On Jan. 17, 2024, S&P Global Ratings published a request for comment (RFC) on its proposed criteria, "Request For Comment: Methodology For Determining Ratings-Based Inputs."

Following feedback from market participants, we finalized and published our criteria, "Methodology For Determining Ratings-Based Inputs," on July 26, 2024.

We'd like to thank investors, issuers, and other intermediaries who provided feedback. This RFC process summary provides an overview of the external written comments and certain other feedback we received from the market on the proposed criteria, the significant analytical changes, if any, we made following the RFC period, and the rationale for those changes.

External Written Comments Received From Market Participants That Led To Significant Analytical Changes To The Final Criteria

We did not receive any external written comments from market participants that led to significant analytical changes to the criteria.

External Written Comments Received From Market Participants That Did Not Lead To Significant Analytical Changes To The Final Criteria

CRA rating use limitation in primary rating drivers

Comment:  We received feedback suggesting that where we use only CRA (credit rating agency) ratings that we had established were sufficiently comparable to our own (primary rating drivers), there should be no limitations to the proportion of such CRA ratings we may incorporate into our analysis.

Response:  The limitations to the proportion of CRA ratings we can use are intended to ensure that the rating we assign reflects our credit opinion and does not merely mirror that of other agencies.

In addition, our determination that CRA ratings are "sufficiently comparable" doesn't establish they are always the same as ours. Rather, it indicates our view that CRA ratings are "sufficiently comparable" for us to incorporate them in our analysis, to a limited extent.

The comparability review includes an analysis of historical default and transition rates and an analysis of overlapping ratings, generally on a broad sector level. The overlapping analysis looks at potential differences between CRA ratings and our own. We do not expect full alignment with our ratings. As a result, there may be differences in credit opinions, which we believe are mitigated by the limited use of such CRA ratings in our analysis, for primary rating drivers. As such, these limitations are an integral part of our framework.

Comment:  Certain comments suggested a single 50% limitation across the board for use of CRA ratings for primary rating drivers, where a comparability review is conducted. Comments also expressed lack of understanding of the reason for the use of varying limits across different sectors, citing collateralized loan obligations (CLOs) and fund ratings as examples.

Response:  Our framework for using CRA ratings considers the materiality of rating inputs to our rating outcome, the potential difference in credit opinions with other CRAs, and the variety of situations that exist. For simplicity, we have established three broad materiality categories along a continuum:

• 1. Inputs that are substantial to our rating analysis, where we do not use other CRA ratings;
• 2. Primary rating drivers for which we believe CRA ratings that are sufficiently comparable to ours can be used to a certain extent; and
• 3. Those that are secondary, where we may use CRA ratings without a limit, on account of the limited impact they have on our ultimate rating outcome.

The second category, primary rating drivers, itself comprises a spectrum of materiality that we seek to reflect with tailored sector-specific limitations.

The potential impact on our rating outcome of differences in credit opinions on the inputs can be larger in certain sectors than in others because of different methodology constructs and the key risk factors we consider in our analysis of credit risk. This is why the limits to using CRA ratings are not the same across all sector-specific criteria under this framework, including within the primary rating driver category. The more material the inputs to the ultimate rating we assign, the lower the proportion of such ratings that can be used. This is to make sure that, in each case, the ratings we assign remain reflective of our own credit opinion.

Comment:  Some commentators suggested that the use of buckets for ratings from other CRAs in the primary driver context effectively means that a minimum percentage of the pool must be rated by S&P Global Ratings. In addition, one commentator was unclear about whether assets carrying CRA ratings above such limits would be subject to any form of "punitive" treatment, such as "notching," in our analysis.

Response:  Nothing in our criteria requires any asset in a pool outside of the established bucket limits to be rated by S&P Global Ratings. We limit our use of ratings from other CRAs in the case of primary rating drivers for the reasons explained in our response to the previous comment.

Beyond these limitations, for assets that are rated by another CRA, we would determine the inputs based on information made available to us and by considering their impact on the ratings analysis. The inputs we determine may be higher, equivalent to, or lower than ratings from other CRAs according to our own assessment of the asset's credit risk. To be clear, we no longer "notch" ratings from other CRAs.

Moreover, assets rated by other CRAs will not be subject to any punitive treatment on account of their source.

Comment:   One commentator suggested that for principal stability fund ratings (PSFRs), we could consider eliminating the limitation on the use of CRA ratings or consider a secondary threshold for assets that carry two CRA ratings.

Response:   Our PSFRs address money market funds' capacity to maintain stable net asset value (NAV) and to limit exposure to principal losses due to credit risk. According to our criteria for this sector, a NAV movement beyond 25 basis points (bps) triggers a downgrade of a 'AAAm' rated fund to 'AAm', and beyond 50 bps, a downgrade to 'Dm', a default under our ratings definitions. Our ratings in this sector are therefore sensitive to spread movements in addition to the credit risk of the assets.

Under our previous criteria, we could use up to 15% CRA ratings in our analysis, but only for U.S. municipal debt. We are now extending the use of this bucket to other asset types, but we are not increasing the bucket size for PSFRs.

Use of private CRA ratings

Comment:  A few comments recommended we use private ratings from other CRAs, as well as public ones, without adjustments. One of these comments suggested this information may be made available to us on a stratified basis. A comment also recommended the use of credit estimates.

Response:  Because we have not done the credit analysis on the assets or entities whose creditworthiness we must consider in arriving at our credit rating, we believe it is prudent to use only credit ratings that are produced by CRAs registered or certified in accordance with relevant regulations and subject to applicable regulatory requirements. This means we would not use other CRAs' credit estimates.

The lack of publicly available information relating to private ratings from other CRAs limits our ability to appropriately assess their comparability to our ratings. As a result, we would not use them as inputs in the context of primary rating drivers.

In addition, unless sector-specific criteria allow for the use of private ratings as part of aggregated data, we don't use them as rating inputs.

Incorporation of sovereign default risk

Comment:  We received a comment that our comparability analysis does not appear to cover ratings assigned above the sovereign. The comment also indicated that the application of a potential adjustment to rating inputs, as a result of the application of our ratings above the sovereign criteria, appears highly complex and introduces uncertainty for CLO managers about the final rating input derived from the CRA's rating.

Response:   We will apply adjustments for sovereign-related stresses only for secondary rating drivers, where no comparability review is conducted. We do so to capture instances where a CRA may predominantly take a different view than we would in our own ratings on the impact of sovereign default risk on rated issuers in a jurisdiction.

However, where exposures are primary rating drivers, as is the case for CLOs, we use CRA ratings that we have assessed as sufficiently comparable to our own. Given the parameters we consider for comparability, as described in our framework, we believe that consistently diverging views, whether on account of sovereign risk or otherwise, would be captured in our review. As a result, we will not adjust CRA rating inputs on account of sovereign default risk stresses.

Issue ratings

Comment:  One commentator observed that the proposal is silent on the comparability of instrument and recovery ratings such that CLO managers would benefit from further clarity on our use of CRA issue ratings when both an issue and issuer rating are available from a CRA (and there is no S&P Global Ratings' rating or credit estimate available).

Response:  Our sector-specific criteria describe which rating types are used--issuer or issue (instrument), for example--depending on the analysis that is conducted.

In the case of CLOs specifically, we seek to capture the creditworthiness of the loan issuers. As a result, we use the issuer credit ratings on the issuers as our rating input, when these are available. When they are not, we have clarified our approach to inferring rating inputs from issue ratings in our updated article, "Guidance: Global Methodology And Assumptions For CLOs And Corporate CDOs."

For the avoidance of doubt, however, we do not use other CRA recovery ratings to determine recovery assumptions in CLOs.

Short-term ratings in PSFR

Comment:  We received a request for further clarity regarding the way we determine short-term rating inputs under our PSFR analysis where an asset carries only a short-term rating.

Response:  We clarified our approach in our updated criteria article, "Principal Stability Fund Rating Methodology" (see the section titled "Investments not rated by S&P Global Ratings but rated by other CRAs").

Other Substantive Market Feedback That Led To Significant Analytical Changes To The Final Criteria

We did not receive any substantive feedback from market participants in a form other than external written comments that led to significant analytical changes to the criteria.

Other Substantive Market Feedback That Did Not Lead To Significant Analytical Changes To The Final Criteria

Comment:  We received feedback questioning our proposal to no longer adjust ratings from other CRAs when using them in our analysis, to account for differences in credit opinions across agencies.

Response:  When we use a credit rating from another CRA as an input, it's to support our overall objective of assigning an S&P Global Ratings credit rating--meaning, a rating that is derived through the application of our own methodologies, a rating that is consistent with our ratings definitions, and a rating that is reflective of our own opinion of creditworthiness. This was the case with our prior approach, which generally involved adjusting other CRAs' ratings, based in part on a statistical analysis, to account for differences with our own ratings, and it remains the case with our new approach.

Both alternatives allow for the use of CRA data as one possible source of information about creditworthiness while taking into consideration the differences in credit risk opinions from other agencies in various sectors so that our rating opinion remains reflective of our own ratings definitions.

Our new approach to using CRA ratings is based on the materiality of those inputs to our overall rating outcomes.

The sectors in which we previously used adjusted CRA ratings fall within the "primary rating drivers" category in our new framework. In this new approach, for primary rating drivers, we use CRA ratings only when they are sufficiently comparable to ours, and subject to a limit based on the sensitivity of our rating outcome to those inputs. In this case, we believe systematic adjustment of rating inputs is not necessary.

The introduction of unadjusted CRA ratings into our analysis for secondary rating drivers reflects their limited impact on our analysis.

Analytical Changes To The Final Criteria That Did Not Arise From Market Feedback

We updated our sector-specific market value criteria ("Methodology And Assumptions For Market Value Securities") to provide that, where less than 50% of the pool carries an S&P Global rating or for assets whose ratings are outside the primary driver/additional liquidity considerations framework, we may be able to apply the principles of the market value criteria if new index data is made available to us that would support the calibration of market value haircuts for such assets in our analysis. This replaces our original proposal in the RFC to apply the credit estimate levels described in Table 11 in these instances. We believe the new approach better reflects our ratings-based input criteria in that it enables the consideration of market data in the analysis of these situations.