Scott Sumner  

100% money?

In Praise of Lord Acton... Lord Acton's Humor...

Alex Tabarrok points out that Switzerland will have a referendum on a 100% reserve banking system. The details are kind of vague, as is the motivation. It's also odd that this is occurring in Switzerland, which supposedly has a "problem" with a bloated monetary base. As far as I can see this would lead to an even larger base, much larger.

In the 1930s, Irving Fisher and Henry Simons advocated a plan where bank deposits would be 100% backed by reserves, and in that case the motivation was clear. The idea was to prevent the sort of big drop in the broader money supply that occurred in 1929-33, due to a falling money multiplier. Under 100% reserve banking, the money supply would be roughly equally to the monetary base, and hence the multiplier would be approximately one. But multiplier instability is not a problem today, under our current fiat money regime. The Fed can offset any fall in the multiplier, keeping the money supply unchanged.

Alex links to an article by Martin Wolf, which provides a different motivation for 100% money:

Banking is therefore not a normal market activity, because it provides two linked public goods: money and the payments network. On one side of banks' balance sheets lie risky assets; on the other lie liabilities the public thinks safe. This is why central banks act as lenders of last resort and governments provide deposit insurance and equity injections. It is also why banking is heavily regulated. Yet credit cycles are still hugely destabilising.

I believe that most people underestimate the role that moral hazard plays in banking instability. Thus I am sympathetic to Wolf's proposal. But how is the plan to be implemented?

First, the state, not banks, would create all transactions money, just as it creates cash today. Customers would own the money in transaction accounts, and would pay the banks a fee for managing them.

Second, banks could offer investment accounts, which would provide loans. But they could only loan money actually invested by customers. They would be stopped from creating such accounts out of thin air and so would become the intermediaries that many wrongly believe they now are. Holdings in such accounts could not be reassigned as a means of payment. Holders of investment accounts would be vulnerable to losses. Regulators might impose equity requirements and other prudential rules against such accounts.

Third, the central bank would create new money as needed to promote non-inflationary growth. Decisions on money creation would, as now, be taken by a committee independent of government.

Finally, the new money would be injected into the economy in four possible ways: to finance government spending, in place of taxes or borrowing; to make direct payments to citizens; to redeem outstanding debts, public or private; or to make new loans through banks or other intermediaries. All such mechanisms could (and should) be made as transparent as one might wish.

The third point is uncontroversial, it's how things are done today, and the last point needlessly complicates the system---they should just buy Treasury bonds. The first two points are the key.

Wolf is not really proposing true 100% reserve banking, as the investment accounts would not be backed by reserves. And I don't understand the point about "creating such accounts out of thin air". The whole subject of "money creation" and the "money multiplier" is surrounded by confusion, and quite frankly, stupidity. All well-informed observers understand what banks do. Whether you prefer to call that money creation, a money multiplier, or something else depends on how you prefer to define terms. I prefer to define 'money' as the monetary base, but most economists prefer a broader definition of money---cash plus bank deposits. No one claims that banks create base money out of thin air. I happen to think the term 'money multiplier' should be dropped as it causes more confusion than it is worth. The more interesting multiplier is NGDP/base, also known as base velocity.

Under Wolf's proposal, banks would be creating deposits "out of thin air" just as much as today. You could deposit $1000 in a bank investment account. Someone might borrow $900 of that money. The borrower might then deposit that $900 in another bank, which then loans out $810, etc., etc. As long as you have bank accounts not backed by reserves (i.e. investment accounts), banks will be creating deposits in pretty much the same way they do today. Whether you want to call that "out of thin air" is an aesthetic choice, not a substantive issue.

Nonetheless, I am attracted to Wolf's proposal because it allows us to remove moral hazard from the banking system. The risky investment accounts would not be insured by the government, and demand deposits that would be insured would be 100% safe, thus shifting no risk to the taxpayer. (Of course you'd still have to address too big to fail, and I'll do a post on that soon.)

