« Liquidity pile-ups on the Wicksellian roundabout | Main | Police, Crime and the Great Canadian Crime Drop »

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

"I agree the CB could choose to make that promise, and practically they may be compelled to make that promise (due to banker's sitting on their governing board), but I don't see why they necessarily legally (by the book) have to make that promise. But then I'm absolutely no expert..."

No, Tom.

It's nothing to do with board representation or other kinds of influence from evil commercial bankers.

The central bank relies on the commercial banking system as the distribution agent for its currency. It must accommodate the flow of currency that is desired by the public and it must facilitate the banking system in accommodating this flow as its agent. It is inconceivable that the CB would refuse a request for additional currency or the redemption of surplus currency - unless it shuts down banking operations altogether, which is a much larger issue. Such a refusal would amount to a breakdown of the monetary system - a gross intervention into the management of bank balance sheets and banking operations more generally. It would be banana republic chaos. That said, the central bank and the regulators and the state authorities determine all these rules - but this particular aspect is in the interests of a functioning banking system at a very basic level. Anything contrary to that would just constitute a fiasco in basic banking operations.

Tom,

This shouldn't be a PKE/MM issue.

It's not theory - it's just the way it works.

The CB does not refuse a bank's request for currency transactions unless it is willing to suspend a bank's operations in total or suspects some sort of legal mischief just short of that. And those are larger issues.

"It is inconceivable that the CB would refuse a request for additional currency"

You mean beyond what current electronic CB-deposits dictate? If CB-deposits balances are at $0, then how will the commercial banks buy currency? The central bank can inject more electronic reserves via asset purchases or they can lend more electronic reserves, or they can perform those operations with currency directly.

Also, say a nation decided to go paperless: no more currency. How would they accomplish that? One way would be to start refusing requests for more currency, no? Operate like a roach motel: currency goes in, but it doesn't come out. Would that destroy the banking system? I don't see why it would. Eventually they'd eliminate all the currency out there. Do think we need paper notes to have an effective central bank?

I agree that it's hard to imagine the a central bank that isn't accommodative to some minimal degree... but I don't think that MMists will concede that point.

The central bank probably doesn't care what the mix is between central bank and reserve deposits, but I don't think you'll be able to convince Nick or any other MMist that it's not possible for the CB to target a fixed dollar amount of MB (currency+reserves). That seems like the fundamental philosophical divide. They'd probably agree it's a bad idea: just like Nick says targeting a fixed overnight rate is a bad idea, but I'd be surprised if they agreed it's not possible.

"The central bank probably doesn't care what the mix is between central bank and reserve deposits"

should read

"The central bank probably doesn't care what the mix is between currency and electronic reserves is"

Tom,

Compare two situations:

a) A bank loses a deposit to another bank - causing a reserve outflow

b) A bank purchases currency for a depositor - causing a reserve outflow (i.e. reduction) - and the deposit disappears just as it does in case a)

Identical reserve effects

If the bank couldn't recover its reserve position in case a), it would be demonstrating a fundamental inability to manage its liquidity position and would be shut down

Banks are doing this sort of thing everyday - moving reserves around - aggregating in the $ hundreds of billions if not trillions

I'm going to stay away from the question of how MM'ers could possibly have a problem with that fact

JKH, I agree w/ your examples, however the bank losing the deposit can always borrow reserves from other banks (if available). They don't have to obtain them from the CB. In your case a) for example, that reserve outflow will likely cause excess reserves to accumulate at the receiving bank (since it's getting reserves in a ratio of 1:1 with the received deposit, instead of 1:10). Thus we know of at least one bank that can lend it reserves after the outflow (and maybe before the outflow!... such that the "outflow" never even takes place... an non-accomodative CB could result in a little bit more flexibility here amongst the banks). In case b) the aggregated banks lose both reserves and deposits. In case a) they are both preserved.

Well I'm treading on thin ice speaking for MMists... but still I think that's the fundamental disagreement: MMists insist a CB can do what it likes w/o regard to the banks (even though that may be a stupid thing to do).

So at this point I should probably shut up and let the resident MMist speak for himself... but when I put my MM hat on, #1 on my list of priorities is to think "The CB can do what it wants!"... then everything else makes sense from there on out. :D

Tom,

"MMists insist a CB can do what it likes w/o regard to the banks"

That seems reasonable within reasonable scope, IMO

But shutting down the normal function of the commercial distribution system for CB currency (while assuming that system in the thought experiment) amounts to operational terrorism - I think there's a limit to what they would consider to be reasonable

P.S.

To see my general point on this, just imagine replacing the existing banking system with a single (central) bank without commercial banks

Deposits would be convertible into currency and vice versa on demand

That's the basic symmetry

The actual banking system facilitates that symmetry by using the commercial banks as distributor/dealers in currency and purveyors of deposits

Same essential symmetry of two-way flow, using an agent function (the commercial banking system) for both deposits and currency

JKH,

I wonder if we are talking about two different things here.

I absolutely agree that, as far as bank customers go, there is symmetry in the conversion between central bank and commercial bank liabilities. The asymmetry I'm talking about though is in the different constraints faced by the central bank and commercial banks.

Take your scenario, where there is a single (central) bank without commercial banks. That bank can set the interest rate at whatever it likes. The interest rate will determine the demand for loans and the demand for deposits and currencies. If it wants, it can set a high rate on loans and a low rate on deposits. This might affect the overall size of its balance sheet. But this bank can never face a liquidity crisis. It can never have the problem of having to pay off liabilities at a time when it cannot realise assets, because the only way that its liabilities can ever be reduced is through loan repayment.

Once we have multiple banks, liquidity becomes an issue. Now banks face the possibility of customers being able to reduce the bank's monetary liabilities in a way that does not involve loan repayment - by exchanging for the liabilities of other banks, either by withdrawing currency or making a payment. This places a constraint on their behaviour. They can no longer set interest rates at whatever they like, without risking running into liquidity problems.

The central bank does not face this constraint. It can operate like the single bank in the imaginary scenario. It can set its rates at whatever it likes without liquidity risk, because the only way anyone can make it reduce its aggregate monetary liabilities is by repaying a loan from it.

None of the affects the symmetry observed by the bank customer who can switch either way between claims on different banks.

Again, I don't know whether "asymmetric redeemability" is the best term for this, but I do think this is equivalent to what Nick was describing.

@ Nick E;

But if the question is only about "setting its rate" due liquidity considerations then the alpha/beta fx-idea doesn't apply anymore? I mean then the banks are not worried about their "market share" and "fx-rate" but risks (like liquidity risk) and profitability (which sounds about right).

I'm not sure the fx-idea can work if the public can always choose between cash (coins and notes) and deposits. Or it could but then it should state that the multiplier is about the amount of cash, which should trade at a premium at the bank desk?

Actually, I was commenting less on this post and more on the asymmetric redeemability post Nick was referring to - http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/10/what-makes-a-bank-a-central-bank.html

Nick E.,

I agree with your description above of the liquidity and interest rate characteristics of banking. I think it’s pretty standard stuff – but in my view it has nothing to do with a concept of “asymmetric redeemability”.

Central banks set rates in the short term. They typically force interest rate arbitrage responses when necessary through excess reserve balance management and/or interest rate channel arrangements. They require a mechanism to accomplish this while at the same time paying zero interest on currency.

If as in some cases (e.g. Nick R.) one wants to assume away the complication of bank reserve balances at the CB, and deal only with currency, then the CB must have the capability of doing OMO with currency. Same difference, I guess. The CB just keeps doing it to keep desired interest rate levels in line. But in neither case does that have anything to do with “asymmetric redeemability” in my view.

My interpretation of “asymmetric redeemability” is based on the following sort of description found at the start of Nick R.’s latest post:

“Beta banks promise to convert their money into the money of alpha banks at a fixed exchange rate. Alpha banks make no such promise the other way. It's asymmetric redeemability.”

And I disagree with that as I’ve explained. I’ll return to it again below.

I looked at Nick R.’s 2009 post on this subject which you referred to in your comment above. I’d forgotten about that one. It’s one example of where he constructs an argument based on interest rate arbitrage, as you say.

http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/10/what-makes-a-bank-a-central-bank.html

He uses two different examples:

a)“Suppose bank A and bank B both issue paper money, but B promises to redeem its notes for A's notes at par, while A makes no such promise. Suppose interest rates are initially at 5 %, and A decides it wants to lower interest rates to 4 %, while B decides it wants to keep interest rates at 5 %. What happens? The interest rate differential creates infinite arbitrage opportunities. Borrow A's notes from A at 4 %, convert them into B's notes at par, and lend to B at 5 %. B would end up accepting an infinite quantity of A's notes, which it could only lend out at 4 %, while paying 5 % on its deposits. It would make infinite losses. To avoid infinite losses, bank B would be forced to lower its rate to 4 % too, or else suspend redeemability at par.”

What makes the particular interest rate arbitrage work there is that B “could only lend out at 4 %, while paying 5 % on its deposits.” That has nothing to do with asymmetric redeemability. It has to do with A’s monopoly control of interest rates – in this case B’s funding cost and B’s lending opportunity. That monopoly power is an arbitrary assumption/assignment to A. A is putting the squeeze on B as a result, but that has nothing to do with any redeemability constraint. The funds just flow that way, creating the interest margin effect according to an arbitrary assignment of interest rate control to A. I see no coherent connection there to an asymmetric redemption constraint.

b)“Suppose A decides to raise the interest rate to 6%, while B tries to keep it at 5%. What happens? The arbitrage opportunity now works in the opposite direction. Borrow B's notes from B at 5%, convert them into A's notes, and then lend them to A at 6%. Bank B would face an infinite demand to redeem its notes, and it could only satisfy that demand by borrowing notes from A at 6%, while earning 5% on its loans. To avoid infinite losses, bank B would be forced to raise its rate to 6% too, or else suspend redeemability at par.

Parallel to a), what makes the interest rate arbitrage work there is that B must end up “borrowing notes from A at 6%, while earning 5% on its loans”. Again, that has nothing to do with asymmetric redeemability. It has to do with A’s monopoly control over interest rates – B’s funding cost and B’s lending opportunity. And that monopoly power is an arbitrary assumption/assignment to A in the example. A is putting the squeeze on B, but the funds just flow that way, creating an interest margin effect according to an arbitrary assignment of interest rate control to A. Again, I see no coherent connection there to an asymmetric redemption constraint.

I therefore continue to interpret the issue from the perspective of the type of general description offered by Nick R. at the start of his latest post.

A similar description was offered at the start of the 2009 post:

“Commercial banks promise to redeem their monetary liabilities for the monetary liabilities of the central bank at a fixed (or at least pre-determined) rate. Central banks do not promise to redeem their monetary liabilities for the monetary liabilities of the commercial banks. This asymmetry of redeemability is what gives central banks their power over commercial banks.”

But the symmetry does exist in my view. The central bank’s promise to redeem exists by way of its promise to deal currency for bank reserves and vice versa, combined with the commercial bank’s promise to do the same with currency and deposits. The commitment is transitive – from central bank to commercial bank to commercial bank customer – in both directions.

For example, in the case of a single (central) bank with no additional commercial banks, the issue of symmetric or asymmetric redeemability is moot – because there is only one bank dealing in currency and deposits, and so there can be no asymmetry of the type being discussed. And the pertinent symmetry of dealing in two different directions on the currency/deposit transaction (without reserve involvement) is assured by design.

In the case of a multi-commercial-bank system, symmetry in dealing in two different directions is also assured. If B promises to convert deposits to currency at par, and if A promises by way of facilitation to convert reserves to currency at par, then A effectively promises to convert deposits to currency at par (and vice versa). As I said before, it is the symmetry of the transaction that is the material issue – not the symmetry of the physical location where the transaction is done. The physical location is simply a function of a reserve system that services multiple banks.

More from that 2009 post:

“It also makes no difference if bank A is in one country, bank B is in another country, and they have different media of account, as long as B promises to redeem its notes for A's notes at some fixed rate, and A makes no such promise.”

In the domestic banking case, A makes a promise via the currency/reserve nexus. If B promises to convert deposits to currency at par, and if A by way of facilitation promises to convert reserves to currency at par, then A effectively promises by transitivity to convert deposits to currency at par (and vice versa).

This case where there is another central bank up the chain, with two different media of account, does in fact introduce a potential asymmetry. That’s because A does not promise to convert its currency to the pegged currency at the targeted rate of the pegged currency. That’s very different from the case where a domestic central bank deals at par in currency and reserve balances with its commercial banks. However, domestic symmetric redeemability continues as before, whatever the outcome of the foreign exchange peg that operates upstream.

A foreign exchange peg introduces a risk level not applicable in the domestic setting. That said, near-symmetric redemption is operative at the FX level - provided that B is successful in maintaining the peg. That means that when A initiates FX transactions (from time to time), the FX rate will be near the peg. The fact that A deals in the FX market constitutes a case of symmetric redemption – but the precise rate will depend on the effectiveness of B’s peg. So there is a pricing asymmetry risk there.

So in my view Nick R.’s post-2009 alpha/beta paradigm is broadly applicable to FX pegs and domestic interest rate transmission, etc. – but with the caveat that it does not indicate asymmetric redemption at the domestic level.

I'm going to leave JKH and Nick E to argue this one out, I think. Because Nick E is better than I am at using the same language as JKH.

Just one observation: I see my 2009 post and my recent posts as arguing basically the same position, it's just that I posed the question slightly differently in 2009. In 2009 I was asking "who sets r?", and now I am asking "who sets M?". The answer is the same in both cases: "the alpha bank".

On the other hand: suppose both A and B issue notes. B notes say "I promise to pay one A note". A notes don't say anything. B can become insolvent. A cannot become insolvent.

JKH, you write the following, or something equivalent, several times:

"But the symmetry does exist in my view. The central bank’s promise to redeem exists by way of its promise to deal currency for bank reserves and vice versa, combined with the commercial bank’s promise to do the same with currency and deposits. The commitment is transitive – from central bank to commercial bank to commercial bank customer – in both directions."

I have no objection, but aren't you forgetting there that the central bank controls the absolute size of its own balance sheet?

I know you hate it, but take my $1 example: that $1 can be any combination of reserves and currency: they turn the lights out for 30 years at the CB but let the banks use their special banks' only ATM machine. That's completely consistent with your paragraph above: the banks can convert excess currency back to reserves, or excess reserves to currency all they want... what they cannot do is alter in any way the total of $1 of liabilities on the CB's balance sheet. This means that it's possible for the banks to create deposits which exceed $1 in total, which puts them at risk should their depositors in turn try to withdraw their deposits in cash. Maybe it's a risk worth taking, but it's a risk the CB isn't experiencing at all.

Maybe you don't like the term "asymmetric redeemability" for that, but whatever it is, it makes the CB an alpha in this situation.

Tom,

“But aren't you forgetting there that the central bank controls the absolute size of its own balance sheet?”

I think we have to be very qualified in making such a statement. It appears to be a statement that monetarists consider sacrosanct. I don’t.

There’s an issue here that I think is parallel to Nick’s point regarding the difference between short term interest rate setting by the central bank - versus long term interest rate determination.

The same tension holds for the supply of currency IMO. The CB does not determine how much currency it needs to supply in the short run. The public has the option to demand it. That becomes a constraint or an input on how the CB manages the size of its balance sheet from there. For example, other things equal, the CB generally needs to expand its balance sheet in conjunction with the expansion of currency it provides on demand.

Regarding your example – it is just too far from real world, IMO. You’re imposing a central bank freeze on currency expansion in order to “prove” asymmetry. Central banks just don’t do that. It’s monetary and economic suicide.

