Monday, January 6, 2020

Dudley on reserves

Bill Dudley, ex President of the New York Fed, has an excellent Bloomberg editorial on reserves.

Reserves are accounts that banks hold at the Fed. The Fed used to pay no interest on these accounts. Accordingly, banks held very small quantities, as little as $10 billion in all, and they managed that quantity very carefully against legal reserve requirements, and having just enough around to make payments. To control interest rates, the Fed used to change the supply of reserves, and then watch the Federal Funds rate, the rate banks charge each other to borrow reserves.
"Each day, the Federal Reserve Bank of New York would assess whether to add or drain bank reserves from the financial system to balance the supply and demand to achieve the desired short-term interest rate. While straightforward in theory, the process was extremely complex to carry out. "
Since 2008, the Fed has started paying interest on reserves. The interest the Fed pays on reserves largely determines interest rates elsewhere, and the Fed doesn't have to go through a hoopla every morning to figure out just how many reserves to offer. The banking system is awash in trillions of dollars of reserves. And there has been no hyperinflation, nor indication that the Fed cannot control interest rates if it wishes to.

There has been a lot of controversy within the Fed about keeping the new system vs. the old.

After the crisis, many FOMC participants favored going back to the old regime. After all, the crisis was over and the old regime had worked reasonably well for several decades. In 2011, for example, a number of FOMC participants wanted to make this commitment clear in the Fed’s statement on its Exit Strategy Principles disclosed in the minutes of the June 2011 meeting. But some others, including myself, resisted. A compromise was struck with language that the “quantity of bank reserves are expected to be reduced to the smallest levels that would be consistent with the efficient implementation of monetary policy.”... 
I fought for that word “efficient” because I didn’t want to commit the FOMC to a return to the old regime before we had gained experience with the new one. I thought it was likely that the Fed would find that the new regime was better. 
Looking back, I think the new one has proved to be superior in several important respects. First, it is much simpler to carry out. ...
Hoopla explained
Second, excess reserves make the payment system operate more efficiently. Banks don’t have to delay payments as often because of timing differences between outgoing and incoming payments. 
This is important, and I think goes to the heart of why didn't the Fed do this years ago. If you want to control interest rates, well, control interest rates. Other central banks as far back as the 1990s said, we borrow at (say) 2%, we lend at 2.1%, come and get it. And interest rates were between 2 and 2.1%. Why didn't we do that?

It's hard for me to explain views I don't agree with, but I think there was some residual monetarist feeling that it is important to squeeze money out of the system directly, not just by charging a higher price. By forcing banks to get along with fewer reserves, they would more quickly lend less and push contraction (or vice versa) through the economy. If you feel that money affects the financial system through payments, the  point  of monetary policy is to force inefficiency in the payment system. It's like slowing a car down on the freeway by draining some of the oil. Monetarists always distrusted the interest rate as a signal of monetary policy.

This is a coherent view, but I think experience shows it just happens to be wrong. Draining the oil from the car will slow it down, and helps speed control, but the Fed shouldn't use every tool that happens to have an effect. Use the throttle.
"Third, and most important, the Fed now has the ability during times of stress to add liquidity to the financial system  .."
I'm less convinced by this one, but no need to fight about it.
"Yes, there have been a few bumps in the road in recent months. Regulatory changes made in response to the financial crisis caused the demand for reserves to be higher than anticipated ...The spike in the fall was not a “canary in the coal mine” signaling bigger problems in the financial system. Instead, it reflects the difficulty in forecasting the demand for reserves given the changes in regulations."
Dudley's too polite, but I think what happened is that the part of the Fed that really wants "smallest levels" "reserves are expected to be reduced to the smallest levels that would be consistent with the efficient implementation of monetary policy” put the level quite a bit too small.

Dudley continues, wisely:
"In my view, the Fed should take four steps to support the new monetary policy framework.  
First, the FOMC should create a standing repo facility that is open to a broad set of counterparties confined to Treasury and agency mortgage-backed securities collateral."
In English, any financial institution and not just banks can come lend to the Fed at the interest on reserves rate. As I said, if you want to set rates you set rates with a flat supply curve. I hope this means narrow banks too!

The "standing repo facility" will now allow financial institutions also to borrow from the Fed, with treasury collateral, without the "stigma" that discount window borrowing has had in the past.  If you really want to set rates, yes, you have to let institutions lend to the Fed at the interest on reserves, and also to borrow from the Fed as well. This summer's hiccups were cases when institutions wanted to borrow. When the Fed wants someday to lower rates, they will have to let people borrow at the low rate to undercut market rates.

Really, this means however, that if the Fed wants to target interest rates with interest on reserves, it can't also limit the size of its balance sheet. It has wanted to target both a price and a quantity, and that is proving unworkable.
"Second, the Fed should explain more clearly to the public, Congress and market participants why the new regime is better. This includes explaining why paying interest to banks on reserves is appropriate and not a handout to the banking industry. "
This is essential. The main argument for a small balance sheet and minimal reserves that I hear at the Fed is entirely political: Congress is too dimwitted to understand that the Fed is not subsidizing banks by paying interest on reserves, and they think the balance sheet is there to be raided for political projects. It is extremely unhealthy for an agency such as the Fed to do things it understands are bad economic policy out of fear that Congress will do what that agency thinks are silly things.

The Fed is just an immense money market fund: It offers interest-paying fixed value assets (reserves) and backs them with Treasuries and Treasury guaranteed mortgages.
"Third, the central bank should consider de-emphasizing the federal funds rate as the primary interest rate target of monetary policy....Why not just set the interest rate on reserves and stop worrying about the federal funds rate target? "
Banks don't borrow reserves from each other any more. I would add, however, the Fed needs to set its lending rate as well as its borrowing rate and then stop worrying.
"Fourth, the Fed should exempt bank reserves from the assets used to calculate a bank’s leverage ratio..."
Yes! A lot of money market pathology happens now because the Fed treats taking deposits and salting them away as interest paying reserves at the Fed as it does a risky activity. There is zero risk to doing this -- narrow banking! -- and it makes no sense to charge capital or liquidity against it.