But if the goal is to get rid of moral hazard, there are more efficient ways of doing so. Why not require 100% backing of demand deposits with base money or interest-bearing government bonds? Then perhaps you would not have to charge a fee to depositors for their checking accounts. Or at least there would be a smaller fee. Yes, there is some risk of a fall in bond prices, but there are two reasons why that risk would be trivial:

1. T-bond prices tend to rise during banking panics.
2. Banks would almost certainly have enough assets backing investment accounts to at least cover the shortfall produced by a fall in T-bond prices that back checking accounts.

And of course you could periodically require banks to "mark to market" the T-bond portfolio. I'm quite confident that deposits backed by T-bonds would be virtually 100% safe.

I've advocated removing banks as much as possible from the monetary system. I'd like to see no reserve requirements, so that during normal times when interest rates are positive, the monetary base would be 99% currency. Then the central bank would control monetary policy by adjusting the stock of currency and coins. Everything in the banking system would then be "endogenous."

So in a sense we are moving in the opposite direction from what I'd prefer. We are moving toward a system with a much larger share of the monetary base being held by banks, due to interest on reserves (IOR), whereas I'd prefer to remove banks from the monetary system as much as possible. Preferably with a monetary policy that leads to positive interest rates, and a system of zero IOR and no reserve requirements.

Comments and Sharing

CATEGORIES: Finance , Monetary Policy

COMMENTS (18 to date)
Ken writes:

How about requiring 100% backing of demand deposits with Capital Conversion (Co-Co) bonds. I think this would be more stable, particularly on a mark to market basis , than even government bonds.

It seems that the investment side of what is suggested is nearly identical to the function of a typical Mutual Fund company like Vanguard.

The investor buys shares in a fund, and the bank simply provides the mechanism of the transaction - serves essentially as a broker.

The moral hazard is elliminated because the investor owns the asset, and accepts the risk - not the bank

Emerson writes:

I have always thought that there should be two types of Banks.

Type I: This banks is FDIC insured but is allowed to invest depositor money in nothing but short term treasuries. The interest paid on deposits would consequently be low, but in return, the deposit would be virtually 100% safe. This would also significantly elliminate the moral hazard issue that currently exists in banking.

This would be nearly identical to the 100% reserve , backed by treasuries suggested by Prof Sumner.

Type II: This bank would not be FDIC insured. And the bank would be at liberty to invest depositor money in whatever it wants. The interest paid on deposits would potentially be higher than Type I but the depositor would have to accept this as the trade off for not being FDIC insured. Lack of FDIC insurance would partially elliminate moral hazard. In Type II banks, capital conversion bonds equal to about 5% of capital would make sense as a means of taxpayer insurance against the need for a bail out.

It seems that Type I banks would fit a particular market niche for safe, low paying deposits. Type II banks would likely find themselves in direct competition with mutual fund companies.

PJ writes:

John Cochrane has essentially advocated narrow banking in his essay "Toward a Run-Free Financial System". I think he makes a strong case. I am also convinced by Chari and Phelan (2014) that improvement in IT and communication technology means that fractional reserve banking is inefficient and will persist unless it is prohibited.

I've asked two leading economists who study banking crises what they think of Cochrane's essay and they basically say he hasn't studied banking long enough to understand why he is wrong. But I was unable to get them to say precisely why he is wrong.

100% reserves seems like a fairly small step given that the banking system currently holds more reserves than there are deposits.

Dan W. writes:


Another control needed on "banks" is I believe they should be partnerships and not corporations. As partnerships the "owners" would have a vested interest in preserving the bank's capital.

Banks should be boring. For those who desire less security and greater reward the free market provides ample opportunity for lenders and borrowers. But it would be understand if one puts $1000 into a corporate or securitized bond that there is no guarantee full value will be payed back.

Emerson writes:

Interesting poing on Banks organized as partnerships.

Thats how things were initially. There's an interesting speech given by Andrew Haldane , Oct 2011. The Wincott Memorial Lecture, London titled "Control rights (and wrongs)".

