Tiered negative interest rates and required reserves

Tiered negative interest rates are just a mirror-image of required reserves with positive interest rates.

Think back to the olden days. Interest rates were always positive, except for currency that paid 0% interest, and banks had minimum required reserves. So for a 10% required reserve ratio, for every $100 of demand deposits banks the banks were allowed to earn positive interest on $90 in loans, but had to keep $10 currency in the basement earning 0%.

Except having all that currency in the basement was an open invitation to robbers, so the bank deposited it at the central bank, in an account that paid 0%. Any bank going below that minimum could either borrow from another bank, or from the central bank, at a positive interest rate.

Required reserves were a tax on banks. They earned 0% on those minimum required reserves instead of some positive interest rate.

Now lets go to Denmark, Switzerland, or Japan. It's just like a mirror image of the olden days. Interest rates are negative, except for currency that pays 0% interest, and banks have maximum allowed reserves. They are allowed to keep a maximum amount of currency or deposits at the central bank earning 0%. Any bank going above that maximum can either lend to another bank, or to the central bank, at a negative interest rate.

Tiered negative interest rates are a subsidy on banks. They earn 0% on those maximum allowed reserves instead of some negative interest rate.

In both cases, what matters is the marginal interest rate. In the olden days we called it Open Market Operations when the central bank bought or sold bonds. Nowadays we call it "Quantitative Easing" when the central bank buys bonds (and negative QE when it sells bonds). Silly new name for a very old thing. By using OMO/QE the central bank can push or pull banks from the 0% interest margin to the positive or negative interest rate margin. With tiered negative rates, plus QE, commercial banks are pushed onto the negative interest margin.

If currency paid a market rate of interest, either positive or negative, none of this would matter. But with the nominal interest rate on currency stuck at 0%, government is collecting a tax on holding currency if market rates are positive, and paying a subsidy on holding currency if market rates are negative. So the minimum required reserves in a positive interest world, and the maximum allowed reserves in a negative interest world, level the playing field between banks and currency, but unlevel the playing field between banks and non-banks.

Everything old is new again, only it's a mirror image.

Comments

[JP Koning]

"Tiered negative interest rates are just a mirror-image of required reserves with positive interest rates."

Great observation, Nick. Demystifies what is otherwise an arcane mechanism.

" With tiered negative rates, plus QE, commercial banks are pushed onto the negative interest margin."

It could still get a -0.1% overnight rate via a non-tiered policy whereby all reserve deposits are penalized. The difference is that with the former, the BoJ is collecting less seignorage since interest is levied on just a portion of the whole. Why subsidize banks at the expense of taxpayers?

[Nick Rowe]

JP: Thanks!

Good question. I don't know the answer. My guess: they don't want bank equity to fall, or banks to go out of business. Or it's some sort of political thing, with a lot of customer resistance to earning negative interest rates on their deposits. (My secret source suggests the latter.)

[Roger Sparks]

"Required reserves were a tax on banks. They earned 0% on those minimum required reserves instead of some positive interest rate."

I think you are misusing the meaning of "tax" in this statement. To me, a "tax" is always a forced transfer of property to the taxing authority. Required reserves, instead of being a "tax", are a requirement to perform in a prescribed way.

You go on to convert the "tax" into a "subsidy" in the mirror image condition. The problem here is that "subsidy" is usually considered as a transfer of property. There is no transfer of property even in the mirror image -- there is only a lack of transfer.

I think if you limit your use of the words "tax" and "subsidy" to describe actual transfers of property, your analogy of a mirror image will be more realistic.

[JP Koning]

Nick, Miles Kimball sort of says the same thing in a recent post in which he advocates tiering:

http://blog.supplysideliberal.com/post/138134320548/goldman-sachss-dirk-schumacher-predicts-the

[Nick Rowe]

Roger: if interest rates are 5%, and I force you to give me a loan of $100 at 0% interest, I am taxing you $5 per year.

JP: good find. Yep, me and Miles seem to be roughly on the same page.

[Tom Brown]

So in Japan, will it work? Will Japan be regularly nailing their inflation target a year from now? Will they be happy with their NGDP and exchange rates and their other macro economic goals?

[Roger Sparks]

"Roger: if interest rates are 5%, and I force you to give me a loan of $100 at 0% interest, I am taxing you $5 per year."

I think you are basing this statement on the assumption that banks are "lending" money to the central bank. Reserves placed at the central bank are not a loan. They are a forced conservation of money that has been placed on account by bank customers.

The basic origin of money may be at issue here.

[Lord]

I would call it OMO at the short end and QE when at the long, though long varies.

[ACEMAXX-ANALYTICS]

Very nice blog post, Nick,

Yes, transmission of the negative interest rate is working well, says the SNB to us.

The SNB calls it „a distribution of sight deposits between banks“.

1)Since Jan 2015, the SNB has been charging neg interest of -0.75% on sight deposits held by banks and other financial institutions at the SNB.

