Wednesday, August 21, 2013

A Bank is not a Financial Intermediary

As always, good discussion on Monetary Realism. There was an interesting piece in the comments where Ramanan linked to a great paper by Tobin which, I think, revealed a lot about the implicit mental models economists refer to when they do their work.

Some additional context: Cullen has been trying to convince Paul Krugman to embrace MMT or PK or whatever you want to call it, unfortunately with limited success to date. He's certainly picked the right target, if Krugman publicly converts then it's a real game changer, but I don't think it's going to happen (why I think this is another story for another time).

Krugman dismissed Cullen by citing Diamond-Dybvig and Tobin-Brainard, the former I'm very familiar with but the latter is new to me. I won't get into the meat of Krugman's dismissal because it's of the form: either this doesn't fit into the model or it's a trivial case the model already covers, it does not grapple with the question at hand because he does not need to. But the papers themselves are interesting for revealing the underlying assumptions.

I'll focus on Tobin because that was the one that was new to me, and zanon, in comments, highlighted a point that I thought was very interesting. Tobin repeatedly refers to banks as "financial intermediaries." What is a "financial intermediary"? Tobin explains:
“…the essential function of financial intermediaries, including commercial banks, is to satisfy simultaneously the portfolio preferences of two types of individuals or firms. On one side are borrowers, who wish to expand their holdings of real assets… On the other side are lenders who wish to hold part or all of their net worth in assets of stable money value with negligible risk of default.”

“…intermediation permits borrowers who wish to expand their investments in real assets to be accommodated at lower rates and easier terms than if they had to borrow directly from the lenders.”
So while Tobin understands that banks make loans which then create a deposit, he sees this function as a market making or coordination activity, something which brings efficiency and eases friction between the actual lender and borrower. It's a middleman role which, if we assume is acting appropriately, one can safely abstract out of models to focus on the primary agents in the exchange, the lender and the borrower. Also, a Ramanan correctly notes, when we look at the composition of national accounts, financial firms similarly play an intermediate role where they sit between initial production at the top, and ultimate consumption at (for example) the household level at the bottom.

But it is precisely this characterization of a bank's role that I think is the problem with modern macro. By seeing them as an essentially coordinating function they are not modeled as agents in and of themselves, and therefore the models are wrong because banks are ultimately active lending agents not lending mediators.

Straight from Google:
Intermediary
A person who acts as a link between people in order to try to bring about an agreement or reconciliation; a mediator.
"intermediaries between lenders and borrowers"


Agent
1: one that acts or exerts power

2a : something that produces or is capable of producing an effect : an active or efficient cause
It's awesome that Google itself has "bank" under their definition of Intermediary.

So, to engage with someone who is taking the standard econ line, you need to approach them at the model level. Both papers Krugman cited are really old, so if they aren't canonical, they are certainly the intellectual touch stone that the profession draws on. So go directly to the source and attack the hidden assumption. If you are successful there, then subsequent papers all become vulnerable because you've undermined the foundation.

Because banks don't occupy a "middle or intermediate" position between people who want to borrow and those who want to lend. Note how careful Tobin is in how he describes what the lender and borrower are doing:
"On one side are borrowers, who wish to expland their holdings of real assets… On the other side are lenders who wish to hold part or all of their net worth in assets of stable money value with negligible risk of default.
Emphasis mine. Does "wish to hold part or all of their net worth in assets of stable money value with negligible risk of default" sound like a lender to you or does it sound like a saver?

So banks do play a middle man or intermediary role, but it is between savers and borrowers and it's primarily one of payment settlement. The lender is the bank itself, and there it's acting on behalf of its shareholders, not its depositors, and you need to look at return on capital and capital constraints.

This then is the conceptual fallacy at the heart of academic macro and what it thinks about banks, and it goes at least all the way back to 1963.

38 Comments:

Blogger Ramanan said...

He wait a sec ...In fact it is quite the opposite. There were the good ideas in the 1960s/70s which have been confused over the years by others.

There existed good ideas - primarily due to Tobin and this is connected to his asset allocation theory where banks mediate the role of portfolio allocation of the non-bank private sector.

That is different from the confused understanding of people writing on banks which I believe followed work such as his about what banks really are.

Somehow to correct the (latter) confused understanding which follows - PKEist tend to say banks aren't "intermediaries".

[There is of course another reason why the phrase intermediary is used - in SNA the main thing is production and in that sense banks are intermediaries to the production process.]

The whole issue is connected with the question - what process does lead to money demand = money supply.

In the simple case consider the case where the households wishes to hold its financial wealth in 50% cash, 30% bonds and 20% equities. On the other hand, there are production firms who borrow from the bank and this creates money deposits. There is no reason for these two Ms to be equal. What processes happen to make them equal? There are many processes involved - one such being banks mediate the portfolio choice of households by holding assets which households do not wish to hold.

"This then is the conceptual fallacy at the heart of academic macro and what it thinks about banks, and it goes at least all the way back to 1963."

The fallacy is on the part of people who misunderstood the genius of James Tobin!

You have to write down stock-flow consistent models of the whole economy to see this and James Tobin himself latter brought stock-flow consistency plus asset allocation theory in economics to show such things.

In other words, you are taking a misunderstood thing and trying to prove it wrong than looking into what the original idea really is.

7:03 PM  
Blogger Ramanan said...

And note - sometimes intermediary is used for lenders who are not banks and who take money from "savers" and lend to "borrowers" and who do not have the privilege of "loans create deposits".

That intermediation role is a bit different from the intermediation role of banks highlighted in Tobin's work. And Tobin obviously knew "loans create deposits" and the money multiplier is wrong.

7:24 PM  
Blogger Ramanan said...

Just registered this:

"A person who acts as a link between people in order to try to bring about an agreement or reconciliation"

That exactly shows the phrase intermediary is right even though it is misleading in some other sense (which you are hooked onto).

And this has to do with the fact that not just loans create deposits, banks' asset purchases also creates deposits and reverse does the opposite.

In most countries except perhaps the United States, banks do take in government securities which the non-bank private sector doesn't hold. Even in the United States, historically banks have held more Treasuries as a percentage of various sizes/flows/stocks than is the case now. As per a US Treasury presentation (which I cannot find now), it is because of the rise in MBSs which they now hold more instead of Treasuries.

So in a simple world with Treasuries as the only bonds and firms borrowing directly from banks, banks will play the role of accommodating the shifts in the non-bank sector's portfolio allocation decisions.

So banks accommodate the shifts in the non-bank sector's demand for money. If the non-bank sector wishes to hold more deposits for government bonds, banks will hold more of other securities and vice versa. In this, it is playing an accommodating role.

This is a quantity adjustment process. There is of course a price adjustment process which is changing loan and deposit rates - different from above.

Relatedly, the Monetarist hot potato process is an attempt to address this theoretical problem but it confuses saving and asset allocation decisions.

Thirdly, there is the convenience lending. It supposes that the deposits held are held non-volitionally by the non-bank sector. If you deny the two (Monetarist is easy to dismiss), you are left with the view of convenience lending which isn't right macroeconomically.

11:41 PM  
Blogger Ralph Musgrave said...

Diamond-Dybvig’s paper is based on the idea that deposit insurance is the best thing since sliced bread. In contrast, I argue here that deposit insurance is nonsense:

http://ralphanomics.blogspot.co.uk/2013/08/government-sponsored-insurance-of-bank.html

Frances Coppola puts a similar argument here:

http://www.pieria.co.uk/articles/lender_beware

So, re you point about going “directly to the source and attack the hidden assumption. If you are successful there, then subsequent papers all become vulnerable because you've undermined the foundation.” . . . . perhaps Frances and I have successfully “undermined the foundations”????

11:51 PM  
Blogger Ralph Musgrave said...

Deposit insurance is also under attack here:

http://www.positivemoney.org/2013/08/the-cancers-at-the-heart-of-the-banking-system/

12:28 AM  
Blogger winterspeak said...

