Home > Uncategorized > Some Second Guessing?

Some Second Guessing?

There seems perhaps to be a little bit of second-guessing going on at the moment. Although the Chicago Manufacturing index was stronger-than-expected at 63.8, and although GDP was near-expectations at 3.2% (Personal Consumption was +3.6%) and +0.6% on the PCE deflator, and although the Employment Cost Index recorded a near-expectations +0.6%, the stock market fell (-1.7%) significantly for the second time in a week and the bond market rallied (+15/32nds on TYM0, 3.66%).

This despite the talk, predicted here on Thursday, that the ECB would shortly announce the rescue plan for Greece. Ordinarily, that produces a Friday rally and Monday disappointing selloff, but on Friday the rally did not materialize. Perhaps it was the rumors that the ECB was asking the Fed to renew swap lines, which some people felt implied the ECB expected a possibility of a tight financing environment for European institutions in the near future. I have no information on the rumor other than to observe that a tight financing environment, until we know which institutions are most vulnerable to a Greek default (and others), isn’t a stretch as far as concerns go. If in fact the ECB made such a pre-emptive request, then good for them!

Goldman fell 9.4% on Friday on the news that the SEC had commenced a criminal investigation into the company. That is a significant development. Remember that Anderson was brought low by the criminal charges, not by any proof of the merit of the charges. Success of a criminal complaint might kill the institution, and the mere threat of it is chilling (albeit not undeserved).

I have trouble believing stocks fell so far, however, merely because Goldman is being shot at or because of the shocking news that European institutions may be at risk if Greece defaults or the contagion spreads. To me this appears to simply be an overextended market where investors are taking some chips off the table due to a change in the preponderance of the risks. This should not be surprising, and I would not be surprised to see it extend further.


I have seen a lot of comments recently suggesting that the ECB or other entities ought to regulate the ratings agencies in some way because the downgrades of Greece and other entities has “fed fuel to the fire” in this developing crisis. Such a suggestion gets it exactly backward, I think.

The ratings agencies are in precisely the same position that real estate appraisers are in, and more importantly were in, when the real estate bubble inflated. That is simply this: all parties have an incentive to see a higher rating rather than a lower rating, just as all parties in a real estate transaction have an incentive to see a higher appraisal. Let’s recall the latter situation first. The home seller has an incentive to see a higher appraisal, because it supports a higher transaction price. Local government likes the higher appraisals and higher prices, because it expands the tax base. The buyer likes the higher appraisal (it was usually the buyer who solicits the appraisal to support the mortgage loan), because it supports a larger mortgage loan and thereby reduces his outlay. The lender likes a higher price, because it makes money in proportion to the size of the originated loan or refi (and, since the loan is securitized, it doesn’t care whether there is a correlation with default). The broker prefers a higher appraisal, because it makes it possible to persuade more market transactions (the value of each of which, since the broker fee is 6% of transaction price, increases with price).  The appraiser has an incentive to appraise properties highly so he gets more business as an appraiser. No one, in fact, has an incentive to see a lower appraisal and the only party that has an incentive to get an accurate appraisal – the investor in the mortgage pass-through security – is the furthest removed from influencing the appraisal. Thus, appraisals are routinely inflated and appraisers are pressured to inflate appraisals by the originator who hires them. There is no mystery about this dynamic, and frankly even in the teeth of the housing bubble some people (ahem) were writing about this dangerous situation.

The same is true in bond ratings. Clearly, the issuer of the security wants a higher rating since it lowers its cost of funding. The buyer of the security, a portfolio manager, likes to see a higher rating since the manager is limited by statute in many cases and by mandate in many others from investing in securities below a certain rating; moreover, if the rating turns out to have been over-optimistic, the buyer can always blame the rating agency. “Hey! It was AA and that’s what I am allowed to buy!” The dealer of course has an incentive to structure securities at the margin where it just gleans the higher rating (and it can do so since the rating agencies make their rating formulas known, for reasons I have never understood), since it is easier to exact a high fee in this case. The rating agency itself wants to make people happy, and it knows people prefer high ratings to accurate ones and they get paid more the more ratings they do – the more bonds that are structured, the more the agency gets paid, if it can be counted upon to rate the bond “appropriately.” The one who cares most about the accuracy of the rating, the investor in the fund, is the one most removed.

What we want to do is to increase the value of high-quality, independent ratings, and we don’t do that by legislating a set of rules that agencies, structuring firms, and investors can just game differently. We need less structure to the rules, not more. How would the ultimate beneficiary of the ratings – that is, the person whose money is invested, look at the problem? They want ratings to reflect the probability of default accurately given the position in the economic cycle. All rating agencies should report publicly the realized default and recovery statistics of the bonds they’ve rated for a forward-looking time period (to prevent sudden re-rating just before default). So, for example, S&P would report that for bonds backed by first-lien mortgage securities that they have rated AAA, 0.05% have defaulted within 1 year of that rating, .11% have defaulted within 2 years; of the same class of bonds they have rated AA, 0.20% have defaulted within 1 year, .22% have defaulted within 2 years, and so on. And then investment managers should be held responsible not for buying bonds of a particular rating but of a particular default probability. “You know that bonds rated B+ by AggressiveCo tend to default at a 2.5% rate, but according to your mandate you should have bought bonds that had less than a 2% chance of defaulting when you bought them.” Rating agencies that gained a reputation for being very astute readers of credit situations would flourish; the hacks would wither. A more finely-grained continuum would also reduce the problem associated when a rating change (for example, to below-investment-grade) essentially produces a run on the credit when its bonds are no longer eligible for inclusion in a mandate.

Because after all, as an investor I don’t want to know whether my manager only bought AAA bonds and they happened to go bust; I want to know why my manager bought this “AAA bond” that yielded 2% more than other AAA bonds and was surprised when it wasn’t as high-quality. What did my manager think about the probability of default? A rating is no substitute for credit analysis, unless the rating is a pretty good proxy for credit analysis. The only way to evaluate that is to look at the history.


On Monday, March Personal Income and Consumption (Consensus: +0.4%/+0.6%, with 0.0% on Core PCE) will be released at 8:30ET and the ISM (Consensus: 61.0 from 59.6) at 10:00ET (also at 10:00 will be Construction Spending, Consensus -0.3%, but that is not a market mover). Although ISM is ordinarily of interest, I believe it pales in market significance to events in Europe. Friday’s Employment report will be meaningful, but right now it is the growing crisis across the pond that matters.


One final note: I was in a meeting with a potential client on Friday and one of them remarked that they had seen my Letter to the Editor printed in Barron’s a couple of months ago. It is funny, because I hadn’t realized my note had been published. I looked it up, and indeed here it is in case you’re curious. I regret including Goldman in that list, but the general point is still true (even with respect to many people at Goldman).

Categories: Uncategorized
  1. George
    May 2, 2010 at 12:21 am

    I like your idea of having statistics on rating agencies, however I suspect that

    “Past Performance is No Guarantee of Future Results”

    still applies.

    For example the Fed printing money may make almost everything AAA until the
    printing stops…

    Or PIK (pay in kind) bonds.

    Clearly it’s just about impossible to default and the statistics would be
    good…. At least at the begining… (Sounds like social security…)

    • May 2, 2010 at 7:52 am

      Yeah, I’m sure there’s some refinement to that idea we can make! Good point.

  1. No trackbacks yet.

Leave a Reply

%d bloggers like this: