Wall Street’s Useful Idiot: Financial Times Shills for CLOs….as Fed Hasn’t Bailed Them Out
It’s not a popular position to point out that a particular financial risk is overblown. But when everyone in Corporate American and investor-land is in “Where’s my bailout?” mode, the usual motivations are reversed. Normally, “Nothing to see here, move along” is the default position when the great unwashed public worries about too much leverage, opacity, and tricky practices. But when central banks are doling out trillions, sounding alarms, whether warranted or not, is the way to get someone else to eat the risks you took for fun and profit. And as we’ll demonstrate, the Financial Times looks to have become an unwitting tool of CLO (collateralized loan obligation) investors who haven’t yet gotten their Fed handout.
The article in question is headlined: CLOs: ground zero for the next stage of the financial crisis? The headline and the breathless tone set the reader up for the idea that these complex structures will blow up the financial system, just the way their cousins, collateralized debt obligations, did in the financial crisis.
But as we’ll explain, the absolute size of the CLO market, and banks’ not-much exposure to the risky parts of it, means that absent fraud (or a systemically important bank and wobbly bank having gotten high and binged, and the pink paper and others would likey have gotten wind of that by now), there’s no risk to the banking system. And in the hoary old days of the crisis just past and its predecessors, that was the justification for throwing official money in big volumes to clean up bad lending decisions: that as much as it would seem proper to let incompetent institutions go tits up, letting banks fail tends to engulf even healthy banks, since no one can tell from the outside very well how solvent a particular institution is, and hurts innocents like depositors (even with deposit insurance, it is pretty much impossible for a business of any size not to have way above the guaranteed amount in its accounts regularly, if nothing else when it issues payroll).
So what the Financial Times piece is effectively getting worked out about is that some investors will lose money. Newflash! Investing involves risk! Who’d have thunk it!
In other words, this Financial Times piece is implicitly selling the idea that the consequences of some deep pockets taking hits is just oh-too-dangerous. This is Greenspan put thinking on steroids. And sadly, it seems to be treated as a reasonable line of thinking. Wealth must be spared. The hell with those who live from labor income.
What CLOs Are and Why Comparisons to CDOs Are Spurious
By way of background, a pet peeve of ours is that the financial crisis is widely depicted as a housing crisis when it was in fact a derivatives crisis. If we had merely had subprime and Alt-A loans go bust, the result would have been on the order of the S&L crisis plus maybe another 50%. Bad but nothing like the seize up of the global financial markets that took place in September 2008.
As we explained long-form in ECONNED, collateralized debt obligations consisting substantially of the risky tranches of subprime mortgage bonds (the BBB and BBB- layers) dressed up the part of of subprime securitizations that no one wanted to buy. The top tranches of those CDOs, comprising 60+% of par value, were rated AAA.
But even that wasn’t sufficient to blow up the global financial system. Demand by subprime shorts and banks seeking protection for the loans they were advancing to the likes of subprime lenders like New Century and IndyMac led to the use of substantially synthetic (made mainly of credit default swaps rather than tranches of actual bonds with actual mortgage loans behind them) as a way to generate artificially cheap insurance on the riskiest rated layer of subprime debt. Those derivate exposure have been estimated at 4-6X the real economy exposures. The reason the financial system blew up is that the side bets were a significant multiple of real economy activity, and the parties on the wrong side of those wagers were systemically important, highly leveraged players like AIG, the monolines, Eurobanks, Citigroup, and Merrill.
It is also important to remember the base line. Banks have repeatedly managed to blow themselves up in entirely conventional ways, with good old fashioned loans. Remember the Latin American debt crisis? The commodities bust of the early 1990s, which crated energy and real estate lenders in the oil patch? The afore-mentioned S&L crisis? The not as well publicized LBO loan crisis of the late 1980s and early 1990s? Oh, and how about one of the mothers of bad lending, the Japanese commercial (and to a lesser degree residential) real estate bubble and bust?
Now let’s look at CLOs. They are intrinsically less risky than asset-backed-securities CDOs, which the press took to calling just “CDOs”.
Those CDOs were resecuritizations. They were created from the riskiest parts of mortgage securitizations made of risky loans, as in subprime or Alt-As. Those tranches usually represented only 3% of the entire deal. They’d be worth 100% if the losses on the subprime RMBS was 8% or less, and totally wiped out if the losses were higher than 11%. Since subprime losses averaged more like 40%, nearly all these CDOs were complete wipeouts.
By contrast CLOs are made from risky corporate loans, so-called “leveraged loans” typically made when a private equity firm is buying a portfolio company. So they are more analogous to a subprime RMBS (residential mortgage backed securitization) in terms of the risk level of the assets.1 Note that even with the high level of subprime losses, AAA tranches of RMBS lost only 0.42% on average.
