Bond yields and disaster risk premia
July 22, 2012 3:46 pm
by Gavyn Davies
.
The last few weeks have seen a further collapse in government bond yields, especially at the front end of some European yield curves, where the ECB’s decision to cut deposit rates to zero has clearly had important effects on interest rates. How low can bond yields now go? ECB board member Benoit Coeure said on Friday that the ECB might cut rates to below zero. He said that the central bank is closely observing events in Denmark, where two year yields have dropped into negative territory, as they have in Switzerland.
Despite the fact that policy rates might drop below zero, it is common to hear in the financial markets that government bond markets are in bubble territory, with equities being “so cheap” relative to bonds that equities are virtually certain to out-perform bonds by a wide margin in the years ahead. Although I have said this myself on many occasions, I am now beginning to wonder whether it is true.
The behaviour of 2 year bond yields in 2012 can be broken down into three groupings. The majority of countries, including the US, Japan, the UK and most of the eurozone, now exhibit yields bunched around zero, having seen significant declines since March.
The second grouping consists of Canada and Australia, two commodity producers with more buoyant economies and higher policy rates. The final grouping consists of the crisis economies in the eurozone, including Italy and Spain, where bond yields have been driven higher because of default risk:
The recent drop in 2 year yields is clearly part of a much bigger pattern in the global bond market. The chart below shows 10 year yields. Again, the stark difference between Italy and Spain on the one hand, and all other countries on the other hand, is immediately apparent. Where default risk is effectively nil, as it is in economies with their own central banks, yields are homing in on zero throughout the curve:
It is tempting to explain this collapse in yields by referring to the central banks’ forward commitments to keep short rates at zero for a long period ahead. But longer term yields have in fact dropped by more than this factor can explain. In the jargon of bond markets, the “term premium” has dropped by about 100-150 basis points in the past 18 months. Many economists attribute this drop in the term premium to quantitative easing by the central banks. Yet this may not be the full explanation for the precipitous drop in global bond yields.
.
A different reason is offered in a paper published last week by Martin Brookes and Ziad Daoud, my colleagues at Fulcrum Asset Management. They draw attention to the impact of the “economic disaster risk premium” which may have been priced into bond markets as investors have realised that the chances of very deep recessions are greater than they had previously realized. (The paper can be downloaded here.)
.
The Brookes/Daoud model, which builds on earlier work by Robert Barro and others, is capable of explaining why most global bond yields appear to be below equilibrium at present, why the equity risk premium appears to be abnormally high, and why Spanish and Italian markets have detached themselves from the rest.
The starting point is a recognition that the perceived likelihood of an economic disaster (defined as a drop of 10 per cent or more in real GDP) has risen markedly since 2008, after many decades in which the risk of such an event had disappeared from investors’ minds. The reappearance of disaster risk increases the probability of a left tail event in economic growth and reduces the valuation of the equity market.
.
What effect does the rise in disaster risk have on government bond yields? Provided that bonds have a low or negligible risk of default, then they have an insurance value relative to other assets. This insurance value rises as the risk of an economic disaster increases, and investors stampede into safe assets. This implies that bond yields should fall as disaster risk rises, because investors attach a greater value to the safety of bonds.
Brookes and Daoud calculate that the rise in investors’ perception of disaster risk in recent years can plausibly account for a 50-100 basis point decline in US and German bond yields. This explains virtually all of the “overvaluation” of these bonds relative to the levels suggested by equilibrium bond models and leaves little or nothing to be explained by quantitative easing by the central banks (which admittedly is something of a puzzle.)
.
.
The Brookes/Daoud model also has some interesting things to say about the sell-off in bond markets like Italy and Spain. They argue that, in these markets, disaster risk is accompanied by a rising risk of outright default on sovereign bonds. When the risk of default becomes sufficiently large, these bonds no longer have much value as “insurance” assets and start to behave like equities. They therefore experience sharp sell-offs as economic disaster risk rises. When this happens, bonds and equities fall together. They become positively, not negatively, correlated.
The graph below shows this process at work in the Spanish bond market in the past few years, with the implied default probability being derived from the CDS market.
A novel observation by Brookes and Daoud is that bonds start to behave like equities when the risk of default in the bond market rises above 3 per cent per annum. CDS spreads indicate that this has already happened in the eurozone crisis economies, and is just on the point of happening in France. Unless this default risk can be brought back down, the outlook for bond yields in these countries, and therefore the sustainability of sovereign debt, looks bleak.
.
Elsewhere, though, bond default risk is still very low. In those circumstances, the presence of economic disaster risk can cause an entirely rational re-pricing of both bonds and equities, and this can disrupt traditional equilibrium pricing models for both assets. Bonds may be “expensive” and equities may be “cheap” for quite a long while.
0 comments:
Publicar un comentario