This is a great editorial. We commenters tend to be grumpy about the Fed, but it does learn lessons from experience and get better over time!

Not everyone agrees, and I hope to see contrary reactions from my fellow bloggers and commenters, which will be interesting to think about.

Updates:  See also very good pieces by David Andolfatto and Jane Ihrig on the standing repo facility, here and follow up here.  Thanks to a few commenters who straightened out an error on the standing repo facility.


  1. A problem with a very large stock of reserves and controlling interest rates by adjusting the interest rate paid on reserves is that it involves rewarding people simply for hoarding money. Or to be more brutal, it involves rewarding the rich for hoarding money.

    To that extent, the MMT idea, namely a permanent zero rate is better. As for controlling demand, that can be controlled by adjusting the deficit, which in turn changes the size of the national debt, which in turn influences demand.

    But it would probably be going too far to say the rate of interest should NEVER vary from zero. Clearly adjusting interest rates is a useful tool to have.

    1. As long as interest rate is below GDP growth, I don't see why you are concerned about the rich hoarding money.

    2. Take total debt (including SS trust fund) from Fred. Divide by interest charges. The division yields 2.5%, growth is between 1.9% and 2.3%.

      What is missed are the rollovers, we generally borrow at yesterdays higher rate for today's spending. We used to pay the ten year rate, using the same calculation. Now we pay the 30 year rate.

    3. "It involves rewarding people simply for hoarding money." Oh, dear, it must be sinful! Do we envy the grocer for hoarding our food, and say he should not be rewarded?
      Why should any discussion pander to the idea that "the rich" are going to benefit more than "the poor"? This is a shoddy, confused sociology and worse economic analysis.
      I appreciate nobody reading this agrees with Marx about social classes and conflict. Unfortunately all modern journalists sincerely follow his fallacy.

  2. "The interest the Fed pays on reserves largely determines interest rates elsewhere"---John Cochrane. we have globalized capital markets and is money fungible? Yes.

    So, money flows easily across borders? Yes/

    So, we really have several major central banks undertaking monetary policy actions, including the ECB, the Fed, the PBoC, the Bank of Japan, and perhaps the BoE and Swiss National Bank? Yes.

    Seems to me long-term interest rates are set globally then, by the global market and perhaps the combined actions of all the major central banks.

    Does it make sense to talk about a lone central bank and any particular national economy, as if that economy is a closed system?

    1. The "world economy" [the 'open part'] is one (closed system) economy at its best, so global arbitrage is what would affect the Units of Accounting used by different governments (exchange rates, domestic price inflations).
      Not "money" but "capital" flows easily across borders. Subtle conceptual shift. (i.e. ownership by a foreigner)

  3. At the heart of all of this is the view that interest rates are used in aggregate demand management. Empirical evidence suggests it does but I've never understood the credible mechanism behind it.

  4. I hope a less-dimwitted commentator than myself would like to explain to me why paying interest on reserves is not a subsidy to banks. Anybody interested?

    Or perhaps more accurately, paying interest on reserves is part of a larger regulatory framework that restricts competition in banking, pays the banks interest, saves them when they take too much risk, and funds it all through seigniorage, which is a tax on holders of nominal-value assets. We shouldn't pick on any part of this framework in isolation, but isn't it correct that the central bank performs a routine and massive transfer of wealth from the general population to banks, by paying interest on reserves, among other things?

  5. "The main argument for a small balance sheet and minimal reserves that I hear at the Fed is entirely political: Congress is too dimwitted to understand that the Fed is not subsidizing banks by paying interest on reserves, and they think the balance sheet is there to be raided for political projects. It is extremely unhealthy for an agency such as the Fed to do things it understands are bad economic policy out of fear that Congress will do what that agency thinks are silly things."

    John, plenty of critics of the floor system, myself among them, understand perfectly well that interest on excess reserves isn't necessarily a subsidy to banks. Nor do most of us want to eliminate it. Instead we favor a corridor system with an IOR lower limit of the sort that has been quite successfully employed by many nations.

    As for the possibility that a Fed balance sheet of potentially unlimited size might be "raided for political projects," taking that risk, and political considerations more generally, into account isn't "bad economic policy." On the contrary: ignoring political reality is bad policy.

  6. "I think there was some residual monetarist feeling..."

    Under Greenspan and early Bernanke, the NY Fed targeted excess reserves (a trivial amount) while letting total reserves float. I don't see how this had anything to do with monetarist thinking.

    Consider: The NY Fed’s model predicted the amount of reserves its member banks would require, and then the trading desk simply added $1.5 bln (the targeted excess) to determine its trades. So total reserves floated, money was endogenous, and the tweaking that needed doing to achieve the fed funds target was highly exaggerated. Fed traders will tell you they only needed to run their model and target the $1.5 bln, since markets automatically adjusted to the FOMC’s published rate target.

    But the main point I wanted to add is that the old method, while not monetarist, has roots in a free-market philosophy that the Fed should only be in the background when it comes to credit creation. With that philosophy, you might rule out IOR while letting others, especially the private sector, determine the total amount of credit in the economy.

    In other words, you might choose to influence credit creation only by setting wholesale interest rates, not by QE or recent policies that some call QE and others don’t. QE and QE-light mean that the Fed plays a bigger credit-creation role through more active and lasting security purchases than with the old-style OMOs and without IOR. That seems relevant on a free-market-oriented blog ;)


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.