Globalisation and financial contagion | VOX, CEPR Policy Portal
Olivier Accominotti, Marie Briere, Aurore Burietz, Kim Oosterlinck, Ariane Szafarz 10 April 2020
In 2007 globalisation reached its peak with global cross-border capital flows amounting to approximately $11.8 trillion. However, the financial crisis that followed contributed to a reversal in this trend, with markets showing the first signs of ‘deglobalisation’ as a result. The recent Covid-19 pandemic is likely to cause another enormous stress test for globalisation, forcing firms and nations to limit traveling and trade, perhaps leading to a reevaluation of the interconnected global economy. Figure 1 shows the evolution of foreign direct investment (FDI) since 1970. FDI has decreased by 60% since its peak in 2007, with cross-border financial flows experiencing a similar trend (Bordo 2017). Deglobalisation could have severe consequences for investors looking for diversification opportunities since it entails more segmented markets, leading to an increase in idiosyncratic risks, as well as in transaction costs.
Figure 1 Foreign direct investment, net inflows (Balance of Payment, current Trillion US$)
Source: International Monetary Fund, Balance of Payment database, supplemented by data from the United Nations Conference on Trade and Development and official national sources
Financial globalisation, financial contagion and portfolio diversification
Financial globalisation is not a linear or irreversible process. Researchers have shown that international capital market integration was significantly more pronounced in the 1880-1914 and post-1971 eras, in comparison to the interwar (1918-1940) and Bretton Woods (1944-1971) periods (Mauro et al. 2002, Goetzmann et al. 2005, Rangvid et al. 2016).
Bekaert and Mehl (2019) study global stock market integration over the 1880-2014 period. They show that, although international stock markets were significantly integrated during the first era of global finance (1880-1914), integration was at its highest in the post-Bretton Woods era. Hence, financial globalisation between 1880 and 2014 followed a ‘swoosh’ pattern (Bekaert and Mehl 2019). However, the authors find no evidence of deglobalisation following the financial crisis.
By definition, globalisation is crisis-insensitive and refers to a general increase in correlations within asset classes, spanning across geographical areas (Berben and Jansen 2005). However, during financial crisis episodes, we frequently observe a significant increase in cross-market linkages which goes beyond what the ‘fundamentals’ can explain (Forbes and Rigobon 2002). This phenomenon, often referred to as ‘contagion’, has important implications for investors. Investors are often searching for the benefits of diversification strategies. However, in the presence of contagion across countries, geographical diversification becomes less powerful during crises. This, in turn, makes investors that are already struggling with low returns even worse off.
Financial contagion in the long run
In a recent paper (Accominotti et al. 2020), we revisit the issue of financial contagion during globalised periods in a long historical perspective (1880-2014). More precisely, we investigate how the level of financial globalisation affects the risk of international financial contagion. Considering the period between 1880 and 2014 allows us to distinguish between four sub-periods with different levels of financial market integration. These sub-periods are as follows:
- The 1880-1914 classical gold standard era, when financial markets were globalised but international stock market integration was more moderate than in the most recent period
- The 1918-1940 years, which first saw a short revival in cross-border capital flows followed by deglobalisation
- The 1946-1971 Bretton Woods period, when stock markets were poorly integrated as most countries implemented capital controls
- The 1972-2014 post-Bretton Woods era, when global stock market integration reached its highest level ever
To overcome the problem of disentangling globalisation from contagion, we follow a sequential process. First, we use an ‘International Capital Asset Pricing Model’ to assess globalisation in the equity market of 17 countries, identifying excess returns over the four stated sub-periods with respect to the international market portfolio. Next, we analyse correlations between monthly equity excess returns, comparing correlation matrices by using the tests proposed by Goetzmann et al. (2005). In sum, we allow for the possibility of globalisation associated with the systematic source of return variation. We then consider overlying contagion.
We show that the intensity of stock market contagion varies with the degree of financial market globalisation, but in a nonlinear fashion. Intuitively, in a globalised world (that is, in a world with high cross-market correlations) the scope for an increase in correlations following a crisis should be more limited than in a world with limited (or no) globalisation whatsoever.
Our findings suggest that the phenomenon of financial contagion was absent from stock markets during both the period of deglobalisation of 1918-1971, and during the era of intense globalisation of 1972-2014. However, we do find some evidence of stock market contagion during the classical gold standard period of 1880-1914, when stock market integration was high but more moderate compared to the most recent period. Capital controls and market segmentation implemented from the 1930s to the Bretton Woods years may explain the absence of stock market contagion during the 1918-1939 and 1946-1971 periods. However, although the 1880-1914 and 1972-2014 eras were both marked by significant capital market integration, contagion was only present in the former period.
Figure 2 illustrates the relationship between financial globalisation and financial contagion. The figure reports cross-country correlations between the monthly stock market returns for Germany, the US, and the UK across the four sub-periods from 1880 to 1914. It also shows the average increase in stock market return correlations between these three countries during identified crisis episodes. As is apparent, international contagion within stock markets only appears to have been a significant problem in the era of ‘moderate’ globalisation in 1880-1914 but was a much less severe challenge in the era of ‘intense’ financial globalisation in 1972-2014.
Figure 2 Cross-country correlations and average correlation increase during crises, Germany, UK and US
Source: GFD, author’s calculations
For contagion effects to occur, markets have to be at least somewhat integrated. When connections between markets are minimal, contagion cannot appear. As a result, short of globalisation, contagion is impossible because it requires some permeability between financial markets located in different countries. Once markets are more integrated, contagion becomes plausible during crises. During the 1880-1914 classical gold standard era, globalisation was reasonably high, highlighting the presence of contagion. However, with peaking levels of globalisation (as in the 1972-2014 period), contagion is doomed to disappear. When stock return correlations are very high, there is little room left for any increase following a shock.
Overall, contagion is more likely to occur when globalisation levels are in the middle range. Our findings are consistent with an ‘inverted U-shaped’ relationship between contagion and globalisation. If cross-market correlations are too high, globalisation can kill contagion. As such, financial contagion might become a significant problem for investors if financial markets return to a more moderate level of globalisation in the near future. Heightened contagion in a less globalised financial system could reduce the benefits of portfolio diversification, making the effects of crises even more dramatic for investors.
Accominotti, O, M Brière, A Burietz, K Oosterlinck and A Szafarz (2020), “Did Globalization Kill Contagion?”, CEPR Discussion Paper No. 14395.
Bekaert, G and A Mehl (2019), “On the global financial market integration “swoosh” and the trilemma”, Journal of International Money and Finance, forthcoming.
Berben, R P and W J Jansen (2005), “Comovement in international equity markets: A sectoral view”, Journal of International Money and Finance 24(5): 832-857.
Bordo, M D (2017), “The second era of globalization is not yet over: An historical perspective”, NBER Working Paper No. w23786.
Forbes, K J and R Rigobon (2002), “No contagion, only interdependence: Measuring stock market co-movements”, Journal of Finance 57(5): 2223-2261.
Goetzmann, W N, L Li and K G Rouwenhorst (2005), “Long-term global market correlations”, Journal of Business 78(1): 1-38.
Mauro, P, N Sussman and Y Yafeh (2002), “Emerging market spreads: Then versus now”, Quarterly Journal of Economics 117: 695-733.
Rangvid, J, P Santa-Clara and M Schmeling (2016), “Capital market integration and consumption risk sharing over the long run”, Journal of International Economics 103: 27-43.