RDP 2009-08: Leverage Constraints and the International Transmission of Shocks 2. Empirical Evidence

We present some empirical evidence that supports our contention that balance sheet contractions may have been an important propagation mechanism for the global financial crisis. First, Figure 1 documents the global nature of the economic crisis, demonstrating a remarkably synchronous collapse in economic growth rates for a sample of Organisation for Economic Co-operation and Development (OECD) countries. It is unlikely that trade linkages alone could account for the simultaneous downturns in all regions. If we take the US economy as the ultimate source of the financial crisis then it would be easy to explain the scale of the downturn in Mexico, for instance. But Figure 1 illustrates dramatic reductions in economic growth in many European economies, only marginally linked to the US through trade flows.

Figure 1: Real GDP Growth

In addition, there is clear evidence that US banks reduced their outstanding claims on the rest of the world. Table 1 contains short-term claims of US banks, for all OECD count ries for which data are available. This is the total stock among US reporting banks of all claims on the destination economy with less than one year remaining until maturity. Under normal circumstances new claims are issued and many maturing existing claims are rolled over each quarter. A rapid decline in less than one year, then, implies little new issuance, and few exposures being rolled over. There is a clear pattern overall that the largest OECD economies (by size of claims) have experienced a substantial fall in short-term US bank claims during 2008. In particular France, Germany, Ireland, Italy, Korea and Luxembourg all experienced major withdrawals over 2008. Further, total claims across all countries declined by more than 20 per cent, with half of that decline occurring in the final quarter.

Table 1: Short-term Claims of US Banks on OECD Economies
$US billion
Destination of funds 2007:Q4 2008:Q1 2008:Q2 2008:Q3 2008:Q4
Austria 4.2 4.2 4.8 3.6 2.3
Belgium 8.7 13.9 17.5 15.8 15.6
Czech Republic 0.5 0.7 0.8 0.9 0.5
Finland 3.2 2.8 2.4 3.0 2.9
France 58.0 69.1 41.8 44.4 55.3
Germany 56.9 65.9 48.4 41.3 39.3
Greece 3.9 4.9 3.0 2.3 2.4
Hungary 0.9 1.0 0.9 1.1 0.5
Ireland 28.3 27.5 28.1 27.8 23.6
Italy 25.2 25.5 26.2 18.6 17.2
Korea, South 26.3 27.4 28.0 29.9 21.5
Luxembourg 26.1 24.7 22.8 21.7 11.9
Mexico 6.5 7.8 7.5 6.8 7.7
Netherlands 43.1 47.0 52.1 47.6 37.2
Poland 2.4 2.3 2.3 2.3 2.5
Portugal 2.9 2.3 2.1 1.7 1.2
Spain 28.3 28.4 25.4 18.7 18.4
Turkey 7.3 6.9 7.0 6.0 5.1

Source: BIS consolidated banking statistics

Aside from bank balance sheets, we can also find evidence consistent with balance sheet contractions in other instruments. Equities in particular were believed by some policy-makers to be a means of contagion, as the following quote by Rakesh Mohan, Deputy Governor of the Reserve Bank of India, indicates:

Our problems are mainly due to the sell-off by foreign institutional investors in the domestic equity markets leading to a sharp reduction in net capital inflows and the sharp slowdown in global economic activity and external demand. (Mohan 2009)

This view is consistent with the data on international capital (Figure 2). The crisis has seen a fall in both US capital inflows and outflows, at the aggregate level. The scale of the fall in flows in early 2009 is unprecedented over the full sample of aggregate US Treasury International Capital (TIC) data going back to 1980. In the model developed later in this paper we will see that the balance sheet contractions implied by this, when combined with binding leverage constraints among financial institutions, can impart an independent international transmission of shocks.

Figure 2: US Gross Capital Flows

2.1 Financial Linkages Versus Trade Linkages

The effects of global balance sheet adjustments should be expected to vary by country. Some economies are more dependent on capital inflows than others, and countries with low credit ratings may suffer more from a sudden reduction in flows than higher-rated countries. Evidence of the effects of a financial channel should be consistent with the difference in vulnerabilities across countries.

We demonstrate the importance of balance sheet contractions as a propagation mechanism for the crisis using regression analysis. As a rough measure of the international effect of the crisis, we use the change in the growth rate of real GDP between the year ended December 2007 and December 2008. The vulnerability of countries to a sudden outflow of capital is measured in terms of total capital inflows (TCI) from the United States, as a per cent of 2007 GDP, using US TIC data. Our sample includes members of the OECD for which these data are available (that is, all of the OECD other than Iceland and the Slovak Republic). We also include trade linkages, measured using exports to the United States in 2007 as a per cent of GDP (X). Finally, we interact each of these variables with the sovereign credit rating of the economy (CR), to capture the idea that capital withdrawals are likely to affect lower-rated economies more heavily than higher-rated ones due to a ‘flight to quality’. Based on Standard & Poor's sovereign foreign currency credit rating in December 2007, we convert the credit rating to a numerical scale where a value of 0 corresponds to a AAA rating, 1 to a AA+ rating, and so on, down to 12 for a BB− rating.

The results are shown in Table 2, and provide support for our argument that financial flows were a causal factor in the propagation of the crisis, while trade channels appear less important. First, the export variables (X and CRX) are never economically or statistically significant, and sometimes enter with the wrong sign. Second, our measure of capital flows (TCI) is statistically significant in all cases.[4] Third, the interactive term between the credit rating and the size of capital inflows from the United States enters significantly, consistent with flight-to-quality, and adds substantially to the explanatory power of the model. Finally, the size of the adjusted R-squared statistics is consistent with capital inflows playing an important role in explaining the downturn, while trade channels are of less importance.

Table 2: Explaining the Slowdown
(1) (2) (3) (4) (5)
X −0.020
CRX   −0.007
TCI     −0.005
CRTCI       −0.046
Adjusted R2 −0.036 −0.072 0.133 0.278 0.294
Observations 29 29 27 27 27
Notes: Dependent variable: real GDP growth rate in the year to 2008:Q4, less the growth rate in the previous year. P-values are in parentheses; bold indicates significance at the 5 per cent level. X equals exports to the United States and TCI is gross capital inflows from the United States, each as a percentage of GDP, in 2007. CRX and CRTCI are interactive terms, where CR is S&P sovereign foreign currency credit rating in 2007. CR=0 corresponds to a AAA rating, 1 for AA+, and so on, to 12 for BB−.

In summary, this evidence suggests the possibility that a financial channel may be important for the international propagation of shocks. Moreover, it is difficult to explain the scale and synchronicity of the global downturn based on trade alone.


The significance of this variable depends on the inclusion of Ireland in the sample. However, this is not true for the interactive term, which is more robust to individual countries. [4]