2)And there is an exemption threshold which is calculated as 20 times the min reserve requirement (two-tired system)

Financial institutions whose sight deposits exceeded the exemption threshold (see: two-tired system) have transferred money from their SNB accounts to banks which had not exhausted theirs threshold and were willing to accept these funds at somewhat less negative interest than -0.75%

i.e. every new created CHF is subject to negative interest (after almost all exemption thresholds are exhausted).

http://acemaxx-analytics-dispinar.blogspot.de/2016/01/bank-of-japan-mit-negativzins-und-die.html

http://goo.gl/wY02Mw

[JKH]

“Everything old is new again, only it's a mirror image.”

Not exactly, at least when you compare it to the old (pre-QE) Fed system.

Under that system, the Fed restricted the quantity of excess reserves supplied in such a way as to effectively target a positive fed funds rate well above the IOER rate, which was zero. You could think of this as an “asymmetric channel” system, because the trading range for the funds target was asymmetrically positioned (usually asymmetrically high) between an upper bound LLR rate and a lower bound IOER rate of zero.

Under the tiered negative system (as I understand it), the CB can be relatively or even absolutely loose about managing the quantity of excess reserves supplied, because IOER acts as a lower bound in such a way that is congruent with the target policy rate. This is very different from the pre-QE Fed system. But it still works through arbitrage in its own context. No bank that has excess reserves earning the negative IOER rate will want to earn an even more negative rate than that. Hence IOER serves as a lower bound. If this is the case, it is essentially a floor system. And I’m pretty sure that’s the way it would be implemented under QE conditions, since QE conditions are generally sloppy when it comes to quantities of excess reserves supplied (compared for example to the pre-QE Fed system). The close juxtaposition of the lower bound or floor under this system is very different from the comparable arbitrage circumstance under the pre-QE Fed system, where typically there is a very large opportunity cost gap between the target funds rate and the zero IOER rate.

(The other possibility in the negative rate system is that the CB does get fancy and restrictive about the way in which it supplies the quantity of excess reserves at the margin, in which case it becomes a symmetric channel system, which is still very different from the Fed’s pre-QE procedure. But I doubt this is the case with negative rates under QE.)

[Sustain]

If anyone still has questions about the implications of NIRP, please read this:

http://www.zerohedge.com/news/2016-01-30/pandoras-box-open-db-warns-japan-may-have-started-silent-bank-run
quoting Deutsche Bank's Japan analyst Mikihiro Matsuoka,
"...

We wonder whether removing the last breakwater of the lower bound at a zero interest rate could end up being an expensive choice in the long run. The factor which pushed the BoJ down this path is probably its view that causality runs from economic activity to wages and then to prices, which we do not agree with. Our view is that causality runs from economic activity to prices and then to wages, and we do not share the argument that inflation does not rise because wages have not risen.

We believe the additional room that the BoJ has to lower rates on bank reserves is smaller than in other countries that have already introduced a negative interest rate, because of the lower net interest margin of commercial banks in Japan. However, if the BoJ pursues this path, we could reach the point of the trade-off of possibly damaging the soundness of the banking system.."

[JP Koning]

"The other possibility in the negative rate system is that the CB does get fancy and restrictive about the way in which it supplies the quantity of excess reserves at the margin, in which case it becomes a symmetric channel system"

I think this is how Denmark works. The Danes have a tiered system, like the Japanese and the Fed, but unlike the Fed and like the Japanese, they have a negative deposit rate. The overnight rate tends to trade above the deposit rate, but below 0%.

https://twitter.com/jp_koning/status/692951550780686336

It's interesting how the Danes have been manipulating the tiers in order to change what is marginal. In the old days the Fed could do this through reserve requirements; the quantity of reserves subject to a tax. Denmark has done this by (in part at least) by the opposite; altering how many deposits qualify for an exemption from the deposit rate penalty. As Nick says, its the same old thing, except a mirror image.

In any case, I like Nick's analogy because it drives home the point that the marginal reserve is the important reserve. Journalists have been mistakenly criticizing the Bank of Japan essentially doing nothing because it has resorted to tiering (ie. see here https://www.washingtonpost.com/news/wonk/wp/2016/01/29/japan-is-trying-to-save-its-economy-with-a-jedi-mind-trick/) so his post is quite useful.

[Sustain26]

"...

We wonder whether removing the last breakwater of the lower bound at a zero interest rate could end up being an expensive choice in the long run. The factor which pushed the BoJ down this path is probably its view that causality runs from economic activity to wages and then to prices, which we do not agree with. Our view is that causality runs from economic activity to prices and then to wages, and we do not share the argument that inflation does not rise because wages have not risen.

We believe the additional room that the BoJ has to lower rates on bank reserves is smaller than in other countries that have already introduced a negative interest rate, because of the lower net interest margin of commercial banks in Japan. However, if the BoJ pursues this path, we could reach the point of the trade-off of possibly damaging the soundness of the banking system..."
Deutsche Bank's Japan analyst Mikihiro Matsuoka
http://www.zerohedge.com/news/2016-01-30/pandoras-box-open-db-warns-japan-may-have-started-silent-bank-run

[JKH]

J.P.

I responded before thinking it all through. I knew there was an asymmetry in there somewhere relative to a comparison against the Fed, and wanted to get that out of my system.

