Even though this is a finance and economics blog, I haven’t written about the disruption in the repo markets. That is in part because the upset is not in any way, shape, or form like the 2008 period when banks were unwilling to repo even Treasuries to each other overnight because they were fearful another major dealer (say Morgan Stanley, which was on the verge of going tits up) would go the way of Lehman. I thought posting on it would feed the false narrative (which sadly is still kicking around) that the repo crunch is a sign of systemic stress, which it isn’t.
The second reason is that pretty much no one seems to have a clue as to why this is happening including most troublingly, the Fed. Perversely, the fact that the Fed is so clearly behind the curve is almost certain to make whatever the underlying problems are worse. Flatfooted central banks waking up to a problem and randomly hitting switches to try to make it go away is not only not a good look, but it makes the Confidence Fairy have a sad, which makes market upsets worse.
We’ll give a high level review of some of the major theories and why they don’t add up (and worse, look like special pleading by banks for regulatory breaks they don’t need) and will turn to an idea from John Dizard of the Financial Times (who regularly has great finds but is oddly buried by the pink paper by relegating him to a weekend wealth section column). Dizard argues that the big banks even with the apparent repo crisis make more money lending to the FX swaps market. This is consistent with big fish like JP Morgan’s Jamie Dimon whining about regulation rather than sounding at all worried.
Mind you, we are not saying there are not problems here. What we are question is whether they are being characterized properly and whether they could become systemic.
If Dizard is right, and Dizard is reviving and amplifying concerns raised by the Bank of International Settlements two years ago, the problem is in FX swaps and forwards, which amount to dollar lending but unlike repos, aren’t booked on balance sheet. That means that this area is a big blind spot for central banks; if you read the 2017 BIS paper, you can see they had to do tons of nitty gritty analytical work to come up with crude guesstimates.
Why This is Not 2008 Redux
We need to say this again and again: Banks are not afraid to lend to each other. There is no fear in the air. In 2007-2008, there were four acute phases of the crisis, starting with the implosion of the asset-backed commercial paper market in July-August 2007. There was a clear explanation as to why the ABCP market (which then was nearly half of the total commercial paper market) locked up. The biggest issuers of ABCP were special purpose vehicles including SIVs (structured investment vehicles) that used commercial paper as part of the funding for longer-term assets…which consisted significant if not entirely of subprime mortgage debt.
Things were so obviously not good that by September 2007, Treasury Secretary Hank Paulson was running around in what proved to be a failed exercise of trying to cobble together a private market solution for all of these SIVs, which were a problem for their big bank issuers. Even though these SIVs were supposedly off balance sheet and therefore not the responsibility of the party that set them up, the SIV investors (which included a lot of heavyweights) let the banks know that they had better act like they were responsible or said institutions would never to do business with that bank again. This sort of “not really off balance sheet” issue later came to bite the issuers of credit card receivables. The investors again made clear if banks ever wanted to be able to sell that paper again, they needed to eat some of the crisis-related losses.
The object of Paulson’s concern was clearly Citigroup, which had $400 billion of SIVs. Only later did we learn that Citigroup was dumb enough to have sold them with an explicit “liquidity put,” which meant the bank would finance the vehicles if the entity could not roll over its maturing short-term debt.
The point of this detour is there is no major category of debt blowing up and leaving investors and banks wondering who is taking big hits and therefore might not be a very good counterparty. Given that we have a lot of leveraged speculation against financial assets that are at sky-high prices, we could easily see some wheels come off in the not-too-distant future. But the repo market tsuiris is not the result of worries about bank or counterparty solvency. However, if the Fed doesn’t come up with better crunch responses than it has so far, its ham-handedness could make a bad situation worse when one develops.
Banks Blame Regulations
At a 50,000 foot level, it is not crazy to wonder if poorly-thought-out, far from comprehensive post crisis regulations might not have something to do with the liquidity crunch. The pre-crisis system suffered from what Richard Bookstaber in his classic A Devil of Our Own Design called tight coupling. That occurs in systems where activities propagate so quickly that humans cannot intervene quickly enough to halt the process.
What is needed in tightly coupled systems is to reduce the tight coupling. Circuit breakers that halt market trading when losses hit a certain level are one example. However, as Bookstabler described, measures to reduce risk in tightly coupled systems that do not address the tight coupling typically wind up increasing risk.
