In December 2014, S&P Global Ratings introduced its CDO Monitor Non-Model framework in order to provide an alternative to its model-based CDO Monitor, and to enhance transparency and operational efficiency for collateralized loan obligation (CLO) market participants while remaining analytically consistent with the model-based approach. (See "Standard & Poor's Introduces Non-Model Version Of CDO Monitor," published Dec. 8, 2014.) In the years since, the non-model approach has been adopted by a large majority of the CLO transactions we have rated. In connection with our updated CLO criteria published on June 21, 2019, we are correspondingly updating some of the assumptions used in the CDO Monitor Non-Model approach, as well as making other changes to further simplify the framework and respond to market feedback.

The CDO Monitor Test (both the model-based and non-model versions addressed in this article) provides an indication as to whether changes to a collateral portfolio during the reinvestment period are generally consistent with the parameters assumed when initially assigning ratings to notes in a CLO transaction. If a CLO fails its CDO Monitor Test, transaction documents typically require the CLO manager to "maintain or improve" the test results for subsequent trades until the test results are brought back into compliance. The CDO Non-Model version replaces the model version with six portfolio benchmarks and some formulas that can be incorporated into CLO desktop compliance systems (for details on the formulas, see the Assessing Portfolio Credit Quality: CDO Monitor Non-Model Test SDR section below). As a result of some of the changes to our CLO criteria, a recalibration of the formula of the non-model CDO monitor is necessary. Additionally, for greater transparency and comparability, two of the six portfolio benchmarks have been revised.

For those CLO portfolios that, in our view, are less diverse relative to a typical broadly syndicated loan (BSL) CLO portfolio (i.e., warehouse CLOs or other atypical portfolios), the CDO Monitor Model (or an alternative regression formula applied using the same six benchmarks) may be more suitable for generating the applicable Monitor Test scenario default rates. For these cases, we encourage market participants to highlight to us as early as possible if an atypical portfolio is expected.

CDO Monitor Background

The non-model version of CDO Monitor relies on the same analytical output as the model-based CDO Monitor approach: the senior tranche scenario default rate (SDR) at the initial rating level and the senior tranche breakeven default rate (BDR). Both approaches are meant to approximate the SDRs and BDRs we use in our analysis of CLO transactions, based on model results produced by CDO Evaluator and S&P Cash Flow Evaluator, respectively. The SDRs from CDO Evaluator represent the expected default rate of the assets underlying the CLO portfolio, based on the economic scenarios we associate with our ratings. The BDRs from S&P Cash Flow Evaluator represent the level of defaults a rated CLO tranche can withstand in our modeling without any impairment. The most powerful drivers of changes in the SDRs typically tend to be the ratings on the obligors, obligor diversity and the weighted average life of the collateral pool; and the most powerful drivers of changes in the BDRs tend to be par gains or losses and the expected recovery levels for the assets in the pool, based on our recovery ratings assigned to the loans and bonds.

Chart 1

As the CLO begins its reinvestment period, the composition of the CLO portfolio may change as a result of manager trading as well as shifting market conditions. The Monitor Test BDR and Monitor Test SDR are typically generated with the current parameters of the CLO portfolio as described below. The "cushion" derived from the CDO Monitor test is the difference between the BDR and SDR, which are both compared at the same rating level. For example, if a CDO Monitor Test for a given CLO transaction were failing and the transaction was required to "maintain or improve" the failing test results for subsequent trades, a trade that reduced the amount of the failure, or achieved passing results from the Monitor Test, would generally be required in order to satisfy the "maintain or improve" language in the CLO transaction documents.

There may be instances when a transaction does not issue a 'AAA' rated tranche, or it does but we did not assign a rating to it. In response to market feedback, as part of the update to the CDO Monitor Test approach outlined in this article, we are adding the ability for the CDO Monitor Non-Model test to rely on the 'AA' SDR and BDR, in addition to the previously available 'AAA' tranche test results. Note that using the 'AA' tranche metrics in the CDO Monitor Test is generally only an option for transactions that do not have a 'AAA' tranche rated by S&P Global Ratings. Table 1 outlines the two calibrations for the test.

Table 1

Applying The CDO Monitor Test To S&P-Rated CLO Tranches: The Highest Rating Level

The CDO Monitor Test will be applied at the highest rating category of the S&P-rated note of the CLO transaction at close (excluding X notes and/or combo notes). For example, if a CLO transaction includes a CLO tranche where the highest S&P Global Ratings' credit rating assigned is 'AA', then the CDO Monitor Test is generated at the 'AA' rating level. Equally, if S&P Global Ratings assigned, for example, a 'AAA' rating to one or more of the CLO notes (excluding X notes and/or combo notes), then the CDO Monitor Test is generated at the 'AAA' rating level. If the rating on a CLO note originally rated 'AAA' is lowered to 'AA' during its reinvestment period, the CDO Monitor Test will continue to be generated at the 'AAA' rating level.

