Monday, August 18, 2008

International economies coupled or decoupled?

One of the largest questions in macroeconomics and for investments in the last few years has been whether global economies have been decoupled. It was the idea of decoupling which has driven much of the international equity diversification. If there is not a link with the US, you can invest in other countries without the fear that US slowdown will take down the rest of the world. You would be willing to sacrifice liquidity for the diversification advantage.

The problem with answering the decoupling question empirically rests on four difficult issues.

One, there are not many recessions so the data available to test whether there is coupling is limited to a few events. The US has only had two recessions since 1990. Abstracting from a small sample of slowdowns makes it difficult to determine whether there is any meaningful relationship.

Two, working with macroeconomic time series is difficult. Assume that there is a US recession and nine months later there is a downturn in Europe. Would that be considered a coupled event? Would the answer change if the time difference was 18 months? How much time between events is necessary before there is a decoupling?

Three, the world has changed. With an increase in trade, there is a greater link across economies so the degree of coupling will be different than 25 years ago. There is a greater likelihood that any event may be transferred across countries either on the trade side or on the financial side. However, this is different from the old adage that if the US sneezes the rest of the world catches a cold. A quarter a century ago the US represented a greater portion of world output so any event in the US would have a greater world impact. World production is more diverse today but the links across countries via globalization are greater.

Four, the very definition of what would be a coupled event is hard to define. Some recessions will be localized. The Mexican crisis in 1994 would be a simple albeit large localized event. The discussion of coupling have to be discussed in the context of macro factors that can be transmitted across region. This, by definition, will reduce the number of coupled events.

Taking note of these issues suggests that we have a coupled event currently: a single risk factor from a credit contraction; a commodity price shock; timing of a slowdown across regions within a nine month time frame; and an event that is affecting trade flows. Coupling of economic events was never dead, we just needed the right set of circumstances.

No comments: