CMBS SASB Dirty Dozen Update- Replicating Ratings Approach

As noted in my original Dirty Dozen  article, I developed a simplified ratings approach based on the existing rating methodology. Using Moody’s LTV/Ratings tables (DBRS are similar), I calculated a tranche specific LTV by dividing the senior certificate balance by the Net Realizable Value (NRV). The NRV was calculated using 90% of appraised value less 10% foreclosure expenses and advances. I used advances to date plus an assumed additional 10 months of further advances. 10 months is probably optimistic given that many of these loans can be stuck in non-performing status for years. After calculating the LTV, I used a lookup table based on Moody’s LTV/Ratings.

The results can be seen below in the smart table (a new feature here at TPE). This table can be sorted, searched, filtered and heat mapped for certain columns. In addition to TPE Ratings, I include columns for actual ratings and ratings actions since I first started publishing on the Dirty Dozen in May/June. As the table shows, over half the deals (Dirty Dozen plus a few more) have since been downgraded and most by multiple notches thereby jumping entire rating categories. As the table shows most rating actions were 4 or more notches thereby jumping entire rating categories (eg AAA to Single-A) 

I used a combination of sources for the appraisal. Generally I used the ratings agency valuations (sometimes the average). For a couple of the deals I used servicer reported appraisals, or external appraisals as reported in news sources. For other deals I adjusted for split loans and similar. 

This article should be viewed as a counter narrative to the rating agency hegemonic ownership of ratings.  I believe more investors and market actors should replicate and have an independent view of ratings. I built this model in the Python programming language and may put it on our website.

The model is an oversimplification and given some complexities in these deals (eg split loans) I likely got something wrong. If you see anything amiss let me know.

Finally, I don't believe a single asset conforms to the core idea of a securitization which on its face requires some degree of diversification (particularly when the highest ratings are multiple categories above the underlying asset). As you will see in my table, I have a number of investment grade ratings. I personally don't agree that a defaulted asset should support an investment grade rating (though in some cases a BBB or Single A may be warranted if the the collateral valuation and resolution is sufficiently clear to assure the notes so rated would be paid off in full with investment grade certainty.)