RAMANAN: The SNA term may be true, but it's irrelevant to this discussion. What we have is Paul Krugman dismissing MMT/PK, and citing two papers. In those papers, at least the Tobin one which is most relevant, it lays out a very specific definition of "asset allocation" and "financial intermediary". It is these definitions that Krugman is using as his touchstone, and he didn't pick these because he himself thinks they are the best, he picked them because it's what his profession uses. So, when trying to understand Paul's language and where he is coming from, we look at what he explicitly cites and sees what mental model is there. I'm sure other definitions are used elsewhere, but that's another topic.

While the Tobin paper may have insights about asset allocation and the role banks play to enable and coordinate that (by holding what non-banks do not want to hold), the role he describes them as playing is a coordination one between lenders (who are really savers) and borrowers. My point here isn't even that this is wrong. My point is that if I wish to formally model an economy, this understanding of a bank's role means I can ignore banking entirely. Thus Paul ignores banking entirely. Thus when Cullen tries to bring up some element of banking, Paul tosses out some old papers which have established, for him and the profession, that banking can be safely ignored.

RALPH: As deposits aren't used to support loans, really, I see no sense in getting rid of deposit insurance. It should be extended to 100% and put an end to the money market business which really is just a regulatory arb strategy.

8:39 AM  
Blogger Steve Roth said...

I've been thinking more and more along exactly these lines.

I just keep coming back to the basic incoherence of loanable funds, the obvious error of composition.

If I transfer $10K from my account to yours for whatever (spend), the banks don't have any more money to lend.

If I *don't* transfer $10K from my account to yours (save), the banks don't have any more money to lend.

In that sense of personal saving, at least, saving does not fund lending or investment.

In the sense of sectoral saving (IOW lending, i.e. real sector lending to the financial sector, for instance), it does make sense. But that's just saying that lending funds lending, which is not a profound or useful insight.

Yes: The "intermediary" formulation is problematic -- at least as it is so widely understood and deployed (including in formal models) -- because it strongly suggests that banks can be abstracted away as *mere* intermediaries between real-sector borrowers and lenders.

This gives a false picture of how economies work, and results in incoherent and misleading models, because unlike real-sector actors (households and nonfinancial businesses), banks:

1. Are not optimizing their intertemporal consumption preferences (do they even have such preferences?)

2. Create new money (deposits) for lending, and burn that money when the loan is paid off

Lenders' and borrowers' inherent incentives and reaction functions are not symmetrical. Butchers aren't lending to bakers, and you can't usefully model an economy as if they were.

I like where Ramanan is going with production -- savings are a result of the surplus from production, not anything to do with "not spending" -- but I don't really understand what he's saying.

9:52 AM  
Blogger Ramanan said...

Winterspeak,

Tobin cannot be said to be the source of Krugman's confusions.

I agree Krugman doesn't get it.

"this understanding of a bank's role means I can ignore banking entirely."

No WS, the quote you quote is perfectly fine from any PKE monetary economics angle. It isn't about loanable funds at all.

I don't know how to say that apart from previous comments. Maybe I should try to write something. The quote is nowhere close to a loanable funds model.

To put it crudely it is something like:

"I am a bank ... oh you don't want to hold the government bond and instead hold deposits, sell me the bond and I shall credit your bank account and create deposits for you"

It actually is a super-cousin of "loans create deposits"

11:06 AM  
Blogger Ramanan said...

Steve,

"but I don't really understand what he's saying."

That production thing was way-off the main points in the post. It was just to highlight a reason for another use of the phrase intermediary. It has nothing to do with the rest of my comments.

11:08 AM  
Blogger winterspeak said...

Ramanan: you're going to need to address the quote more directly. I assume you mean:

“…the essential function of financial intermediaries, including commercial banks, is to satisfy simultaneously the portfolio preferences of two types of individuals or firms. On one side are borrowers, who wish to expand their holdings of real assets… On the other side are lenders who wish to hold part or all of their net worth in assets of stable money value with negligible risk of default.”

There are three entities in this quote: a borrower, a lender, and an intermediary. Seems like loanable funds fundamentally to me.

Tobin seems pretty clear. I don't see where Govt bonds come into it.

