One worry that people have about ratings agencies is that they have conflicts of interest in which they are paid by bond issuers to rate bonds, so they have incentives to give bonds -- particularly those of frequent and high-paying issuers -- high ratings. One solution to this conflict of interest is not to have the agencies paid by issuers. That happens in sovereign ratings. European countries, which pretty much hate the heck out of the ratings agencies, do not pay them for their ratings. So you'd think that would avoid the main conflict of interest.

But not if you're the European Securities and Markets Authority, which today released its report on how evil the credit ratings agencies are about downgrading European sovereigns. The backdrop is that "when compared to historical trends, sovereign ratings assigned to EU member states have experienced high levels of volatility" -- because EU member states have experienced high levels of volatility in (market perceptions of) their creditworthiness, you'd think -- and that some people "have argued that sovereign rating changes helped exacerbate the financial crisis." So ESMA was sent to sort the agencies out and find flaws in their process of downgrading European sovereigns. Which it sort of did?

The first "key finding" of the report is that, for one or more credit rating agencies, "members of the CRA's Board of Directors were involved in the rating process" by discussing potential ratings actions with senior managers, and that in other cases "specific rating actions had been driven by senior management, with limited or late-stage involvement by the lead analysts." This can present a conflict of interest because of the following assortment of words:

Even where certain Board members are not directly involved in the commercial operations of a CRA, they are nevertheless involved in business and managerial decisions at the highest levels, including discussions on the business targets and the revenues of the company. This could affect their independence when voting on the ratings or when providing guidance on rating decisions.

But … how? I do not know.1 ESMA's view of conflicts of interest is pretty weird: Senior management can't have a say over ratings, because they might have commercial concerns, though not commercial concerns about those particular ratings. Elsewhere there are complaints that junior staff have too much of a say over ratings. It's almost as though they set out to find fault with the agencies without worrying too much about logic.

There are other conflicts of interest. ESMA doesn't like that ratings analysts also publish research, because the research is revenue-generating (from subscriptions) and so there might be a conflict because, I don't know, "research activities and the time-consuming work involved in drafting and approving these reports" might "divert analysts from their core analytical tasks" or "influence the way information on the rating and future rating actions is disclosed to the public." And ehhhhh I guess?

I don't know, I mean, there really was a thing where credit ratings agencies got paid a ton of money to develop mathematical models that overstated the creditworthiness of structured credit securities so that they could be sold by the banks on which the agencies depended for their livelihoods to unsophisticated investors who relied on the ratings. Like that kind of happened and was bad, and you can tell a very clear conflicts-of-interest story about it. The thing where the ratings agencies downgraded some European sovereigns when the markets got really worried about their ability to repay their debts does not seem like quite that sort of conflict of interest. But, y'know, ratings agencies, conflicts of interest, those are words that go well together.

Anyway, though! The other thing that ESMA worries about is sort of ratings insider-trading: "In one or more CRAs, ESMA observed several instances of disclosure of upcoming rating actions to an unauthorised third party, before publication and, in some cases, before the rating committee had taken place." My initial reaction was, hahahaha, who cares, nobody would trade on a ratings action. Ratings actions are lagging indicators: They don't tell the market what to do; they tell the public what the market has already done.

Then I went and sort of halfheartedly looked at the data, and while it does nothing to dispel the view of ratings as lagging indicators,2 it does suggest that markets might react to ratings changes. Just taking 34 post-2008 ratings actions (downgrades, negative watches and one lonely removal-of-negative-watch) for Spain, France and Italy, it seems that credit default swap markets did have some tendency to move on ratings actions: CDS widened after a downgrade or negative watch in 20 out of 33 cases, with an average move (over all 33 cases) of about +5.6 basis points.3

I wouldn't go around arguing that that's causal, but there it is. Perhaps ratings do matter. I can't find much evidence that they leaked, though. On average CDS tightened by 8 basis points in the week before a ratings action was announced, but the data is all over the place.

Horizontal axis shows CDS moves (- tighten, + widen) from 1 week through 1 business day before ratings action (downgrade or negative watch). Vertical axis shows move from 1 day before to 1 day after ratings action. CDS data from CMAQ, ratings data from Bloomberg CRPR.
Horizontal axis shows CDS moves (- tighten, + widen) from 1 week through 1 business day before ratings action (downgrade or negative watch). Vertical axis shows move from 1 day before to 1 day after ratings action. CDS data from CMAQ, ratings data from Bloomberg CRPR.

One possibility is that, if the ratings agencies are leaking their sovereign ratings actions to the market, the market doesn't care that much. European governments, on the other hand, seem to care quite a bit.

1 Is the subtext "these jokers don't pay us so let's lower their ratings"? That doesn't seem very plausible. You can't rate all the sovereigns junk. The report says "the unsolicited nature of several sovereign ratings need to be carefully considered in assessing potential conflicts of interests faced by CRAs when issuing sovereign ratings" but doesn't say how ESMA did that.

2 For instance, Standard & Poor's downgraded Spain to AA- in October 2011, when its five-year credit default swaps were trading at around 380, and then downgraded it a few more times until reaching BBB- in October 2012, when its CDS was at around 370, suggesting that the market was at least a year ahead of S&P on that one.

3 And tightened in 12 cases, and was unchanged in one. Data here. Also, Spain tightened dramatically after losing its negative watch, so that is nice for them.