Table 2 summarizes the impact on CDO Monitor parameters, and subsequently BDRs and SDRs, following a decline in the CDO Monitor Test cushion.

Table 2

Assessing Portfolio Credit Quality: CDO Monitor Non-Model Test SDR

As with the previous CDO Monitor Non-Model approach, this updated approach is built upon six simple portfolio benchmarks that are intended to provide insight into our view of CLO portfolio characteristics and credit quality. Each benchmark captures a component of portfolio risk taken into account by our corporate CDO criteria. The portfolio benchmarks enable CLO market participants to compare metrics across CLO portfolios, as well as assess changes within a given portfolio over time. A description of each of the six benchmarks is provided below.

The six benchmarks described below, and the remainder of the approach, can easily be incorporated into CLO compliance systems. Alternatively, market participants can build a simple spreadsheet that allows for calculation of the test results (see "How to Build Your Own CDO Monitor E8 (Non-Model Version)," published on June 21, 2019).

S&P Global Ratings' weighted average rating factor (SPWARF):   The SPWARF of a CLO portfolio provides an indication of the overall credit rating distribution of the portfolio weighted by each assets par balance. The rating factor for each of the portfolio assets is determined by S&P Global Rating's credit rating (or implied rating) and the rating factor table (see Appendix). The SPWARF is calculated by multiplying the par balance of each collateral obligation (with a rating from S&P Global Ratings of 'CCC-' or higher) by S&P Global Ratings' rating factor, then summing the total for the portfolio, and then dividing this result by the aggregate principal balance of all of the collateral obligations included in the calculation.

Default rate dispersion (DRD):   This is the weighted average absolute deviation of individual asset rating factors in the portfolio relative to the SPWARF, weighted by asset balance. This is essentially a measure of how dispersed a portfolio is with respect to the asset rating factors around the SPWARF; a higher DRD benchmark means a wider dispersion of derived S&P Global Ratings' credit ratings. The DRD is calculated by multiplying the par balance for each collateral obligation (with a rating from S&P Global Ratings of 'CCC-' or higher) by the absolute value of the difference between the S&P Global Ratings' rating factor and the S&PWARF, then summing the total for the portfolio, and then dividing this result by the aggregate principal balance of the collateral obligations included in the calculation.

Weighted average life (WAL):   This is the portfolio's average life, weighted by asset balance, based on the final maturity of each loan in the portfolio. The WAL is calculated by determining the number of years between the current date and the maturity date of each collateral obligation (with a rating from S&P Global Ratings of 'CCC-' or higher), then multiplying each obligation's principal balance by its number of years, then summing the results of all obligations in the portfolio, and then dividing this amount by the aggregate par balance of collateral obligations (rated 'CCC-' or higher).

Obligor diversity measure (ODM):   This is a measure of the "effective" number of obligors in a collateral pool, taking into account the size of each obligor as a proportion of the overall portfolio. For purposes of calculating ODM, multiple loans from a single obligor in a portfolio should be rolled up together. The ODM is calculated by determining the aggregate principal balance of the collateral obligations (with a rating from S&P Global Ratings of 'CCC-' or higher) from each obligor and its affiliates, then dividing each of these amounts by the aggregate principal balance of obligations (rated 'CCC-' or higher) from all the obligors in the portfolio, then squaring the result for each obligor, and then taking the reciprocal of the sum of these squares.

Industry diversity measure (IDM):   This is a measure of the "effective" number of industries in a collateral pool, taking into account the size of each industry as a proportion of the overall portfolio. When calculating IDM, industry categories from our corporate CDO criteria should be used. The IDM is calculated by determining the aggregate principal balance of the collateral obligations (with a rating from S&P Global Ratings of 'CCC-' or higher) within each S&P Global Ratings industry category in the portfolio, then dividing each of these amounts by the aggregate principal balance of the collateral obligations (rated 'CCC-' or higher) from all the industries in the portfolio, then squaring the result for each industry, and then taking the reciprocal of the sum of these squares.

Regional diversity measure (RDM):   This is a measure of the "effective" number of regions in a collateral pool, taking into account the size of each region as a proportion of the overall portfolio. When calculating RDM, the country and region classifications from our corporate CDO criteria should be used. The RDM is calculated by determining the aggregate principal balance of the collateral obligations (with a rating from S&P Global Ratings of 'CCC-' or higher) within each S&P Global Ratings region category, then dividing each of these amounts by the aggregate principal balance of the collateral obligations (rated 'CCC-' or higher) from all regions in the portfolio, then squaring the result for each region, and then taking the reciprocal of the sum of these squares.