You may be right about Denmark; I haven’t studied the individual European cases.

IF there is QE, then I think the key is whether a QE quantity effect is captured within the required reserve portion, or whether it “leaks” into the excess reserve portion. If the latter, then the excess portion must behave according to a floor system, because the CB loses the necessary leverage over the excess reserve quantity in order for it to be otherwise.

It seems to me that under QE it must work that latter way - unless there is a very impressive co-ordination between QE implementation and a parallel required reserve ratcheting to absorb the excess reserves being added otherwise.

Ironically, I’m more familiar with the pre-QE Fed system and the old Bank of Canada system (which worked according to the same interest rate model) than I am with current day BoC operations. I think it’s a symmetric channel, unless QE effects (if they exist) are implemented, in which case they must work within excess reserves because required reserves are zero.

The part I like most about Nick’s presentation is the anti-symmetric fiscal effect of required reserves. Required reserves paying zero interest under otherwise positive interest rates constitute a tax on banks (and “free funding” for the CB). Required reserves paying zero interest under otherwise negative interest rates constitute a subsidy to banks (and funding at a relative “penalty” rate for the CB).

[JKH]

And of course - required reserves earning zero interest are a min constraint under positive rates, but a max constraint under negative rates.

[Nick Rowe]

Thanks all for comments. I feel I've done my bit on this topic though, and probably don't have anything useful to add. Both JP and JKH know more than me about the remaining issues.

[Roger Sparks]

Nick: Thanks for writing about this timely subject (negative interest rates).

Here is a new (I think) pattern: Link these observations into a logical framework:

Central Banks are an arm of government.
Reserves may be cash or government bonds.
Negative interest rates are designed to "tax" only the largest of "excessive" reserve accumulations.

It seems to me that the cumulative effect of these observations is " negative interest rates are a penalty designed to encourage the purchase of government bonds" .

If this is correct, the bank (with excess reserves) will consider two options:

1. Place excess reserves on account at the central bank and pay the negative interest rate "tax".

2. Buy a government bond, thus converting cash-in-reserve into government term debt.

Is there a hole in this chain of economic coercion?

[Bill Woolsey]

Where do government bonds serve as reserves? Not in the U.S. It is deposits at the Fed and vault cash only.

[Roger Sparks]

I need to do more analysis!

"Is there a hole in this chain of economic coercion?"

Yes, a big one! Thanks to Bill Woolsey for pointing it out.

My second "observation" was incorrect, very incorrect. Banks are not allowed to count government debt as reserves. As Bill said, reserves must be cash or deposits at the Fed. Much more information on this aspect of reserves can be found at
http://www.federalreserve.gov/monetarypolicy/reservereq.htm

So if the goal of negative interest rates is to extract a tax to incentivize behavior, who's behavior is a target?

I understand that the negative rates apply to large deposits. This would narrow the number of suspects.

I understand that money only disappears when debts are paid. The JCB wants inflation, not a shrinking of debts. The goal of negative interest rates should not be to reduce the amount of money taxed by negative rate. There are no suspects here.

Maybe the JCB wants money to turn over faster. Maybe it wants to increase the velocity of money. I don't see that this would be a result of negative rates unless the increased velocity can be achieved by actually reducing the total amount of money in the economy, which would take us back to reducing the amount of debt, which we have already ruled out as a likely goal.

So what might the mechanics of effective negative interest rates look like? Who's behavior will change and what form will the change take?

[JKH]

excellent article on Japan (HT - JP on twitter)

http://blogs.ft.com/gavyndavies/2016/01/31/bank-of-japan-tries-another-flavour-of-qe/

"The vast bulk of their holdings (the “basic balance” accumulated during previous rounds of QE) will continue to earn interest at +0.1 per cent. Reserve requirements above that level (the “macro add-on” balance) will earn interest at zero. Anything on top of that level (the “policy rate” balance) will be paid -0.1 per cent."

"But the BoJ says that this new mechanism will succeed in reducing money market rates into negative territory, because these rates are determined by the policy rate paid on marginal increases in bank assets, not on the entire aggregate of these assets accumulated in the past."

(NR - note use of the term 'money market')

"But, under the new BoJ plan, any shift by banks into cash will be penalised by increasing the scale of the reserves at the central bank that will be charged the negative rate. So the problem of a drain into cash as policy rates dive into increasingly negative territory is mitigated."

clever

[Nick Rowe]

JKH: "(NR - note use of the term 'money market')"

EeeeeeeeeeeeeeeK!

[JKH]

Got your heart started for a new month!

[D]

"...
Keynesian central bankers leave time out of their calculus. While they think they are lending money, they are really lending time. Borrowers purchase the use of time. Hazlitt reminds us that the old word for interest was usury, “etymologically more descriptive than its modern substitute.”

And as Mises explained above, time can’t have a negative value, which is what a negative interest rate implies.

Borrowers pay interest in order to buy present assets. Most importantly, this ratio is outside the reach of the monetary authorities. It is determined subjectively by the actions of millions of market participants.."
http://fee.org/articles/when-zeros-too-high/