Wednesday, January 26, 2011

Response to S&P Proposals

 26 January 2011

Regarding Aspects of Fundamental Analysis and Rules in S&P’s Proposed Bond Insurance Criteria.

To: Standard and Poor’s

Various measures have been developed in the course of time to assist the financial analyst in her quest for valuable insight. Five years ago, S&P was using quantitative determinants for ratings in structured finance. However, the run-up to the financial crisis saw these parameters supplant fundamental analysis.

Your proposal for bond insurer rating criteria includes large swaths of fundamental analysis. I commend you for including stressed scenario analysis (32-35 and 50-56) and related analysis to answer the questions: “What are the chances of default? What would recoveries likely be?” But I encourage you to consider making your models more flexible. In January 2008, would it have been appropriate to model 3 or 4 years of growth in NPO before a stress period as the measure of capital adequacy? A crisis does not come with a four year warning, please allow your analysts to look at the most appropriate horizons as dictated by changing circumstances. I also encourage your analysts to focus on liquidity bottlenecks, or peaks in large amounts of concentrated (or correlated) lower quality insured debt service. I understand this is a more labor intensive method that may be called for only in extenuating circumstances. In short, enable and encourage S&P analysts to think like a chief risk officer; do not relegate them to administration of data and rigid parameters.

Such rigid parameters are witnessed in S&P’s newly proposed leverage limits (29-31). These static limits would account for a mezzanine tranche and a super-senior tranche of the same deal equally. As stated, exposure size would also rate zero-coupon and premium bonds equally. Given the impartial treatment of par outstanding and considering that these are absolute limits, S&P should believe there is no feasible institution that could be both AAA and in violation of these limits. However three months ago, S&P considered such an institution to exist in Assured Guaranty at multiples of the proposed limits. This will result in lower ratings for the only insurance company that was able to write business throughout the recent experience. If such inflexible limits are applied, they should be a high-limit safety net and not a low-limit Maginot Line.

I have found your S&P grading systems helpful when considering specific facets of a credit. Your proposed grading scales (22-28) will be no exception. However, it offends logic to think the inflexible relationships between these scores will not greatly benefit from common sense fundamental analysis.

Frustrated athletes often revert to bad habits, S&P must remember that rigid models have been gamed and train-wrecked before. A kaleidoscope of numbers easily produces an illusion of precision. But such a structured illusion will only offer an opportunity for your analyst to defer to a Rube Goldberg project of scores connecting to matrices to limits. This class of machine appears to do much but in the end only fries an egg; and normally the egg is burnt. Your good cooks can do better!

The remaining leaders of the bond insurance industry have asked for a more stable ratings regime. They may not agree with my rejection of overwhelming reliance on inflexible models and limits. As an analyst, investor and observer, I suggest S&P look within; reexamine the sectors in which S&P has the best and worst ratings track records. Municipal bond ratings have been stable, predictive and founded on key tenants of ability and willingness to pay. Then look at RMBS.


Mark Tapley

The Blue Dragon

Connect with me on LinkedIn:

No comments:

Post a Comment