Arnold Kling  

Bob Hall on Quantitative Easing

PRINT
Health Care Costs and Wages... Where's the State of the Free?...

He was giving a talk on unemployment. Along the way, he said that with quantitative easing, the Fed borrows short and lends long. So the Treasury could do the same thing by issuing less long-term debt and more short-term debt. So he does not understand why quantitative easing is anything to get excited about.

If the Treasury did this, and short-term interest rates started to go up, then the Fed would have to print money to maintain the short-term interest rate, which would make it clear that this is the same thing as quantitative easing. Anyway, food for thought.


Comments and Sharing


CATEGORIES: Monetary Policy



COMMENTS (18 to date)
ThomasL writes:

Do you happen to know if his talk will be available online later?

Scott Sumner writes:

He also indicated (last year) that negative interest on reserves would definitely provide monetary stimulus to the economy.

Felix writes:

I'm confused. I thought "quantitative easing" was another name for "printing money". Is that not true?

Paul writes:

Felix, I'm confused too. If the fed really "prints" money then why don't people focus on the net open mkt operations as a measure of currency debasement.

gaohan_cn writes:

why if Treasury issue more short term debt would causing short-term interest rates go up?

Jeff writes:

The Fed is "borrowing short" only because it is paying interest on reserves. Otherwise, QE would just be printing money. Over on his his blog, Jim Hamilton has some estimates that show how much the Fed can lower long term rates by buying long Treasuries and paying with excess reserves. This works for as long as the banks are willing to sit on the reserves, like they're currently doing.

Simon K writes:

In what sense does the Fed borrow in a QE regime? Surely it just creates reserves and uses them to buy stuff?

JPIrving writes:

He is saying that the Treasury could manipulate the Fed into printing w/ this strategy. I guess the idea is that 3month tbills and fed funds are close substitutes. Greatly increase the supply of 3 month tbills and you drive the yield up (price down) on all short yielding assets, including federal funds. The Fed will react by printing and buying enough of the new tbills to counteract the rise in the fed funds rate.

I guess it could work as long as the Fed is not secretly perusing a 1% inflation target, or otherwise doesn't mind being played the fool.

Charles R. Williams writes:

When the fed buys T-bonds it creates excess reserves which earn an above market interest rate. All the fed is doing is fiddling with the term structure of the federal debt - as long as we think of excess reserves as just another denomination of federal debt.

For QE to do anything, it must involve propriety information sensitive debt. In other words, the fed has to imitate what shadow banking does when it finances AAA securities with repos. The difference is that no one will question the value of the fed's collateral. So if they buy used car loans, they create money which the private sector can't create, whereas if they buy T-bonds, they convert near-money into idle bank reserves.

Bill Woolsey writes:

Consider a central bank issuing paper money subject to redemption in gold. The issue of paper money is borrowing by the central bank. Those holding the paper money are lending to the central bank. With hand-to-hand currency, the interest rate on the loan is zero.

If the central bank instead targets the price level, inflation, or nominal GDP, the situation is fundamentally the same. Only if the central bank targets the quantity of base money and makes no commitment to the value of that money in terms of gold or anything else, does this borrowing approach break down.

When a central bank on a gold standard issues paper money it is borrowing very short. It is like a commercial bank borrowing by creating checkable deposits. If a central bank holds T-bills it is lending short, but not as short as it is borrowing. A central bank targeting the price level or inflation is fundamentally doing the same.

If the central bank buys Treasury bonds or notes with 5 years to go before maturing, and issues gold redeemable currency, it is borrowing very short and lending long.

In my view, a liquidity trap, or the zero nominal bound, involves a spill-over from an excess demand for the bonds with the near zero nominal interest rate to an excess demand for money. In that special case, open market operations that create money by purchasing the very same bonds with the near zero nominal interest rate really do expand the demand for money more or less with the quantity of money. Expanding the quantity of money by purchasing zero interest bonds is ineffective--at least in the short run.

And so, the obvious alternative is to purchase bonds without zero yields, and long term government bonds work.

If the government were to sell T-bills and use the proceeds to pay off long term bonds, this would tend to relieve the underlying excess demand for T-bills. This would reduce the spillover to an excess demand for money, and result in higher money expenditures.

