House of Debt's recent piece on full reserve banking opens
with the following line.
"So both John Cochrane and Martin Wolf are advocating
100% reserve banking."
I found this interesting, not so much for the idea of the
same proposal coming from rather different sources, as for the differences in
their approach.
Wolf's article describes the system proposed by Positive Money (PM). This involves strictly
limiting bank deposits to two types, effectively transactions accounts and
savings accounts. Banks would have to
hold 100% reserves against the balance of transaction accounts. Savings accounts, being all bank accounts
that are not transaction accounts, would then have to have a minimum notice period
of, say, a month.
By requiring banks to hold reserves in full cover of their
demand liabilities, it is intended that this structure would reduce the risk of
bank runs. A bank could still find
itself unable to repay savings accounts when they fell due, but because of the
notice period on such accounts, it is hoped that this risk would be easier for
the central bank to manage. It should be
noted that the proposal would not prevent additional liquidity requirements
applying to savings accounts, such as the BIS Liquidity Coverage Ratio.
Because of the 100% reserve requirement against transaction
accounts, additional lending by banks would deplete their liquid asset holdings. As they would be constrained by the need to maintain minimum liquidity levels, additional lending would have to wait until the liquid assets get redeposited again. The intention is that this would slow up the
pace of credit creation.
So it's not unreasonable to refer to this proposal as 100%
reserve banking. However, it is
important to recognise that the full reserve requirement is only proposed for transaction
accounts. This differs from, for
example, the full reserve proposal of Benes and Kumhof.
Reserves appear on the asset side of the bank's balance
sheet. John Cochrane focuses on the
liability side. His concern is with what
he calls run-prone liabilities. This is
principally short term fixed value debt, but he also worries that longer term
liabilities can still be problematic. He
would like to see an incentive system that pushes financial institutions (he does
not treat banks as special here) towards drawing a greater portion of their
funding in the form of capital. As he
sees it, the problem is the extent of funding taken in the form of fixed value
liabilities that are shorter in term than the underlying assets and the
solution is to reduce this. This
proposal has some similarities to the Limited Purpose Banking concept of Chamley, Kotlikoff and Polemarchakis.
So, in a sense there is quite a big difference between
Cochrane's proposal and that of PM. Cochrane
says little about reserves and is instead concerned with bank
capitalisation. It's not clear that this
is really 100% reserve banking at all. In
PM's proposal, on the other hand, reserves play a central role; they are less
concerned about the capital structure of the intermediation aspect of banking.
One of the interesting differences between the proposals, I
think, is in the way they conceive of money.
Cochrane sees much less of a need for fixed value liabilities in the future. He describes how technology substantially changes the requirement
for conventional transactional account balances.
"With
today's technology, you could buy a cup of coffee by swiping a card or tapping
a cell phone, selling two dollars and fifty cents of an S&P 500 fund, and
crediting the coffee seller's two dollars and fifty cents mortgage-backed
security fund."
This is a world with payments and balances, but where the
division between money and non-money is less clear. This fits quite well with my own way of
looking at things. PM's concept of money
is rather more conventional and their analysis relies on a clearer concept of
what constitutes money.
Yet, despite the differences, there is clearly a common
theme in reducing the mismatch in bank balance sheets. The concept appears to be attracting support from a variety of commentators. If anything does come of it, it will likely be a watered down version, but it will be interesting to see where it goes.
"Because of the 100% reserve requirement against transaction accounts, additional lending by banks would deplete their liquid asset holdings. As they would be constrained by the need to maintain minimum liquidity levels, additional lending would have to wait until the liquid assets get redeposited again. The intention is that this would slow up the pace of credit creation."
ReplyDeleteI haven't read the specific proposals, but what do you mean when you use the word "deplete" in "additional lending by banks would deplete their liquid asset holdings"?
My initial interpretation of the concept of "100% reserve banking" is to assume that the 10% reserve requirement of today (in the U.S.) is changed to 100%. But in a simple, friction-less theoretical model, I'm not exactly sure why this, in and of itself, would force banks to "wait" (as you put it) any longer to lend than they do today. I'd presume we'd have a similar reserve maintenance period structure, and banks would still be participating in money markets to acquire the optimum amount of reserves.
You referred to Woodford's "Monetary Policy in the Information Economy" paper in a previous post - I'm thinking the interbank model he details stays the same here. The only thing changing the reserve requirement from 10% to 100% does, ceteris paribus, is to shift the reserve demand curve way over to the right, which would cause an increase in the interbank rate (probably to the ceiling). To counter this, the central bank would have to flood the reserve market with reserves in order to bring rates back down to its target. Or it could change its ceiling rate to the target rate (effectively running a "ceiling system"). In the former case, banks trade long-dated for short-dated securities. In the latter case, banks would have to lever up their balance sheets by essentially taking on massive overdrafts from the discount window.
Rates-wise and lending causality-wise, it'd seem we'd be in a similar place. The reserve market is now just operating in a different location of the 'supply/demand' graph. If there would be effects on bank lending, perhaps it would be due to the secondary effects of changes in bank balance sheet composition, leverage, equity, etc.
As I understand it, the intention with this proposal is that the monetary authority would be looking to stabilise the stock of reserves, rather than the interest rate. So the monetary authority would decide the level of reserves from time to time, rather than supplying them on demand, or providing the amount needed to stabilise the interbank rate.
DeleteBanks would then hold reserves: a) as backing for transaction accounts; and b) for their own account. If banks made loans to non-banks, that would then involve a reduction in reserves held for the own account of banks and an increase in that backing deposits, so effectively a reduction in the liquid asset position of banks overall. Ultimately, that can only be replenished by the banks taking more funds into savings accounts.