One of our big beefs about inadequate crisis reforms is virtually nothing was done to curtail derivatives, and we explained long-form in Chapter 9 of ECONNED why the crisis was a derivatives crisis, not just a bad housing loans crisis. The only measure taken, to move some derivatives clearing to central counterparties, has been regularly criticized as merely creating new too-big-to-fail entities. Note that had customers been required to post high enough margin to the CCP, that in fact would have cut into derivatives use. But no one wanted to mess much with this Wall Street golden goose.
Separately, we are not persuaded that this level of over-the-counter derivatives activity is necessary or virtuous. The really high margin, custom derivatives are used almost entirely for tax and accounting gaming. Many of the plainer-vanilla derivatives are used in connection with investments, or what it more technically called secondary market trading. More and more studies have found that the level of finanicaliztion in advanced economies is a negative for growth, and the amount of brainpower and resources devoted to secondary market trading and asset management is one of the biggest perps. So a simple missed reform opportunity would have been a transactions tax. It would have cut secondary market trading and in particular, the use of derivatives.
Oh, and derivatives are a big source of demand for repo funding. From ECONNED:
Brokers and traders often need to post collateral for derivatives as a way of assuring performance on derivatives contracts. Hedge funds must typically put up an amount equal to the current market value of the contract, while large dealers generally have to post collateral only above a threshold level. Contracts may also call for extra collateral to be provided if specified events occur, like a downgrade to their own ratings.17 (Recall that it was ratings downgrades that led AIG to have to post collateral, which was the proximate cause of its bailout.) Cash is the most important form of collateral.18 Repos can be used to raise cash. Many counterparties also allow securities eligible for repo to serve as collateral.
Due to the strength of this demand, as early as 2001, there was evidence of a shortage of collateral. The Bank for International Settlements warned that the scarcity was likely to result in “appreciable substitution into collateral having relatively higher issuer and liquidity risk.”
That is code for “dealers will probably start accepting lower-quality collateral for repos.” And they did, with that collateral including complex securitized products that banks were obligingly creating.
Notice how the use of repo increased starting in 2017 and more decisively in 2019 (the original New York Fed chart is interactive):
One way to look at this is that the Fed did undo the tight coupling by flooding the financial system with liquidity. The central bank for years has been trying to back its way out of the super-low interest rate regime it created, and has been finding that hard to do. Recall how the Fed lost its nerve in the 2014 taper tantrum.
And that increased use of repo roughly parallels how the Fed started shrinking its balance sheet, which is tantamount to reducing liquidity. FOMC minutes from March 2017 showed it planned to start trimming the size of its holdings by the end of that year. It actually started in October 2017.
Remember that prior to the crisis, the Fed had intervened daily in the repo markets from its New York Fed trading desk to manage money market rates. It abandoned that practice after the crisis and moved instead to paying interest on reserves when the banks were awash with reserves. It has not abandoned that approach despite also launching specific repo initiatives, although they were size limited, rather than using the former approach of doing what it took to hit the target rate.1
The initial repo market upset in September was quickly attributed to a series of demand for cash at banks, such as a corporate tax due date. The Fed being awfully slow to react was not helpful. And repo rates kept spiking up. Big banks started pointing fingers at rules that made them hold large liquidity buffers: “See, if you didn’t make us hold all this extra cash, we’d be lending to the market and you wouldn’t have this mess. So cut us loose.”
Former FDIC chairman Shiela Bair debunked this whining for deregulation (emphasis original):
So why didn’t the banks lend into the repo market?
Perhaps because they didn’t want to. Some of those caught in the repo squeeze were their nonbank competitors. As one industry insider with a top 10 bank told the Financial Times, “We have plenty of liquidity. We are just choosing not to lend it out overnight to hedge funds.” Or, harkening back to the 2008/2009 bailouts, maybe they thought that calming market squalls is no longer their responsibility, but that of the government. They were certainly quick to bash the New York Federal Reserve Bank for not being prepared to step in….