2:03 PM  
Blogger Ramanan said...

Winterspeak,

Try to be a bit more open minded.

You are reading it wrong.

Perhaps ".. the essential function" is read by you as the most essential function.

It is clear in Tobin's second paper that loans make deposits etc. Yet you want to keep thinking he is a loanable funds guy.

Try changing "... the essential function..." to "... an essential function ..."

Does that make more sense to you?

I sort of understand the words should have been different and if I were to write it, I will present it differently. However, there is a good idea which is getting missed out because of being misled by wordings.

In the government bond example, the government is the debtor and the holder is the creditor. If the non-bank sector wishes to hold more deposits and reduce its holding of stock of government bonds, the banking system is prepared to accommodate this.

If you can't see with government bonds, replace it with some other bond.

In fact this mechanism is not loanable funds at all and another blow to the Monetarist hot potato mechanism.

6:43 AM  
Blogger Ramanan said...

The thing is loans create deposits is just the starting point.

The idea is related to the following question:

Suppose for some reason, borrowers borrow a lot from the bank. This creates deposits because loans create deposits. However, suppose wealth owners wish to reduce the fraction of their wealth kept as deposits. There is now a discrepancy. The banking system will then accommodate this and do this reconciliation perhaps by selling their holdings of existing bonds to the private sector who wish to buy bonds for money.

6:51 AM  
Blogger Ralph Musgrave said...

Winter,

There are very good reasons for getting rid of deposit insurance, which I set out here:

http://www.positivemoney.org/2013/08/deposit-insurance-is-a-farce/

Briefly: if the insurance premium is charged to depositors (which it should be) that will wipe out the extra interest they get from placing money in a commercial bank rather than in something near 100% safe, like a money market mutual fund that invests just in short term government debt.

Alternatively, if it’s the taxpayer that pays the premium, then that’s a subsidy of banking, and we trying to get rid of bank subsidies.

8:12 AM  
Blogger Detroit Dan said...

Winterspeak-- I've been letting your post sink in and it's making an impact. I think you've hit on a fundamental concept, as you say. Banks lend their own money, not the money of their depositors.

8:30 AM  
Blogger JKH said...

Tobin’s paper is fine.

Banks are intermediaries.

So are insurance companies.

Banks are special.

So are insurance companies.

It’s only logical that different elements in a classified group have things in common and things that are unique.

The primary role of banks is not necessarily payment settlement. You might think that way if you thought loans creates deposits describes banking. But banks have assets and liabilities with a non-trivial duration and risk mix. There is an intermediation effect. Banks manage that effect through centralized function such as asset-liability management, risk management, and capital management. And if people don’t know how those functions work, they should check their bearings before being too aggressive about calling mainstream economists ignorant about banking. There’s some truth to that, but there’s also a question of degree. And in that context, Tobin’s paper is fine.

Yes, return on capital is a priority.

But you can’t get there without providing a service to borrowers and depositors, obviously.

Everything gets priced.

Return on capital is a constraint on asset-liability pricing, but in a sensible way market competition is a constraint on return on capital expectations. You can’t make 20 per cent ROE on a business where the market is at 15 per cent ROE. Otherwise, you end up with no banking business if you think that way about everything.

“…the essential function of financial intermediaries, including commercial banks, is to satisfy simultaneously the portfolio preferences of two types of individuals or firms. On one side are borrowers, who wish to expand their holdings of real assets… On the other side are lenders who wish to hold part or all of their net worth in assets of stable money value with negligible risk of default.”

Nothing wrong with that.

Tobin is basically describing asset-liability management, which is an internal function performed by banks in fact.

Nothing there implies loanable funds.

Tobin fully understands loans create deposits and that the money multiplier is wrong. It’s right there in the paper.

The reason Tobin appears to refer to bank depositors as lenders is because they are lenders in the counterfactual without banks. They become depositors instead with banks. This is what he means and it is a non-issue.

Total confusion on the internet about this idea of intermediary. Like it’s some virus to be avoided.

Incredibly silly.

Its just more of moving definitions and vocabulary around from one corner of the room to another.