As the amounts in transaction accounts would be fully backed, banks should never be long or short reserves as regards clearing balances, although clearly that could happen as a result of adverse events on the intermediation side. I'm not sure how it is expected the central bank would respond to this or to variation in non-bank demand for liquidity.
So, it's sort of trying to make medium of exchange money truly exogenous, if you like, rather than what we have at the moment. I agree that simply changing the reserve requirement from 10% to 100%, but continuing to supply reserves on demand wouldn't achieve very much, but I don't think any advocates of 100% banking are suggesting that.
trying to make medium of exchange money truly exogenous
DeleteI was quite surprised to see that statement coming from you. While it is also what I undestand the PM people to be saying, I had always considered the implication of endogenous money to be that that was impossible. To my mind, loans create transaction deposits (how else would one repay the loans?), which means that settlment balances are by definition endogenous, period.
Depends what you mean. I'm not sure that endogenous money people would necessarily say that it was impossible for the central bank to strictly control any monetary measure - just that it doesn't usefully describe how our current system works.
DeleteWithin the PM system, loans create transaction deposits but when depositors invest into savings accounts, that destroys transaction deposits. And any variation in total transaction deposits (held by non-banks) is matched by an offsetting variation in the balance of reserves held for own account by the banks. So the total is the same and it is that total (which is what PM views as the supply of "money") that the monetary authority would be looking to hold constant.
Note also, that I'm not saying that the monetary authority would be able to do this in practice. It might be difficult to reconcile it with maintaining financial stability. I'm not sure.
There is a divide among the endogneous population (pardon the pun) between those who say it is a necessary feature of any mondern monetary system and those who say it is a necessary feature of any monetary system, period.
Deletehttp://www.elgaronline.com/view/journals/roke/1-2/roke.2013.02.04.xml
I guess the division can be on the exact demarcation of a monetary system vs. subsidiary means of payment within the same 'pyramid of liabilities'. But I think both 'schools' would agree on that no modern payment system could function with an exogenously determined quantity of the means of settlement.
(Sorry, I was somehow under the impression that you were a card carrying member of either one or the other school.)
I guess the division ...
Deletewas meant to read:
I guess the division DEPENDS ON the exact demarcation of a monetary system vs. subsidiary means of payment within the same 'pyramid of liabilities'. But I think both 'schools' would agree on that no modern payment system could function with an exogenously determined quantity of the means of FINAL settlement.
I'm a little bit wary of describing myself as a supporter of endogenous money, just because it seems to means different things to different people.
DeleteFor a start, very little is truly exogenous. Things that have already happened are, and therefore also the balance sheets we start with. Things like the weather may be exogenous. Governments and central banks react to what's happening in the economy and wider political issues. Their actions are only exogenous in models, because we don't want to model them.
I believe that the more important causal direction is from the level of activity to the money supply, rather than the other way around. I think that the impact of the stock of money (particularly the balance of transaction accounts) on activity is not very important. This happens to be what I understand the Kaldorian position on money endogeneity to be. However, I think some people understand it differently.
Rather than refer to exogeneity, I wondered whether anybody thought that it was impossible for the central bank to strictly control any monetary measure. I believe it is the aim of the PM proposal to create a monetary measure that the central bank can strictly control. Victoria Chick has supported PM, so I presume she thinks this is possible. I can see how it is possible in principal within what they propose. I am less sure that the central bank might not be forced to adjust the money supply at times in response to events. In that case, you might argue that the money supply is really endogenous. But then, by the same token, what is exogenous in our current system? The interest rate? The inflation target? All seem to be amended when circumstances dictate.
Was it true that under the free banking/gold standard era in the USA in the 1800s, gold was a sort of monetary base that couldn't really be increased in response to requirements. So deflationary crises caused a reset of prices such that the size of the monetary base once again was enough to serve the economy?
DeleteThe Mpesa system in Kenya is interesting because the means of payment is mobile telephone talk time -and 30% of Kenyan GDP is transacted with Mpesa.
Nick Edmonds, you write:
Delete"I'm a little bit wary of describing myself as a supporter of endogenous money, just because it seems to means different things to different people."
That seems to me to be a wise move!
Here's an example (and keep in mind my status as a lowly amateur hack who's just trying to learn)... Last August I had noted how different folks had used the word "endogenous" and "exogenous" wrt money, first encountering it on pragcap, with Fullwiler and Keen, etc, and then seeing how the MM types used it: Glasner, Sumner, Rowe etc. So I'd made myself the following note (on a page I just keep to make notes to myself):
"Nick Rowe's comments on "The supply of money is demand determined." ... Not even not even wrong, and all that. Also some bits about "perfectly inelastic wrt" and perfectly elastic, etc. Basically one of two talking about money being endogenous in the short term (between six week meetings of the BoC or Fed) but not in the longer term (like two years out) where inflation targeting makes it exogenous. [here I updated my note to indicate that Nick Rowe had corrected me]
http://worthwhile.typepad.com/worthwhile_canadian_initi/2012/04/the-supply-of-money-is-demand-determined.html"
Much to my surprise Nick Rowe read that note, and made the following comments:
Nick Rowe:
"Tom: thanks, but that's not quite right. Under inflation targeting the quantity of money is endogenous in both the short run and the long run. It's the nominal rate of interest that is exogenous in the very short run (6 weeks or less, for the Bank of Canada anyway), but endogenous in the long run. "
Tom Brown:
"Here's what I mean Nick. This from Beckworth a week or two ago:
"I suspect you view all changes in the monetary base as endogenous. I only view short-run changes for a given interest rate target as endogenous. Over longer horizons the Fed is changing interest rates according to something like a Taylor Rule. These policy changes in target interest rates mean exogenous changes in the monetary base."
http://macromarketmusings.blogspot.com/2013/08/a-permanent-expansion-of-monetary-base.html?showComment=1375901514604#c7523207141783165074
Glasner here:
"So while I think that bank money is endogenous, I don’t believe that the quantity of base money or currency is endogenous in the sense that the central bank is powerless to control the price level."
http://uneasymoney.com/2012/04/11/endogenous-money/
And Scott Sumner here:
"In the long run banks are constrained, as the Fed will adjust the monetary base to prevent economic overheating. The endogenous money folks, who are right about the period between Fed meetings, overlook this longer run problem with their theory. Six weeks is not a long enough period to have major macroeconomic consequences. But in the very short run the banks are not constrained by a lack of reserves, if the Fed is targeting the short term interest rate. The base is endogenous during that period."
http://www.themoneyillusion.com/?p=21786
Nick Row:
"David Glasner is explicitly using "endogenous" in a different sense than normal. I agree with David there, once we take that into account.
I disagree with David Beckworth on this bit: " These policy changes in target interest rates mean exogenous changes in the monetary base." (Unless he too is using "endogenous" in a different sense that normal.)
I don't disagree with anything Scott actually says there, but he maybe implies something about the long run I might disagree with."
Those comments from Nick Rowe really helped me! However, I'm still not on terra firma with the concepts of endogenous and exogenous. I bounced the idea off of Frances Coppola and David Beckworth that both are a matter of degrees (for example that we might be able to measure exogeneity on a scale of 0 to 1, and then endogeneity would thus be the complement of this: i.e. endogeneity = 1 - exogeneity. I also bounced that off of David Glasner. None of them laughed at the notion... and in fact provided some positive feedback... but thinking about it more, I'm not sure how useful that idea is. BTW, I even bounced a proposed way of conducting that measurement off of Glasner and I was again encouraged that he didn't laugh at me (and yes, he did respond). I'm still trying to come up with a good way to think about these concepts. Here are some of my unorganized thoughts along the lines of a "measure" in case anybody's interested:
Deletehttp://banking-discussion.blogspot.com/p/partial-endogeneity-and-exogeneity.html
I take a "dynamic systems" and "feedback controls" approach, since that's my background.
And here's my "note to self" page (unfortunately I can't link to anything other than the whole page, and it's a huge mess, but Rowe made the 1st comment):
Deletehttp://brown-blog-5.blogspot.com/p/links-to-remember.html?showComment=1398807202031
Oliver - I happened to see Victoria Chick yesterday, so I asked her what she thought of my characterisation of the PM proposal as trying to make the money supply exogenous. She agreed with this, but cautioned about being careful with the use of the word "exogenous" - along the lines of what I've said above.
Deletestone - I don't know. I'm afraid my knowledge of economic history is not very good.
Tom - Interesting. I think I'd agree with Nick's position on each of Glasner's, Beckworth's and Sumner's comments. More generally, "exogenous" and "endogenous" are really just features of models. You can ask how useful a model is, but there's no proper answer if you try to ask what is exogenous in the real world.
@ Stone (from the link above)
Delete‘[g]eology being limited, the [gold] stock above ground did not grow nearly as fast as the demand for liquid balances, and banks came into existence to supplement the supply’ Joan Robinson
@ Nick
I happened to see Victoria Chick yesterday...
Now that's what I call service!
Anyway, I reread my own link from above and can see how Chick can accommodate the PM proposal within her historical view, to the extent that it is accurately described.
I also understand your unease with the terms themselves, in that if they're taken all too literally, what starts as a modeling assumption tends to take on the quality of a natural law, i.e. something truly exogenous.
Nevertheless, since as you yourself say, the causal direction is probably best thought of as running from the level of activity to the money (means of payment) supply and not vice versa, any model that starts with the money supply first is quite literally putting the cart before the horse. And personally I find the PM proposal to be an extreme example of such reverse causation in that it seeks an absolute control over the volume of the means of payment. There's something authoritarian about it I very much disagree with and somehow my logic tells me it wouldn't work at all. I guess I'm closer to Joan Robinson than to Victoria Chick on this. Maybe next time you run into her... :-)
... Joan Robinson, I mean.
Delete@ Tom
DeleteI think the real dispute is about the effectiveness of monetary policy as such (not the endogeneity of the money supply for a given interest rate) the most extreme position being that it is completely ineffective. If the supply curve for money is flat, the CB can follow all the reaction functions it wants (or not), it won't make a (positive) difference. The CB rate becomes an arbitrary choice - it is exogenous to the system.
Oliver,
Delete"the CB can follow all the reaction functions it wants (or not)"
What do you mean by "reaction functions?" Do you mean like a Taylor rule?
Nick Rowe had a post a while back entitled: "The sense in which the stock of money is "supply-determined"" Did you see that one? He set up a simple example with a CB which did all the banking in an economy and which targeted an exogenously fixed rate of interest. No deposits: everything was cash, and all borrowers were creditworthy and thus would not be turned down for loans. What he wanted to show was that the demand for money did not determine the quantity supplied. He argued that the quantity of money supplied was determined by two things: the borrowers and the CB's money supply curve. Elsewhere (in comments to a previous post) he described the borrowers as the suppliers of loans which the CB buys. I like that description, but when another commenter asked him "two supply curves?... was that a typo" he responded that it was not, that the two supply curves interact to determine the stock of money.
