Excess Spread — Banks buy loans, the WBS of the future
- Owen Sanderson
Banks buy loans
There’s a clear divide between the investment banks which predominantly provide financing against asset pools, and those which predominantly own asset pools.
It’s not a totally bright line — banks which are mainly principal shops still quote on financing mandates (though probably at wider spreads than the top flow banks), and some institutions, like JP Morgan and Citi (sometimes) do both.
But you can still distinguish Goldman Sachs, Morgan Stanley and Jefferies, mostly principal finance banks, from say…. Bank of America, BNP Paribas, Barclays, Deutsche Bank, Santander, HSBC, NatWest, UniCredit and Société Générale.
All of these are active securitisation banks, with greater or lesser proportions of own-originated collateral and very differently shaped franchises. But all mostly in the financing business, rather than the asset-owning business.
Jefferies and Morgan Stanley have even sought senior financing for their purchases from commercial banks; they’ve acted more like hedge funds than banks (Jefferies, to be fair, is not a bank). That’s not to cast shade; it’s not taking deposits and its cost of capital is high.
But the distinction is becoming more blurry. We touched on the Barclays forward flow for Funding Circle (and the potentially fruitful collaboration with Angelo Gordon) but we should also note the entry of Bank of America into the principal finance business, through the purchase and securitisation of the Auburn portfolio.
Bank of America’s European securitisation business has long been the odd one out among its US investment bank peers.
The bank has a top tier mortgage and RMBS franchise, underwriting everything from the cleanest low LTV owner-occupied portfolios to beaten-up legacy portfolios stuffed with county court judgements and self-certified mortgages.
But unlike its peers, it never buys portfolios itself, and never holds the “risk retention” in RMBS deals — a role which can put a bank on the hook for portfolio representations and warranties, potentially for as much as 50 years, but allows clients such as Pimco and M&G to trade actively out of the portfolio without being hamstrung by regulatory obligations.
Principal finance in mortgages, or “mortgage trading” effectively mean the same basic activity — buying a portfolio of mortgages to be securitised.
It’s possible, in theory, to buy a portfolio of mortgages low, and turn around and sell them high.
But it is quite unlikely. The original portfolio sale was probably a competitive process among the same 10-20 funds. None of these buyside firms (or investment banks) are likely to want to leapfrog the winning bid on a portfolio they’ve already been shown, simply so an investment bank can make a turn. The loan portfolio market trades by auction, not by intermediation.
Securitising them can allow the portfolio exposure to be sold for more than the purchase price, or more commonly, can provide lots of cheap leverage. If market conditions or portfolio performance allow the owner to sell high, it’s easier to sell a small slice of subordinated securitisation bonds than to run a comprehensive auction for a package of whole loans.
Not buying portfolios doesn’t mean Bank of America is shy about committing balance sheet to RMBS — it will buy senior bonds for its own account, offer repo financing in size, and its trading desk has even backstopped various legacy RMBS deals in the past.
But at a very senior level, the bank has historically been a little leary of committing to risk retentions, to buying portfolios outright, or to owning originators directly to feed its mortgage machine. That stems partly from the legacy of the financial crisis.
Bank of America’s unfortunate bet on buying US subprime lender Countrywide cost it north of $40bn in repurchased loans and bonds that failed reps and warranty tests.
Its European travails were less-well publicised, but Merrill Lynch, merged with Bank of America during the financial crisis, owned several UK non-conforming (mortgages to adverse credit borrowers) lenders including Freedom Lending (subsequently Wave), Mortgages plc, and Springboard Finance in Ireland. When securitisation markets froze in 2008, Merrill (and then Bank of America) were stuck with these loans in hung warehouses.
