Winter is surely coming. One might hope it will arrive without the sorcery, murder, mayhem and intrigue of that memorable HBO show, but surely it will be freighted by its own quantum of trauma and anxiety. Actually, what am I saying? Winter is coming? Winter is already here, but many have elected to not yet get out of the comfy confines of our Barcaloungers and walk outside to notice.
For those who are still wondering, you can stop. We are on the cusp of a material wave of distressed debt. It’s baked into the cake and cannot be avoided by any combination of monetary and fiscal policy, insouciant disregard or magic dust. I won’t repeat the case I’ve been making on this point recently, but higher for longer is our reality. If you think you see a bridge returning to the zero bound, it’s a pier; it’s an illusion. It’s time to embrace the fact that we are confronting a substantial volume of mortgage debt across the CRE space that neither works nor will work for a very considerable period of time. It’s time to adjust strategies and think hard about finding the pony in that manure-filled stall. That is our reality.
That’s what we’re talking about here today. NPL securitizations must be a part of that plan. NPL securitizations were a quotidian part of our post-GFC life and the tools to deal with them can indeed be found at the dusty bottom of everyone’s toolbox.
NPL securitizations developed as a means of obtaining match term back leverage on non-performing loans (NPLs) and sub-performing loans (SPLs) and indeed on REO. They involve a technology that could mix and match different asset categories, performance status and, indeed to a certain extent, could tolerate a small amount of non-CRE assets in the collateral pool. The ultimate leverage wasn’t terribly high and it wasn’t cheap, but it did facilitate the purchase, accumulation and finance of assets that needed to be traded and needed to be repriced. It helped clear out the zombies.
While NPL securitizations came in more than one flavor, the most common structure was called a liquidating trust. This title is a bit of a misnomer, as typically the structure involved a sponsor-owned SPV issuer which issued notes under an indenture for the benefit of noteholders. In many respects, it looked a lot like a modern CRE CLO with considerably dodgier assets.
The notion underlying this structure was that the financial assets held by the issuer would, between current period interest, repayments, prepayments, liquidation and resolution proceeds, net of expenses and operating costs, provide sufficient funds to service the debt. Typically, the technology was capable of producing a modest amount of low investment grade paper at attachment points in the mid-50s against fair value. In most cases, this structure paid off handsomely for the sponsors. Perhaps this was more attributable to timing and luck than structure and the quality of assets, but as cash was recognized earlier and at higher levels than was projected when the transaction was assembled and rated, the results were outstanding.
So, welcome back folks. Moody’s, KBRA and DBRS have criteria for this structure currently available in the market. Fitch is refreshing its criteria and will publish soon.
Here’s what you need to know to build your very own NPL securitization:
- The fundamental legal structure will be pretty straightforward and familiar in the CRE-CLO securitization space. There is a seller, and perhaps an originator which conveys assets into an issuer. There will be an asset manager which is probably an affiliate of the sponsor, a servicer, a special servicer, a custodian, a note administrator and a trustee (and, of course, let’s not forget the placement agents, lawyers, accountants and attendant hangers-on).
- There will be the regular panoply of cash management arrangements (more on liquidity below).
- There will be an MLPA, but not a lot of reps. There may be a guarantor of that limited set of reps. The assumption is that property level data will not be great. Assets may come off multiple platforms and, in many cases, the loans will have never been intended for securitization. However, there’s a core set of data fields which will be required and the agencies have all made clear that in some cases the data may be so bad that no rating is possible. If you can actually get close to completing the CREFC IRP, you will be golden. The core data will include such things as loan term, UPB, payment status, lien status, property type and basic property information, a rent roll (which might be far from complete or current), an appraisal or BOV (again, which may not be current) and sponsor’s cost basis. Environmental, property condition and seismic information will obviously be very desirable.
- Typically, there will be a simple cap stack with rated senior bonds (probably BBB-) and either equity or a subordinate note held by the sponsor. There’s no reason that the structure couldn’t be more complex, more nuanced in its segregation in risk and yield, but, at least in the early days, there will surely be a return to simplicity here.
- The structures are available for NPLs, SPLs and REOs. The structure will even accommodate a little bit of non-real estate collateral tucked in around the edges. All asset classes are grist for this particular mill, including… (gasp!) land. Deals might include performing loans to sweeten levels. They also might include performing loans which are in some way non-conforming, including re-performing loans. Late in the post-GFC world, loans with future funding also made it into these deals without full reserve funding in place and we should probably expect that to happen again.
- Except with respect to very large pools of small assets, the fundamental analysis will be highly granular, built up from the asset level. This will be done using any data that’s available. Each asset will have an attributed cash flow model which will take into account any principal and interest to be paid on the run prior to maturity and resolution proceeds, all net of operating expenses. Operating expenses are clearly things like enforcement costs, payment of property protection amounts, and particularly for REO, even things like capex. The model takes into account the timeline of the resolution and includes an NPV analysis; early resolution is obviously better than a later resolution. From that information, an aggregate cash flow model with a pool is constructed and hair-cutted by the agency based on its internal modeling paradigm. KBRA, for instance, calls this a Resolution Path for each asset.
- The quality of the sponsor and the asset manager will be critical. This analysis is not widely dissimilar from the inquiry into the competencies of a collateral manager, special and primary servicer in a regular-way deal, and includes a review of people, experience, capabilities, policies and procedures. But there’s perhaps a sharper point on the stick here, given the dodgy nature of the assets and the need to manage aggressively.
- One of the key structural issues in these transactions is how cash is managed and how liquidity is obtained. Is there cash leakage, and is there any sort of liquidity facility in the structure? Leakage is (shockingly) not favored by the agencies nor investors, albeit loved by sponsors. Note however that leakage became a regular component of these deals as the market matured. Where leakage occurred, there was typically one or more fast pay triggers which would cut off leakage in the event of the pool’s deterioration.
- Liquidity typically took the form of interest and working capital reserves replenished on an ongoing basis during the lifetime of the transaction. In some cases, funded liquidity facilities may also be viewed as a form of credit enhancement if not subject to leakage, and improve the levels for the transaction as a whole. A third-party P&I advancing is unlikely to be part of the structure (but certainly could be).
- Expect servicing fees to be higher in the NPL deals than regular-way transactions given the nature of the assets.
- From that, we move to the math and the agencies’ haircutting sorcery, which I won’t pretend to understand, and will leave to those more numeric than I. Voila! Levels and attachment points are found.
Each of the agencies has a somewhat different approach to rating these types of transactions. Not to do my Captain Obvious routine here, but it behooves any potential user of this technology to interact with each of the agencies, delivering preliminary tape information and structure to discern whether the variables in one’s proposed tape or structure will be important to the agency and how. By way of example, some of the agencies’ criteria appear to take recourse into account in a much more significant way than others. There are different approaches to diversity (some might argue that diversity is irrelevant in a distressed debt pool). There are differences to the extent the type of legal foreclosure jurisdiction in which the assets are located matters.
So, there’s a path forward. Based on my conversations with the agencies, there is a fair amount of interest illustrated by the number of folks who have asked about criteria. Now, that does not a market make, but in a macroeconomic environment characterized by cyclically high interest rates, a growing conviction that interest rates will remain higher for longer, increased pressure on bank and non-bank balance sheets and the probability of ongoing liquidity constraints, this product might become the best game in town. Leverage is needed (as it always will be). As the saying goes, you may not get what you want, but if you try, you may get what you need.