I think that's confusing and prefer to think of the "money supply curve" as actually a "loan demand curve" for an exogenously fixed rate of interest, which is determined by the CB. This interacts with a "supply of loans" curve determined by the borrowers. I write up the whole thing here:
http://banking-discussion.blogspot.com/2014/03/nick-rowes-example-from-sense-in-which.html
And rather than "supply and demand for loans" it's even more clear to me to think of it as "supply and demand for loan principal dollars": with the borrowers determining the supply curve (which they have to option to buy back at par from the bank) and the bank determining the demand curve. Also, borrowers and deposit holders are not the same thing. That's the way Mishkin describes it in his text book on money too, saying four entities determine the supply of money: borrowers, deposit holders, banks and the CB.
What's funny is that Scott Sumner read Nick Rowe's post and declared in his own post:
Delete"Nick Rowe has a new post that argues the money supply is fully exogenous, even when the central bank is targeting interest rates."
http://www.themoneyillusion.com/?p=26400
I get into a long discussion with Sadowski saying that Rowe is NOT saying that, he's saying that borrowers and the bank "together" and "in conjunction" BOTH determine the supply of money. Sadowski resists me on this.... but eventually I figure out that he's just pulling my chain. However when I argue that to Scott... eventually Rowe comes over and says
"Scott: I think you understand me right."
http://www.themoneyillusion.com/?p=26400#comment-325167
Lol... Frances Coppola made the comment that Scott really has a big picture view, and that's why he said that. Sadowski liked her comment, and reposted it:
“Typically, Sumner simply avoids the problem by ignoring the endogeneity and focusing on the exogenous “envelope” which constrains endogenous monetary base creation and interest rates. He’s very “macro”….sometimes I wonder if he realizes that woods are made of trees.”
http://www.themoneyillusion.com/?p=26468#comment-326468
@ Tom
DeleteI don't follow Nick Rowe's blog (any more), nor any of the other M&Ms. Not my cup of tea, I've found.
Oliver, no prob. I only mentioned all that because, on the surface anyway, it seemed to match this statement from you, including this part "If the supply curve for money is flat"... which is what I drew in my post.
DeleteTo my mind separating away the payments system from the lending system would prevent the payments system from being held as a hostage for the banks to demand bailouts. The politicians in 2008 justified the bailouts on the premise that we couldn't have the payment system fail. I think we need a payment system that is just like M-Pesa or PayPal with no connection to lenders and 100% backed by bank reserves -just a very low cost utility. I thought the positive money proposal was a similar idea to that.
ReplyDeleteI also think the lending system should be funded with liabilities with a maturity equal to the term of the loans made so that there is never any risk of a run and also so that funding is priced appropriately so that equity finance can compete on an equal footing with debt finance. Again I thought that positive money wanted the savings accounts to have a maturity match to the loans such that as a savings account came free to withdraw, the loan it matched was paid off.
Where I had trouble with the positive money idea is that I thought it didn't make sense to ban a secondary market in debt liabilities. Positive money claim that by making savings accounts illiquid they would prevent debt fueled asset bubbles. But I think people would borrow money on the basis that they soon had a savings account maturing.
Is it true that Wonga is 100% funded with capital?
I think I've probably linked this before http://directeconomicdemocracy.wordpress.com/2013/04/09/what-full-reserve-banking-could-and-couldnt-achieve/
To my mind it doesn't really matter how much of the credit risk gets taken on board by those holding debt securities and how much gets taken on by the share holders of the lending institution just so long as there are never any bailouts and it is clear that there will never be any bailouts AND the debt securities have the same time to maturity as the loans made. So a lender could fund the loans by selling junk bond style debt and have very little capital and those holding the bonds would know that if the loans did not get repaid, then they would suffer a default. OR a lender could have lots of capital and fund the loans with high quality debt that had very little credit risk. OR a lender could be like Wonga and be purely capital financed, only lending out retained earnings and the starting capital.
DeleteI agree that there would seem to be significant advantages in separating the payment system from the intermediation system, at least as it operates now. Interestingly, this is one aspect where the approaches of Cochrane and PM differ. PM situate their payments system within the large stock of public debt, so it is not backed by private sector obligations. Cochrane is happy to use private assets (well he would, wouldn't he, as he's not much of a fan of public sector debt), but to allow the value of such assets to vary.
DeleteI don't think PM's proposal involves match funding assets, but I may be wrong.
I would be very surprised if Wonga was 100% equity funded - I would expect it to have bank debt.
Nick, Robert Peston has something about Wonga:
Deletehttp://www.bbc.co.uk/news/business-23078746
"In many ways Wonga.com is an impressive, even admirable business (and please resist your temptation to send me hate mail - I am feeling delicate).
It is, for example, funded exclusively with equity capital, or £100m genuinely at risk of being lost if things go wrong.
So, unlike a bank, it has no depositors or creditors who can pull their money out in a panic and bankrupt it."
You're right. I just had a look at their 2012 accounts. Apparently, they do have a revolving credit facility, but it's undrawn.
DeleteI'd assumed they would have debt just for tax reasons, because of the deductibility of interest. On reflection though, I can see that whilst this makes sense for most financial institutions, it applies less to Wonga. For most lenders, interest will represent a significant part of their expense, so funding with non-deductible equity is costly. Wonga's main expense, by a long shot, is loan losses - an impairment charge of £126m on year end net current assets of £188m. If you've got a write of rate of 67% pa, then deducting a few percentage points of interest is neither here nor there.
Obviously for normal lenders, with lower loss rates, that doesn't apply.