Even though there was a tidal wave of risky “leveraged loans” to fund deals at sky-high prices right before the last crisis, and a lot of them wound up in CLOs, while their value traded down afterwards there were no losses. Admittedly, the Fed dropping interest rates and manipulating long-term rates lower allowed many of the pre-crisis loans to be refied at lower rates. But the Fed didn’t engage in measure intended to rescue corporate borrowers; they were just lucky beneficiaries.
Nevertheless, CLO structures have been made more conservative since the last crisis. See here for geeky details.
But with the bottom dropping out of the entire economy, a lot of former sure-looking bets won’t work out so well.
But should we even care about CLOs? First, the market isn’t all that large. S&P in early 2020 pegged it at $675 billion. By contrast, the subprime market, depending on whether or not you included Alt-As, was estimated back in the day at $1.3 trillion to a bit over $2 trillion. And US GDP was $14.5 trillion in 2007 versus $21 trillion in 2019. Using Pimco’s forecast of 5% contraction for the full year, you still get roughly $20 trillion, showing that the relative importance of subprime lending has fallen even further. And that’s before you factor in the way that credit default swaps greatly magnified the real economy exposures and concentrated them at systemically important, fragile players.
This chart from Guggenheim Investors is a little dated, since it shows a smaller CLO market size but it gives a good idea of who buys what:
The key bit: what do banks own? They are the ones that have to be saved even when they do really dumb things; the other players are supposed to be savvy investors who can take losses.
You can see that banks are the big owners of the AAA tranches, which have shown over time to have enough loss protection that you’ll get all your money. And they are listed third as owners of other investment grade tranches, which means they are less important players than insurers and money managers.
More detail from S&P in March 2020:
CLO holdings at U.S. banks increased by roughly 12% in 2019, to $99.5 billion, with a number of banks growing their CLO securities exposure by double-digit rates, according to year-end filings with the Federal Reserve….
What is perhaps most striking about the results is the high degree of concentration of CLO debt inside the three banking entities, relative to the rest of the banking sector. At $80.2 billion, the three largest holders make up nearly 81% of all U.S. bank CLO holdings.
However, the size of CLO holdings as a percentage of the top-three banks’ investment portfolios is relatively small. For JPMorgan Chase, Wells Fargo, and Citigroup, CLOs as a percentage of total securities on book are 7.5%, 6.9%, and 6.0%, according to Y-9C data.
Now, at just under $100 billion, total U.S. bank exposure represents roughly 15% of the $675 billion CLO market. Assuming that U.S. CLO AAA supply is around $400 billion and that most of U.S. bank CLO holdings are AAA, then U.S. banks can be said to hold roughly 25% of all AAA supply, according to researchers from Wells Fargo.
Japanese banks can also be expected to hold near the same amount, or around 25% of the total AAA supply of U.S. CLOs, according to Wells Fargo.
So who is at risk? Apparently the equity tranche, the bottom 8.5% in the chart above:
Waterfall of payments in CLOs, via Morgan Stanley, which estimates that 85% of outstanding CLOs in the U.S. may fail their junior over-collateralization (OC) tests as leveraged loans default. pic.twitter.com/cx2GFlAXL5
— Tracy Alloway (@tracyalloway) May 7, 2020
To translate: securitizations have risk buffers in the form of overcollateralization and excess spread. “Overcollateralization” results from the fact that the loans in a securitization have a higher principal amount that than the total par amount of all the tranches. Say $100 of loans have $103 of loans. The $3 is the overcollateralization and the CLO documents describe who gets how much benefit from it.2
What the tweet is saying is that the losses on the loans look like they’ll get to be high enough that the equity layer in some (many? most?) CLOs will stop getting interest payments. But again, why should we care? If all the equity tranches stop paying out, that was shy of $60 billion in par value. And it’s not a total bust since they got payments in full before their investment crapped out. And the owners are primarily hedge funds. This is just not worth getting lathered up about, particularly in comparison to all the other exploding loss-bombs.
Likely Reason for the Financial Times’ Hyperventilating
Despite being a “Long Read,” meaning it presents itself as having more reporting behind it than a typical story, it’s remarkable to see the Financial Times duck the real story, private equity leverage, which predictably leads lots of lenders holding the bag at the end of every financial cycle.