Ramanan is right that Tobin is not responsible for Krugman’s confusion. Nothing wrong with Tobin (or very little). Krugman is responsible for his own confusion.

4:53 PM  
Blogger NeilW said...

I think the main problem is not realising that the asset side and the liability side are asynchronous with each other. The connection is that that price the assets are sold at has to be higher than the price the liabilities are incurred at and the two sides adjust around that.

Usual problem of not understanding the business dynamics and relying heavily on static serialised analysis in a artificial model universe.

Krugman needs to go and work in a bank for a bit.

4:52 AM  
Blogger winterspeak said...

JKH/Ramanan: Let's revisit the original context.

Krugman clearly believes loanable funds. To support loanable funds, and dismiss MMT/PK he cites Diamond-Dybvig and Tobin-Brainard. I think the first point is pretty uncontroversial and the second is just a fact.

We're trying to figure out how to constructively engage with mainstream econ, and that means engaging with their models and literature.

So, if Tobin does not support loanable funds, but Krugman thinks he does, that's a promising angle of attack. On the other hand, Tobin does get the role of reserves wrong, so if that's the source of Krugman's error we can focus on that.

Regardless, Krugman cited Tobin, so a good rebuttal cites Tobin back.

2:13 PM  
Blogger Ramanan said...

"Tobin does get the role of reserves wrong,"

No money multiplier there.

Reserve requirements impose a cost on the banking system and consequently a higher interest rate on loans than otherwise. Depending on interest elasticity of borrowing on bank loans, bank borrowing is reduced a bit.

This effect needn't be high but nowhere in the article does Tobin say that it is so.

2:37 PM  
Blogger winterspeak said...

Ramanan: Don't let your feelings towards Tobin blind you to what he wrote. He's not infallible.

And I never said he believed in the money multiplier, but nonetheless, he got the role of reserves wrong. His model leaves you looking at the wrong margin.

3:36 PM  
Blogger Ramanan said...

Winterspeak,

I am a fan but I know his mistakes - he was a neoclassical economist. Whatever said his contribution to economics is tremendous and PKEists themselves have used it.

But what about the role of reserves - what role did he get wrong?



11:03 PM  
Blogger NeilW said...

"Reserve requirements impose a cost on the banking system and consequently a higher interest rate on loans than otherwise."

Do they? If the central bank is providing those reserves to demand, and those reserves don't impact the capital ratio, then the only cost to the bank is if they are remunerating deposits at a higher rate than they are earning on the reserves. I can't see why they'd do that.

It doesn't affect the amount of loans at all.

If capital is C and the loan limit is therefore CL, then without reserve requirements the asset side is essentially CL at maximum extent, whereas with reserve requirements the asset side is CL + R - where R is the reserve requirement.

The only difference is that there are slight less government bonds in circulation and slightly more reserves - which means that some actor in the private sector is getting less income.

Since bonds tend to be held by non-banks in the main, it is that non-bank entity that is getting less income because of the reserve requirement - not the bank. Instead they have the extra deposit with the banks that matches R.

12:42 AM  
Blogger Ramanan said...

"Do they?"

Yes they do.

Imagine reserve requirements being changed from 5% to 25%. Banks will first have to borrow this from the central bank and pay interest on it. Plus their collateral goes to the central bank and is tied up.

Later the central banks can provide these by doing open market operations with non-bank counterparties so that borrowed reserves come down. Even here, the reserves earn no interest and hence that is the case.

"those reserves don't impact the capital ratio,"

Yes true but irrelevant to the reserves discussion really.

2:59 AM  
Blogger Ramanan said...

"It doesn't affect the amount of loans at all."

Assuming the interest elasticity of borrowing is ZERO with no doubt whatsoever.

3:05 AM  
Blogger NeilW said...

". Banks will first have to borrow this from the central bank and pay interest on it. "

No they won't - because the central bank has committed to provide the reserves required by the banks and target an interest rate. That is in their framework policy documents.

So when the reserves are overdrawn due to interest in reserve requirements, the central bank will purchase the necessary government bonds from the system to inject the required reserves - since the overall amount required has gone up systemically due to the reserve changes.