Nick, I just checked the positive money site again
Deletehttp://www.positivemoney.org/wp-content/uploads/2013/04/The-Positive-Money-Proposal-2nd-April-2013.pdf
-and you are right, they don't say that the time to maturity of the "investment accounts" as they call them would need to match that of the loans. And also the "investment accounts" would take some of the risk of losses from the outset rather than the shareholders being wiped out first. To be honest I think the lack of maturity matching is a really bad mistake on their part because if people simply chose not to lend, then there would be big losses that would be born by the current crop of "investment account" holders who had no control over what the bank was doing to recruit the new lenders to rollover the funding. That would be a vicious circle with no one wanting to be the odd one out, opening an investment account at a troubled lender and being exposed to all of the rollover risk.
Funding circle does do exact maturity matching now and is an up and running lending system. But again, then account holders bear the credit risk of the loans. I think it needs to be the professionals who are assessing the borrowers that stand first in line with capital to absorb losses from loans that don't get repaid.
It's sounds to me like maybe you'd have more time for Cochrane's proposal than PM's.
DeleteNick, you're totally right, I'm just reading the Cochrane link you gave now and much of it seems very much along the lines of the fix that I thought was needed. To my mind though, the key point in Cochrane's piece was not that in principle people could hold current accounts in the form of variably priced SP500 funds but rather that maturity transformation should be eliminated. Cochrane does say that people could use current accounts that were 100% backed by bank reserves or by short term treasury bills and so had a fixed value.
DeleteI guess the thing that Cochrane doesn't seem to think matters is having the payments system mixed up with the lending system or the financial system in general. I really think a total institutional separation is needed because the lending system is such a crucial utility for the entire real economy. I think the payment system needs to be utterly robust and isolated from any likely source of "financial innovation". It needs to be something that "finance people" have no more interest in than say telecoms or sewerage provision.
"To my mind though, the key point in Cochrane's piece was not that in principle people could hold current accounts in the form of variably priced SP500 funds but rather that maturity transformation should be eliminated."
DeleteI agree. I only commented on that bit because it interested me, not because I thought it was his main point.
Nick, you ever look at the accounting in this?
ReplyDeletehttp://www.ucl.ac.uk/~uctpa36/3equation_2005_withtitle.pdf
It's basically what Simon Wren-Lewis says he teaches. I would expect it to be correct. BTW, I think it's funny that econ types apparently regularly use lower case Greek letter pi as a variable. I wonder why that is?... is it because the concept of natural modes and bandwidths (having frequencies expressed in radians / time unit) just never comes up in econ? I'd think any discipline exploring auto-regressive type models and the evolution of mathematically defined, and feedback laden systems through time would find the concept of modes and bandwidths useful. Maybe you do and I'm just not seeing it when I skim these articles. Nevertheless, I find it a bit unsettling to see pi with a subscript. Lol.
Is there any accounting in that?
DeleteEconomists use all sorts of greek letters. There's not enough roman letters for all the variables and parameters that need labeling.
"accounting" ... probably not. Have you seen it? Is it SFC? (It's fair to ask that I suppose).
DeleteYeah, a lack of greek letters can be be problem. I've even seen people turn to Cyrillic.
Hi Nick,
ReplyDeleteUnfortunately, the Positive Money system you describe is not really what PM are advocating. Rather than reserves backing ‘Transaction Account’ deposits, so that account holders hold claims on their bank, which are back with reserves held at the central bank, instead Transaction Account holders would instead directly own central bank liabilities themselves. From the individual’s perspective they would still bank with HSBC, RBS or whoever, however, the legal relationship would be fundamentally changed – banks would for all intents and purposes simply be administering individuals’ accounts at the central bank, rather than “borrowing” from the customer in question. The change is therefore not regulatory in nature (an increase in reserve requirements) but rather legal (restricting the types of liabilities “banks” can issue). Essentially, banks will only be able to issue liabilities which have maturities above a certain duration (we went for 28 days, but it could be anything over a few days really). I have added a SFC (but not in the Godley sense) PowerPoint presentation here (http://www.positivemoney.org/2014/05/positive-money-proposal-balance-sheets/) which should helpfully make things clear.
As well as being important from a legal standpoint, the difference has potentially important economic effects. Namely, there is some doubt as to whether 100% reserve systems can adequately constrain bank behaviour with regard to lending and the simultaneous creation of deposits. Essentially, the argument is that 100% reserve requirements will not constrain bank behaviour, for the same reason 10% reserve requirements do not. That is, in the words of Alan Holmes, "in the real world, banks extend credit, creating deposits in the process, and look for the reserves later.” In a 100% system, as in a 10% one, if banks lend in excess of their reserves, and they can’t obtain the reserves they need for 100% backing from other banks, then their only other obtain is to get them from the central bank. In this situation the central bank could be forced to provide reserves to the bank in question, on demand, in order to maintain the 100% ratio. Now of course sanctions and the like may be forthcoming from the authorities, but if the banks decided en masse to ignore the reserve requirements (not impossible to imagine in a country like the UK where 5 banks make up 85% of the market) there may be little the central bank could do.
Contrast this with a system in which essentially everyone banks with the central bank, and non-central bank liabilities are unable to function as money. Banks cannot (without breaking the law) issue new on demand liabilities. And even if they did, they would very quickly run into liquidity problems they would not be able to deal with. For example, imagine a bank illegally makes a loan and creates a deposit, and the regulators are asleep at the wheel. The bank then pays away this money to wherever the customer wants. This would require a transfer from the bank’s own account at the central bank to the receiving customers account at the central bank. In the current system this money would be transferred to one of the other banks reserve accounts at the central bank, and could be instantly borrowed back by the original bank to make more payments.
However, in the PM system this is not possible, as rather than entering possibly 1 of 40 odd bank accounts at the central bank, it can enters any one of 20 million plus accounts at the central bank, which the other banks have no ownership or control over. Consequently, the bank cannot easily borrow back central bank money that they use to make payments to other banks, and will very quickly run out of money in their own accounts and face liquidity problems.
