One remarkable feature of the current financial crisis has been the speed and apparent synchronicity with which it has spread around the globe. While it originated in the United States, it has affected not only economies that shared similar vulnerabilities, in particular the exposure of financial institutions to toxic assets, but it spread to virtually all economies, advanced and emerging alike. Moreover, the crisis has not been limited to the sphere of financial markets but has had a major impact on real economic activity, inducing the largest global recession since the Great Depression. Even after an initial decoupling of emerging market economies (EMEs), global economic activity became temporarily highly synchronized in the second half of 2008 and the first half of 2009.
Different hypotheses have been put forward as to why the crisis has become truly global in reach. A first hypothesis is that of liquidity, and the fact that credit markets and in particular interbank markets became highly illiquid, leading to the collapse or near collapse of numerous financial institutions and severely curtailing the capital available to the real side of the economy (e.g. Adrian and Brunnermeier (2009), Brunnermeier and Petersen (2010), Shin et al. (2010), Borio (2009)). A second hypothesis relates to the pricing of risk. While financial institutions in North America and Europe were highly leveraged and exposed, financial institutions in many EMEs, in particular in Asia and Latin America were not. Moreover, the financial crisis triggered a massive reversal of private capital flows globally or what has been dubbed a "flight to safety" phenomenon with capital exiting in particular EMEs and being shifted from relatively risky financial assets into safer assets such as US treasuries. Such a reallocation of global capital related to a re-pricing of risk may thus have spread the crisis, and even to countries and regions that had been less exposed through the liquidity channel.
A third hypothesis is linked to the collapse of global economic activity. The economic slowdown in the US and Europe of late 2007 and early 2008 quickly intensified and spread strongly to other parts of the world after the collapse of Lehman Brothers in September 2008. Declines in GDP growth rates in EMEs, even in many Asia and Latin American countries, in that period were as strong as those in advanced economies. This decline has been in large part been related to the severe recessions in advanced economies and the ensuing collapse in global trade, which had been significantly stronger even than that in GDP. This affected adversely in particular EMEs, which tend to be relatively more open and more dependent on trade than many advanced economies (e.g. IMF (2009)).
The paper sets out to explore the role of these three different mechanisms in spreading the crisis, both to advanced economies and to emerging markets. What complicates such an analysis using standard macro models is that the crisis comprises a relatively short period and that it is inherently difficult to identify meaningful measures of shocks to liquidity, to risk and to real economic activity at quarterly or monthly frequency. We tackle this issue by taking a financial market perspective, analyzing the response of short-term interest rates as a proxy for financial market conditions, and the response of equity markets as a proxy for the impact on the real economy. Using a Global VAR (GVAR) approach allows us to identify these three types of US-specific shocks: shocks to liquidity and to risk appetite (using the US TED spread between US short-term money market rates and US treasuries, and the US VIX index of implied volatility of the S&P500) and shocks to US economic activity measured as surprises to high-frequency announcements of key US economic activity variables. Using weekly data, this enables us to trace the effect of these three types of shocks to a broad set of 26 economies worldwide.
The empirical approach we employ allows us to deal with the challenge of identification and in particular with the large dimensionality problem. We resort to a novel methodology introduced by Chudik and Pesaran (2010) and later extended by Pesaran and Chudik (2010) in the context of the analysis of VARs of growing dimensions (so-called infinite-dimensional VARs), a methodology which also establishes conditions under which the increasingly used Global VAR model developed by Pesaran et al. (2004) is applicable. In this set-up, all variables are treated as endogenous, which is arguably a very important advantage for our purpose. Restrictions to overcome the dimensionality problem are based on an intuitive concept, namely that of neighborhood effects. The restrictions employed in this paper allow for rich spatial and temporal interactions among variables. In particular, we allow for the US to potentially have a dominant influence on other countries, other sources of strong cross-section dependencies besides the dominant US variables (i.e. we allow for the presence of unobserved strong common factors), and an unspecified weak-form cross-section dependence of residuals (see Chudik, Pesaran, and Tosetti (2010)). The dominance of the US in financial markets also helps us distinguish US shocks from shocks to other economies. To distinguish between different types of US shocks and to separate them from other global shocks, we implement a standard sign restriction approach combined with a partial ordering of variables in the context of our high-dimensional VARs.
Download
The Role of Liquidity, Risk and Economic Activity in the Global Transmission of the Financial Crisis