And the article also sidesteps another key question: did CLOs make the situation worse than the old-fasioned approach of syndicating leveraged loans (which foreign banks took down in size in the first LBO wave in the 1980s)? Quite honestly, I suspect you can argue it both ways. On the one hand, unless the benefit of CLO risk slicing and dicing went entirely to the sponsors and packagers (possible), investors presumably were willing to accept lower interest rates to get more finely tuned risk exposures. That would somewhat lower the cost of private equity lending, but it’s not clear that this made anything more than a marginal difference.3
On the other, CLOs per the discussion above really did move a lot of the risky exposures out of the hands of usually self-destructive banks and over to investors who bill themselves as sophisticated and able to take risk.
The Financial Times account does provide market size (using JP Morgan data, which gives a slightly higher level that S&P did), albeit a bit of the way into the piece, but no distribution of exposures by rating or who the various investors are. And before the authors get to that, they hand wring over a debt restructuring of medical staffing company Envision Health. As vlade pointed out by e-mail:
The headline is fearmongering and a lot of comments lapped it up.
Thinking about it, the picking of the health company and saying how it’s going to suffer is also there, implying “if you don’t bail out the debt, look at how bad things will happen.”
Vlade is not exaggerating. The article waves the “saving us from coronavirus” flag:
In the midst of a global pandemic, emergency rooms across the US have fallen strangely quiet as patients with other illnesses have stayed away for fear of contracting Covid-19. As a result, one of the surprising corporate casualties of the coronavirus crisis could be some of the companies that provide staff for hospitals.
Envision, one of the largest medical staffing companies, completed a restructuring of its roughly $7bn of debt this month as it moved to stave off bankruptcy.
US readers of this site know better than to see Envision as a good actor. As we’ve written several times, based on the gumshoe work of private equity expert Eileen Appelbaum, KKR-owned Envision and Blackstone’s Team Health have been lead players in the “surprise billing” and other price gouging schemes. The fact that Congress and some states have been working on legislation to end surprise billing is why Envision’s bonds are in trouble. And that reaction also shows how important the fleecing is to the company’s bottom line.
In fact, one has to wonder if the Envision staffing cuts mentioned later in the piece are a shot across legislators’ bows: “Nice ERs you have. Shame if something were to happen to them.”
This chart also serves to exaggerate how bad things are:
The viewer’s eye naturally goes to the top line, for the BB tranches the yields have risen the most. Those BB instruments are a mere 3-4% of the total value of the CLO market! They are rounding error. By contrast, the AAA tranches are yielding pretty much what they did in later 2019. Yes, they aren’t trading the same as other AAA instruments, but seasoned structured credit investors ought to know that these AAA creations can and often do trade down in crises, unlike Treasuries and government guaranteed credits.
Moreover, eyeballing the charts, yields have improved since mid-late April, meaning investors are less edgy, when even then, Reuters reported that new CLO deals were getting done on reasonable terms. From an April 22 story:
There has been US $39.4bn of US CLOs arranged this year through April 19, inline with the US$38.7bn sold during the same period last year, according to the data from LPC, a unit of Refinitiv. A record US$128.1bn of US CLOs was arranged in 2018.
US CLO issuance this year has been challenged as spreads on Triple A tranches, the largest and most senior piece of the funds, have continued to widen, hitting an average of 136.2bp in March, slightly tighter than the more than two-year wide of 138bp in February, according to the data. The wide spreads can cut into returns paid to equity holders, who are paid last after all other debtholders receive their distributions, and may cut into overall issuance.
“Spread levels on US CLOs still look relatively attractive compared to other securitized products and corporates,” Collin Chan, a CLO strategist at Bank of America Merrill Lynch, said in an email…
“Although the arbitrage for new-issue deals has compressed, (CLOs) are still getting done because equity investors still find the levels acceptable,” Chan said. “Insofar that we do not see significant spread tightening in the loan market while CLO spreads stay wide, dealflow could certainly continue.”
Now as the Financial Times article points out, some CLOs are more wobbly:
More than 100 CLOs were failing at least one trigger brought on by escalating triple C loans, according to April data compiled by Barclays, with 40 failing tests for their triple B rated tranche or higher — debt that is regarded as investment grade. Twelve have also failed tests up to the double-A rated tranche, according to BofA.
It’s hard to know how significant this is without having dollar values. The Financial Times also warns that one-third of the BBB tranches are on review for a downgrade, but again, I find it hard to be sympathetic. If you bought BBB paper, you knew there was decent risk it would decay into junk. However some of these bonds are held by life insurers, who are restricted in how much they hold in non-rated/non-investment grade assets. They usually max out that bucket with real estate. So if their BBBs get downgraded, they’d probably have to sell.
