Economists have come under a lot of criticism lately for failing to predict the 2008 financial crisis. Perhaps we should humbly admit the limits of economics, even if many ask “What is economics for?”. This is the implication of recent research by Andrew Rose of the University of California (Berkeley) and Mark Spiegel of the Federal Reserve Bank of San Francisco. They looked at economic indicators in the year 2006 for 107 countries and find none of them to be related to the severity of the 2008 crisis. They conclude that it is impossible to construct a useful ‘early warning system’ for financial crisis.
We examine a large number of potential explanatory variables for the crisis that have been discussed in the literature; these cover a host of “fundamentals” including the regulatory framework, financial conditions, and the macroeconomic, institutional, and geographic features of a country. However, we found almost none of our posited variables seem to be statistically significant determinants of crisis severity ….
Negative results like ours in a cross-section make us dubious about the accuracy of an early warning model that will have all the problems we have encountered and, in addition, the problem of predicting the timing of future crises.
Andrew K. Rose and Mark M. Spiegel, “Cross-Country Causes and Consequences of the 2008 Crisis: Early Warning”, Federal Reserve Bank of San Francisco Working Paper 2009-17, July 2009.
An important caveat is that data on housing prices are available for only 44 of the 107 countries in Rose and Spiegel’s cross-section. The authors confess, in footnote 33 of their paper, “Since real estate prices are available for a smaller set of countries, the inclusion of our real estate variable does cut down our sample and it may be the sample truncation that is precluding statistical significance for real estate appreciation.” Housing bubbles have accompanied the financial crisis in more than a few countries, so this is a serious limitation of the study.
An alternative explanation for the negative results is that the financial crisis could have been transmitted contagiously across counties, like an infectious disease, infecting ‘healthy’ and ‘unhealthy’ countries alike. Rose and Spiegel test this hypothesis in a subsequent paper, and are unable to confirm it.
Indeed, countries that were more exposed to the United States — those that held disproportionate amounts of American securities or depended heavily on exports to the United States – seem if anything to have experienced smaller crises, holding other factors constant. Overall though, we find remarkably little evidence that the intensity of the crisis across countries can be easily modeled using quantitative techniques and standard data that is either country-specific or links countries to the source of the crisis. This negative finding in the cross-section is powerful since we know, with the benefit of hindsight, both the approximate timing and the epicenter of the 2008 crisis. It makes us skeptical of the ability of “early warning systems” which must be able to predict the incidence of future crises across both countries and time.
Andrew K. Rose and Mark M. Spiegel, “Cross-Country Causes and Consequences of the 2008 Crisis: International Linkages and American Exposure”, Federal Reserve Bank of San Francisco Working Paper 2009-18, September 2009.
A tip of the hat to William Watson for the pointer.