Otherwise the central bank loses control of its policy rate rapidly on shortage of reserves - and the transaction system starts to gum up as clearing is delayed.

"Assuming the interest elasticity of borrowing is ZERO with no doubt whatsoever. "

Nope. Assuming none of that. The borrowing doesn't change. All that happens is that bank deposits exists that wouldn't exist in the other state - offset by the extra reserves.

What you are suggesting happens because you have implicitly changed the interest rate from the policy rate to the penalty rate by restricting the amount of reserves contrary to the published framework

Obviously a change in central bank policy affects loan amounts. Just changing the reserve requirements *and keeping the policy the same* does not - because the central bank has committed to providing sufficient reserves.

If you increase reserve requirements massively then you are effectively doing a form of QE which will bring down the yield on government bonds - and that may affect loan amounts. By making them cheaper and more available.

8:23 AM  
Blogger Ramanan said...

"No they won't - because the central bank has committed to provide the reserves required by the banks and target an interest rate. That is in their framework policy documents. "

Torture with words. Reserved borrowed to satisfy reserve requirements can be provided at the policy rate.

In fact a lot of countries have an overdraft banking system in which the banking system as a whole is indebted to the central bank - so in the aggregate reserves are borrowed.


"What you are suggesting happens because you have implicitly changed the interest rate from the policy rate to the penalty rate by restricting the amount of reserves contrary to the published framework "

Even in the case where central bank makes sure that nonborrowed reserves is minimal by doing open market operations with the non-banking sector, banks will have assets which do not earn anything. (Unless of course you assume that interest is paid on all reserves which is not true in general).

And that is effectively a cost to the banking system.

So banks do a markup. They target a profit and compared to the counterfactual in they have an asset earning less and hence they would raise loan rates by a little.

Your arguments is tantamount to saying reserve requirements have no effect at all. This is beyond silly especially when I started by saying the effect is less.

Go away. Stupid comments.

9:56 AM  
Blogger Ramanan said...

Plus consider the case around Nov/Dec 2011 when the Eurosystem/ECB reduced reserve requirements from 2% to 1%. It was a huge relief to the banking system.

10:01 AM  
Blogger Ramanan said...

http://moslereconomics.com/2010/01/13/china-raises-banks-reserve-requirement-ratio-50bps/

"All reserve requirements do is raise the cost of funds for the banks, not the quantity of funds they can loan."

From your own MMTers mouth by a random google search.

Of course it is true that doesn't change the quantity of the funds the can loan but once it is accepted reserve requirement is a cost it is a simple matter of markups to say it increases loan rates as well a bit.

Given higher loan rates, and a non-zero interest elasticity, it affects borrowing.

Also:

"The lesson should be changing reserve requirements for a non convertibility currency, as the yuan is domestically for all practical purposes, doesn’t alter liquidity, but does alter balance sheet composition and pricing necessary to hit return on equity targets."

http://moslereconomics.com/2010/10/12/comments-on-chinas-temporary-hike-of-reserve-requirements/

So who is right, Mosler: cost or Neil Wilson: no cost?

10:14 AM  
Blogger Ramanan said...

My argument was based on Moore's book Horizontalists and Verticalists which says (page 65):

"Contrary to conventional wisdom, changes in reserve requirements imposed by the central bank do not directly affect the volume of bank intermediation. A change of required reserve ratios influences the volume of bank intermediation only indirectly, by affecting the required reserve markup or spread. A rise (reduction) in reserve requirements raises (lowers) the cost of obtaining funds to place in loans financed via additional reservable deposits, in the manner of an indirect tax. The banks will therefore raise (lower) the markup of their lending rates over their borrowing rates. As a result, depending on the interest elasticity of demand for bank credit, the volume of bank intermediation will be indirectly reduced (increased)."

10:30 AM  
Blogger Steve Roth said...

So Neil, Ramanan, you're disagreeing over direct versus indirect effects. 'Nuf. Don't make me pull this car over!

10:37 AM  
Blogger winterspeak said...

