Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
People get very cross about credit ratings.
Issuers get cross when the individual ratings they’re assigned are, in their opinion, too low. Politicians can throw tantrums, or even issue arrest warrants for analysts daring to dish out a downgrade. And not every firm with a spotless payment history is thrilled to be handed a rating at the low end of junk (although there is one notable exception we can think of).
Meanwhile, everyone gets cross when ratings turn out to be collectively too high. As the final report of the Financial Inquiry Commission, set up to investigate and report the causes of the global financial crisis, put it (their emphasis):
We conclude the failures of credit rating agencies were essential cogs in the wheel of financial destruction . . . Investors relied on them, often blindly. In some cases, they were obligated to use them, or regulatory capital standards were hinged on them. This crisis could not have happened without the rating agencies.
An increasing chorus of worrywarts including the IMF, the BIS, the Bank of England, and the chair of UBS have been publicly fretting that the explosion in private credit ratings might be understating credit risk in similar ways. Check out today’s MainFT’s deep dive into the smaller agencies behind the private credit boom for more.
This got us wondering just how good credit ratings are anyway, and how much has default experience has differed across the different agencies?
Payments are made, or not made
Here’s some data from S&P’s annual default and transition study about their experience with ratings and defaults, chartified:
It’s hard not to be just a bit impressed. With the angels-on-a-pinhead exception of AAA-rated and AA-rated corporates, cumulative default rates for lower ratings have been higher over both short and long time horizons, which is exactly what you’d hope.
Of course, individually, pretty much every rating is a dud: for the 97 per cent of triple-B ratings for which no default occurred, the payment profile was, ex post, triple-A. And holders of bonds issued by the 3 per cent of ratings that went on to default will have felt their experience was pretty junky. But collectively, ratings look to have described credit risk pretty well.
How does this stack up against other credit rating agencies?
We planned to grab data from each of their swanky reports but, unfortunately, not all agencies produce them, and those that do provide them in formats that make it hard to compare numbers. But you know what they all do produce? Form NRSRO Exhibit 1 documents online. Because it’s the law.
A Form NRSRO Exhibit 1 document is a hefty pdf packed with rating transition matrices. And a rating transition matrix is a big grid showing where ratings started the period and where they ended the period. Here’s one of Fitch’s:
(High res)
For those who can’t be bothered to squint, the grid is titled ‘Corporate issuers — 1-yr Transition and Default Rates’. Corporates are just one of twelve SEC-designated categories of rating, along with financials, insurance, CMBS, RMBS, CLO, CDO, ABS, other securitised, sovereign, US and international public finance. And, as well as one-year matrices, Form NRSO Exhibit 1 documents contain these for three years and 10 years.
This particular matrix shows that Fitch started 2024 with 3,118 corporate ratings ranging from triple-A all the way down to single-C. And over in the third-from-the-right column is a series of numbers labelled ‘default’. From this we can see that nothing with a rating of BB+ or higher defaulted in the year, unlike every one of the eight entities rated single-C.
To compare the default experience across agencies, we pulled together data from the largest seven agencies’ 10-year matrices for each of the corporate, financial and insurance categories of rating and then grouped these into broad rating buckets. This is what we found:
At first glance, the bars look kind of all over the place. But this funkiness is (mostly) a function of two things. First, look at the scales on each of the mini charts (double-A weirdness aside): they’re rising, and rising steeply. The scale for single-A defaults goes up to 0.4 per cent; for triple-Bs goes up to 4 per cent. Within the context of these rising scales, differences between default records across distinct agencies’ same broad rating buckets have been pretty minor.
Second, the outliers are usually explained by their being the product of dividing one very small number by another very small number. Hover your pointer over each of the bars to see how many ratings were in the category and how many ended in default. Or, you can just feast your eyes on the Marimekko version of the chart:
That DBRS spike in AA-land? Of the paltry 94 double-A rated DBRS corporate, financial or insurer ratings at the end of 2014, one went on to default. And 1/94 is a big number. The KBRA triple-B spike has vanished as Kroll’s rating footprint was so tiny in the space in 2014 that it doesn’t register on the chart.
What happens if we also chuck in all the securitised, public finance and sovereign ratings? This brings the number of ratings up to just shy of 190,000 (because we’ve left single-Cs out from our chart-fest):
And in Marimekko form:
Does the data tell you there’s no rating shopping going on? Sadly, it can’t shed pretty much any light on this question. All it can speak to is what happened to the ratings that were outstanding at the end of 2014.
Lacking access to a crystal ball, rating analysts are just giving their opinion on how robust a business will be to economic, regulatory, and financial shocks. The data does seem to suggest that differences between rating agencies in terms of default experience per broad rating category over the past decade has been fairly modest.
To find out if rating shopping is going on now, check back in on FTAV in 10 years and we might have the answer. Although there’s still one thing we’re trying to get our heads around, and that’s what happened to all of the ratings that were issued in 2014 and subsequently withdrawn. And with one in six ratings withdrawn within a decade of issuance, this is no small detail.
Further reading:
— The new crop of rating agencies behind the private credit boom
(MainFT)
— Why are people being mean about Egan-Jones? (FTAV)
— What’s up with private credit ratings? (FTAV)
