VoxEU Column Global crisis Global economy

Who fell in 2009: Those with current account deficits or with extra froth?

Virtually no country was untouched by the crisis. But which countries saw the sharpest declines in GDP – and why? This column shows that those with higher growth rates before the crisis fell harder while relatively closed economies were somewhat insulated. In contrast, the relationship between current account deficits and the decline in growth rates is fuzzier.

The full force of the global crisis was felt in 2009. Relative to 2008, generally regarded as the first full year of the crisis, the fall in 2009 growth rates was breathtaking (Table 1).

Table 1. Change in the 2009 GDP growth rate relative to 2008

 
Number of countries
Mean
Standard deviation
25th percentile
75th percentile
 
 
 
 
 
 
Advanced
33
-4.8
2.1
-5.9
-3.7
Upper-Middle Income
34
-6.9
4.3
-8.7
-3.7
Lower-Middle Income
36
-4.1
4.2
-6.1
-1.1
Low Income
34
-2.5
2.8
-4.1
-0.6
Average
 
-4.6
3.1
-6.1
-2.1
 
 
 
 
 
 
 
 
 
 
 
 
Note: Growth forecasts for 2009 are from the IMF’s October World Economic Outlook. The sample does not include countries with current account surpluses greater than 25% or less than -25% of GDP, or small island economies. “Upper Middle-Income countries are those with per capita incomes greater than $4000. The “Lower Middle-Income” countries are those with per capita incomes between $900 and $4000. “Low-Income” countries have per capita incomes below $900. These definitions correspond to the World Bank classifications.

On average, a country’s GDP was about 4.5 percentage points lower in 2009 than in 2008. The devastation was widespread. Virtually no country was untouched by the panic in financial markets and the collapse in world trade.

But the crisis also had diverse effects. The 25th percentile in the change of growth rates was -6 percentage points and the 75th percentile was -2 percentage points. The diversity is also evident across country groups; the fall was large among advanced economies (4.8 percentage points) and, especially upper-middle-income economies (6.9 percentage points) and more moderate among lower-middle-income (4.1 percentage points) and low income countries (2.5 percentage points). Within these groups, the experience varied considerably, particularly among the middle-income countries.

A widely held view is that countries with large current account deficits were living beyond their means and suffered disproportionately during the crisis (see DeLong 2008). Latvia’s large deficit and its growth collapse symbolise this view, although it is common to generalise to all of Central and Eastern Europe. How empirically well-founded is this relationship? Figure 1 reports bivariate relationships reflecting country access to foreign capital and institutional credibility in managing the flows (see the appendix table for details).

Figure 1. Change in growth in 2009 relative to 2008 vs (a) growth in 2007 (b) current account in 2007

GDP growth, 2007                        Current A/C, 2007                                 GDP growth, 2007                          Current A/C, 2007

GDP growth, 2007                             Current A/C, 2007                            GDP growth, 2007                          Current A/C, 2007

Froth

The consistent evidence is that the greater the pace of growth in 2007 – at the peak of the upswing—the sharper was the fall in 2009. It is as if the strong growth rates in 2007 reflected a global boom that was unsustainable. The fact that this relationship holds in a cross-country regression implies that there was a global component of unsustainable growth. Thus, even countries that did not have bubbles in the obvious sense of overheated property sectors or rapid credit growth shared in the global froth. In addition, among middle- and low-income economies, if growth did not moderate sufficiently in 2008, the penalty was paid in 2009.

Current account deficits

There is much weaker support for the idea that higher current account deficits contributed to a sharper fall in 2009. The bivariate relationships are relatively fuzzy, reflected in the lack of precision on the current account variable in the regressions. With the conventional positive sign on the current account balance, a larger surplus in 2007 is associated with a greater growth deceleration in 2009 among the advanced countries. Though the coefficient is not significant, it does reflect the differing forces operating. The decline in US growth was, for example, less than that of Germany and Japan, as these surplus countries were particularly hit by the contraction of world trade. This, in turn, reflects the difference in the composition of growth in the surplus and deficit economies. Countries running surpluses relied heavily on world trade, which took a major hit in late 2008 and the early part of 2009, sharply curtailing growth rates in 2009. The deficit countries relied more on domestic consumption, which also fell but less than did world trade.

Among non-advanced countries, deficits did hurt, but the coefficient is significant only for the lower-middle income group, that at a 10% level. Once again, this reflects the fact that both net exporters and net importers were hurt. Even among low-income countries, commodity exporters with sound current account positions suffered as commodity prices fell. There is some evidence that lower-middle-income and low-income countries that consolidated in 2008 were less vulnerable in 2009.

