14/2/2013: New Compensation Model for Rating Agencies

I wrote yesterday about two studies on the effectiveness of the rating agencies (link here). Another interesting study on the agencies and their performance was published in Spring 2012 issue of the Journal of Structured Finance (vol 81 number 1, pages 71-75). Authored by Malesh Kotecha, Sharon Ryan, Roy Weinberger and Michael DiGiacomo and titled Proposed Reform of the Rating Agency Compensation Model, the study looks at the current model of rating agencies compensation - the so-called issuer-pay model whereby issuer of securities being rated paying for the rating delivery - in light of the apparent conflicts of interest implicit in the model. The authors propose an alternative model based on fee levied on new issues and secondary markets trade. Fees would be deposited in a dedicated fund which will pay these out to the rating agencies. The agencies will be rotated on a performance basis, taking into account accuracy of their ratings over time.

The criticism of such an approach to compensation model - as noted by the authors - stems from 

-- the disincentive to rating agencies under the proposed changed model to innovate in ratings models development (higher potential errors etc)

-- the fact that the US bonds market is global in coverage and thus requires competitive pricing systems, 

-- difficulty of devising a merit-based rotating system and setting appropriate levels of fees; and

-- the new model introducing a transactions tax.

Overall, the weakest point of this proposal is undoubtedly its failure to consider the US markets role as global issuance platform with any transaction tax eroding cost competitiveness and the failure in recognising that global scope of rating agencies and the US markets also implies severe limits on new compensation model ability to capture the activities of these agencies.

13/2/2013: Rating Agencies Role & Effectiveness: 2 recent studies

In light of all the cases being filed against the rating agencies and in light of the general controversy surrounding them, here are two recent papers dealing with the topic of rating agencies performance and the links between their ratings and investors' decisions.

One interesting paper was published in the Journal of Banking & Finance )vol 36, number 5, May 2012, pages 1478-1491) by Thomas Mahlmann titled "Did Investors Outsource Their Risk Analysis to Rating Agencies? Evidence from ABS-CDOs".

The paper looks at the floating-rate tranches from collateralised debt obligations (CDO) backed by asset-backed securities to test whether yield spreads at the point of issuance (origination) act as good predictors of future performance. The paper found that, once we control for rating at issuance and other deal-specific data, yield spreads do indeed indicate future performance. However, this result is primarily due to tranches that were initially rated below AAA.

The study concludes that at the issuance / origination, investors did not rely only on ratings, when pricing the CDOs studied. The study econometric evidence also indicates that ratings were inadequate for CDOs because these instruments are much riskier than corporate bonds.

Another paper, published in the International Journal of Finance and Economics (March 2012) by Eduardo Cavallo, Andrew Powell and Roberto Rigobon, titled "Do Credit Rating Agencies Add Value? Evidence from the Sovereign Rating Business" look at the credit downgrades made during and after the financial crises asking whether rating agencies opinions provide any incremental information to the market.

The answer to the question is yes. Rating agencies opinions on sovereign risks do explain a portion of variation in three macroeconomic variables, relevant to the market: exchange rates, stock market indices and future sovereign bond spreads, once the authors control for observed bond spreads.

The study also found that rating agencies upgrades (downgrades) are correlated with:

1. decreases (increases) in the spreads one day forward;
2. increases (decreases) in the stock market;
3. nominal exchange rate appreciation (depreciation) relative to the USD.

Top level conclusion: the ratings do contain information about specific credit and the rating's informational content is in addition to other publicly available market data, such as credit spreads.

23/10/2012: Article in Expresso

A link to the Portugal's Expresso article on sovereign debt risk premia quoting me.