Much Ado About Nothing

The Fed's interventions in the repo market are attracting considerable comment. A lot of people seem to think the Fed has embarked on another QE program without Congressional approval. And the usual suspects are complaining that the Fed is pumping up stock prices and debasing the dollar.  Stocks are indeed heading for the moon - though so is the dollar, which rather undermines those who think it is being debauched. But the Fed's interventions in the repo markets have nothing to do with stock prices. They are all about banks.

Last September, sudden spikes in the Fed Funds Rate (FFR) and its repo market equivalent, the Secured Overnight Funding Rate (SOFR), caught the Fed off guard. It  acted quickly, injecting copious quantities of reserves to bring the rates down. But this was by any standards a seat-of-the-pants operation. The Fed simply hadn't expected banks to run out of reserves. After all, despite the Fed's balance sheet reduction, total reserves were still far more than banks needed to meet reserve requirements. There should have been ample reserves in the system for banks to draw on when settling sudden large payments. But for some reason, there weren't.

"Excess" reserves used to mean any reserves the banks held over and above their reserve requirement. But since the introduction of Basel III liquidity rules and Fed "living wills", that has changed. Banks need to keep far more liquid assets than their reserve requirements suggest. Admittedly, Basel III requirements don't force them to keep these in the form of reserves. But "living wills" do. Additionally, the big banks now have to pre-fund their expected liquidity requirements for the following day, rather than relying on intraday overdrafts at the Fed. So they tend to hold much larger quantities of reserves than they used to. For example, Jamie Dimon estimated that JP Morgan had to hold at least $60bn of reserves to meet its expected intraday liquidity requirements. That is one heck of a lot of reserves withheld from the market.

"Excess reserves" has thus acquired a different meaning. It now means whatever reserves the banks are willing to unload, either by lending them to each other or by lending cash in the repo market (which triggers a reserve movement). But when banks are facing enormous payment demands, they won't do anything that reduces their reserves. So when corporations started making their quarterly tax payments, banks simply stopped lending excess reserves to each other. And because cash lending to non-banks causes reserve outflows, they stopped that too.

But quarterly tax payments happen, er, every quarter. Why was September different? Well, it wasn't. The Fed made an error of judgement. It assumed that the big banks would continue to act as dealers of last resort in the repo market even when the reserves on which they depend for liquidity were being drained by Fed balance sheet run-off, overnight reverse repos, and Treasury debt issuance to finance the growing US budget deficit. But banks' primary duty is to their customers and shareholders, not to the Fed. If market liquidity becomes stressed, banks will protect themselves. Rather than lending more cash to maintain market liquidity, they will stop lending and hoard reserves, thus making the market liquidity problem worse. So the SOFR and FFR spikes had two causes. Firstly, a sudden drain of market liquidity, mainly due to quarterly tax payments and Treasury debt issuance; and secondly, reserve hoarding by big banks.

After the September spikes, the Fed continued to inject liquidity into the repo market by means of intermittent asset purchases. In December, Zoltan Poszar at Credit Suisse predicted that the Fed would have to start QE to prevent the repo market freezing at year end. This was incendiary. When the Fed made substantial liquidity injections over the year end, thus preventing serious market disruption, lots of people decided Poszar was right. The Fed had started QE again, and it was pumping up stock and bond prices.

But the Fed's liquidity injections are not like QE. They are more like the open market operations (OMOs) that the Fed used to do before the financial crisis, though on a much larger scale, because the repo market is much larger than the pre-crisis Fed Funds market and not restricted to banks.

Like OMOs, the Fed's liquidity injections put downwards pressure on short rates, thus keeping the FFR within its target range and the SOFR within a few basis points of the FFR. And by relieving funding pressures, they enable the big banks to supply liquidity to the repo market, preventing it from freezing as it did in 2008. Keeping the repo market liquid prevents stock and bond prices from crashing due to fire sales by distressed non-banks.

But preventing a disastrous stock market collapse due to liquidity failure is emphatically not the same as "pumping up" asset prices with QE. And the Fed's asset purchases this time are fundamentally different from QE purchases. In QE, the Fed bought government bonds and agency MBS of varying maturities, thus flattening the yield curve. The idea was to encourage investors to diversify into higher-yielding corporate bonds and equities. But this time, the Fed is buying Treasury bills and TIPS - short-term government bonds of similar liquidity to reserves. The effect is to change the composition of repo market liquidity, reducing the quantity of liquid collateral in circulation and increasing the cash and reserves held by banks. Reducing the quantity of liquid collateral depresses short-term interest rates, leaning against any tendency of the yield curve to invert (note that this is the opposite of QE's effect). And increasing the cash and reserves held by banks should encourage them to lend. As repo market lending is entirely short-term, none of this should have any effect on longer-term rates.

So the Fed's asset purchases aren't QE, and they aren't pumping up stock and bond prices. They are a purely technical operation to enable the Fed to control short-term interest rates as it has done for decades. If the Fed were intervening in the Fed Funds market in this way, no-one would say a word.

