Contagion

From Foreign Exchange Research, Barclays Capital

The key lesson from the ERM crisis of 1992 and the Asian crisis of 1997 is that contagion can emerge quickly and often in unpredictable ways. Unwinding of leveraged positions by distressed market participants, herding behaviour among investors, and loss of liquidity that gives way to general flight to quality can all lead to heightened correlations between markets and, in extremely circumstances, set off a self-filling crisis on a regional/global scale. There have been clear signs over the past week that the distress in the Greek government bond market is increasingly being felt in other euro area countries such as Spain and Portugal. The most likely explanation of this development is the “demonstration effect” – the Greek crisis is likely to have caused investors to re-evaluate the fundamentals of these countries. Spain and Greece may not have strong financial or economic links, but their fundamentals have a lot in common.

The possibility of contagion of the Greek crisis may not end with Spain. There is a presumption among investors that in the worst case scenario (and we are not there yet) the EU will give Greece financial assistance. If this is an isolated event, the effect on the overall public finances of the EU is unlikely to be substantial. However, a bailout of Greece may make aid to other countries in similar situations more likely (moral hazard). A series of open-ended bailouts would not only undermine the fiscal positions of core euro area countries such as Germany and France but, more dangerously, would weaken their political commitment to the continuation of the euro area project.

This is probably why Greece is exerting what may seem to be a disproportionate influence on the markets of other euro area countries. We can see this by using the principal component analysis (PCA) to identify the common drivers of the EU government bond markets. Our exercise focuses on Germany, France, UK, Spain, Italy, Sweden and Greece. Figure 3 presents the factor weights of the first two principal components for the past six months. The first two principal components can be interpreted as latent factors capturing the majority of the variation (the r-square is a cumulative 80%) in bond yields across the  various European countries. They illustrate that while Greece has been a relatively small part of the primary common driver of the major EU government bond markets, it has been the most important component of the secondary common driver.

It is often found in PCA that it is difficult to attribute a particular principal component to a specific economic driver. However, while we can probably view the first principal component as capturing the common factor for the European interest rate cycles, one interpretation of the second principle component is as a reflection of the common sovereign risk. The fact that the factor weights for the likes of Greece, Italy, and Spain carry a positive sign, while Germany, Sweden, UK and France carry negative weights, would support this interpretation. We also note that this second principal component can explain 20% of the joint variation of returns of government bond yields over the sample period.

Global dimension of sovereign crisis

If it was a mistake three months ago to view the Greek crisis as a localized event, it would be a mistake now to see it as a euro-area-specific event. Indeed, global sovereign risk premia, which declined steadily in the first half of 2009, have been all moving higher at varying speeds since last September (Figure 4).

Sovereign CDS has a short history. To get a sense of where sovereign credit risk premium is now relative to history, we need to find alternative measures. One such proxy is swap spreads (differential between the fixed rate leg of interest rate swaps and government bond yields). While the usual interpretation of very wide positive swap spreads is heightened counterparty risks and flight to quality, the natural interpretation of very narrow positive swap spreads (or wide negative spreads) is that it reflects high sovereign credit and refinancing risk premium. Figure 6 plots the average 10y swap spreads for the 10 largest developed economies in the world (US, Japan, Germany, France, UK, Italy, Spain, Canada Australia and Sweden). It demonstrates that the level of the spread is currently a two standard deviations event relative to the past twenty years. In this sense, it may not be an exaggeration to say that general developed country sovereign risk premium is at a 20-year high.

The same chart shows that swap spreads were negative back in January 2009 when the market had to digest the news of President Obama’s $787bn stimulus program but these concerns eased through most of 2009. Can that happen again? We would argue that the easing of sovereign risk is due in large part to the Fed’s large-scale asset purchase program, which we have argued in the past (The case for a stronger USD in 2010, 27 November 2009) extensively created a bond shortage (Figure 5). With the BoE deciding to pause in its asset purchase program this week and our view that the Fed will follow suit in March, the bond market will struggle to digest the record sovereign issuance in the pipeline globally this year. This means that all else being equal, sovereign credit risk premium can remain at the current elevated levels or even rise further. From this point of view, the risk is that yield curves could continue to steepen, inflation breakevens expand (Figure 7), and swaption vols rise again (Figure 8).

Implications of sovereign credit risk crisis

• Increased general sovereign credit risk premium will have obvious market implications:It is clearly negative for cyclical and risky assets. Over the past year, the markets have become used to the idea that policymakers are not only all willing but they are also all powerful when it comes to crisis response. In so far as higher sovereign credit risk premium will constrain the ability of the policymakers to respond aggressively to future crisis via fiscal policy, at a minimum this calls into question the assertion that they remain all powerful. Higher sovereign risk premium also implies that crowding-out effects associated with more expansionary fiscal policy will be greater.
• Increased sovereign credit risk premium can lead to flight to quality where more-liquid government bond markets and countries with stronger fiscal positions will benefit at theexpense of less-liquid markets with weaker fiscal positions. The USD will benefit from the relative liquidity of the US Treasury markets. In addition, as we have been arguing since December, the USD will also benefit from the relative improvement of the US fiscal position in 2010.
• The JPY has had few friends over the past few months as investors have focused on Japan’s deteriorating fiscal story and its high debt overhang. This is why short JPY positions were very popular until the past few weeks. What would be the effect of a general increase in sovereign risk premium for the JPY? We believe such a development may turn out to be ironically JPY positive. This is because of Japan’s low dependence on foreign financing. One way to see this is by comparing JGB yields with Treasury and Bund yields, but adjusting for currency risk. By this measure, we find that both US and German sovereign risk premia have increased relative to that of Japan over the past year and their relative risk premia are near their highs of the past 10 years. Thus, it is not that the market is not paying enough attention to Japan’s fiscal challenge, but it appears that US and Germany’s fiscal problems have brought them much closer to Japan.