Should central banks become banks?
By Scott Sumner
The Economist has an article that discusses the possibility of allowing ordinary people to have checking accounts at central banks:
A RECESSION strikes. Central banks leap into action, cutting interest rates to perk up investment. But what if, as now, there is not much cutting to do, with rates already at or close to zero? In such cases the manual calls for purchases of government bonds with newly printed cash–quantitative easing, or QE–swelling the reserves each bank keeps at the central bank. Imagine instead that people also kept accounts at the central bank. New money could be added to their accounts, providing a direct, equitable boost to spending. That is one of several potential benefits of individual central-bank accounts, which are among the more intriguing of the radical policy ideas in circulation.
Interesting idea. Let’s start with the effectiveness of distributing new money to depositors, as a method of stimulating the economy during a recession.
If the money were paid out in proportion to the current size of deposits at the central bank, then this would be equivalent to paying interest on base money, which is contractionary. But of course there’s much more to be said:
1. The subsidy would be provided to deposits at the central bank, not to currency holders. Thus expectations of this payout would create a situation analogous to expectations of currency devaluation under a gold standard, when the dual media of account (cash and gold) are suddenly expected to earn different rates of return. In my study of the Great Depression, I discovered that expectations of currency devaluation led to gold hoarding and therefore were typically contractionary, even though currency devaluation itself was expansionary. Similarly you might have a situation where expectations of a payout to depositors is contractionary, even though the addition of the new money eventually has an expansionary effect.
2. If the payouts were lump sum, and not in proportion to the size of deposits at the central bank, then the effect might be more expansionary. Especially if everyone already had a deposit there, and hence expectations of a payout did not induce more people to put their funds into a central bank account.
Even so, it seems safer to inject money the old fashioned way, by buying bonds. Why risk boosting money demand? So why is this “radical” proposal being made?
This leads to the second claim in the article, that the policy would provide a “direct, equitable boost to spending”. I’m having trouble understanding exactly how it would be more equitable. The pre-2008 system of open market purchases of government bonds seems extremely equitable, almost as neutral as one could imagine:
1. The purchases are at market prices, so there is no subsidy to bondholders.
2. Even if the Fed bought something other than bonds, the price of bonds might rise as the injection of new money leads to lower interest rates, via the liquidity effect.
3. Even if the Fed buys T-bonds, the price of bonds often falls in response, as the Fisher effect usually outweighs the liquidity effect in the long run. Bondholders were devastated by the huge monetary injections of 1965-81, but they would have been equally devastated if the money had been injected in some other way.
4. The commissions on these bond purchases are tiny, and at competitive rates, so there is no subsidy to bond dealers.
5. Yes, OMOs often cause other asset prices (such as stocks) to rise, but that would be true of any policy that prevents a depression, not just monetary stimulus. If you don’t want asset holders to do well, then create a big depression.
So I’m have trouble understanding what sort of inequity this proposal is supposed to address.
And in some ways, payments to accounts held at the central bank seem less equitable than normal OMOs. Those who do not have accounts at the central bank (disproportionally the poor) would get none of the new money. The proposal is a sort of helicopter drop, which combines fiscal and monetary stimulus (although the monetary part is of doubtful effectiveness, for reasons cited above) but it’s a helicopter drop where most of the new money falls over Beverly Hills, not Compton.
Nonetheless, I’m not willing to dismiss this idea out of hand. It’s not obvious to me why banks should have deposits at the Fed, while I cannot have a deposit at the Fed. It would be possible to allow ordinary people to have deposits at the Fed, and still do monetary policy the old fashioned way, by buying T-securities. It would not even be necessary to use the current “floor system”, heavily criticized by experts such as George Selgin. The deposits could pay below market interest rates (even zero), so that monetary policy would return to the pre-2008 regime (before the interest rate on reserves was set above market rates.) What about allowing the central bank to pay depositors a rate 50 basis points below the 3-month T-bill yield?
One criticism of this system is that it would socialize banking, and socialism is bad. Because I’m a libertarian, I’m certainly willing to consider that criticism. But on close inspection, our system is already heavily socialized. Due to deposit insurance, when you put money in the bank you are basically lending it to the Treasury, which re-lends the money to your bank at the same interest rate. That’s a very inefficient system, which encourages excessive risk-taking (especially in the 1970s-80s, and 2000s)
Another criticism is that it would hurt private banks, which are an important source of financing for small businesses and homebuyers. By why should the government subsidize lending?
Maybe the central bank would not provide as many conveniences as private banks. But in that case, people would continue to use private banks. Of course, for a level playing field, private banks should be allowed to issue currency notes.
I’m willing to support private accounts at the central bank if combined with one other policy reform, abolition of FDIC–at least for deposits not 100% backed by government bonds. The public could be told that if you want a safe bank deposit, put it into an account backed up by Treasury securities. Otherwise you can take more risk, and earn a higher return, by lending the money to banks that make commercial and mortgage loans.
Under my proposal, the additional socialism of deposits at the Fed is less harmful than the harm currently being done by deposit insurance on accounts where money is lent out to risky borrowers.
Finally, I would quibble a bit with this claim:
Money is commonly considered a liability of a central bank. Accountants would frown at distributing new money without obtaining assets in exchange (like the government bonds purchased when banks carry out QE), since they would create a huge negative position on central-bank balance-sheets. But an institution that can create its own money cannot go bankrupt. As long as a central bank is keeping to a policy target (like a 2% inflation rate) an ugly balance-sheet is not a problem.
Technically that’s true, but it seems a bit misleading to me. The problem here is that a central bank may not be able to keep inflation at 2%, precisely because of a negative position in its balance sheet. The Economist is right that this outcome would not be default (it can print unlimited amounts of money) but it could lead to economic instability.
This is one of the many reasons why I oppose helicopter drops. There is no problem that can be solved with helicopter drops, that can’t be solved much more effectively and equitably with ordinary open market purchases, especially if combined with reforming the monetary regime—NGDPLT. And if we don’t adopt NGDPLT, even helicopter drops may not work—as the Japanese discovered during 1993-2013.
PS. Technically, when you deposit money at a bank, you are lending it to the Treasury, which lends it to FDIC, which lends it to the commercial bank, all at the same interest rate.
PPS. Why do I say IOR is contractionary, when I also argue against “reasoning from a price change”? Because taxes and subsidies are not ordinary price changes. If we know that the price changes due to a tax or subsidy, we know the impact on quantity. IOR is a subsidy and negative IOR is a tax. They are not ordinary market prices.
PPPS. One final comment on my previous post. I’d encourage people to read Richard Rorty, who points out that saying something like, “Two plus two is four” is equivalent to saying “I believe that two plus two is four.” Similarly, saying “Joe is not morally responsible” is no different from saying “I regard Joe as not being morally responsible”. People seem to want to draw meaningless distinctions. Don’t forget; “That which as no practical implications, has no philosophical implications”.
Now admittedly it’s possible to regard someone as being not being morally responsible, and pretend that you hold them morally responsible. But it’s much simpler to just assume that people who hold other people morally responsible actually believe they are morally responsible. That’s my view. Ockham’s razor.
All statements about the world are statements about how we view the world. Don’t look for deeper realities, beyond what we believe to be true. It would be like saying “She’s actually pretty, even though every single person (and God if you wish) regards her as being ugly.” That means nothing.