Changing the risk-based capital framework is likely to increase charges on BBB, BB and B rated tranches of collateralized loan obligations (CLOs).
- The NAIC perceives a capital structure arbitrage between owning underlying leveraged loans and a vertical slice of the CLO tranches.
- Eliminating this arbitrage is likely to have skewed impacts on CLO rating tranches: A or higher rated tranches could be advantaged, while BBB or lower rated tranches might be disadvantaged.
- For insurance portfolios, Voya continues to emphasize A or higher rated tranches.
A proposed solution to a perceived problem
The National Association of Insurance Commissioners (NAIC) has proposed a process whereby risk ratings of CLOs would be modeled by its Structured Securities Group instead of relying on nationally recognized statistical ratings organizations. The NAIC’s view is that there is currently a capital structure arbitrage between owning the underlying leveraged loans and a vertical slice of the CLO tranches. To that end, they have proposed adding new risk-based capital (RBC) factors ― 6A, 6B and 6C ― which would have base charges of 30%, 75% and 100%, respectively.
We are sympathetic with the premise that there shouldn’t be capital arbitrage between owning a vertical slice of a CLO deal and the underlying loans. But where should the charges change across the debt stack?
Skewed impacts for CLOs
For A or higher rated CLO tranches, moving to a modeling process may result in a capital uplift. AAA and AA tranches, which have never taken a loss across credit cycles, should model out as zero loss and receive an NAIC 1A designation. That probably also holds true for single A CLOs, which have had lower historical losses than A rated corporates and would prove demonstrably “loss remote” in a severe stress modeling approach. If higher rated tranches got lower charges, lower rated tranches would have to get much higher charges to get back to a profile that was capital-charge neutral versus the underlying loans.
We generally view BBB rated CLOs as beta neutral to the underlying leveraged loans, since the credit enhancement in these tranches is offset by the detachment point, the level at which losses would wipe out a levered tranche. They also tend to trade at similar spreads to the spread on the underlying loans, but the BBB tranches currently have a better standalone capital treatment. We calculated an analytical beta — i.e., the covariance of CLO tranche returns and leveraged loan returns divided by the variance of leveraged loan returns — and found that BBB and lower-rated CLOs have a meaningfully higher beta than the underlying loans, which speaks to the concern about those tranches getting favorable capital treatment compared with below investment grade loans.
This calculation is by no means decisive in the risk distinction between the two asset types ― CLO returns can be highly sensitive to technical factors, and the NAIC modeling likely will care more about risk of loss than return volatility.
As of 10/22. Source: Voya Investment Management.
Diving deeper into stress testing
The NAIC’s paper on CLO stress testing for year-end 2020 (link at left) includes an embedded link to the NAIC’s stress testing methodology. Importantly, A rated CLOs only took a 2 basis point (bp) principal loss, even in their most severe scenario. BBB rated CLOs only took a 10 bp loss in a scenario that included historical default rates and a step down in recovery rates from secured loan level to unsecured bond level. The BBB securities did take losses in the most severe scenario, averaging 18% principal losses. We know how the NAIC would think about stress testing CLOs, and single A CLOs still appear to be quite loss remote.
NAIC Capital Markets special report on CLO stress testing: https://content.naic.org/sites/default/ files/capital-markets-special-reports-clo-stress-test-ye-2020.pdf
Potential challenges for lower rated CLO tranches
Considering this background, we expect that in an amended framework, BBB, BB and B rated tranches may be disadvantaged. That seems like the path of least resistance for the NAIC as well, since regulated insurance balance sheets have been smaller players in those market subsegments ― only 13% of total CLO investments on insurance company balance sheets have been in BBB tranches.
There is still a lot to untangle here. The NAIC takes a modeled approach to commercial mortgage-backed securities (CMBS) and residential mortgage-backed securities (RMBS), but those are static pools without the potential for manager intervention. These differences raise important questions for evaluating CLOs:
- Would Tier 1 managers with a successful track record of “building par” over previous credit cycles be granted a favored status in the modeling?
- If they take an approach that models a stress period over a three-year period, similar to CMBS and RMBS, would CLOs closer to the end of their reinvestment period and structural deleveraging receive better capital charges?
- Should there be a distinction between broadly syndicated loan-backed CLOs and middle market CLOs, which currently does not exist explicitly in the ratings factors?
- What happens with the growing commercial real estate (CRE) CLO market?
CLOs are a big enough part of insurers’ balance sheets that a change is likely to happen here, but it will not be a year-end 2022 event. Public comments from regulators have suggested the NAIC’s review could extend into 2024.
The NAIC’s Valuation of Securities Task Force (VOS) recently released a memo describing its proposed rating methodology for CLOs. While the NAIC eventually will model all tranches of broadly syndicated loan CLOs held by US insurance companies, at this stage, its modeling will exclude commercial real estate and middle market CLOs. The NAIC memo seeks comments on the methodology and suggestions for modifications, but leaves undefined the scenarios that will be used in the risk rating process. It therefore remains to be seen how this proposed methodology will affect CLO capital charges.
It’s too soon to take action on any potential changes to CLO RBCs. In the meantime, Voya continues to bias insurance portfolios to A or higher rated tranches, which in our view, potentially offer better risk- and capital-adjusted returns.