LOL! Called to rights on whether reserve requirements are a cost! Nevertheless, I think the important question is whether they are costs on the margin, since the question about what reserves actually do comes up in the case of the marginal loan.

So here's the Tobin quote on reserves I have problems with:

"Without reserve requirements, expansion of credit and deposits by the commercial banking system would be limited by the availability of assets at yields sufficient to compensate banks for the cost of attracting and holding the corresponding deposits. In a regime of reserve requirements, the limits which they impose normally cuts the expansion short of the competitive equilibrium. ...Reserve requirements...are effective, the marginal yield of bank loans and investments exceeds the marginal cost of deposits to the banking system. In these circumstances additional reserves make it possible and profitable for banks to acquire additional earning assets. The expansion process lowers interest rates generally."

So, two problems -- one maybe stylistic and the other more substantive.

The stylistic one is "banks will borrow/take deposits at x% and then make loans at x+1%". Tobin may be saying nothing wrong, but boy does it obfuscate the key mechanism.

The substantive one is that reserve requirements cuts expansion short by acting as a ceiling, and when additional reserves are added, then banks can lend more, with this expansion lowering rates. So here, rates are determined by a competitive bidding process in the market which heats up with additional loans, which were capped by reserves but are now uncapped.

In reality, reserve requirements factor into average cost of capital, but the decision point is marginal cost of capital vs marginal return on credit risk. I think it's misleading to say that reserves are playing the role of a ceiling here, and I don't think that lower rates are a consequence of credit expansion, I think that credit expansion is a consequence of lower rates as the margin moves down.

2:52 PM  
Blogger Ramanan said...

Tobin is talking in the context of competition between banks and non-banks.

Loans create deposits in his paper and reserves can be borrowed from the central bank. Money multiplier is wrong too there.

"The substantive one is that reserve requirements cuts expansion short by acting as a ceiling,"

The ceiling refers to interest rate ceilings on deposits that was around the time.

11:58 PM  
Blogger NeilW said...

An awful lot of appealing to authority there again. You'll forgive me here but where is the actual cost?

(I accept the permanent overdraft model will impose reserve costs - since the banks are always paying the central bank a fee all the time - and I think that explains the Eurozone which uses that model IIRC).

"And that is effectively a cost to the banking system. "

Sorry, still don't see the net cost unless the central bank has changed its targeting policy - particularly for very large commercial banks.

There will be enough reserves in the system - because the central bank says there will be and buys bonds (ie forcibly creates bank deposits) to do so. Very large banks will have customers that are government bond holders. Therefore even obtaining collateral is easy for such a bank. It just buys the bond and that expands its own deposit base (ie forcibly create deposits) and gets a bond with which it can play collateral games.

Don't forget few current account deposits are remunerated either, and if there are more deposits and less bonds then there are more deposits chasing the income earning opportunities out there which ought to bid them down.

Banks buying bonds to use as collateral for reserves lowers the yield of bonds QE like. As does the central bank injecting reserves.

So I'm seeing the adding of the required reserves by the central bank lowering the yield curve because there are less bonds in the system and less income going around. Isn't that good for loans?

If reserve requirements impose costs on banks, then surely so does QE for the same reason. And I don't think we're seeing that.

So I would argue whether overall there is a marginal cost at all - particularly for large banks. If anything the impact of the policy change is distributional - favouring larger banks over smaller ones.

12:07 AM  
Blogger Ramanan said...

Appealing to your authority not mine.

A trick of trade on thinking on such matters: start with an overdraft banking system like the Circuitists do. The results then apply to asset-based banking system.

In an overdraft banking system, if the central bank increases reserve requirements, the size of its MRO will be higher in the next refinancing operation which will provide reserves at the policy rate.

The banking system's borrowed reserves rise and since they have to pay interest to the central bank, banks' costs rise.

Now think about an asset based system. Most systems are a mix of two but let's just think of the pure case.

If the central bank raises reserve requirements and does an open market operation with a non-bank counterparty, banks' borrowed reserves are low and it seems this point has been taken by you to think that there are no costs.