For its part, JPMorgan Chase says it was sitting on about $120 billion in reserves when the repo market ruptured. Dimon claims that it did not channel those funds into repos because regulators require that the bank maintain that level of reserves to prepare for stressed conditions, or worse, the bank’s failure. It’s impossible to know based on public disclosures whether Dimon’s interpretation of regulatory requirements is accurate. In any event, it does not seem unreasonable for regulators to want a bank with nearly $3 trillion in assets to keep $120 billion in ready cash on hand. The problem would seem to be that JPMorgan Chase — a dominant player in the repo market — was managing its liquidity too close to its regulatory minimums. With more ample reserves, it could have easily stepped in. Notably, Chase had allowed its reserves to drop by over 50% since the beginning of the year.
To be fair, it may be that the financial system needs more banking reserves to function properly. “Excess reserves”— that is, funds banks keep on deposit at the Fed that exceed regulatory requirements — are reported to be about $1.4 trillion. This sounds like a lot, and it is by historical standards, but it represents nearly a 50% drop from the peak in 2014. In its efforts to unwind its own balance sheet, the Fed may have drained bank reserves too far given the massive amount of cash the Federal government is pulling out of the markets to maintain its $16+ trillion-and-growing pile of public debt.
Bair’s take seems correct: the Fed drained liquidity too quickly. Izabella Kaminska pointed out another contributor: the Fed’s reverse repo facility:
It’s been a while since we checked in on what’s percolating through the mind of financial-plumbing specialist Zoltan Pozsar at Credit Suisse….
Here’s the gist from the opening of his latest note (with our emphasis, oh, and RRP stands for reverse repurchase agreement):
The FOMC should forget about r* for the moment and focus on Sagittarius-A* – the supermassive black hole at the center of global dollar funding markets.
The black hole is the foreign RRP facility, which has seen close to $100 billion of inflows since the beginning of the year.
The driver of these inflows is the curve inversion, and the longer the inversion persists the more inflows will follow. The trade war is also contributing to the inflows – given the inversion, as foreign central banks weaken their currencies they “buy” the foreign RRP facility and not Treasuries like in the past.
Foreign central banks are rate shopping… …and an uncapped foreign RRP facility is what enables that. Like the matter that enters a black hole, the reserves that are sterilized by the foreign RRP facility are gone for good – like the reserves “shredded” via taper.
However, I like the Dizard theory even better because it gives a direct explanation of why banks that had liquidity weren’t taking advantage of the opportunity to make some dough in the repo squeeze. That does not make Bair or Izzy wrong; they have important pieces of the equation, but are missing what looks to be a central driver.
FX Swaps and Hidden Dollar Debt
Dizard argues that the repo market spikes are symptoms of liquidity stress elsewhere, specifically in the FX and currency swaps markets.
Even though parties seeking to engage in currency hedges can engage in economically equivalent transactions using FX swaps, spot and forwards, or foreign currency repos, only the repo would be accounted for on the balance sheet of the financial firm. That means literally nobody knows how much in the way of FX swaps, currency swaps, and forwards are outstanding.2
The BIS was worried in 2017 about this hidden foreign currency debt, since the gross numbers were large and it was hard to infer much about it:
Every day, trillions of dollars are borrowed and lent in various currencies. Many deals take place in the cash market, through loans and securities. But foreign exchange (FX) derivatives, mainly FX swaps, currency swaps and the closely related forwards, also create debt-like obligations. For the US dollar alone, contracts worth tens of trillions of dollars stand open and trillions change hands daily. And yet one cannot find these amounts on balance sheets. This debt is, in effect, missing….
Focusing on the dominant dollar segment, we estimate that non-bank borrowers outside the United States have very large off-balance sheet dollar obligations in FX forwards and currency swaps. They are of a size similar to, and probably exceeding, the $10.7 trillion of on-balance sheet debt. On the other side of the ledger, as much as two thirds of the dollar-denominated bonds issued by non-US residents could be hedged through similar off-balance sheet instruments. That fraction seems to have fallen as emerging market borrowers have gained prominence since the GFC.
The implications for financial stability are hard to assess. This requires a more granular analysis of currency and maturity mismatches than the available data allow. Much of the missing dollar debt is likely to be hedging FX exposures, which, in principle, supports financial stability. Even so, rolling short-term hedges of long-term assets can generate or amplify funding and liquidity problems during times of stress.
It’s not as if all of the FX activity is hidden, but derivatives accounting treatment is way more forgiving than repos. As Dizard put it:
FX swaps contracts allow a non-US entity to exchange non-dollar cash flows for dollar cash flows. Thanks to the magic of derivatives accounting, only the variances in the relative value of the exchanged currencies (or “replacement cost values”) need to be disclosed by the banks mediating these transactions.
