Who doesn’t like a good dataset?
As markets wade into the second quarter of 2012, reflect for a moment on the fact that the crisis sparked by subprime mortgages now has nearly five years’ worth of observational data. This is very exciting
for nerds, since the conclusions from studying the period will get progressively more meaningful and insightful — something not lost on the credit strategy team at Deutsche Bank when they published their 2012 Default Study on Monday.
Concerning the last five years, here’s the high level conclusion from the more recent dataset when it’s compared to historical norms (emphasis ours):
Remarkably we show that the defaults seen in the 5 years between 2007 and the end of 2011 have been pretty much ‘average’ relative to history albeit with some large sector divergence (e.g. financials vs. non-financials). This is all down to unprecedented intervention from the authorities. Those of a bullish persuasion might conclude that if defaults have been so contained in such a hostile environment then the longer-term investor in credit should currently take great comfort at what are essentially ‘average-to-cheap’ levels of credit spreads historically.
As an investor, it is a comfort that intervention by authorities prevented the situation from being as bad as it would have otherwise been. As a taxpayer, it’s been a little less ‘comforting’… But wait, that was the good news:
… the bad news is that the crisis is unlikely to be over and it is not guaranteed that the worst is behind us. As we’ll see in the section on Sovereign CDS, spreads are still in full flung crisis mode. Indeed Portugal is now in a worse spread position than Greece was when we compiled last year’s report and most European Sovereign spreads are much wider. Within SovX WE the market is currently implying 4.3 or 7.4 credit events out of the 15 sovereigns in the index assuming either a 40% or 20% future haircut. Last year we were pricing in 2.8 to 5.1 such events.
One needs to take the default prediction-powers of all credit default swaps with a pinch of salt. The models that use CDS spreads to arrive at probabilities of default are very simplistic. Furthermore, trying to get logic to stick to a sovereign restructuring or default, particularly in a currency union, is futile. Apply politics liberally, it creates a Teflon like surface. We’ve seen recently that: (default probability) ≠ (probability of a credit event), and that can mess up the maths.
Nonetheless, as a general trend, sovereign credit spreads widening to the extent described above is concerning.
Back to the corporate universe now. The historical context for defaults is below:
It’s from this graph that one can say that defaults during the crisis haven’t been much different from long run historical averages — again, thanks to intervention by authorities.
One of the more unique characteristics of the crisis though was the increased incidence of default of highly rated credits, particularly Aa and A. It takes something very sudden and dramatic for a corporate with that rating to “jump” to default from there, rather than to get migrated to a lower rating first.
An even more recent, and also concerning, trend has been for recovery rates to be low even though default rates have fallen, as this graph shows:
As the team at Deutsche Bank explain:
As the chart shows recoveries normally rise in a low default environment and vice-versa, however that has not been the case in the past year as the relationship has diverged about as much as at any point in the last 10-15 years. Could it be that whilst authorities have dramatically reduced the tail risk and helped lower defaults, they have as yet failed to produce the normal post recovery economic strength that would normally boost the demand for assets? Is the anaemic recovery weakening the relationship between defaults and recoveries?
Good questions, and ones that should concern those going long credit. Answers welcome in the below comment box.
Full note in the usual place.