Andrew
ReplyDeleteThanks for the clarification.
I think I'd understood what you describe. I had been paraphrasing a bit, in the interest of brevity, and I had in mind a position where banks held reserves in excess of the 100% required against deposits, but where the 100% was a strict minimum. This would mean they could only create new deposits where they already held an excess. I think this is economically equivalent to what you describe, if not legally, but maybe it would be easier to game as you describe.
I appreciate though that it is a common misperception that proposed 100% reserve schemes would operate just like 10% reserve schemes with reserves provided on demand. My brief paraphrased description probably didn't help this, but hopefully my clarification in my replay to ATR's comment helped a bit.
Anyway, it's good to have the more precise description on here, so thanks again.
If the central bank has 20 million accounts, is the central bank going to end up being lumbered with in effect running the payment system? I don't know about this at all. Is it a trivial matter to administer 20 million accounts or are we asking a lot of the central bank? Basically nationalising a major utility? When telecoms were nationalised they were run in a shoddy manner.
DeleteCan some analogy be drawn between pooled nominee accounts (called street name accounts in USA I think) for shares and bonds and 100% reserve backed transaction accounts? When someone has a nominee account with a stock broker, isn't it true that the stock broker must actually hold the shares? And yet the central securities depository only has an entry for the stockbroker not for every client that every stockbroker has. Is that system really fragile? If not, then why couldn't 100% reserve backed transaction accounts work in the same way except with money in place of stocks and bonds?
DeleteI think the idea is that accounts would be managed by commercial banks, even if they were legally with the central bank, and off-balance sheet for the commercial bank. You can think of it as each commercial bank having a customer account with the central bank, and managing customer sub-accounts in that main account. That's pretty similar to what you describe.
DeleteThanks Nick.
DeleteAs you say, given similar policy choices by the authorities, there is little difference between a 100% reserve system and the Sovereign Money system advocated by Positive Money (assuming away the problems in regulation etc.) I wrote a paper (well some bullet points really) outlining the key differences for Michael Kumhof at one point, I’ll see if I can find it if you are interested.
Regarding the difficulty in running the accounts, you could do something similar to pooled nominee accounts, or what you said Nick (we have also advocated this method in the past, although the account holder has legal ownership). Personally now I’m of the opinion individuals should each have a separate account at the central bank. All that is required of the central bank is the provision of payment services. And this is essentially just reducing the balance of one account and increasing the balance of another (in a spreadsheet), upon receipt of an instruction from a bank. Which of course is effectively what banks do now when you make a payment to someone at the same bank. Consequently the level of computing power required is unlikely to be large.
I don’t think the analogy with telecoms is valid, as the provision of the utility is basically just running a scorekeeping system. And the central bank already does this for banks reserve accounts through the RTGS processor. Of course, banks will still be able to innovate in the ways in which they can present payment services to the public (e.g. contactless payment) but the behind the scenes stuff (settlement) will be handled by the central bank. The actual running of the payments system (in terms of the messaging etc.) could be public or private. Last time I checked the BoE we thinking of taking a far more prominent role in its administration - it is such a crucial utility its failure would be catastrophic. In Greenspan’s autobiography he actually says that if terrorists wanted to attack a country the most effective way would be to go after the payments system!
Thanks for the reassurance. I simply didn't know whether or not it would amount to a big new job for the central bank.
DeleteThe other thing that puzzles me is why funding loans using liabilities over 28days is thought to constrain bank behavior so much more than funding loans using deposits. I understand the principle that matching duration of liabilities to that of loans makes a lender robust against bank runs. But as far as I can see funding thirty year mortgages with 28day liabilities would still be very prone to a bank run type event in much the same way as funding with deposits is. And I don't see why a credit fueled asset bubble couldn't be blown even using very long duration liabilities to fund the credit. You say that bank behavior would become constrained, but in what manner would that constraint become manifest? Imagine banks collectively go on a lending binge and a vast number of 28day liabilities get sold and some asset bubble gets madly inflated. What slows that down?
I guess what I'm saying is that it is all very well saying that the government will have control over the amount of money in transaction accounts rather than it being a result of bank lending. BUT if the banks go on a lending binge and the amount of 28day liabilities soars and we have inflation such that there is no longer enough government money to avoid transactional frictions in the payment system -Then presumably there will be sharply spiking interest rates for the 28day liabilities? Won't the central bank then simply provide more government money as a way to relieve those sharply spiking interest rates?
DeleteIt really doesn't seem that more of an obvious constraint on the banks than we traditionally have had. In principle the central bank could always have dug in its heels and refused to accommodate increased demand for bank reserves but much of the point of the central bank was to prevent such "tight money" financial crises.
Please correct me if I'm in a muddle over this.
I'm not sure if you're asking me or Andrew.
DeleteMy only observation was that the proposal was intended to slow the pace of credit creation, rather than create a limit on it. This is effectively because the banking sector as a whole is constrained to lending only out of reserves held for own account, and every time a loan is made it reduces that balance. So the ability to keep making loans depends on how fast those reserves flow back - by being deposited into term accounts.
How effective this would be, I'm not sure.
Thanks Nick, I was asking anyone who could explain it to me!
DeleteSo basically the break on the extent of lending is due to the populous leaving money lingering in transaction accounts rather than shifting it into the term accounts. But basically that boils down to something like the "monetary base control" idea? It sounds like what Margaret Thatcher was imploring the Bank of England to do when she first came to power. Monetary base control was abandoned because it amounted to very high and volatile short term interest rates and that wasn't what was hoped for.