So the Auburn move is quite a bit change. Cerberus, the previous owner of the Auburn portfolio, has been a long time client. It was one of the largest buyers of legacy assets in the UK following the financial crisis, and the nationalisation of Northern Rock. It bought the £13bn “Project Neptune” portfolio, the first major sale from the UK’s bad bank UKAR, as well as portfolios from GE Money, and Capital Home Loans, a buy-to-let lender. There was also an unfortunate sale and repurchase episode from Metro Bank, but let’s not dwell on that for too long.
Bank of America arranged many of the subsequent securitisations under the Towd Point shelf, and committed other balance sheet to funding these investments; where Cerberus was unable to hit its intended selling price levels for securitisation bonds, for example, Bank of America’s trading desk stepped in to finance these bonds on repo.
The trading desk also backstopped one of the refinancings of the Neptune / Granite portfolio — the deal execution collided with one of the crucial Brexit votes, raising the risk of market volatility for UK assets.
The mortgages Bank of America bought were securitised in three deals, Towd Point Mortgage Funding (TPMF) 2018-Auburn 12, TPMF 2019-Auburn 13 and TPMF 2020-Auburn 14. These deals had fairly weak call incentives, and, indeed, Cerberus did not call them when the FORD came due.
This, however, made things more difficult long term. The hike in interest costs after the call date meant that there was no longer sufficient cash flowing through the structure to pay for the interest costs on the bonds which are junior in the capital structure.
That doesn’t trigger a default, but the unpaid interest piles up — making it more and more expensive to eventually call the deal. Some investors expected that Cerberus would effectively walk away from the structures, leaving them to amortise down as the underlying mortgages paid back, down to legal final maturity in 2045.
Instead, as part of a broader attempt to monetise its CHL investment (Cerberus also ran a portfolio sale process for post-2021 origination from the lender), Cerberus marketed the call rights (effectively, the control and equity exposure) in the three portfolios.
I’m not smart enough to work through the economics, but BofA bought the rights, and securitised all three portfolios into a new transaction, Auburn 15 — dropping Cerberus’s characteristic Towd Point prefix.
The deal includes some structural tweaks allowing for the fact that Auburn 13 can’t be called just yet, but needs to wait for the next IPD — there’s a variable funding note (provided by Bank of America itself) which can be drawn down to fund redemption of the final deal in the series.
It also features a liquidity facility, rather than a reserve fund — the former tends to be more efficient for banks, as it is an undrawn contingent exposure, while the latter ties up deal collateral, but often works out cheaper than sourcing a liquidity facility from a third party.
So what’s the mandate of the new BofA business? Is this the start of something new where BofA should be first call? Or just a special trade for a special client? How’s the journey from financing counterparty to true counterparty? Watch this space for more.
We also have to mention Deutsche Bank’s hire of Pankaj Soni, an ex-MD at Goldman Sachs Asset Management. Soni is joining next month, and, though the bank is staying tight-lipped about its plans (decline to comment), this could be a signal that it’s heading back towards the “owning assets” end of the spectrum. Soni’s disclosed trades at GSAM include investments in Fleximise, a UK SME lender, and Zilch, a UK BNPL platform, with no capital structure info, but likely in the deep mezz or subordinated end of things.
Deutsche has a long pedigree as a principal finance bank, and was one of the most active institutions buying legacy portfolios in the 2010s. Turn to Deloitte’s Deleveraging Europe report for 2021, and it lists Deutsche as the buyer of some €20bn in portfolios from 2014, largely in the years 2014-18 (the pre-Christian Sewing “fun Deutsche” era). That puts it just behind Davidson Kempner, but ahead of Fortress, Apollo, Oaktree, Pimco and all other banks including Goldman.
Things have slowed down a bit since then. Looking at the same Deloitte data a little further on, and DB isn’t even top 20 for post-2016 portfolio purchasing. That doesn’t mean the securitised products group has been idle, far from it, but the balance has tilted towards more client business. There was a time when Deutsche was barely seen in flow assets; this year it’s done three out of three UK master trusts.
