Excess Spread — CMBS shoots, trade finance (again), straying from the light
- Owen Sanderson
Excess Spread is our weekly newsletter, covering trends, deals and more in structured credit and ABS — subscribe to this newsletter here.
CMBS might be back
The European commercial mortgage-backed securities market has been essentially dead for the past five years, with nothing more than a trickle of transactions heavily concentrated by sponsor (Blackstone) and asset class (logistics).
But so far this year we have FOUR different sponsors: Blackstone with a logistics deal, Carlyle with a logistics deal, and Bank of America’s new Taurus 2025-2 transaction, which contains loans to Lone Star and Henderson Park. I’m not going to count the small balance deals from Finance Ireland or Together as true CMBS for our purposes.
This rate is unlikely to continue, but it still points the way to a considerably healthier market. Even before the period of Blackstone dominance, the market was concentrated around single borrower prestige assets, such as London’s Citypoint and Aldgate Tower, or Frankfurt’s Squaire. Some large esoteric portfolios (motorway services, life science labs, cold storage facilities) also made their way to the market.
While these deals were all distributed, it’s not clear how widely they’ve been bought. CMBS in Europe is a very specialist subject, and requires quite particular organization, desire, and energy.
Stepping back, what’s the natural buyer base for large loan CMBS? I think we can safely assume, given the dealflow over the past five years, that very few institutions these days employ people specifically to buy and trade European CMBS.
Part of the reason it’s been such a dead market is partly a reflection of low property transaction volume; if sponsors aren’t selling buildings, nobody needs acquisition debt to buy them.
But it’s much more an effect of the growth of CRE debt funds, who are happy to participate in loan-only syndications and have no need for their positions to be transformed into securities (securitised!). While this market is tighter and deeper than CMBS execution, why bother with the hassle and expense of structuring and rating a CMBS transaction?
Meanwhile, the investment banking teams that used to be CMBS desks have spent more of their time distributing these whole loans and providing back-leverage against CRE loans, or portfolios thereof.
While CRE debt funds don’t need bond-format deals, they may not be averse to it; CMBS structures allow a more tailored exposure. The brain damage of buying a senior loan and back-levering it can be circumvented by just buying mezzanine notes in an existing CMBS transaction.
An open question is to what extent regular securitisation investors want to play CMBS deals. Statistical approaches to granular consumer or resi portfolios don’t necessarily equip investors to underwrite large cap CRE risk; that said, there’s probably a price at which it does become interesting.
An example: triple-A notes in Taurus 2025-2 are talked at 140bps-150bps. That’s 100bps back of prime RMBS, or around 70bps back of the last specialist resi print — for a less levered bond sitting on a whopping 19.9% debt yield.
Admittedly, the only actual triple-A principal loss in European securitisation since the GFC was in a CMBS (Elizabeth Finance 2018, to be specific) and no IG resi bond has even come close. But still, 100bps!!!
Securitisation investors who’ve been in the market a long time will have done some CMBS back in the day, and can dust off their notes; perhaps the point of the CMBS revival is to get some price tension playing off the CRE funds against ABS investors.
The BofA deal is particularly interesting because it represents a partial revival of the conduit model. It bundles two loans with very different characteristics into a single transaction — Henderson Park’s loan, Silverburn, is backed by a single Glasgow shopping centre, while Lone Star’s Wildcat loan funds 35 properties across industrials, offices, and retail, spread throughout the UK.
The loans have different terms, different covenants, different events of default, different cash trap triggers, and wildly different collateral. They are both relatively granular-looking through to the tenants (125 in Wildcat, 100 in Silverburn) but essentially this deal means underwriting both portfolios in detail.
Both loans (sized at £247.8m and £111.6m) are probably too small for their own CMBS takeout, so there’s some efficiency in pooling them, and the resulting larger deal size should mean better secondary market liquidity. It’s up for debate exactly how many CMBS buyers are active in the market because of its strong liquidity characteristics, but it’s all relative: it’s much more liquid than a subparticipation in a whole loan, and that has to have some value.
Bundling together unrelated (and hopefully uncorrelated) loans is a step towards a CMBS market which resembles its pre-crisis pomp, or indeed the US market, which is still functioning. If banks can run industrialised loan-bundling CMBS conduit programmes, the product can become once again a commoditised and ordinary-course-of-business route for banks to bundle and distribute CRE exposures.
There are only two loans in this particular deal, and it’s only one deal, so perhaps this is getting too excited too early. But any green shoot in this direction is encouraging for the sustainable revival of the market.
Oh dear it’s happened again
I took my eye off the ball for a few minutes and whoops, another securitisation-funded non-bank — namely Artis Finance — has collapsed in a puff of disappearing receivables. Rob Smith at the Financial Times has the full story, but you can also take a look at KBRA’s rating of Artis LoanCo 1 (or rather, withdrawal of same).
It’s not pretty. The performance of the portfolio has been poor, and breached triggers, leading to a deal event of default. Per KBRA:
- 24.3% of the Portfolio was in arrears by more than 60 days (includes any performing Loan Asset that has undergone a Pre-Approved Restructuring or an Eligible Restructuring), which amount exceeded the 20.0% threshold.