I think we have to draw the line somewhere in these thought experiments. The most radical ones end up “proving” nothing IMO, because we’ve so altered the nature of the thing we started to talk about that we’re not talking about it anymore.

JKH, don't let the $1 limit be a problem: it could be $X ... and again, it's not currency: it's total currency and reserves. But I think I see where you stand. I just pointed that out because it seemed in your response to Nick E, you brought up that point several times: that the CB is willing to trade at will currency for reserves and vice versa... but that says nothing about the total of currency and reserves in existence.

But I know you're not agreeing to the monetarist claim that CB's have the power to fix the liabilities on their BS's at a specific dollar amount ($X): you've made that clear!

... and X could be plenty big enough to accommodate the banks and the customers for the next decade... but at some point X will be an active constraint. If the banks know ahead of time that the CB only consists of an ATM machine, they'll plan accordingly.

JKH: "It appears to be a statement that monetarists consider sacrosanct."

I first heard it said by a Deputy Governor at the Bank of Canada. I think he said: "the *only* thing we control is our own balance sheet".

Even the rate of interest on reserves is a promise about how it will change the balance sheet in response to the current state of the balance sheet.

I find it hard to imagine a central bank that did not freely swap currency for reserves at par. It lets the public+banks determine the *composition* of the base. But any option on the *total base* is an option the Bank of Canada chooses to give them, and it also chooses the terms on which it will allow them to exercise that option.

Looking at the two examples (a) and (b), in both cases the arbitrage involves converting one bank's notes to the other's. In both cases, it is Bank B that is having to the conversion, which is why it is incurs the loss. If it was Bank A having to do the conversion, Bank A would incur the loss. It is because Bank A does not have to do the conversion, that it can force Bank B to follow suit on the interest rate setting. So the asymmetry is in who has to do the conversion.

That's all it is. I'm not sure that's actually any different from what you describe as A's monopoly control of interest rates. On the whole, I'm not seeing anything I really disagree with you on, in relation to mechanical steps. So maybe this is just what you call it. And "asymmetric redeemability" is not my term (although I can at least see why NR has used the term).

(I'm sure when I first read this post, it included a link to the earlier post explaining asymmetric redeemability, because that was how I found it - by clicking the link. That link doesn't seem to be there now. I may not have helped this discussion, by being rather focussed on the content of that older post.)

Nick E: the link is not in this post, but in the post this post links to!

Here's the link again to my 2009 post on "asymmetric redeemability".

In case a) as described by Nick R., the alpha bank A drops its rate to 4 per cent. The alpha customer lends to the beta bank B at 5, requiring conversion of alpha currency to beta currency. On the return trip, the beta bank lends to Alpha at 4, requiring conversion of beta currency to alpha currency.

Nick R. says Beta will drop its rates to avoid such a lending loss in the return trip. Fine. But the counterfactual if it doesn’t do that requires currency conversion. Moreover, given Nick’s construction of a 2 bank system, the interbank clearing system will force the conversion of beta currency to alpha currency if the beta bank does nothing else. That is really what the beta bank fears if it does nothing.

The important point though is that nothing in this case depends on which bank does the currency conversion. It could be either alpha or beta. It doesn’t matter. Which means it doesn’t depend on asymmetric redeemability. The question of who needs to do it is moot. The interest rate arbitrage holds no matter who does the conversion. Nothing depends on the beta bank converting beta currency to alpha currency (or vice versa), as opposed to the alpha bank converting alpha currency to beta currency (or vice versa).

Similar in case b).

Alpha has interest rate control. Interest rate control is not a function of asymmetric redeemability.

So this case fails to prove the dependence of the interest rate arbitrage on asymmetric redeemability. It doesn't illustrate asymmetric redeemability as a necessity.

And earlier, I described symmetric redeemability in the real world case.

"But the symmetry does exist in my view. The central bank’s promise to redeem exists by way of its promise to deal currency for bank reserves and vice versa, combined with the commercial bank’s promise to do the same with currency and deposits. The commitment is transitive – from central bank to commercial bank to commercial bank customer – in both directions."

