An interesting post at VoxEU by Andrew K. Rose and Mark M. Spiegel does a series of analyses looking to explain how the crisis evolved internationally, but the obvious connections don’t provide an answer:
The 2008 financial crisis is sometimes characterised as one where financial difficulties in the US spread to the rest of the world. But is there clear evidence of such international contagion? This column reports research indicating that neither financial nor trade linkages to the US help explain the cross-country incidence of the crisis. If anything, countries more exposed to the US seem to have fared better….
The case for contagion seems superficially clear. In this column, we discuss our recent research that probes more deeply into the 2008 crisis. In particular, we ask if countries that were more heavily exposed to toxic US assets suffered deeper losses during the crisis of 2008.
Surprisingly, our answer is negative – countries that had disproportionately high amounts of trade with the US in either financial or real markets did not experience more intense crises. If anything, countries more heavily exposed to the US seemed to fare better than others. And our results are robust; we examine over 40 different linkages between countries, in both trade and capital. This negative finding makes us sceptical of the ability of “early warning systems” – such as the one discussed in the de Larosiere Report (2009) – to successfully predict the incidence of future crises across both countries and time.
A bit of further thought shows why this finding (or lack thereof) is not as surprising as it might seem.
There is evidently more to the research than their high level summary at VoxEU shows, but it does not appear that they found data on foreign country exposure to toxic assets; indeed, that information would take an extraordinary amount of effort to compile. How do you identify, say, which foreign banks hold collateralized loan obligations or CDOs that have tanked? You need to find which ones have done badly (already a huge task, this information is available only from dealers) and then you have to trace back ownership. Thus the researchers presumably could not develop the most relevant data.
And using US asset exposures as a proxy for toxic asset exposures is wrongheaded. Consider Japan and China. Both are huge trade partners of the US, and both have holdings heavily weighed towards US Treasuries and agencies. They were not buyers of exotic debt instruments. I believe the same is true of Taiwan.
Another factor this analysis fails to capture is the vulnerability of a country to financial crises. Relevant metrics would probably be private debt to GDP and size of the financial sector relative to GDP. Switzerland has limited trade relationships to the US, but has an outsized banking sector, and one that held a lot of dollar denominated assets that turned out to be rotten funded by dollar liabilities. If a country has a banking crisis in a country’s own currency is comparatively manageable (albeit painful and costly). But when you have a large external debt, it creates all sorts of financial instability and the issue of solvency (even for a supposedly rock-solid country like Switzerland) becomes a real concern.
So I have no doubt that contagion factors can be found, but they may not lie in clean data sets that economists like to use in regression analyses.