Let us stress that much the way credit default swaps were not bona fide derivatives (they were unregulated insurance contracts), it beggars belief that FX swaps are booked like derivatives. As the BIS explains:
These transactions are functionally equivalent to borrowing and lending in the cash market….
Why such a difference in accounting treatment? One reason is that forwards and swaps are treated as derivatives, so that only the net value is recorded at fair value, while repurchase transactions are not. Since the value of the forward claim exchanged at inception is the same, the fair value of the contract is zero and it changes only with variations in exchange rates. Yet, unlike with most derivatives, the full notional amount, not just a net amount as in a contract for difference, is exchanged at maturity….
Yet, despite this basic equivalence, the amounts borrowed through swaps and forwards never show up on any balance sheet while those related to repos do.
Recall that Shiela Bair quoted a banker saying that his bank had lots of liquidity but elected not to lend it out overnight. Dizard points out why they wouldn’t bother: they could make more money lending in the FX swaps market:
Why have the Fed’s large interventions since September 17 not calmed the money markets? Perhaps it is because that liquidity is being sopped up by the demands of the FX swaps market.
Ralph Delguidice, a money market observer and global macro strategist at Pavilion Global of Montreal, points out that for a yen/dollar FX swap, “a dealer receives a total of about 230 basis points for what is a synthetic yen-dollar repo. That includes the three-month dollar Libor rate, the negative rate on Japan Libor, a charge by the dealer called the ‘basis’, and the negative yield on the JGBs which are the best place for the dealer to park the yen they receive in return for dollars.
“Compare that to dollar repo which was 190 earlier this week, and Treasury bills, which are about 150. You can see why the dealers are incentivised to lend dollars in the FX swap market rather than the domestic money market.”
Now of course this raises the question of why the FX swaps market, and specifically dollar-related swaps, are in demand right now. Part of it is that the dollar dominates these markets. As the BIS paper helpfully explains:
The dollar reigns supreme in FX swaps and forwards. Its share is no less than 90% (Graph 2), and 96% among dealers (Table 1). Both exceed its share in denominating global trade (about half) or in holdings of official FX reserves (two thirds). In fact, the dollar is the main currency in swaps/forwards against every currency.
Financial Times reader Stimpy put the problem in more colloquial terms:
The point of the piece (and the BIS as well) seems to be that non-US bank USD lending has exploded (12T) and that those lenders are sourcing some (~1 1/2T) of the USD they need to fund the loans in the wholesale FX-swaps markets at huge maturity mismatches. The currency risk is nil, but the rollover risk is huge as demand for USD crowds out other borrowers, like repo, because swaps pay more and get better accounting treatment.
We still have the question of why the apparent increased demand for dollars. Is extra hedging going on due to tariff tantrum worries? Brexit wobbles? Readers who have insight are encouraged to pipe up.
In the meantime, Dizard believes that the FX swap cash demands could rise to being a serious issue, particularly since liquidity dries up around year end:
But the FX dealers and the borrower-counterparties can get spooked even more easily than repo market participants. As Delguidice puts it: “On the 23rd of December they will be concerned they have to roll over and they get a dial tone. Your risk manager is on line two and he wants to talk to you.”
Cynical people working for the dealers and regulators believe that the C-suites of very large global banks would use such a panic to induce the Fed and Treasury to push through a loosening of regulation and lifting of mandatory charges on capital.
So if and when an FX swaps crisis arrives, expect the demands for regulatory waivers to get even louder. But as we pointed out, they would just amount to yet more stopgaps.
1 Warren Mosler thinks the Fed senior staff is likely well aware of these issues and said historically, political appointees would constrain action.
2 We discuss only FX swaps, which the BIS treats as currency exchanges of less than a year; longer than a year is a “currency swap”. The BIS paper also contends that FX swaps are generally used for trade and currency swaps, generally for investments like foreign bond purchases. However, it would be nice if practitioners piped up, since to fully hedge a foreign currency bond, you’d need to do swaps or forwards to cover the interest payment, and the nearest two would be in less than a year, putting them in FX swap terrain. I also wonder how much these instruments are used for short term currency speculation. Forwards certainly are a staple on currency dealer desks, and most multinationals have for over a decade been treating their Treasuries as profit centers, meaning they get to speculate too.