I am not sure what Robert Hall's point was supposed to be. Quantitative easing won't work? Or was it, why do some see it as something about to create hyperinflation?

Jeff writes:

Simon, a bank's reserves are deposits at the Fed. But they're like checking accounts,not time deposits, in that the bank is free to pay for stuff with them at any time.

If the Fed is paying interest on those deposits, they are effectively an overnight loan from the bank to the Fed, a loan that gets rolled over every night. The loan is riskless because the Fed cannot go bankrupt so long as it has ink for the printing press. Paying interest on reserves also reduces the stimulative effect from the QE by lessening the incentive to lend to the private sector. Why lend to a risky private business when you have the riskless alternative of lending to the Fed?

Jeff writes:

Bill Woolsey, you can say it more simply than that. Bank reserves on deposit at the Fed are essentially overnight loans to the federal government. When the Fed buys short term Treasuries, which are also short term loans to the government, it's just swapping one close substitute for another. To actually make a difference, reserves have to be exchanged for something that they're not a close substitute for, like long term bonds.

Charles R. Williams writes:

The issue here is how close a substitute T-bonds are for cash. 30 yr T-bonds are liquid. Default risk is zero. They bear a small interest premium over T-bills that covers the risk of interest rate fluctuations. Under current conditions T-bonds are effectively another denomination of cash.

Excess reserves that bear an above market interest rate may not really be money. The banks are effectively being paid to sit on money.

Rebecca Burlingame writes:

@Charles R. Williams,
That's the explanation I've not yet understood. Why are the banks being paid for those reserves, especially given how that dampens the effect of QE2?

Bill Woolsey writes:

The nominal interest rate on 30 year treasury bonds is 3.98 percent. The interest rate on T-bills due in four weeks is .14%. If you want to call 3.84 percent a small premium, I guess it just depends on your standard.

10 year are 3.62%. Now, 5 year are only 1.2% so the premium is only 1.06%. And 3 year are at .59 percent. Now, that is getting to be a small unmber to me-- .45%.

If Treasury bonds really are a close substitute for money (and for T-bills,) then modest quantitative easing should drive their nominal yields to zero as well.

The notion that paying money on reserves means that the reserves aren't really money is wrongheaded. From the point of view of banks, anyway, they serve as media of exchange. They are money.

Balances in checkable deposits are also money. Many checkable deposits pay interest. The form a very important part of the medium of exchange.

Identifying money with hand to hand currency is an error. Hand to hand currency is money, but the fact that it has a zero nominal yield isn't what makes it money.

Finally, the fact that bank reserves pay .2 percent does push the zero nominal bound up from zero. And there is nothing good about that.

Jeff writes:

@Rebecca Burlingame,
There are two reasons for paying interest on reserves. Elimination of the "reserve tax" has been on the Fed's wish list for many years. Commercial banks that have to keep some part of their assets in a non-interest bearing reserve account are at a competitive disadvantage to financial intermediaries like hedge funds and investment banks who are free to invest all of their funds. Over time, more and more intermediation has been moving out of the commercial banking system and into less-regulated institutions. The Fed wanted to stop, or at least slow, this migration.

The second reason is that paying interest on reserves lets the Fed expand or contract its balance sheet without affecting the federal funds rate. If the Fed wants to buy a bunch of, say, MBS, it pays with reserves. Once there are enough reserves in the system to satisfy everyone's reserve requirements, creation of much more of them drives the funds rate to zero. But if the Fed is paying interest on reserves, the funds rate will drop only slightly below the rate the Fed is paying, since banks can always just leave their funds at the Fed.

Simon K writes:

Jeff - Yes, I see. I guess I overlooked the fact the Fed is paying interest on reserves because I think it shouldn't be! If anything it should be charging interest on reserves before embarking on any QE.

Charles R. Williams writes:

The issue with interest on excess reserves is not the paying of interest but the paying of a rate of interest that exceeds market rates for overnight money and T-bills. The banks are being paid a premium to sit on the money.

It is important for the fed to pay a market rate on required reserves held by the banks so that they do not have a competitive disadvantage relative to shadow banking.

Comments for this entry have been closed
Return to top