The Deutsche team certainly have the chops to do principal finance work; inside every flow trader there’s a prop trader waiting to get out. But this hire also looks like a statement of intent.
The ideal securitisation franchise probably has the “buy portfolio” and “finance portfolio” tools available in equal measure. Which other banks are working on risk approvals for the buying end?
Just resting
If whole business securitisation didn’t exist, would one want to reinvent it? There’s a lot of pearl-clutching about Thames Water’s serpentine/labrynthine capital structure at the moment, and CFOs that find themselves inheriting aged WBS financing platforms would generally rather put them out of their misery at the cheapest available price.
Dignity Finance is an interesting case of the above — it’s a whole business securitisation dating from 2002, and (perhaps) not long for this world. NatWest Markets ran a consent solicitation last year seeking bondholder permission to unwind the structure, as a precursor to a (possible) refinancing in 2024.
Bondholders were not promised make-whole (class A would get par and class B 84.25, a big uplift on previous levels) and the consent passed with strong support.
In effect, it’s a sort of contingent unwind — the WBS stays in place, no cash need be found immediately, but you bake in capital structure flexibility. The business can be sold or merged or radically transformed, without the WBS sitting there as millstone around the neck of any potential buyer. Getting rid of this uncertainty gets rid of the valuation complexity as well; paying the bonds at make-whole puts interest rate risk onto the company (or any future buyer), when it might be better borne by fixed income investors, and the valuation gap between
The current shareholders, who took it private in May 2023, but had been invested for several years before that, clearly know it well, and perhaps there’s a sort of financial services / asset management play — funeral home operators like Dignity run funeral plans, which act a little like a mini life insurer, paying out mortality-related liabilities and taking in a steady stream of savings. But Dignity is the largest funeral home operator in the UK already, and it’s shedding assets rather than scaling up.
Also, while bondholders wait for the big moment and the promised refi, the business inside the securitisation is doing poorly. EBITDA margin is at its lowest ever level, and £35m LTM EBITDA isn’t much for a business with an £833m debt stack.
The two big covenants, Free Cashflow DSCR and EBITDA DSCR are both in breach on the basis of business performance; equity is still dripping money into the structure. This is to comply with a previous round of consents and commitments, in which bondholders agreed to waive certain breaches in return for equity cures.
Still, one has to assume the Dignity shareholders have a plan. But it’s been monstrously difficult to execute. UK take-privates, even by an existing anchor shareholder, are long and expensive, and this was followed by several consent rounds culminating in the “permission to refi” last November. The offer announcement was in January, so if there’s value in the business, it took nearly a year for the shareholders to even get themselves into a position to startrealising this value. Stick on an M&A process or further transformation in the back end and it’s a very tiresome journey.
So you can see why CFOs want rid of WBS. On the other hand, it’s undeniably a useful way to get leverage points higher on infrastructure and hard asset businesses, and open up highly levered assets to the deep IG-focused pool of UK real money.
The question for the smart structurers out there — is it possible to do a version of old-style WBS which does have the flexibilities required? How far can you expand the list of possible corporate actions without compromising bondholders?
Staying with call optionality for the moment, there’s a sound (but annoying) regulatory reason to go for overengineered make-wholes / spens structures.
Investments for insurers that are supposed to go in the Matching Adjustment bucket (meaning better Solvency II capital treatment) need to have certainty on redemption date. Flexibility to add extra call optionality is limited, because the redemption of Matching Adjustment instruments can’t be discretionary.
The regulatory point of the make-whole or spens structure is to allow insurers a way around this issue. If they get paid basically the same money and can reinvest it risk-free, then it’s fine to count the callable issue as Matching Adjustment eligible.
Help may be at hand, though. Reforms to the UK’s Solvency II framework will give insurers another bucket, for “highly predictable” cashflows. This is expected to help infra and CRE financing in general, but raises the possibility that the hardwired redemption conditions in future WBS could be relaxed.
WBS but flexible? Surely something can be done.