Regional diversity

The regional diversity among the non-advanced economies suggests that regions that were relatively closed to international trade and financial flows remained somewhat insulated from the global crisis. On the whole, African and Middle Eastern countries, excluding Iraq and those with current account surpluses greater than 25%, performed better than Asian economies, the benchmark for the analysis. In contrast, countries belonging to the former Commonwealth of Independent States did markedly worse. The lower-middle-income Latin American countries fell behind. Among the Central and Eastern European economies, there was considerable variation. Those hurt were lower-middle-income economies such as Albania, Bosnia and Herzegovina, and Macedonia, and the Baltic nations. The rest did only marginally worse than the Asian economies. The Latvian example is inappropriately extended beyond its borders. Even by the standards of Central and Eastern Europe, Latvia was going beyond the speed limits (Fabrizio, Leigh, and Mody, 2008). But the results do caution that generalising from Latvia is not warranted. The more general finding is that economies of various stripes were overheated to varying degrees, and the crisis was, in part, a correction of that overheating.

Some policy speculations

The evidence suggests that within relatively homogenous country groupings, the decline in GDP growth in 2009 relative to 2008 was greater the higher the growth rate in 2007. While there are many reasons to celebrate that expansion, including bringing into the fold many low income countries, it also accumulated an unsustainable element.

Understanding the reasons for the global coordination of country bubbles is key to thinking about future systemic vulnerabilities. Clearly, running a large current account deficit is a source of vulnerability. But the evidence cautions that pointing the finger in that direction may distract from the more potent underlying reasons for the “Great Recession” of 2009.

Disclaimer: The views expressed here are those of the author and not necessarily those of the IMF, its management, or its Executive Board.

References

DeLong, J. Bradford (2008), “The wrong financial crisis”, VoxEU.org, 10 October.

Fabrizio, Stefania, Daniel Leigh, and Ashoka Mody (2008), “The Second Transition: Eastern Europe in Perspective,” Paper prepared for the Workshop: “Five years of an enlarged EU—a positive-sum game” Brussels, 13-14 November, forthcoming in a European Commission edited collection.

Appendix

Appendix table: Explaining the drop in the 2009 GDP growth rate relative to 2008

 

 
Dependent Variable: Change in GDP growth rate between 2008 and 2009 
VARIABLES
Advanced
Upper-Middle-Income
Lower-Middle-Income
Low-Income
Upper-Middle-Income
Lower-Middle-Income
Low-Income
 
 
Change in growth rate between 2007 and 2008
-0.29
-0.18
-0.60***
-0.7***
-0.71***
-0.33**
-0.6***
(-1.27)
(-0.66)
(-2.79)
(-5.98)
(-5.51)
(-2.41)
(-9.49)
Change in the current account between 2007 and 2008
-0.11
-0.22
0.28**
0.21***
0.03
0.01
0.23***
(-0.70)
(-1.13)
(2.60)
(3.51)
(0.39)
(0.18)
(7.25)
Current account deficit in 2007
-0.02
0.08
0.09*
0.03
-0.05
0.06
0.11***
(-0.35)
(1.08)
(1.88)
(0.80)
(-1.46)
(1.61)
(5.32)
Growth rate in 2007
-0.78***
-0.57*
-0.39*
-0.5***
-0.70***
-0.51***
-0.5***
(-3.44)
(-2.01)
(-2.00)
(-5.51)
(-5.52)
(-4.06)
(-9.70)
Per capita Income in 2007
0.00
-1.01
-2.62***
-0.22
-1.71**
-1.26**
0.57*
(0.00)
(-0.64)
(-3.40)
(-0.38)
(-2.20)
(-2.14)
(1.84)
Africa
       
3.02**
0.39
0.66*
 
       
(2.36)
(0.51)
(1.89)
Middle East
       
8.28***
0.52
 
 
       
(4.94)
(0.59)
 
Western
       
1.84
-3.02***
 
Hemisphere
       
(1.62)
(-3.79)
 
CIS
       
-0.82
-13.6***
-2.9***
 
       
(-0.61)
(-12.47)
(-5.84)
Baltics
       
-12.2***
   
 
       
(-8.48)
   
Other Central-Eastern Europe
       
-2.17
-3.61***
 
       
(-1.62)
(-2.92)
 
Constant
-2.37
5.87
19.02***
2.73
11.14
10.10**
-1.92
 
(-0.23)
(0.40)
(3.35)
(0.76)
(1.57)
(2.35)
(-0.96)
               
Observations
32
34
36
34
34
36
34
R-squared
0.41
0.24
0.52
0.69
0.94
0.93
0.9

 

1. t-statistics in parentheses, *** p<0.01, ** p<0.05, * p<0.1

2. Estimation uses the robust regression procedure to minimise the influence of outliers. A regression using robust standard errors gives virtually identical results.

3. The sample does not include countries that current account greater than 25% or less than -25 percent.

4. “Upper Middle-Income countries are those with per capita incomes greater than $4000. The “Lower Middle-Income” countries are those with per capita incomes between $900 and $4000. “Low-Income” countries have per capita incomes below $900. These definitions correspond to the World Bank classifications.

5. The omitted region is Asia.