But why is the Fed intervening in the repo market instead of the Fed Funds market? The reason is that in this strange post-crisis world, the Fed Funds market is fast becoming an anachronism. Banks now prefer the safety of collateralised lending, even to other banks. The interbank market has embedded itself in the repo market, and the repo market has effectively replaced the Fed Funds market as the primary vehicle through which the Fed must transmit monetary policy. The Fed wanted to restrict its interventions to the Fed Funds market and rely on the big banks to transmit policy to the repo market. But the big banks wouldn't play the game. So the Fed now has to backstop the repo market itself in order to keep control of interest rates.

However, the Fed's backstop is problematic. Firstly, as I noted before, it means that the Fed is now implicitly guaranteeing dollar liquidity to non-banks and foreign banks as well as US regulated banks. Admittedly, foreign banks already enjoy an effective guarantee of dollar liquidity via their own central banks because of Fed swap lines. But guaranteeing dollar liquidity to non-banks is a different matter. The last time the Fed did that was in the financial crisis, and it was severely criticised for it. Firms that know the Fed will dampen interest rate volatility are much more likely to take excessive risks. If Elizabeth Warren finds out what is going on there will be hell to pay.

Secondly, leveraged trading is a feature of the repo market, unlike the Fed Funds market which is a pure cash market. Leveraged trading inevitably amplifies rate movements. Transmitting monetary policy via such a market means that the Fed must constantly intervene to dampen rate movements. But the Fed is talking about ending liquidity injections in April. Colour me sceptical. Unless it can persuade the big banks to take over as lenders of last resort to this market, it will have to provide funds forever.

What would the Fed have to do to persuade the banks to take back responsibility for providing liquidity to the repo market? Well, it would have to guarantee them unlimited liquidity. You would think they already have such a guarantee, especially since the consequences of central banks failing to provide adequate liquidity to banks were so catastrophically exposed in the financial crisis. But you would be wrong. They don't - at least, they do, sort of, but they are systematically discouraged from using the facility that is supposed to provide them with liquidity at need. I refer, of course, to the Fed's broken "discount window", which is a fine example of how perfectly sensible institutions can be rendered useless by ridiculous beliefs.

Non-banks such as hedge funds, pension funds and shadow banks are the customers of regulated banks, which act as their gateways to the payments network. Non-banks are also prone to runs. When non-banks are run on, the money flows through regulated banks can be far more than the banks hold in reserves: during the systemic runs on shadow banks in 2008, trillions of dollars flowed through the big banks. The Fed must provide unlimited liquidity to banks to enable those flows to happen. If it doesn't, then disturbances in the non-bank (shadow bank) sphere can cause regulated banks to fail, with potentially disastrous consequences for the real economy.

The discount window is intended to enable banks to obtain liquidity at need, thus eliminating the risk that a bank will fail due to exceptionally large outflows of customer funds - whether because of a run on the bank itself or runs on its customers. But the discount window is expensive, and - more importantly - stigmatised. Banks are very unwilling to use it, because it signals to the market that they are in distress, making their funding problems even more acute and raising the risk of failure. An emergency lending facility that no-one will use because using it makes matters worse is surely worse than useless.

David Andolfatto of the St. Louis Federal Reserve has proposed that the Fed should create a "standing repo" facility (SRF) which would enable banks to help themselves to new reserves from the Fed whenever they need them, provided they have adequate collateral. A bank that meets its Basel III requirements for high-quality liquid assets (HQLA) would be unlikely to lack collateral. So a SRF would effectively eliminate funding stress for banks.

The quid pro quo for this arrangement would be that the Fed would no longer backstop the repo market as it has since September. The Fed would guarantee liquidity for solvent regulated banks, and the banks would be expected to guarantee liquidity for creditworthy non-banks.  Monetary policy would be transmitted through the banks, not through enormous open market operations in an international money market populated by foreign banks and non-banks. In short, normal service would be resumed, despite the imminent demise of the Fed Funds market.

However, not everyone agrees that a SRF is the solution. The Fed's Randal Quarles agrees that banks need to be provided with unlimited liquidity, but thinks that fixing the discount window would be sufficient. The discount window could be extended to provide intraday liquidity against HQLA collateral, and the interest rate could be reduced to a few basis points above SOFR, which would keep a lid on SOFR and (by extension) FFR.

But the problem is the discount window stigma. Stigmas are notoriously difficult to remove. The Fed can signal that using the discount window is now to be considered "normal", but it has no way of ensuring that markets will see it that way. I fear that improving the discount window's facilities won't be sufficient for banks to want to use it. And if banks won't use it, then the Fed will be unable to stop intervening directly in the repo market.

Rather than attempting to persuade banks that the Fed now wants them to use a facility that it has spent decades trying to stop them using, it would probably be simpler to create a new facility. As Claire Jones at FT Alphaville says, "It would be neater 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)." I concur. The discount window is beyond repair. Time to replace it.

And it is also time to stop worrying about the repo market. The Fed's backstop is effective. Interest rates are under control, and no-one is suffering serious liquidity shortages. The liquidity injections are not creating a dangerous bubble in stocks and bonds. And the moral hazard that the backstop creates will eventually be eliminated when the Fed replaces its market liquidity injections with unlimited funding for banks, whether that is via a reformed discount window or a new SRF. There is nothing to look at here. Time to move on.

Related reading:

The blind Federal Reserve 
More "money" Treasuries would calm repo markets - FT Alphaville
Anatomy of a Bank Run
Liquidity Matters

Image from David Strine,CC0, via Wikimedia Blog