- 14.2% of the Portfolio was in arrears by more than 60 days (does not include any performing Loan Asset that has undergone a Pre-Approved Restructuring or an Eligible Restructuring), which amount exceeds the 10.0% threshold.
On 18 February 2025, KBRA received confirmation that the at-risk credit insurance, covering 100% of the outstanding loan amounts, remained in place. The servicer subsequently informed KBRA on 27 February 2025 that insurance policies covering buyer receivables for two borrower positions were no longer in effect.
This was followed by Artis Finance entering administration last week, in somewhat awkward circumstances. KBRA again, quoting the notice:
“Amendments were made to recent Servicing Reports which misrepresented the true performance of the Portfolio against various metrics, including that certain financial covenants were reported as met when they were not in fact met. The Joint Administrators have commenced their forensic investigation into these matters.”
The portfolio appears to be mainly commodity companies supplying China, which seems like a lucrative and important business but comes with obvious risks. Or at least, commodities from difficult-to-enforce jurisdictions going to another difficult-to-enforce jurisdiction would seem to present obvious risks — especially when, as in the case of Artis, some of the insurance comes from a related party.
To pull back a bit…why does this keep happening? By ‘this’ I mean ‘non-bank securitisation-funded trade finance provider is embroiled in some kind of suspicious situation’. We have Greensill, the granddaddy of failed trade financiers, but we also have Petra Management, Stenn, Entrepreneur Invest, and Gedesco.
If I ran a legitimate trade finance business, I’d be mad as hell about all these guys letting the side down! The banks in Stenn may be taking a big bath, and frantic credit committees are going to be unwilling to underwrite new deals. Maybe it won’t be forever, but non-bank trade finance is going to be a tricky sell for a long time. There are a couple of big reasons:
- Adverse selection is a major one. Trade finance of various kinds is a big business for many commercial banks; if a company can’t get a bank line, perhaps this is because it is small and poorly served, but also perhaps it is because it is problematic, for reasons of jurisdiction, business line, or proximity to fraud. Adverse selection in consumer assets is well understood — you simply charge more interest to compensate for customers who can’t get a credit card or mortgage from a mainstream bank. But it’s a good business because generally adverse-credit consumers are not actively seeking to scam their lenders; the same may not be true for companies.
- Ease of fraud is another. The general format of trade finance is rapidly revolving credit, supported by bills of lading or, more commonly, invoices. Due diligence will spot-check these documents at the inception of a facility, but nobody is going into ports and kicking bags full of copper on an ongoing basis. Get through the early approvals, and there’s a lot of…flexibility baked in.
Straying from the light
UK specialist lender buy-to-let RMBS is the meat and drink of European securitisation these days, and there are few more storied names in BTL RMBS than Paragon. Admittedly it’s been out of the market for half a decade, but it did 60 deals in the quarter century before that!
One of the earliest trades, as far as I can tell, was in 1994, when a little known Manchester band called Oasis was getting plaudits for its debut album. Paragon’s reporting is so good you can go back and take a look at the trade, Homer Finance No. 3, including £128m of Class A notes across three classes at spreads of 15bps, 20bps and 32.5bps, arranged by Goldman Equity Securities.
If I’ve missed one and you were involved in an even earlier Paragon trade, please write in! Presumably a Homer Finance No. 1 existed. Let’s have a requiem for a once-great securitisation issuer.
Now Paragon has issued its debut covered bond, a three year £500m floating rate note at 60bps backed by BTL mortgage loans. I’m not a covered bond expert, but most reporting around it suggests it’s the first BTL-backed covered bond, with most of the outstanding UK covered bond programmes backed by prime owner-occupied collateral.
This is a little like running out of a 30-year marriage to shack up with the personal trainer, but in truth, the relationship had been souring for a while. Securitisation was once essential to Paragon’s business — shares popped around 10% following the pricing of its post-crisis market comeback, Paragon 16 — but it’s been marginal for a long time since.
RMBS issuance slowed down after Paragon gained a banking licence in 2014, but it continued to print transactions. However, its last placed deal, Paragon Mortgages 26, was in 2019, and deals since then have been retained at closing — as a bank, it now benefits from having own-issued collateral to repo if necessary.
The three year floating rate format of the new covered bond makes it easy to comp to RMBS (and probably means it was sold to the same mix of bank and builder treasuries that you’d see in an RMBS senior note).
At 60bps, it’s still outside prime master trust RMBS spreads, and it’s not a million miles off the 72bps achieved by Enra for senior notes in its Elstree 2025-1 RMBS (which receives much less favourable regulatory treatment for bank buyers).
Enra’s deal was a pure first-charge transaction, mixing owner-occupied and BTL (65%). You can argue the toss over collateral quality: Enra’s deal mixes collateral types, but it is a pure front book transaction, while the Paragon cover pool includes some 2007 and 2008 vintage legacy mortgages, and lots of multi-family / HMOs.
When comparing the products, a traditional pro-RMBS point is to note the greater collateral efficiency in securitisations — but this isn’t totally clear cut.
Paragon has a £900m cover pool supporting a £500m transaction, an advance rate any securitisation structurer would consider obscenely low, while Enra was at 89%. The theoretical Paragon maximum, however, according to Fitch’s report, is 92.5%, and Fitch would give a triple-A down to 95%.
Theory, though, doesn’t equal practice. Covered bond issuers generally steer clear of actually getting near their maximum asset percentages, whereas in a securitisation the number is the number for the actually existing leverage point on the actually existing deal. Is it better to have an imaginary 95% advance rate, or a real rate of 89%?
Depends what you’re after! Most non-bank specialist lenders are in the business of maximising leverage and capital efficiency, eking out holdco notes, financing risk retention and selling down the stack. This is not the game regulated banks are playing. They like capital efficiency, but it’s usually regulatory capital that is the binding constraint, and banks need to keep a certain credit rating if they want to fund through senior or covereds.
From first principles, you might assume a bank that encumbers lots of assets into its cover pool would pay more to issue unsecured funding of whatever kind, but in practice the case is far from proven; placing a big portfolio in the cover pool just isn’t that much of a constraint.
With Paragon blazing the trail, will others follow? There are several well-liked securitisation issuers with big prime BTL books and a desire to fund cheaply. OneSavings Bank would be one example, while newly minted Vida Bank (Belmont Green, sponsor of the Tower Bridge BTL RMBS shelf, gained a banking licence in November last year) is another.
Setting up a covered bond programme takes time, but if the other constraints can be overcome, it still probably delivers a few basis points of savings compared to RMBS. Let’s hope it doesn’t catch on.
Model ‘Y is this happening’
Tesla stock is the retail YOLO security par excellence, with all the attendant volatility you’d expect from shares that are absolutely lousy with zero day-to-expiry options punting — in other words, a million miles away from the stability that generally characterises European ABS.
This week, the Tesla volatility has been mostly negative; shares are down 30% over the past month, and really puked on Monday (down 15%) before retracing a bit later in the week. With a market cap of roughly $800bn, Monday’s drop was worth roughly one Unilever, or a GM and Ford put together. Or three Barclays. You can do the maths.
There’s not much direct read-across from this into European asset-backed markets, but there is, of course, some financing out from European institutions to fund Tesla Financial Services, allowing customers to lease their shiny EVs straight from the dealership. Lloyds and Credit Agricole fund the UK arm, for example, through a facility signed last year (Awesome Range UK Limited is the SPV).
These institutions sit on top of a Class B position retained by Tesla, but we do wonder how Tesla residual values are going to hold up. No car company has ever been this tied up in political drama, and residual value calculations and reporting aren’t really set up to respond dynamically to such radical sentiment changes.
This time last year, Tesla vehicles were desirable prestige vehicles — yes, the company’s ‘technoking’ was a bit of an oddball with a penchant for tweeting from the hip, but he wasn’t actually dismantling the US government.
Now they’re being called ‘swasticars’ and set on fire, and Tesla dealerships are being attacked (and defended by the forces of the US government). New car sales in some European countries are down 70% year-on-year. Tesla residual values were already volatile, with price cuts in 2023 driving EV volatility across all used car markets, so presumably a certain amount of cushion is baked into every Tesla-linked financing.
But this rapid change in Tesla-linked sentiment is unprecedented. Strap in!
Gold top
We’ve got a particular fondness for the agricultural in this newsletter, ever since reporting on Alantra’s privately marketed cow-backed transaction.
Looking through Finance Ireland’s investor deck for its latest auto ABS transaction, we see there’s been an innovative milk-backed transaction staring us in the face for the best part of a decade — and you’d butter believe the credit quality looks excellent.
The milk funding package is for a product called MilkFlex, which provides a loan to dairy farmers secured on the offtake agreements with dairy co-ops. These loans pay 4.5% above Euribor, for lending that has suffered 9bps average expected credit losses in the years since 2020, and had zero defaults. Trust me, yoghurt paid for doing this off-the-run origination!
What makes it attractive from the farmer’s perspective is that it has milk pricing triggers built in. If milk is below 33c per litre including VAT for three months, interest and principal payments are cut by 50% for the following six months; this can be activated four times during the life of the loan. If milk prices are below 31c including VAT, the cut is 100% for the following six months, a feature which can be activated twice.
These loans are then extended, up to a maximum of two years.
The facility is provided by Rabobank, which makes a specialty of agricultural lending, and by Ireland Strategic Investment Fund, an Irish state vehicle that was formerly a Finance Ireland shareholder. Lending is provided through a fund structure rather than a true securitisation.
Finance Ireland described it as an on-balance sheet warehouse facility, and says it has had the same funding parties since 2018. It’s a decent size: €288m of funding provided since 2016, a current €138m portfolio, with €33m of origination in 2024.
The company offers no data on whether these milk-related payment interruptions are hedged away. A sufficiently large reserve fund could surely cope with the volatility, and at 450bps on the asset spread, there should be enough excess spread to make the previously semi-skimmed investors whole.
Our customers receive this content ahead of the crowd — find out more about 9fin’s news and analysis.