Monday, February 23, 2009

Economic Crises not as rare as we may think

There has been so much talk about the current crisis and the potential for it being the worst since the Great Depression that there has been little voice as to the number of crisis across the globe in the last century. The chance of a crisis should not just be measured by what has happened in the US but relative to what has occurred across many countries. To look at only events in the US is egocentric as if the rules of the rest of the world do not apply to this country. It also does not allow a clear analysis of what is the potential for downside economic events. Albeit crises are rare, they are not “Black Swans” with infinitesimal probability.

These issues have been addressed in a number of research papers by Robert Barro, most recently with a piece focused on macroeconomic crises since 1870. Looking at both consumption and GDP over the last 100 plus years, he found 95 consumption crises in 29 countries and 152 GDP crises in 36 countries. A crisis is a cumulative decline of at least 10%. The worst crises were as expected World War I and World War II followed by the Great Depression, the 1920 influenza slowdown, the LATAM debt crisis and the Asian crisis. The problems are greater for non-OECD countries and have been worse before WW II, but almost every country is likely to have a crisis. The probabilities while small are certainly greater than 1% and the severity may be quite large. Most of these crises last for years and do not have a quick solution.

Barro does not talk about the policy implications of this work but has focused on the fact that rare events of large magnitude do occur helps explain the large equity premium in stock markets. The risks have been priced in the market. Perhaps one of the largest reasons for the current crisis has been the complacency of investors who got used to the period of the Great Moderation. This complacency caused greater risk seeking behavior which lead to higher leverage and larger loses once a crisis did occur. This may not help us exit the problem, but it will shape our future behavior if we are better aware of low probability risks.

No comments: