MORE "MONEY" TREASURIES WOULD CALM REPO MARKETS / THE FINANCIAL TIMES

More ‘money’ Treasuries would calm repo markets

By: Claire Jones


$414bn is a big number – even if you’ve got the right to print as many dollars as you like.

When repo rates spiked at 10 per cent in mid-September, it shook the US Federal Reserve. So much so that the central bank has in the intervening months bloated its balance sheet by more than $414bn to $4.2tn:




The Fed hasn’t liked doing that at all, nor the accusation that the expansion marks the restart of its quantitative easing programme. Yet until last week, the bank had kept quiet on what it planned to do instead.

Last Thursday, we heard the most comprehensive response to date.

Randall Quarles, the Fed’s man for supervision, signalled a series of changes the Fed hopes will both lower demand for central bank cash, and make banks more willing to lend out their reserves. In short, the fixes work by making treasuries behave more like cash.

To judge the success of Quarles’ fix, which we explain in more detail below, we need to answer two questions.

First, would it have stopped the turmoil of last September? To which, we think the answer is yes. Second, does it fix some pretty major flaws in the Fed’s operations with markets? To which the answer is no.

Let’s back up a little and explain why this matters.

The reason the Fed cares so much about the spike in overnight repo rates is because it wants the impact of its monetary policy to transmit throughout the entire financial system.

The repo market reflects the health of this transmission and the Fed is looking to have a spread between it and the central bank-agreed federal funds rate that they can measure in single, or perhaps low double, digit basis points. So for repo jump to about eight percentage points above the federal funds rate in mid-September caused a great deal of concern. And if it had remained as elevated as that from days on end, it could also have been disastrous for the health of the financial system.

At the same time, the Fed wants to wean the banks off its crisis stimulus by lowering the amount of reserves in the system – which for the best part of a decade has been a shockingly large amount:




The Fed can do this by offloading its stock of treasury bills in return for cash. This in turn shrinks the Fed’s balance sheet by lowering lenders’ reserves held at the Fed – something the central bank is keen to do in order to show that monetary conditions are returning to normal.

Months on, we still don’t really know exactly what caused the spike, with myriad factors at play. But one often mentioned culprit is the new liquidity rules for banks, known in supervisor parlance as the liquidity coverage ratio, or LCR.

The LCR is intended to give banks a bigger buffer of assets that are either cash, or cash-like (ie, can be easily sold in order to access cash), to cushion them against a bout of turmoil in which riskier avenues of funding would close. The ultimate aim being to reduce reliance on taxpayer support.

The result is that banks have almost doubled their holdings of high-quality liquid assets (such as central bank reserves, and Treasury securities – dubbed Level 1 HQLA). Which is to the Fed’s liking, given that it makes the financial system more resilient against bouts of panic.

What’s less to the Fed’s liking (but also of the Fed’s making) is the way in which this has led to lenders wanting to hold massive amounts of central bank reserves.

We mentioned earlier that the Fed wanted to shrink its balance sheet by offering Treasuries in exchange for cash. But the trouble is that while both Treasuries and central bank reserves count as Level 1 HQLA, they are not quite perfect substitutes – especially, as Quarles acknowledges, in times of panic:


It may be difficult to liquidate a large stock of Treasury securities to meet large “day one” outflows. For firms with significant capital markets activities, wholesale operations and institutional requirements (such as hedge funds), this scenario is not just theoretical. In the global financial crisis, several firms experienced outflows exceeding tens of billions of dollars in a single day.

The result of this is that, for the purposes of liquidity stress tests that banks conduct*, Treasuries and central bank reserves are far from perfect substitutes, leading the banks to have to hold huge amounts of assets to meet regulatory requirements and intraday liquidity obligations. Via a recent Brookings Institution speech by Quarles’ predecessor Daniel Tarullo, now of Harvard:

During JPMorgan’s Q3 earnings call in October, Jamie Dimon identified these requirements as the constraint that prevented JP Morgan from doing more repo and said his bank is required to maintain at least $60 billion in reserves at the Fed on an intraday basis. Although one would expect that JPMorgan’s size and complexity mean that its requirement is considerably higher than that of other banks, some back of the envelope extrapolation would suggest that the aggregate requirement for all banks could be significant.

Of course, none of this mattered during the era of QE, when lenders were flush with central bank cash. But as the Fed has tightened, it has become more and more of a problem. Without making treasuries more cash-like, a repo spike was perhaps inevitable.

And this is where Quarles’ idea comes in. Here’s the money quotes:

My suggested options are grounded in the principle that the Federal Reserve has an important role in providing liquidity to depository institutions. Today this role is played by the discount window, through which Reserve Banks provide fully collateralised loans to healthy institutions. While the range of eligible collateral for such liquidity provision is very broad, it may be worthwhile to focus for the moment on the Federal Reserve’s potential to provide liquidity that is collateralised only by Level 1 HQLA. With firms posting Level 1 assets as collateral, the Fed would be well positioned to provide liquidity to bridge the monetisation characteristics of HQLA securities versus reserves. Acknowledging this potential role in stress scenarios, the Fed may promote efficient market functioning while assuming very limited risk. If firms could assume that this traditional form of liquidity provision from the Fed was available in their stress-planning scenarios, the liquidity characteristics of Treasury securities could be the same as reserves, and both assets would be available to meet same-day needs…

...We have also already publicly clarified in the 2019 resolution planning guidance that firms can assume discount window access in their Title 1 plans if they can meet the terms for borrowing, such as recapitalizing the bank subsidiary.

We could build on this approach by also allowing firms to rely on the discount window in their ILSTs as a means of monetizing, for example, Treasury securities in their scenarios. This approach would acknowledge a role for the discount window in stress planning, improve the substitutability of reserves and Treasury securities in firms’ HQLA buffers, and maintain the overall level of HQLA that firms need to hold.


Under this regime, then, treasuries would become much more of a substitute for cash, as in theory they could be exchanged for cash over the course of the day at the discount window.

The impact would be twofold. Lenders would need to hold fewer reserves at the central banks, thereby enabling the Fed to shrink its balance sheet. Banks would also be more willing to lend cash, knowing that they had regulatory cover to meet intraday obligations.

(As an important aside, Quarles also said that for the purposes of assessing compliance with additional rules for the biggest lenders, the Fed was thinking of looking at balance sheets over the entire quarter, rather than focusing on a year-end snapshot. This is something that we think would further ease tensions given that at least one big bank had threatened to stop lending dollars at this time.)

If this regime had been in place in mid-September then a spike to eight percentage points above the federal funds rate would have been unthinkable.

We’d note too, that it has a good chance of happening regardless of who wins the race for the White House later this year. While Quarles was appointed by the Trump Administration, Tarullo, who is regarded as being close to leading Democrats, homed in on a remarkably similar theme:

In stressed circumstances, when liquidity providers generally become more cautious even as borrowers need more funding, monetization through other market actors may not be immediately possible. Thus, if the aim of resolution planning and liquidity requirements is to assure funding self-sufficiency by banks under all circumstances, there will be a preference for reserves. So long as reserves have been plentiful, this preference is a feasible regulatory choice. If they prove not to be, though, these requirements might contribute to the kind of [repo markets turbulence] that motivated this conference. Given the pace of increase in the supply of Treasuries created by the post-2017 budget deficits, the importance of the asymmetric regulatory treatment of Treasuries and reserves may become more significant. This, I think, is a regulatory issue worth revisiting, though – as already noted – the absence of publicly available firm-specific and aggregate information may make it difficult for those outside the regulatory agencies to debate the matter in anything but conceptual terms.

Quarles’ fix also clarifies that, as we have previously argued, the current balance sheet expansion is not QE. It indicates that the actions of the past few months do not represent a policy shift to a permanently bigger balance sheet. The Fed clearly wants to squeeze its size to levels more in line with those seen pre-crisis (although there is still a long way to go).

We are not sure why the Fed is so intent on sticking with the discount window – Quarles says in the speech “there may be benefits to working first with the tools we already have”, rather than creating a new facility.

It would be a neater fix to start afresh with a standing repo facility designed with the specific purpose of exchanging treasuries for reserves, rather than trying to use a discount window long seen as a last resort (complete with the stigma one would expect). While allowing the discount window to be used in this way in theory has the same effect, in practice we still do not know whether lenders would view the stigma of actually accessing it as barrier.

Still by making treasuries more like cash, we think Quarles’ fix is a big step forward in consigning the repo market’s troubles to history.


*The article was changed here to reflect that the ILSTs are internal tests conducted by banks, and not the Fed.