However, since the open market operation with the non-bank has created deposits, the bank with the higher deposits as a result will have to pay an interest rate to keep the depositor. This has resulted in a rise in cost for the bank.

And it proves Tobin's general point about Widow's Cruse. Implicitly you are committing a mistake of assuming Widow's Cruse.

3:11 AM  
Blogger NeilW said...

"However, since the open market operation with the non-bank has created deposits, the bank with the higher deposits as a result will have to pay an interest rate to keep the depositor."

Will it?

Because no other bank needs the depositor. All that has happened is the supply of depositors has gone up. So why will anybody bid them away when there are now more depositors chasing the same amount of overall income?

As I said the total amount of deposits has gone up and the yield on bonds has gone down.

I think you're missing that effect in your analysis. QE switched a lot of bonds for deposits and there has been no deposit 'bidding away' frenzy.

I only see a cost rise if the reserves are borrowed - and even then that cost is recycled via the remittance to the government and government spending (say interest payments on bonds then returns the amount as a deposits again). So I'm not even sure that remains for long. There is a timing difference though and that differential may be sufficient to generate an effect.

I see no overall effect unless there is a shortage of reserves and therefore an amount of money has to be remitted to the central bank as a fee. And even then it appears to be a small timing effect - as the money comes back into circulation via Treasury spending.

4:00 AM  
Blogger Ramanan said...

"Because no other bank needs the depositor."

What is this Widow's Cruse mania. First the deposit is created and it shifts out of the bank. It is in the interest of the banking system to keep the deposits.

Banks compete hard to attract deposits.

"So why will anybody bid them "

Widow's Cruse.

And note deposits from the point of view of depositors is determined by their preferences. So the equivalence is brought about by changes in prices and quantities in the financial markets (not talking of prices in goods and services market).

"I only see a cost rise if the reserves are borrowed - and even then that cost is recycled via the remittance to the government and government spending (say interest payments on bonds then returns the amount as a deposits again). So I'm not even sure that remains for long. There is a timing difference though and that differential may be sufficient to generate an effect."

Interesting, but as banks need to pay more interest, they also become more indebted. The recycling is a wash with this.

From the government's viewpoint, the expenditure out of central bank profits remitted is an income flow but from the point of view of the bank this payment flow is a financial flow. Banks hence are still paying interest to the central bank. They are not getting it back. The reserve level is different from the case if we had ignored this of course but in no sense is this an income flow for the bank and will rather appear in the flow of funds.

To digress a bit - have a look at Godley/Lavoie's book and reserve requirements there. It does have a small effect on loan rates. From a long term viewpoint however many other interesting things happen.

6:10 AM  
Blogger NeilW said...

"What is this Widow's Cruse mania"

What an odd phrase.

There are enough reserves in the system to fulfil the additional reserve requirement. The central bank ensures that as a matter of policy.

Therefore some banks are up reserves and some are down reserves - as ever. It's a distributional issue. There is no overall extra cost. One banks cost is another banks profit. Banks overall need not pay any more interest on deposits. In fact they will likely pay less overall because they are competing with less bonds in circulation at lower yields - those that the central bank has removed and swapped for reserves/deposits. Just like now with QE.

So I think the central bank has to go to a shortage of reserves policy before imposing higher reserve requirements can have any cost implication on loans - and that's because a fee is then paid to the central bank, which is essentially a tax.

Anyway getting way off topic here.

8:18 AM  
Blogger Unknown said...

Ramanan

Banks compete for deposits because they receive reserves when they accept the deposit liability. The reserves can be used to fulfil capital requirements.

The deposit itself is just a liability that any bank can create. But banks cant create reserves.

1:44 AM  
Blogger Ralph Musgrave said...

I regard that Diamon Dybvig paper cited by Winter as a waste of ink and paper. It’s typical of what some academics produce: pages of near incomprehensible English and maths all designed to keep themselves employed and hide the fact they’ve nothing interesting to say.

Basically they’re just saying that government organised deposit insurance can improve things. They could have said that in so many words. Or if I’ve missed something, can someone tell me what?

3:43 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home