers in econometric modelling, such as Tinbergen (1962) and Pöyhönen (1963), observed that trade between country A and country B followed a simple formula. Exports from A to B correlate very well with the size (e.g. GDP) of country A's economy multiplied by the size of country B's economy divided by the distance between them. For a long time, this relationship lacked a theory. More recent work, particularly by Anderson (1979) and Deardorff (1998), tied this relationship to standard models of international trade. In Eaton and Kortum (2002), two of us developed a particular model that allowed for production of a large number of goods that can be traded but at a cost. A feature of this model is that an increase in exports can occur at both the intensive margin (selling more of the same good) and the extensive margin (selling a broader variety of goods).1

Recently we have adapted this framework to address the US current account question (Dekle, Eaton, and Kortum 2008). Fitting the model to 2004 data on GDP and bilateral trade flows among 42 countries, we solve for the new equilibrium in which trade in manufactures (the major component of the current account imbalances of the big players such as the US, Japan, Germany, and China) adjusts to eliminate all current account imbalances. While achieving exactly this outcome would be a remarkable coincidence, the exercise gives some sense of the magnitudes rebalancing would entail.

The effects of correcting international imbalances

We perform this exercise making different assumptions about the flexibility of national economies in adapting to a rebalanced world. How easily can productive resources (most importantly workers) move between the production of non-traded goods and manufactures? How easily can countries expand exports by increasing the range of products that they can produce and sell abroad?

Table 1 presents a synopsis of our results (for a handful of countries) in the two most extreme cases.

  • Flexible case: Economies are fully flexible in both respects. Workers can seamlessly change sectors, and countries can seamlessly change the portfolio of products that they sell in different markets.
  • Inflexible case: Workers are stuck in their initial sectors and exports to a market can adjust only at the intensive margin, by selling more or less of the same set of goods.

One can think of the first scenario as reflecting the ultimate long-term consequences and the second the immediate effect of a sudden change.

For each scenario, Table 1 reports in the first column the percentage change in the country's GDP relative to world GDP. This change is likely to correspond most closely to the change in the country's exchange rate. The second column reports the percentage point change in the share of manufacturing in the country's GDP. The third column reports the percentage change in GDP deflated by the change in local prices.

Table 1. Changes in outcomes under different adjustment scenarios

              Maximum Flexibility Minimum Flexibility  
  GDP Mfg. shr Real GDP GDP Mfg.shr. Real GDP
Canada -0.9 -0.9 0.3 -5.7 -1.2 0.1
China 1.5 -0.4 0.0 8.2 -0.5 -0.2
Germany 2.5 -1.4 0.2 18.8 -1.8 0.8
Japan 3.3 -1.7 0.2 26.2 -2.1 0.3
Mexico -2.2 0.3 -0.1 -14.7 0.2 -1.7
US -4.5 3.0 -0.4 -29.9 3.5 -2.0

Note: First column is percentage change in GDP, holding world GDP fixed. Second column is percentage change in manufacturing share of GDP. Third column is percentage change in real GDP.

Interpreting the results

Beginning with the second column, we note that in either scenario adjustment requires a substantial increase in the size of the US manufacturing sector, between 3 and 3.5 percentage points. The reasons behind the change in the two scenarios are different, however. In the flexible case US manufacturing expands because resources move there. In the inflexible case, the wages of workers in manufacturing rise relative to those in the rest of the economy.

Looking at the implied change in GDP (first column) and considering the flexible case, we see that the change in the relative sizes of the different economies under the flexible case is quite modest. The US as a share of the world economy falls by just 4.5% while Japan's rises by 3.3%. The inflexible case, however, requires a much more radical realignment in the relative size of the major economies. The US declines by nearly 30% relative to the world while Japan grows by over 26%. (Combining the numbers, the adjustment would require over a 50% devaluation of the US dollar in terms of the Japanese yen).

Turning the implications for the change in real GDP (third column), we see that large changes in relative GDP translate into much more muted changes in real GDP. For instance, the real GDP of the US falls by only 2%. The reason is that the more the US relative wage (and hence relative GDP) needs to decline to make US exports (e.g., tractors, wide-bodied aircraft) more competitive abroad, the lower the price of what Americans produce for themselves (e.g., medical services, personal training, auto repair), which comprise the lion's share of what Americans (and other people) spend money on.

The outcomes for the large surplus economies (Japan, Germany, and China) are the reverse image of those for the US. Note that in either scenario the US pulls down the relative GDPs of Canada and Mexico, even though Canada starts out running a surplus and Mexico only a small deficit. The reason is that these countries' largest foreign customer shrinks substantially. Despite the decline in the size of the Canadian economy, Canadian GDP can buy more, since goods from its largest foreign supplier have gotten much cheaper still. Hence its real GDP rises.

To summarise, the realignment that is necessary depends on flexibility, with more flexibility requiring less adjustment. Even if movements in relative GDP's are substantial, however, once price changes are taken into account real effects are much more modest.

The adjustment in progress

In fact, there are signs that the correction has already begun. From 1 March 2007 to 1 March 2008 the value of the US dollar declined by nearly 18% against the Canadian dollar, over 16% against the Mexican peso, by nearly 14% against the Euro, and by over 8% against the Chinese yuan. Various trade-weighted exchange rates reported by the IMF show a US dollar decline of 10 to 13% from the first quarter of 2007 to the first quarter of 2008. During this same period US merchandise exports grew 18.4% and merchandise imports grew 12.7%. Some of this growth is the consequence of the commodity boom. But even removing soybeans, corn, and wheat from exports leaves growth in the remaining categories of US exports at a hefty 16.8%. Moreover, if imports of crude oil are taken out, US spending on imports grew by only 5.9%.

Much larger changes than these are needed to bring the US current account into balance. How much more of a dollar decline is needed depends on how adaptable the US economy is at moving resources into the production of goods that are exported or used to replace imports and on how successfully it expands the range of products it can produce and sell abroad.


Alvarez, Fernando and Robert E. Lucas. (2007) “General Equilibrium Analysis of the Eaton-Kortum Model of International Trade,” Journal of Monetary Economics, 54: 726-768.

Anderson, James E. (1979) “A Theoretical Foundation for the Gravity Equation,” American Economic Review, 69: 106-116.

Dekle, Robert, Jonathan Eaton, and Samuel Kortum (2008) “Global Rebalancing with Gravity: Measuring the Burden of Adjustment” forthcoming, IMF Staff Papers.

Deardorff, Alan V. (1998) “Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World?” in Jeffrey Frankel, editor, The Regionalization of the World Economy. University of Chicago Press, 7-32.

Eaton, Jonathan and Samuel Kortum (2002) “Technology, Geography, and Trade,” Econometrica, 70: 1741-1780.

Pöyhönen, Pentti (1963) “A Tentative Model for the Volume of Trade Between Countries," Weltwirtschaftliches Archiv, 90: 93-99.

Tinbergen, Jan (1962) Shaping the World Economy: Suggestions for an International Economic Policy. New York: Twentieth Century Fund.


1 Alvarez and Lucas (2007) extended our framework to incorporate the production of non-traded goods as well.



Topics:  Global economy

Tags:  US current account deficit, international imbalances, gravity model

Professor of Economics, University of Southern California

Professor of Economics at New York University

Professor of Economics, University of Chicago