He discusses the history of banks as partnerships. The problem with a partership is that there is unlimited liability. If things go south, the partners must cover all the loses - even the loses that go beyond their ownership interests. He relates a situation of the failure of Glascow Bank in 1878 where the depositors were made whole, but the partners (including some spinsters and widows) were left destitute.

From this evolved partnerships of extended liability meaning liability of initial investment plus an extra defined amount.

But the problem was that even with extended liability, investors were fearful of being a partner in a bank. As a consequence start up capital was hard to come by so banks were few and credit was scarce.

Finally banks evolved to limited liability corporations. This solved the problem of obtaining capital. But it helped create (along with FDIC, low capital requirements etc.) the problem of moral hazard.

Gordon writes:

"We are moving toward a system with a much larger share of the monetary base being held by banks, due to interest on reserves (IOR), whereas I'd prefer to remove banks from the monetary system as much as possible."

Scott, do you have any concerns about the Fed's plan to increase the federal funds rate by raising IOR to .5%?

Scott Sumner writes:

Everyone, I think several of the ideas mentioned are promising. I'll do a post on Co-Co bonds soon.

The key is that FDIC backed deposits need to be safe.

PJ, Yes, but 100% reserves could be a big deal if we ever went back to 5% interest rates, or even 3%.

Gordon, I don't like the idea of IOR, but I doubt that an increase in December would do much harm. Still, it's at least slightly risky, and I see almost no benefits.

Mike Sproul writes:

Luckily, the 100% reservers refuse to believe that people create money by using their credit cards, so the ridiculously tight money conditions that 100% reserves would create will be mostly offset by the use of credit cards.

Jose Romeu Robazzi writes:

As I understand it, the idea of 100% reserve is to avoid liquidity mismatch. Demand deposits backed by credit may lead a bank to not be able to redeem deposits, even if it is solvent (i.e. its assets have enough value to pay for its liabilities). Investment accounts might work if they are not redeemable "on demand", and there is som rule to manage liquidity mismatches between assets and liabilities. I would really like to see the payments system totally disconnected from the investment system. Prof. Sumner, I don't know if you wrote about your idea of "removing banks as much as possible from the monetary system" before. If you did, I would be interested in reading it. I don't see how no reserve requirements would achieve that, since with no reserve requirements banks could have increasing levels of liquidity mismatch, which in my view is the real danger. With high liquidity mismatch, in a time of increased perceived risk in the system, a lot of participants trying to liquidate assets will inevitably lead to a quick monetary contraction, which, if the monetary authority tries to offset, will lead it to end up bearing credit risk (the assets from the banks that it will have to hold, at least for a while).

bill writes:

I believe that if people had the choice between a guaranteed savings account (even with 0% interest) and a fund where their local bankers got to invest the money in business loans and construction loans, etc., that almost no one would choose the local banker loan investment fund. At least not for an extra percent or two. It's a blind pool. Maybe it would be a good thing. It probably would be. But it definitely would be a dramatic change.

Jose Romeu Robazzi writes:

Pre-2008 invstment banks were your type II "banks", they were levered 50+ times, and most of their funding came from short term repos, which type I banks provided them ...

Emerson writes:

So what is the answer to a Type II / Investment bank that over leverages and becomes insolvent?

1): Tax payer bailout

2): Co Co bonds to provide an " internal bailout" via the private creditors that purchased the Co Co's ( my personal favorite).

3): No bailout- let the chips fall where they may. This would work provided that the Type II bank has been restricted in size such that it would not be " systemically important" currently defined as assets less than $50B

Scott Sumner writes:

Mike, For some reason I can't tell if you are joking, probably my fault.

Jose, I think you missed the point, I am arguing that banks should be made safe by requiring them to back up any insured deposits with T-bonds. As far as investment banking, I will do a new post in a couple days on how to avoid banking crises. But reserves don't really do anything to make banks safer, that T-bonds couldn't do just as well.

