It goes without saying that 2002 was a tough year for commercial lenders. Not only did more loans default, but lenders also recovered less value from the defaulted debt – less than at any other time in past 14 years, according to ratings agency Standard & Poor’s. The combination of the anaemic recovery rate and the elevated default rate hit banks as a “double whammy”, sharply increasing their realized credit losses.

Analysts say it’s no coincidence that loan recoveries plumbed new depths at the worst possible time. The picture was much the same for defaulted bonds. In fact, there are good reasons to believe that the default rate and recovery rate are negatively correlated, so recoveries always fall when defaults rise. As such the recovery rate is an important component of expected loss calculations and should also play a part in credit pricing – but opinions are divided on how easy it is for banks and investors to anticipate falling recovery rates. 

So what evidence is there to suggest a link between default and recovery rates? Fitch Ratings’ data shows that the total dollar value of bonds defaulting last year in the US was more than $100 billion, far outstripping the previous record of $78.2 billion in 2001. The weighted average recovery rate of 26% was well below the 35% long-run average. (Recovery rates for commercial loans tend to be higher than publicly-issued bonds, thanks to covenants and collateral arrangements, but are harder to compile than the recovery rates on bonds, whose cycle is much the same).

But 2001-2002 is one of only two recent periods during which default rates have spiked significantly, says Richard Cantor, managing director for ratings research and analysis at Moody’s Investors Service in New York; the other was the downturn of the early 1990s. That means there’s not much empirical evidence with which to confirm or refute the link. “We have two periods of high default rates, both associated with low recovery rates, but that’s not a huge amount of history to go on,” he says.

Nevertheless, Cantor, like other credit analysts, believes that recovery rates and default rates go hand-in-hand. The arguments are largely a matter of common sense. Supply and demand pressure has an impact, as do borrowing levels at individual companies and market sentiment about their prospects after default.

Increased default rates produce a glut of debt for sale. What’s more, there may be fewer potential buyers because interested parties are likely to be similar to the distressed company, and thus may well have pressing financial concerns of their own. This combination of excess supply and low demand results in low prices for defaulted companies, or their assets – and that means that creditors recover less of the face value of debt.

This is most evident when a particular industry is under pressure – as recently seen in the telecom and aviation sectors. “Say you’re an airline, you default and you need to sell a fleet of 747s,” explains Don van Deventer, president of risk model vendor Kamakura. “But everyone else in the industry is defaulting at the same time, then buyers are going to be fewer. Prices – and recoveries – will be lower.”

Simultaneous (but opposite) spikes in default and recovery rates can also be a hangover from periods of excess. During periods of credit expansion, companies borrow more; when things go sour, that not only increases the chances they will default, but means that the ratio of realizable value to outstanding debt is smaller – leading to a lower recovery rate.

One reason that prices for distressed debt were depressed last year was because borrowers were over-leveraged, confirms Mariarosa Verde, Fitch Ratings’ New York-based head of credit market research. A majority of the bonds that defaulted last year were issued at the height of the late-1990s boom, and many companies have subsequently found that they could not support the level of debt that they took on.

Verde argues that a third factor is market sentiment. When a company reorganises in bankruptcy, the market re-assesses the amount of debt it can sustain as a going concern. “In an environment when company values are declining, recovery rates will decline as well, simply because the market has a depressed view of future company earnings.” General economic gloom – fostered by, among other things, increased default rates – therefore leads to lower recovery rates.

“If asset values are declining, that makes defaults more likely and also lowers the recovery rate,” summarises Kamakura’s van Deventer, who contends that any model that uses asset values as a key input will therefore produce results that incorporate expected recovery rates. He says there is “no question” that recovery expectations are routinely priced into bonds and credit at the point of sale. Although recoveries may not be included as a discrete element of the price, the impact that recoveries have on expected loss is “certainly replicable by tools using the reduced form model or the Merton model”.

Others aren’t so sure about anticipating recovery rates for pricing purposes. David Kelson, managing director at Fitch Risk Management in New York, says this kind of prediction “would not be straightforward at all”.

“In order to accurately price for loss given default one would have to predict when the credit would default and what the default rate would be at that time for companies in that particular industry and geography,” explains Kelson, “and other factors, including collateral amount and quality and capital structure, also impact recoveries”.

Moody’s Cantor is also wary of suggesting that recovery rates can be anticipated with any precision – over periods of years, at least – but he says that some models are capable of providing good estimates of recovery rates over periods of a single year.

Fitch Ratings’ Verde says that the relationship between defaults and recoveries is variable, and that makes predictions hazardous. For example, if last year’s defaults had been spread more evenly across a range of sectors, rather than being particularly concentrated in a handful, recoveries would not have declined as steeply, while the default rate would have remained the same. Fitch’s statistics for 2002 show that the telecom sector produced an average recovery rate of 11.5%, while cable debt had a recovery of 53.4%.

That kind of variation doesn’t rule out the possibility of modelling recoveries – but it means that predictions need to be specific to individual companies and particular sectors. Different sectors have different recovery rates, depending, for example, on how rapidly their asset value erodes. The assets of merchant energy companies, for example, become obsolete more slowly than those of telecom companies.

The rate at which this happens also changes over time. If the business environment changes rapidly, as it did for many “fallen angels” – once highly rated companies that have fallen abruptly from grace - then it can do so again. That makes it difficult to rely on current asset values as a proxy for future recovery values, says Moody’s Cantor. “There’s a lot of uncertainty for periods of more than a year. You can’t know what the environment for a particular industry will be like, so the negative correlation between default and recovery rates just isn’t very useful.”

Just because telecom debt has offered “miserable” recoveries over the last year, Cantor says, there’s no reason to expect the same results from telecom companies in the next credit downturn. Recoveries are determined more by the “clustering effect” of defaults in one industry than by more general economic factors.

So the use of historical observations about recoveries as a basis for future assumptions is a long way from being a perfect solution. But Fitch Risk’s Kelson says it’s about the best that banks can do right now, and he holds out hope for improvements. “Work is only just starting on how to be more precise,” he says. “Maybe down the road, we’ll be able to anticipate changes in recoveries more efficiently.”

©2010 Sungard. All rights reserved. Legal Information

Home | About Ambit ERisk | Solutions | Collateral | Contact Us | Search
Risk Reports Case Studies RMA Ambit ERisk Whitepaper Economic Capital Training And Workshop Events	About Ambit ERisk Solutions Collateral