Rating Method CSC
CSC is a traditional credit rating agency (CRA) which rates credit issuers—borrowers-and credit instruments — the stream of contractual future cash flows that borrowers repay over time.
Investors in credit instruments—lenders—have to worry about two main types of risk: market, or changes in the price of money; and credit, the likelihood that a borrower can’t or won’t repay the instrument on time and in full. Market risk is thoroughly explored in academic finance, whereas credit risk has been relegated to CRAs.
CSC differentiates from the rating agencies by starting with the fact that the same credit instrument can be held to maturity on one balance sheet or repackaged, sold and warehoused on another balance sheet. To promote consistency, CSC’s analytical starting point is point-in-time ratings from cradle to grave in a holistic, traceable approach.
Whereas corporate finance is top-down, tending to infer receivables’ performance based on the balance sheet where they are warehoused, CSC’s approach is bottom-up. The balance sheet risk is influenced by the quality of receivables it is currently warehousing.
The CSC framework considers the entire credit-granting circle (CGC) as a whole. First, a loan is originated. Then, it may be held to maturity or pooled with other loans and tranched. Securities produced this way can be re-rated using time-series data from origination to the present. The capital structure, the data and the rating scale together form a cybernetic or control structure where actual performance is measured, again and again, against theoretical.
Structured and Loan Ratings: Scale & Methods CSC’s ratings on structured securities and loans is a measure of expected impairment based on the promised rate of return. It can incorporate changes in the index or reflect pure credit impairment.
Ratings are not reproducible by third parties without a critical element: the rating scale.
CSC ratings are anchored to an unchanging numerical scale, where the values are uninfluenced by the cycle. The scale represents a system of credit grades, which are measures in basis point intervals of expected impairment across a large number of scenarios. The measures represent the distance between the security’s payment promise (the ideal) and its empirical average yield-to-maturity (YTM) as of today.
The mapping to alphanumeric ratings is as follows:
Bankruptcy-remoteness is a separate legal analysis. For issues from a bankruptcy-remote SPE:
- CSC simulates expected future collateral cash flows from “t” to maturity “t=T,” iterating a large number of times. The appropriate number depends on how many variables simulated and how wide the dispersion. If focusing on loan collateral PD, 2500 usually produces a nice distribution.
- At each iteration, simulated cash flows are passed through the precise capital structure and the resulting YTM is recorded for each tranche.
- Compute the YTM with a risk-free (not risk-weighted) hurdle rate because these cash flows already reflect the risk.
- Give each recorded result an equal weighting and compute the average YTM.
- The payment promise net of the YTM is the measure of expected impairment. Map it to the rating scale, interpolating the distances where necessary. Remember that the scale is not linear.
Because CSC ratings “follow the money,” the subroutines start with assets and end with liabilities:
- The collateral analysis in (1) above simulates the composition of accounts likely to pay as expected, prepay or default, and the recoveries. Defaults are assumed to follow a logistic process. Prepayments and recoveries are estimated from recent trends.
- The security (liability) analysis (2) above proceeds from modeling the waterfall, a logical architecture, in its entirety, i.e., precisely reflecting the payment priorities and triggers. Because this is difficult to do in Excel without taking shortcuts, CSC uses a proprietary software, WaterfallEditor©, that replicates deals in their language.
- The final rating is close to the mapping in (3) above.
Ratings at Origination versus Re-Rating Seasoning ABS*
When CSC rates the transaction at origination:
- The key unknown and driver of credit risk in ABS*—the collateral Expected Loss (EL)—is a guestimate, no different than for any other credit rating agency. Uniquely, thereafter, CSC will update the EL based on new data using its proprietary machine-learning tool, the ABSTRAK®.
- ABSTRAK® liberates the rating from the stagnant analysis that results when ratings are overdependent the initial EL guestimate, with all its shortcomings. Besides accurate point-in-time ratings, multi-dimensional metrics facilitate precise, multi-dimensional value-from-credit analysis.
When CSC rates a seasoning transaction, we assume that whatever has happened post-origination to the senior class (the entire stack) they merited the original rating on Day One.
- CSC first performs a calibration exercise to solve for the collateral short-rate volatility that makes the senior class (or classes) output an average reduction of yield close to the numerical value of the associated rating. EG:
- If the seniors have a Aaa rating, the collateral should have an instantaneous volatility that produces a YTM, which, net of the payment promise, results in a loss of 0.025 BPS—the numerical definition of Aaa.
- If the seniors were wrapped at a Baa1 attachment point, the short rate volatility is the number that ultimately produces an average 23 BPS reduction of yield on the security. If the security has a 6% yield at par, the expected YTM will be 6.00% - 0.23% = 5.77%.
- If it is impossible to solve for the short-rate volatility, the deal as structured was infeasible. (This does not mean it will not repay in full. It means the source of funds from repayment will not come exclusively from the deal, and the investor has not been adequately compensated for the full risk of the deal.)
- Having solved for the short-rate volatility, the analysis proceeds as above.
The algebra of cash flow simulations and waterfall allocation are detailed exhaustively in "Elements of Structured Finance"  and "The Analysis of Structured Securities"  by Oxford University Press.