Bill, In the 1920s people could choose between uninsured bank deposits and safe T-bills, and they often chose bank deposits.

Emerson, I have a piece on Co-cos that I'll post early next week.

Jose Romeu Robazzi writes:

@Prof. Sumner
Ok, I got that wrong, 100% T-bills and T-bonds backed deposits are ok, IMHO. Thanks.

Bill Woolsey writes:

I found this post a bit confusing.

Requiring transaction accounts to be "backed" by T-bills or reserves is not at all the same thing as requiring FDIC insured accounts to be "backed" by T-bills or reserves.

It would be possible to have transaction accounts that are not FDIC insured. They could be used for transactions but could be "backed" by a variety of assets. In my view, trying to prevent runs by regulating bank assets is unlikely to be effective. It just drives financial innovation aimed at regulatory arbitrage.

And not all FDIC insured accounts are transactions accounts. Savings accounts and Certificates of Deposit that are FDIC insured would be matched on bank balance sheets by reserves or T-bills under such a proposal.

Jeff writes:

What you want is to separate money, the means of payment, from credit. There are two means of payment, currency and checkable deposits. Currency is a liability of the Federal Reserve System and has nothing to do with credit. But checkable deposits are liabilities of private depository institutions, and they grow or shrink along with the assets that back them. Those assets include bank loans and other credit market instruments, along with reserve accounts at the Fed and physical assets like bank buildings and the land they sit on.

In the current system, there are two routes through which credit market disturbances can affect the means of payment. The first is that if any of a bank's assets or other liabilities change in value or quantity, checkable deposits are likely to be affected because the balance sheet has to balance. The Fed can offset this for the banking system as a whole by changing the amount of reserves outstanding.

The second way a credit market event can affect the means of payment is harder to offset. If a bank is perceived to be insolvent, or close to it, checks drawn on that bank may not be honored. Changing the amount of reserves outstanding in the system as a whole does not affect the solvency of any particular bank. So we have things like capital requirements to try to avoid insolvencies in the first place, and deposit insurance to avoid or at least mitigate panics.

A system of 100 percent reserves required against checkable deposits closes both routes. The simplest way to do this is to disallow institutions offering checkable accounts from issuing any other kinds of liabilities, and limit their assets to the safe ones. These are narrow banks. If you allow non-narrow banks to offer other liabilities as well, you will need a bunch of additional legal rules and regulations to keep things like sweep accounts and solvency concerns at bay. Better to avoid the arms race with bank lawyers and just do narrow banking.

Another virtue of narrow banking is that, since transactions money is always equal to the amount of Federal Reserve liabilities outstanding, the Fed can have very precise control over it. If it wants to offset an aggregate demand shock, it can do so easily through open market operations, guided, if our host has his way, by market-based forecasts of things thought to be closely related to aggregate demand. Maybe even expected NGDP.

Floccina writes:

Why is money considered a public good? Is it both non-excludable and non-rivalrous?

Market Fiscalist writes:

"Under Wolf's proposal, banks would be creating deposits "out of thin air" just as much as today. You could deposit $1000 in a bank investment account. Someone might borrow $900 of that money. The borrower might then deposit that $900 in another bank, which then loans out $810, etc., etc. As long as you have bank accounts not backed by reserves (i.e. investment accounts), banks will be creating deposits in pretty much the same way they do today."

I do not follow this. If I deposit $1000 in a "transaction account" and the bank lent out $900, then both I and the borrower could spend that money - you could see why people might say the bank created $900 out-of-thin air.

If however I deposit $1000 in a "investment account" and the bank lent out $900, then if I understand correctly what an investment account is, I can not spend the money until after the $900 is paid back. I find it hard to see how this could count as creating money out-of-thin air.

The initial $1000 can multiply up if some of loans themselves end up in "investment accounts", but there will never be more than $1000 available to be spent at any time.

Comments for this entry have been closed
Return to top