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/220868/foi_monetarycontrolgreenpaper_9.pdf
Under the PM system, when the lenders need to roll over a tranche of 28day term accounts and customers are being tardy about shifting money from transaction accounts into term accounts, the lenders will have to offer very high interest rates in a scrabble to not have a funding shortfall for the refinancing.
It seems to me that the way to keep lending at a level that serves the real economy (rather than damaging it) is to have debt appropriately priced (by eliminating maturity transformation and keeping all loans on the lending company's books) and to have tax shifted to being an asset tax.
On the face of it, interest rate volatilty would be an important issue to consider. I confess I don't know enough about the proposal to know how that is proposed to be addressed.
DeleteNick, interest rate volatility, would suggest changes in the demand for funds from or supply of funds into Investment Accounts. Hence, volatile interest rate would need to be offset by the central bank lending to banks, so they could on lend the money.
DeleteSorry, posted this in the wrong place before: To answer your first post (stone3 May 2014 07:15) - Yes, the point of the reform proposals was to stop banks creating money in the process of making loans. If banks could issue demand deposits, these could be transferred between people, and the banks would be able to create a substitute for money. Of course, there could still be runs on deposits of 28 day duration. If this happened banks would probably push up interest rates. If this wasn't enough to stem the flow one of two options would be available, either temporarily suspend convertibility of time (Investment Accounts) into sight deposits (Transaction Accounts), or the central bank could lend to the bank in question (but only if it thought the underlying investments were solvent). This would have to be short term, - i.e. on the condition the bank would secure funding quickly. If not, the regulator would have to put the bank (or at least that part of it) into insolvency proceedings.
DeleteTo answer your second post, under the proposals banks cant lend unless they already have cash. So they couldn't go on a lending binge and then force the central bank to lend to them. See the second and third paragraphs of my post above (2 May @ 10:23) for why. However, problems could occur if lots of people tried to cash out of their time deposits at the same time, as I mention in the previous paragraph.
As to your third post, regarding monetarism, I think it is very different! Monetarism attempted to control bank money creation through control of base money. This was impossible - the causality was all the wrong way round for one thing: loans create deposits. However, the PM system is very different - there are no longer two circulating forms of money (base and broad) with banks monopolising deposit base money (i.e. central bank reserves). Instead there is one integrated quantity of base money used by banks and non banks alike - money only exists on the balance sheet of the central bank. And because banks no longer monopolise the means of payment, they cannot lend, create money in the process, and obtain the reserves later, or force the central bank to acquisition to their needs as they did in the 80s. As soon as money is removed from banks balance sheets, and the payment system is no longer being held hostage, banks can be allowed to fail at no macroeconomic cost. Hence the central bank can actually refuse requests from banks, etc. Obviously we would want to avoid high interest rates though, at least on lending into the real economy...
Andrew Jackson, thanks for your explanation.
DeleteI'm still left scratching my head though because I thought that the 2008 crisis was driven by credit expansion (and a subsequent run) primarily in the shadow banking system. Doesn't that fly in the face of your assurance that things would be different if lenders were not allowed to fund lending by creating bank deposits? Shadow banks are not allowed fund lending by creating bank deposits. In effect your proposed lending system seems very much like the shadow banking system that fueled the credit boom and subsequent collapse in 2008.
The idea that whenever clouds gathered in the economy, savings accounts would be frozen, seems to me a recipe for simply making runs even more of an accident waiting to happen. People would all run from savings accounts whenever there was a whiff of fear that an account freeze was on the cards.
I also think that saying that the central bank would supply more reserves as and when they are needed to relieve interest rates really amounts to acknowledging that lenders still will determine how much money there is. The money may be all created by the central bank but it will be the actions of the lenders that determine when and how much is created.
No worries! To confirm, the proposal is not meant to completely stop bubbles from happening - people will always be susceptible to irrational periods of exuberance etc. The point is to make a banking system in which banks can be allowed to fail without pulling down the rest of the economy with them. And the proposal will also probably help with asset bubbles, by limiting credit expansion to an extent, as well as by aligning risk and reward so that for banks it's not "heads we win, tails the taxpayer bails us out."
DeleteFair point on the freezing of savings accounts. Perhaps a better option would be to not allow people to cash out until someone else puts in, as standard. This is how some of the peer to peer lenders work, I believe (and the stock and bond markets, but prices vary). Anyway these are just ideas...
Regarding the point about lenders determining reserves in the PM system, in the main it will be unanticipated withdrawals of money from Investment Accounts that will lead to increasing interest rates. And as i put above the lending would be on a temporary - on a 'get more funding within a month or two, sell off your loan book, or prepare to be shut down' type of basis.
Re the shadow banking system, as well as the bit above I wrote something replying to Krugman's blog on this point:"Krugman rightly identifies that there weren’t many runs on traditional banks. This is true, for the simple reason that traditional bank deposits are insured (i.e. guaranteed) by government, whereas funds with other types of financial company are not.
Perhaps more importantly, however, Krugman misses the way in which the actions of the traditional banking sector led to the run on shadow banking assets. The increase in house prices from the late 90s to 2007 were in large part the result of traditional banks lending huge amounts of money into the housing market. This lending increased both house prices and the amount of private debt in the economy, without increasing incomes. As private debt can’t continue to increase in excess of income forever, eventually things came to a head. In 2007-08 individuals started to default on their loans, resulting in the financial crisis. However, by this time the traditional banks had sold many of their mortgage assets to the shadow banks. Hence many of the problems of the 2007-08 crisis manifested themselves in the shadow banking sector. Yet it is important to remember that the housing bubble probably could never have happened without the private banks lending in the first place."
I'm totally in agreement with you on the benefits of separating the lending system from the payment system.