So why all this unseemly whining? The most obvious explanation is that the Fed isn’t doing much for CLOs. One has to wonder if money men have been playing up how bad things supposedly are to get more rescue money. From the Wall Street Journal on April 22:
The Federal Reserve will lend $2.3 trillion to support the economy…The Fed excluded some securities tied to corporate loans and commercial real estate that were among the newest, fastest-growing segments of the bond markets…
The [$100 billion] TALF program won’t include the vast majority of a popular structured security known as a collateralized loan obligation, which is made up of corporate debt generally used to finance buyouts. The program does include so-called static CLOs which don’t allow for reinvestment of loan proceeds, but those account for only a small share of the market. The vast majority of CLOs allow such reinvestment and aren’t included.
The central bank threw some small additional bones to the CLO market earlier this week, but still left most CLOs ineligible. From Bloomberg:
The Federal Reserve revised its Term Asset-Backed Securities Loan Facility to allow CLOs that hold a broader range of leveraged loans to be used as collateral.
The Fed will now accept new AAA CLOs with leveraged loans, including refinanced loans, that priced as far back as January 2019, according to a statement Tuesday on the central bank’s website. Previously, eligible collateralized loan obligations could only hold newly-originated loans….
Yet the changes will only go so far, market watchers say. The terms still require eligible CLOs be static vehicles wherein managers can’t actively trade the loans underpinning the deals, a structure that makes up only a small portion of the market.
It is really a shame to see what has happened to the Financial Times. It was the only major media outlet before the crisis where quite a few writers, particularly Gillian Tett but also John Authers, Martin Wolf, and John Dizard, saw that risk was being underprices across all credit markets and sent clear warnings that things could end very badly, when the US financial press was uniformly in la-la land.
Since the crisis, the paper has gone firmly neoliberal, and perhaps due to the lack of time to do reporting on top of being now the US editor, Tett veers from doing extremely insightful pieces reminiscent of her glory days as capital markets reporter, to doing far too many articles that have all the hallmarks of her having spent too much time with people talking their book. Similarly, this CLO article does show the authors did a lot of legwork, yet somehow missed or chose not to address the most important questions.
1 And remember, at least so far, we don’t have evidence of widespread lending fraud, a big feature of the subprime mania. However, a key difference is that with mortgage-backed securities, the mortgages are transferred through several intermediaries to a trust and they stay there. Certificates are then sold that represent the rights to interest and principal payments made to that trust. With CLOs and CDOs, they can be “static” meaning the loans (or RMBS tranches) are set at the get-go and don’t change, or “active,” meaning a manager can trade assets in and out from time to time, subject to overall restrictions. Most CLOs are active.
2 I haven’t seen any actual CLO documents, so some of my interpretations from CDOs and RMBs may be a bit wide of the mark. RMBS had distinct waterfalls for principal and interest payments. This chart covers only interest payments. It doesn’t mention that the deals almost certainly had “excess spread”, meaning that loans that paid, say, an average of 7.5% across the CLO would have total interest payments of only 7.2% across all CLO tranches. This chart presupposes that the excess spread is already gone and required interest payments are at risk, so the equity tranche takes the hit first.
The article does have good detail on this issue. Basically, downgrades of loans are producing breaches of collateral quality tests, and that looks set to cut interest payments to the equity layer:
Typically, CLOs are permitted under their own rules to hold up to 7.5 per cent of their assets in triple C rated debt. If they stay within this threshold then managers of the CLOs can treat the loans they own as if they are worth 100 cents on the dollar.
This is because when a CLO is created, the underlying portfolio of loans is larger than managers need to pay off debt investors. This is known as over-collateralisation. So long as the number of loans that are near to defaulting remains below the 7.5 per cent threshold, the portfolio is treated as though there will be enough money left at the end to pay investors.
However, if a CLO exceeds its triple C bucket, as many now have, then the lowest priced loans above the threshold are required to be valued at a market price. In a crisis, with loan prices having fallen sharply, this lowers the overall value of the portfolio of loans reported by the CLO. That in turn can impact how investors are paid.
If the value of excess collateral slips by more than a few percentage points, managers typically first cut off 50 per cent of any interest payments on the underlying loans that would have gone to equity holders and use it to buy more loans, with the aim of increasing the value of the portfolio and correcting the breach.
If the value of the underlying loans falls further, then all money is cut off to equity investors. That money is used not only to pay interest to debt investors but to start paying back the principal of the debt, beginning with the triple A rated bonds.
3 Even if private equity firms could achieve somewhat higher leverage levels, that’s not a big bennie. Private equity has for many years been sitting on tons of “dry powder,” meaning uncalled capital. So their equity isn’t a scarce commodity. Thus the gain is a marginal decrease in interest charges. Helpful but hard to see as significant.