Review & Outlook

Our take on the investing, financial, & economic themes of the day

The Beauty Pageant where the Contestants Pick & Pay their Judges: Rating Agencies are Ripe for Disruption

This remarkable beauty pageant is also known as the bond market.  If a municipality or corporation wants to borrow by selling bonds, it needs to hire a ratings agency to evaluate the riskiness of its debt.  Since borrowers pick and pay their rating agency, there’s a massive conflict of interest for the agency between satisfying its client’s desire to borrow cheaply (and getting hired next time) and to evaluate its client’s finances rigorously.

Today Barron’s passes along a recent note by analysts at Janney Montgomery Scott (the former home of Sally Sulcove, a senior portfolio manager here at O’Brien Greene) who are “increasingly skeptical” of the accuracy of municipal debt ratings performed by Standard & Poors.  S&P, of course, is one of the troika of rating agencies that control the market, along with Moody’s and Fitch.  Janney believes that S&P has been inflating its ratings and investors may be unwittingly led to take excessive risks by simply taking S&P’s word for the health of a municipal bond issuer.

During the financial crisis the ratings agencies notoriously gave the highest credit ratings to mortgage-backed securities that turned out to be junk.  They haven’t really been punished by the market for proving themselves unreliable, however, because the federal government requires “investment-grade” bond issuers still to get ratings from one of the agencies.  This results in the worst of both worlds: there is a measure of competition among the several agencies to get more business by pleasing the issuers, but there is government support for the market that makes the existing agencies indispensable and shelters them from suffering the consequences of unreliable ratings.

The whole market seems ripe for disruption.  Ben wrote a few days ago about the mania surrounding the theory of disruption, which is deployed to justify wild stock market valuations in certain tech companies.  Nevertheless, it seems like big data does really offer the possibility for disrupting at least part of the ratings market.  A few days ago Paul Murphy wrote in FT Alphaville about an interesting start-up called Credit Benchmark:

Credit Benchmark, a London-based start-up, on Wednesday completed a $7m venture capital financing round, led by Index Ventures, to fund its initial expansion into the $6bn credit risk information market.

The thinking behind Credit Benchmark is beguilingly simple. All sizeable entities in financial markets — be they banks, brokers or whatever — assess each other for counterparty risk. They rate each other on a near constant basis. When it comes to the largest banks especially, this data represents a mine of extremely high quality credit risk information, put together by actual market participants, using assessment models validated by regulators.

So, why not collect all that information, make it anonymous and then aggregate it, producing dynamic consensus credit ratings on individual market participants?

Credit Benchmark has already signed up 12 “global” (but unnamed) banks to share information and has completed a so-called proof of concept, according to chief executive Elly Hardwick, who joined the venture from Thomson Reuters. The founders are Mark Faulkner and Donal Smith — the guys behind Data Explorers, which (now part of information group Markit) has used the same data sharing approach to monitor stock borrowing.

The firm will now be producing ratings on everything from sovereigns to emerging market corporates. Aggregated data sets will be available publicly, with deeper information restricted to the contributor banks.

Hardwick reckons they can move into new asset classes, geographies, types of institution, and so on, very quickly since it is simply a matter of having the technical infrastructure in place to tap information that already exists inside the banks.

Are Moody’s et al miffed? Apparently not (yet). Hardwick says they’ve been having discussions about displaying traditional ratings information alongside the newly-liberated counterparty info.

It’s not clear that the Credit Benchmark concept will really work, or if it does, how broadly applicable it will be across credit markets.  It’s intriguing nonetheless, and it looks to be an example of our economic system producing a solution to a problem that is in large part self-created.