Peter Stella, former head of the IMF Central Banking and Monetary and Foreign Exchange Operations Divisions, clarifies his postion on bank reserves and lending, emphasizing that banks don't lend reserves, that the quantity of base money has nothing to do with bank lending, and that negative reserve rates would be contractionary rather than expansionary.
Right in line with MMT.
Read it at The Financial Times | FT Alphaville
The base money confusion
by Izabella Kaminska
(h/t Andy Blaltchford and Kevin Fathi via email)
Warren Mosler links to the post, too. Peter Stella on QE
Warren Mosler links to the post, too. Peter Stella on QE
13 comments:
I am not sure why Izabella is so resistant to understand banking the Post-Keynesian way. It's not even a divisive MMT issue, and she even has an affinity for MMT!
The Stella comments seem very good, but the Kaminzka comments seem problematic to me.
In a couple of places, she seems to suggest that not only do changes in the monetary base not induce changes in the broad money supply, but also that changes in the broad money supply do dot induce changes in the monetary base.
So broad money supply rises, but the amount of base money remains the same — irrespective of how much lending takes place.
...
As Stella points out, the US banking system could increase lending a thousandfold this year without any change in their deposit holdings at the Fed.
This seems in conflict with the MMT picture according to which increases in bank lending are automatically accommodated by the Central Bank, in order to maintain its target rate.
Maybe under current circumstances, banks can dramatically increase aggregate deposits without any additional central bank accommodation because the system already contains so many excess reserves. But that is not a general state of affairs.
I also don't fully understand the statement on negative rates:
It thus stands to reason that negative rates — by reducing the central bank’s balance sheet — are contractionary rather than accommodative when it comes to credit supply.
Just as increases in the quantity of reserves do not induce an expansion of lending except insofar as they influence price (a lower Fed Funds rate), shouldn't we also say that a decrease in the quantity of reserves doesn't induce a contraction of lending unless it also influences price. As I understand MMT's economists on IOR, the latter is just an additional tool for targeting a Fed Funds rate. We can't draw conclusions about expansion or contraction from premises about whether aggregate reserves are increasing or decreasing.
She quotes Stella who says:
All this brings me back to the issue of imposing a negative interest rate on commercial bank balances at the central bank. I cannot understand how anyone could think this would spur lending. First, as we have already said, bank reserves do not “leave” the doors of the central bank when credit expands. Second, banks would respond like any business whose inventory is subject to a new tax. They would reduce their inventory, that is, shrink their operations, not expand them.
But that inference doesn't seem airtight. I agree that there is no reason to think the tax on holding excess reserves would spur lending. But I also don't see why it would necessarily cause bank's to "shrink their operations". The costs of holding reserves would increase. But if they decreased their lending in response their loan assets would decrease as well. And as Stella has already recognized, since banks don't lend their reserves, it is very misleading to think of reserves as a bank's "inventory".
Banks need reserves to make payments without incurring overdraft charges, and thus a tax on reserves is analogous to a fee charged for making payments. Businesses never like it when their costs of doing business go up, but that doesn't mean they will respond to these higher costs by doing less business. They might try to pass on these added costs to their customers and thus drive away some business. But they might absorb the costs themselves. It seems hard to predict, with no one answer applicable to all market conditions.
Dan:
Maybe under current circumstances, banks can dramatically increase aggregate deposits without any additional central bank accommodation because the system already contains so many excess reserves. But that is not a general state of affairs.
There is also the important factor of trust, which is what underlies the shadow banking system's rehypothecation, repos, and collateral chain process that can expand the quantity of credit without a corresponding increase in reserves. These institutions are deposit-free.
One angle from the article that left me confused was the idea put forth by both Stella and Kaminsky that a bank could dodge paying negative rates on excess reserves by holding reserves in the form of physical currency in their vaults.
For the purpose of determining a given bank's reserve level, doesn't the central bank look at vault+reserve account balance? If so, it doesn't seem that keeping more reserves as physical notes in the vault would make any difference at all.
Dan,
"As Stella points out, the US banking system could increase lending a thousandfold this year without any change in their deposit holdings at the Fed.
This seems in conflict with the MMT picture according to which increases in bank lending are automatically accommodated by the Central Bank, in order to maintain its target rate."
I think what she means the increase in reserves is not automatic, but has to be performed by the CB, and it will, to defend the target rate. She stresses that reserves neither trickle in nor out of the banking system, have to be supplied from the source: CB.
"I also don't fully understand the statement on negative rates:
It thus stands to reason that negative rates — by reducing the central bank’s balance sheet — are contractionary rather than accommodative when it comes to credit supply.
Just as increases in the quantity of reserves do not induce an expansion of lending except insofar as they influence price (a lower Fed Funds rate), shouldn't we also say that a decrease in the quantity of reserves doesn't induce a contraction of lending unless it also influences price."
Negative IOR is just a tax - central bank money flows to the CB as the payment of the negative interest, and nothing (like bonds) flows out (unlike with QE where CB adds reserves and subtracts securities), so this is really fiscal policy - contractionary one, a net transfer from the private sector to the CB, so a tax (money destroyed).
Peter P,
I don't think I agree with the idea that because negative IOR does not involve bonds, that makes it like fiscal policy. MMT's beef with monetary policy isn't just based on the idea that monetary policy is about securities swaps. Even if the bank was just handing money to banks for free and putting it in their reserve accounts - as it does when it gives positive interest on reserves - that would not necessarily have an expansionary effect on bank lending. Similarly, charging a tax on reserves would not necessarily have a contractionary effect on lending.
The real issue is not repos and reverse repose vs. taxes and giveaways. It is how the money is injected into or extracted from the economy, and who receives it. If the government sends you $1 million and you deposit it in a bank, then not only will the end result include a $ 1 million transfer from a government account to your bank's reserve account, but you - a consumer or firm - will have a nice fat $1 million balance in your ordinary bank account, for free, and ready to spend.
But if the government simply credits your bank's reserve account with $1 million dollars, there is no guarantee at all that that action will result in a single consumer or firm having more money to spend. If the bank had no incentive to lend to you before, it probably doesn't have all that much incentive to lend to you now.
Banking is a unique and limited type of business. Basically all banks buy is loans. That is, they give borrowers money in exchange for promises by the borrower to pay more money later. That's why government swapping liquidity for other assets with banks, or giving money to them outright, is so much less effective than providing the money directly to the real, non-financial part of the economy. Why give money to a firm that is only permitted to buy promises when you could give it instead to consumers or firms that are permitted to buy cars, televisions, airplanes, factory equipment, tractors yoga lessons? To me, that is the essence of why Central Bankism is so stupid.
Applying the MMT "price not quantity" dictum, I think we should say that just as stuffing free money into reserve accounts will not boost lending unless it drives down interest rates, for the same reason taxing money out of reserve accounts will not depress lending unless it raises interest rates.
geerussell, good point. I tried to make the same point today over at Sumner's blog.
Pete, good point about shadow banking.
Dan and geerussell -- Does the Fed actually look at "reserve+vault cash" or just reserves held at the cb? I don't know the answer myself, but if "reserves never leave the cb" then the vault cash would seem to be something different from reserves. Since it can't be used for interbank lending for reserve purposes, perhaps it would not be counted in calculating a "reserve tax".
Dan Kervick,
I was thinking about banks being capital constrained, not reserve constrained. Negative IOR is a hit to capital, directly. Less capital is less lending (in a good economy, now I would guess banks have way more capital than they need). If the CB gives a bank 1M in reserves, the bank can pay its CEO additional 1M in deposit, not as a loan, but as a gift ("bonus") and its capital is intact. So I would imagine injections of reserves can trickle into deposits easily. Same thing in opposite direction - if the Fed debits 1M of reserves as a negative IOR charge, the bank to preserve its capital could choose to lower salaries by 1M.
Thanks, Tom.
@ John Z
Yikes. I see that I left the "NOT" out. It should read "In my reading on negative rate proposals, the presumption has been that the tax on reserves would NOT apply to vault cash. I haven't seen a counter-instance."
I'll take down the wrong statement.
Post a Comment