I agree that the symmetry exists both ways.

Alpha money is eventually 1 to 1 convertible to beta money both ways. Here is another point. The different beta monies are 1 to 1 convertible to each other. For example, a JPM demand deposit is 1 to 1 convertible to a BAC demand deposit both ways. They are both 1 to 1 convertible to currency both ways.

Let's say the banking system of the USA suddenly produces $5 trillion in demand deposits from loans. Everyone panics and wants currency. That demand will be met with an "elastic" currency. Demand deposits down, currency up. Next day the opposite happens. Demand deposits up, currency down. Basically, currency and demand deposits have unlimited supply. There is no gold limiting either one.

JKH,

I think we're going to have to put this one in the relatively small pot of things where we agree to disagree, I'm afraid. But it relates to something I've been thinking about a bit, so it may crop up again in one of my own posts. The discussion has at least been helpful for me in thinking about how I might frame the issue.

Too Much Fed, the sum of reserves and currency in circulation (MB) has a potentially unlimited supply, but ultimately it's the choice of the central bank, since MB will show up as a direct CB liability, no matter the mix between currency and reserves. So in a case where the CB has fixed MB, then MB is the upper bound on the amount of currency.

Nick E.,

OK

Helpful for me also

I may touch on it as well in a post sometime

Thanks for discussion

"So in a case where the CB has fixed MB, then MB is the upper bound on the amount of currency."

Most CB's are not pegged/fixed to anything. Where is the fixed MB?

"Too Much Fed, the sum of reserves and currency in circulation (MB) has a potentially unlimited supply, but ultimately it's the choice of the central bank, since MB will show up as a direct CB liability, no matter the mix between currency and reserves."

If the CB stopped saying there was 1 to 1 conversion of demand deposits to currency thru the commercial banks, then the economic system changes drastically.

JKH KEEPER quotes:

"Regarding your example – it is just too far from real world, IMO. You’re imposing a central bank freeze on currency expansion in order to “prove” asymmetry. Central banks just don’t do that. It’s monetary and economic suicide."

With emphasis on the too far from the real world part.

"But the symmetry does exist in my view. The central bank’s promise to redeem exists by way of its promise to deal currency for bank reserves and vice versa, combined with the commercial bank’s promise to do the same with currency and deposits. The commitment is transitive – from central bank to commercial bank to commercial bank customer – in both directions."

Basically, JKH is saying that currency plus demand deposits is the MOA and MOE.

zyxzyxooxwo

“If alpha and beta money were perfect substitutes for each other, people would be indifferent about the proportions of alpha to beta monies they held. The desired share or ratio of alpha/beta money would be indeterminate,”

I believe that is how the real world works.

“Commercial banks are beta banks. They fix the exchange rates of their monies against central bank alpha money. It seems reasonable to assume that commercial bank money is an imperfect substitute for central bank money. Chequing accounts are an imperfect substitute for currency.”

I believe checking accounts are a perfect enough substitute for currency. When I go to spend on financial assets or goods/services, I can use either one. I believe the exchange rates are fixed both ways.

“If beta banks issued more beta money, holding constant the stock of alpha money, the total stock of money would be higher than desired, and there would be an excess supply of both monies against all other goods. But no individual would choose to go to the beta bank to convert his beta money into alpha money, because, by assumption, he doesn't care about the share of alpha/beta money he holds. The Law of Reflux will not work to eliminate the excess supply of alpha+beta money against all other goods.”

In two places there, it says against all other goods. You have left out the possibility of against all other financial assets.

Too Much Fed,

"You have left out the possibility of against all other financial assets."

Goods include financial assets, alpha money, and beta money.

"However, once they have a legal contract with a depositor, they are obliged to abide by that contract, which includes redeeming deposits for central bank money."

In the US, the major criminal banks have started inventing scams like "account closure fees", which they apply retroactively to existing accounts, in order to refuse to redeem their deposits for central bank money.

Any analysis of banking has to pay attention to this sort of behavior. Really, study of banking today is criminology, the study of frauds.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad