Sunday, October 11, 2009: 11:30 AM
The recent coordinated recessionary outburst has strengthened the importance of understanding international linkages and propagation mechanisms of shocks. The research on international business cycles linkages has generated a substantial amount of results that could be of use for policymakers as well.
Ideally, linkages take place either in the form of common shocks, or of spillover effects. With the former term, we refer to a shock hitting simultaneously all economies, while the latter refers to international transmission of shocks which were originally idiosyncratic to each country. There is widespread evidence that real developments in developed economies are driven by a large extent by a common factor, but the extent of spillovers has become stronger in the recent years.
While real linkages have been studied extensively, less attention has instead been devoted to linkages of different species. At the current juncture, it is not only important to assess how real developments in the US spilled over to the other economies, but also how the confidence and financial channels can play a role. Evidence of strong spillovers may not only shed more light on the causes of the recession, but also contribute to the shaping of the recovery.
In this paper, we tackle the issue of international spillovers by using the index recently proposed by Diebold and Yilmaz (2009). Such index is built in the setting of a cross-country VAR model, more specifically upon the notion of variance decomposition. The underlying intuition is that, if spillovers across countries are negligible, most of the variance of innovations should come from shocks that are either common or idiosyncratic to each country. This implies that the variance-covariance matrix of forecast errors should be dominated by its diagonal elements; the spillover index is computed by dividing the sum of off-diagonal elements by the sum of all elements. By looking at sums by rows and by columns, the methodology also allows identifying directional spillovers transmitted from one country to others, as well as the spillovers across country pairs.
We apply the index in the setting of a VAR model, including industrial production, confidence and financial variables for a set of developed economies, with the idea of examining real spillovers net of other effects; this requires the use of structural restrictions on the covariance matrix of the VAR. In particular, we work under the assumption that financial and confidence shocks have no immediate impact on industrial production.
Results suggest that United States are a leading exporter of real shocks, but their role appears to dampen when financial and confidence variables are taken into account. Furthermore, spillover effects appear to have intensified in the recent years.
Ideally, linkages take place either in the form of common shocks, or of spillover effects. With the former term, we refer to a shock hitting simultaneously all economies, while the latter refers to international transmission of shocks which were originally idiosyncratic to each country. There is widespread evidence that real developments in developed economies are driven by a large extent by a common factor, but the extent of spillovers has become stronger in the recent years.
While real linkages have been studied extensively, less attention has instead been devoted to linkages of different species. At the current juncture, it is not only important to assess how real developments in the US spilled over to the other economies, but also how the confidence and financial channels can play a role. Evidence of strong spillovers may not only shed more light on the causes of the recession, but also contribute to the shaping of the recovery.
In this paper, we tackle the issue of international spillovers by using the index recently proposed by Diebold and Yilmaz (2009). Such index is built in the setting of a cross-country VAR model, more specifically upon the notion of variance decomposition. The underlying intuition is that, if spillovers across countries are negligible, most of the variance of innovations should come from shocks that are either common or idiosyncratic to each country. This implies that the variance-covariance matrix of forecast errors should be dominated by its diagonal elements; the spillover index is computed by dividing the sum of off-diagonal elements by the sum of all elements. By looking at sums by rows and by columns, the methodology also allows identifying directional spillovers transmitted from one country to others, as well as the spillovers across country pairs.
We apply the index in the setting of a VAR model, including industrial production, confidence and financial variables for a set of developed economies, with the idea of examining real spillovers net of other effects; this requires the use of structural restrictions on the covariance matrix of the VAR. In particular, we work under the assumption that financial and confidence shocks have no immediate impact on industrial production.
Results suggest that United States are a leading exporter of real shocks, but their role appears to dampen when financial and confidence variables are taken into account. Furthermore, spillover effects appear to have intensified in the recent years.