Sunday, May 04, 2008

What are corporate bonds worth in a recession?

Corporate bonds just finished one of their best months in recent history. Coming off the historic wide spread levels of March, investment-grade corporate bonds tightened by 38bps during April, according to Lehman Brothers, outperforming Treasury bonds by 245bps. That's the best monthly performance for corporate bonds since 1988.

And yet its widely acknowledged that the real economy is quite weak, and corporate defaults will invariably rise as the economy falls into recession. So how can corporate bonds tighten given the economic backdrop?

The short answer is that there is one price for corporate bonds under a liquidity crisis and another under a bad economy. So we're transitioning from the former to the later.

See, the reality is that for investment-grade corporate bonds, liquidity matters much more than run-of-the-mill economics. To see what I mean, consider this simple valuation model for a generic Baa-rate corporate bond.

According to Moody's, the average 10-year cumulative default rate for a Baa-rated issue is 4.6%. In other words, buy a random portfolio of Baa-rated bonds and hold them blindly for 10-years, you'd expect 4.6% to default. The average recovery rate given default for unsecured bonds is 38%. So over 10-years, you'd expect 4.6% of your Baa bonds to default, giving you expected credit losses of 2.85% (4.6% defaults times 62% loss given default).

If default risk were evenly distributed over time, one would have an expected credit loss of 0.285% per year for 10-years. This implies that if corporate bonds were to pay a mere 28.5bps above the risk-free rate, that would, exactly compensate an investor for default risk. Of course, we know that corporate bonds don't trade anywhere near that level, but hold that thought for now. We'll get back to that.

In reality, defaults do not occur evenly over time, but are more likely during period of weak economics. So let's say that the 4.6% default rate is the right cumulative number, but that all those defaults happen in a single year. In other words, over our 10-year horizon, one of those years is a recession, and all the defaults happen in that one year. We could then do a simple discounting calculation to figure the NPV of the expected loss.

We assume that for every $100 par, the investor will suffer $2.85 in credit losses in the recession year. Let's further assume that under normal circumstances, investors assume that there will be a recession at some point during the 10-year horizon, but are unsure as to when it might occur, and therefore assume it will happen in the 5th year. Using the current 10-year swap rate (4.48%) as the base rate, the portfolio of corporate bonds would have to pay a spread of 27bps above the base rate to compensate for credit losses in the 5th year.

Here is my math, in case you want to check it.

Now let's assume that the recession happens in the first year. If the losses are accelerated into the first year, the required spread rises. But only by 7bps. The required spread goes from 27bps to 34bps.
I know, I know. You are getting ready to fire off a nasty comment, claiming some kind of witchcraft over these numbers. OK, think about it logically. Investment-grade companies typically have reasonable balance sheets and relatively stable business models. Otherwise they wouldn't be investment-grade. Take companies like Comcast or Kraft or FedEx. All Baa-rated. Have the odds of these companies going bankrupt really increased by a large degree just because we've entered a recession?

Of course, 34bps isn't even in the ballpark of where corporate bonds have historically traded, even in good times. That's because in real life, investors demand a spread that compensates them for expected credit losses and the uncertainty surrounding credit losses. In a recession, its that uncertainty that rises. So realistically, the spread should change by more than just 7bps as we enter recession.

But don't kid yourself about how high spreads should be just because GDP growth is weak. According to Lehman Brothers, the average investment grade bond spread is now 218bps, 50bps tighter than the all-time wides hit on March 17.

But spreads are still 115bps above their long-term average. Despite how far corporate bonds have come, there is plenty of room to move tighter. Plenty.

But spreads are still 115bps above their long-term average.

Despite how far corporate bonds have come, there is plenty of room to move tighter. Plenty.


PNL4LYFE said...

It's amazing how such a simple and straightforward analysis can put things in perspective. Thanks for that. However, I can't agree with your assertion that investment grade rated companies are investment grade because they have reasonable balance sheets and stable business models. If it weren't for the "too big to fail" assumption that was validated by the Bear Stearns bailout, I think all financials would be drastically wider (rightfully so).

Accrued Interest said...

"Reasonable" balance sheets and "stable" business models are obviously qualitative statements. And who trusts the ratings agencies anyway? My point is merely to say that the tendency is for investment-grade companies to be pretty well capitalized.

Put another way: In 2006, the market seemed to assume that credit would never be a problem again. In 2008 the market seemed to assume credit will always be a big problem. Neither is right.

Anonymous said...

This was an excellent article - clear and comprehensible even to non-bond-geeks like myself. Keep up the good work.

Anonymous said...

Thank AI - as always, very insightful but i thought you might be interested in some of my analysis. If you go back to 1960 (based on Moody's data) and track the 'actual' defaults that occur and how they cluster around recessions, how their recovery rate drops as recessions hit, and incorporate ratings transitions (lots of stuff i know) we get to around an expected 140bps spread expectation for a CDS (so similar to a libor spread though obviously different). This 140bps recessionary expectation ties in with recent IG9 wides at 199bps and current 86bps and 18mth ago 30bps. For my money, we should be closer to 200 than 50 currently and are likely to see another leg wider before spreads settle down again - simply based on actual defaults happening.

Extend your analysis to HY and its a different picture - historical wides at around 1100bps compared to current 550bps, recent wides at 750ish and recession expectations around 650. As the recesssion starts to impact leverage and bank lending (see today's SLO survey) continues to pro-cylically tighten, defaults are likely to rise (see RECSAP's comments today). We'd expect to see HY under-perform as real defaults hit and IG will get dragged wider on risk aversion rather than implicit 'real' default risk.

Reasonable and stable balance sheets can quickly collapse if refinancing is tough and the recent strength of the new issue market still leaves botha decent concessions and a significant cash-cds basis out there thanks to funding issues.

One more final thought to consider (sorry for going on) but credit cannot be judged on an aggregate basis and then applied to single-names - note the approach that ratungs agencies used for rating CDOs! Not only is the performance of the aggregate corporate bond market very different from a portfolio of credits and it is monitoring for 'blow-up' risk that pays off in this downturn phase of the cycle. The other risk premium is simply a reflection of MTM risk - the last few months have seen the link between spread vol, beta, and risk premia become much more dependent and well worth doing some regression (nudge, nudge).

Sorry for the lengthy comment and keep up the great work on your blog.

Accrued Interest said...


I'm going to post a graphic in a few minutes, then I want you to explain a bit more about your process. I can't really picture how you'd get to the numbers you have. It sounds like you've really done the work, tho, so I'd really love to hear more about it. Maybe we're using different data sets?

Anonymous said...

Interesting post. BTW, what is LUCROAS (Index) on your BB screengrab? Who provides it? BB says "Access to this security is restricted by its supplier"

Accrued Interest said...

Its Lehman's Credit Index. I'm probably allowed in because our firm trades with Lehman.