Summary of CLO portfolio benchmarks 

Chart 2

Converting the six portfolio benchmarks into the Monitor Test SDR

The six benchmarks are intended to provide transparency into the various portfolio characteristics, allowing for comparison among different CLOs and collateral managers (note that the benchmarks, along with other metrics, are published for a large proportion of our rated CLO book each quarter). When expressed along with the regression formula below, the six benchmarks also serve to provide the Monitor Test SDR for the CDO Monitor Test. Which of the two formulas below is used for a given CLO transaction will depend on the ratings of the most senior credit class rated by S&P Global Ratings, as described in table 1 above.

Monitor Test SDR (AAA) = 0.247621 + (SPWARF/9162.65) – (DRD/16757.2) - (ODM/7677.8) - (IDM/2177.56) - (RDM/34.0948) + (WAL/27.3896) 

Monitor Test SDR (AA) = 0.137223 + (SPWARF/8829.01) – (DRD/20413.6) - (ODM/9556.72) - (IDM/2256.55) - (RDM/40.2751) + (WAL/26.7396) 

The formulas presented are applicable for CLOs with well-diversified portfolios to generate the Monitor Test SDR.

Setting The Test Threshold: CDO Monitor Test BDR

In our rating process, the BDRs are calculated using our cash flow model, Cash Flow Evaluator, under stress scenarios outlined by our corporate CDO criteria. For the BDR Monitor Test of CDO Monitor, we use the deal specific cash flow model at closing to generate a range of BDRs across different weighted average recovery rate (WARR) and weighted average spread (WAS) values, representing a broad range of potential outcomes, and then create a transaction-specific regression formula based on the results.

WARR:   This is the weighted average recovery given S&P Global Ratings' recovery ratings on the underlying loans from issuers with a performing S&P Global Ratings' credit rating and the recovery table from our Corporate CDO criteria, at the applicable rating category ('AAA' or 'AA'). The Monitor Test cushion is reduced if the WARR of the portfolio is reduced, all else equal.

WAS:  This is the weighted average nominal spread of floating-rate loans from issuers with a performing S&P Global Ratings' credit rating.

Deal-specific 'AAA' BDR coefficients or 'AA' BDR coefficients will be provided at close of the CLO transaction:

'AAA' Monitor Test BDR = C0 + (C1*WAS) + (C2*WARR) 

'AA' Monitor Test BDR = C0 + (C1*WAS) + (C2*WARR) 

WARR for the 'AAA' Monitor Test BDR is the weighted average S&P Global Ratings' recovery rating at the 'AAA' level as per our CLO criteria. WARR for the 'AA' Monitor Test BDR is the weighted average S&P Global Ratings' recovery rating at the 'AA' level as per our CLO criteria.

Monitor Test BDR par adjustment

There is one final step required before we can incorporate the BDR into the CDO Monitor Test. We need to adjust for changes in par compared to the CLO's effective date target par balance, to give credit for par gains, or to account for par lost. This is done using the BDR scaling formula provided below, which is the same across CLO transactions:

Monitor Test Adjusted BDR = BDR * (OP/NP) + (NP-OP) / [NP*(1-WARR)] 

Where BDR = Monitor Test BDR ('AAA' or 'AA'); OP = target par balance on the effective date; NP = current par balance (performing collateral + principal cash + redemptions to the senior most class during the reinvestment period + lower of market value or S&P Global Ratings' 'AAA' or 'AA' recovery assumptions for non-performing collateral (as the case maybe)); and WARR is the applicable weighted average S&P Global Ratings' recovery rating (see BDR formula above).

Putting It All Together: Monitor Test Result/Cushion

Given a deal-specific BDR formula, we can incorporate changes in par, recoveries, and spread of the portfolio to generate a Monitor Test BDR. Given the SDR formula (above), we can incorporate changes in the credit and diversity related parameters of the portfolio to generate a Monitor Test SDR. Putting it together to calculate the Monitor Test cushion, we use:

Monitor Test Cushion = Monitor Test Par Adjusted BDR - Monitor Test SDR 

As long as the cushion of the CDO Monitor Test is positive--i.e., the expected default rate of the portfolio is less than what the CLO tranche can withstand without impairment, or the cushion is otherwise maintained or improved--the CDO Monitor Test is considered satisfied.

Related Criteria

• Global Methodology And Assumptions For CLOs And Corporate CDOs, June 21, 2019

Related Research

• How To Build Your Own CDO Monitor E8 (Non-Model Version), June 21, 2019
• Credit FAQ: Key Considerations For September 2018 Updates To CDO Evaluator Industry Codes, Sept. 26, 2018

Appendix

The S&P Global Ratings' rating factor of an individual asset is the five-year default rate given its S&P Global Ratings' credit rating and the default table in the Corporate CDO criteria, multiplied by 10,000.