DeleteI'm still puzzled on why you don't favor the lending system being funded with maturity matched liabilities. So thirty year loans being funded with thirty year bonds and so on. I had thought that that would be the way to make everything run-proof and the Cochrane link in Nick's post said the same. Basically savers could keep savings in an ETF of such bonds and sell whenever they wanted spending money. An ETF of thirty year bonds would, as you say, have a fluctuating value but there could be ETFs of shorter duration too that would have more stable price. My impression was that Positive Money considered it a great virtue that the savings accounts were illiquid unlike ETFs that can be sold at any point. If you are saying the main aim is to avoid financial fragility and acknowledge that it is messy to try and constrain lending behavior by constraining liquidity, then having maturity matched tradable liabilities seems the way to go IMO.
Does it boil down to whether or not people will have "better" spending and borrowing behavior if they are holding an illiquid savings account versus holding savings in an ETF of bonds?
DeleteMy guess is that no "bubble dampening" advantage would be gained by insisting that savings accounts could not be traded. But I'm happy to be persuaded otherwise.
If the idea is that some restraint on lending is exerted by insisting that savings accounts are illiquid, then I guess the only mechanism by which such an effect could occur would be by way of causing savers to keep over more of their savings as "ready cash" in the transaction accounts.
DeleteI really wonder whether that would actually help. My worry would be that in slump periods people might keep lots of money in transaction accounts but in booms they would shovel everything into the savings accounts and if they wanted the cash due to an unforeseen expense, they would simply borrow themselves, reassured that they were soon due to get a maturing savings account to pay off the loan with.
Yes, you could do it through bonds, if you wanted, it certainly has some benefits. I don't think this necessarily solves all the problems though. One of the concerns was that these liabilities could start to function like money, as the Money Market Mutual fund shares did in the US, for example. And then the distinction between risky investments and money gets blurred, and the the government might end up getting sucked into guaranteeing them, as they did in the US. This is why we went for making the accounts illiquid and non transferable. But the system is equally viable too.
DeleteIn response to your last point, presumably the interest rates on borrowing (i.e. through an overdraft) would be larger than the ones on Investment Accounts (i.e. the bank needs to make a spread). So it would not make sense to do what you suggest.
I totally agree that those liabilities would function somewhat like money. The crucial line to draw IMO is that it needs to be made utterly clear that their price could fall. So long as everyone takes that on board, everything is robust and there will never be any need for bailouts. Money Market Mutual funds had an ethos that they "would never break the buck". Bond ETFs are something the public is used to and the public accepts that they do often fall in value.
DeleteI'm not saying that anyone would plan to borrow simultaneously with holding money in an illiquid savings account. I'm simply saying that people would put all of their savings into an illiquid savings account and then, if some unforeseen expense cropped up, they would borrow to tide themselves over.
I suppose it all boils down to whether the aim is simply to eliminate the fragility that comes from run prone bank liabilities OR the aim is to prevent monetary expansion from being driven by lending behavior.
DeleteIf the aim is simply to eliminate the fragility that comes from run prone bank liabilities, then insisting on maturity matching of loans and liabilities is necessary and sufficient.
If the aim is to prevent monetary expansion from being driven by lending behavior, then that is a whole other kettle of fish. I wonder whether it is even realistically possible. I totally agree that it is a good idea to prevent credit fueled asset bubbles. Policies that could help would be insisting that all loans are kept on the books of the lender who originated the loan and also shifting taxes to being a tax on gross asset value. Monetary expansion would still then be a passive consequence of lending behavior BUT lending behavior would not be deranged.
To answer your first post (stone3 May 2014 07:15) - Yes, the point of the reform proposals was to stop banks creating money in the process of making loans. If banks could issue demand deposits, these could be transferred between people, and the banks would be able to create a substitute for money. Of course, there could still be runs on deposits of 28 day duration. If this happened banks would probably push up interest rates. If this wasn't enough to stem the flow one of two options would be available, either temporarily suspend convertibility of time (Investment Accounts) into sight deposits (Transaction Accounts), or the central bank could lend to the bank in question (but only if it thought the underlying investments were solvent). This would have to be short term, - i.e. on the condition the bank would secure funding quickly. If not, the regulator would have to put the bank (or at least that part of it) into insolvency proceedings.
ReplyDeleteTo answer your second post, under the proposals banks cant lend unless they already have cash. So they couldn't go on a lending binge and then force the central bank to lend to them. See the second and third paragraphs of my post above (2 May @ 10:23) for why. However, problems could occur if lots of people tried to cash out of their time deposits at the same time, as I mention in the previous paragraph.
As to your third post, regarding monetarism, I think it is very different! Monetarism attempted to control bank money creation through control of base money. This was impossible - the causality was all the wrong way round for one thing: loans create deposits. However, the PM system is very different - there are no longer two circulating forms of money (base and broad) with banks monopolising deposit base money (i.e. central bank reserves). Instead there is one integrated quantity of base money used by banks and non banks alike - money only exists on the balance sheet of the central bank. And because banks no longer monopolise the means of payment, they cannot lend, create money in the process, and obtain the reserves later, or force the central bank to acquisition to their needs as they did in the 80s. As soon as money is removed from banks balance sheets, and the payment system is no longer being held hostage, banks can be allowed to fail at no macroeconomic cost. Hence the central bank can actually refuse requests from banks, etc. Obviously we would want to avoid high interest rates though, at least on lending into the real economy...
The Socialist Anti-Semitic Myth of the Creation of Money out of Thin Air http://iakal.wordpress.com/2014/07/04/the-socialist-anti-semitic-myth-of-the-creation-of-money-out-of-thin-air/
ReplyDelete