Arnold Kling  

Monetary Thought Experiments

CBS Sixty Minutes: Insider Tra... Canadian Recalculation?...

Scott Sumner writes,

Why would doubling the supply of currency cause people to hold twice as much currency in real terms?

Here is my thought experiment. Suppose that the government retired lots of $10 bills, giving $5 bills in exchange, at the fair rate of two fives for one ten. Suppose that this doubled the supply of $5 bills in circulation. In order to keep the real supply of $5 bills constant, the price level would have to double. Do we think that people would go on a spending spree with their excess $5 bills and cause the price level to double?

I don't think so. I think people would adapt to having more fives and fewer tens.

Next, suppose that instead of retiring $10 bills with $5 bills, the government retires 30-day Treasury bills with $5 bills. Suppose that this doubles the supply of $5 bills in circulation. Do we think that people want to hold only a constant real quantity of $5 bills, so that the only way we can get back to equilibrium is to have people spend to the point where the price level doubles?

Again, I don't think so. Exchanging $5 bills for 30-day Treasury bills probably will be more consequential than exchanging $5 bills for $10 bills. But I don't think it will be so much more consequential.

The other issue where Sumner and I differ is the formation of expectations for inflation. I do not believe that there is any policy that the Fed could have followed that would have resulted in expectations for high inflation from September of 2008 through September of 2009. Perhaps the government can convince people that the price level will be high ten years from now (it has convinced me of that). But it cannot manipulate near-term inflation expectations.

Yes, if an asset is now priced at $100 and I peg the price at $150 three years from now, the price has to go up now. But the prices that go into the Consumer Price Index or the GDP deflator are not asset prices. The prices of ordinary goods and services will not start rising until more people get used to them rising. It takes a while for expectations of inflation to become self-reinforcing.

Note that because I have been arguing that monetary policy is rather impotent, I cannot agree with David Beckworth's recent post arguing that monetary policy set the stage for the financial crisis. He writes,

There are at least two reasons why monetary policy was important here: (1) it helped create macroeconomic complacency and (2) it created distortions in the financial system via the risk-taking channel.

Comments and Sharing

COMMENTS (10 to date)
Mike Sproul writes:

Another thought experiment:

Two central banks have each issued 100 currency units in exchange for 100 oz. of silver. Each currency unit will be redeemable for 1 ounce in 1 year. At a 5% interest rate, they would each sell for .95 oz at the start of the year if the cost of issue were zero. But suppose the currency units cost 5%/year to keep in circulation. That way they start and end the year worth 1 oz each, yielding zero interest. The cost of issue eats up the interest.

One central bank issues another 200 currency units in exchange for bonds worth 200 oz, while the other issues 200 units and throws them to crowds in the street.

All through the year, the first bank's currency will be worth 1 oz., while the second bank's currency will be worth 1/3 oz, because the second bank will have only 100 oz with which to buy back 300 currency units at the end of the year. The first bank will have 300 oz worth of assets with which to buy back its 300 units.

It's not how much money is issued. It's how much backing each unit of money has that matters. If anyone can think of any important way in which modern central banks differ from the first bank above, I'd like to hear about it.

Bob Murphy writes:

Dr. Kling,

I am always surprised by how unqualified you make your claims. If Bernanke had announced that he would begin writing $1 million checks for every gallon of milk or loaf of bread shipped to various Fed warehouses, you don't think that would push up CPI pretty quickly?

I grant you, Obama would send in the Delta Force to take out Bernanke if it came to that, to thwart such a policy. But you continually state that it is impossible for the Fed to move people's short-term price expectations very quickly, and I think that's not true.

So do you mean "within the bounds of reasonable policies" or are you sticking to your guns?

Philo writes:

"The prices of ordinary goods and services will not start rising until more people get used to them rising." People were used to 2% inflation; they revised that expectation *very quickly* last year. This refutes your claim that inflation expectations can change only slowly.

"Perhaps the government can convince people that the price level will be high ten years from now (it has convinced me of that). But it cannot manipulate near-term inflation expectations." Yet you admit: "Yes, if an asset is now priced at $100 and I peg the price at $150 three years from now, the price has to go up now." If, as you say, this doesn't happen in the consumer economy, it must be that consumers and, especially, retailers--while expecting the $150 price in three years--are too stupid to do the backward calculation. That's not plausible.

As Scott Sumner says, old-time bankers and economists would have laughed at the idea that the monetary authority *over a fiat currency* couldn't produce lots of inflation *very quickly*.

Simon Kinahan writes:

But 2 $5 bills are a (almost) perfect substitute for 1 $10 bill. Its value in cash is not perfect substitute for a 30 day treasury - its just the amount of cash at which the seller is willing to sell it. X*(1+r) in 30 days is clearly not the same as X now, even if r is actually zero (which it isn't quite). 5+5 now really is the same as 10 now and always will be.

There are a host of reasons why someone holding a 30 day bond might choose to take cash for it instead, and no reasons at all why anyone should care whether they have 2 fives or a ten. At the very least someone taking cash for a treasury bill will turn around and buy private sector debt of equivalent value. Which is after all the whole idea.

woupiestek writes:

The FED should not have created expectations of high inflation, but just sustained expectations of normal inflation. Isn't that much easier to do?

Also, I think a lot would spend more money if they had more money to spend, or expected to have more money to spend. The whole idea that for no good reason at all, everyone decides to hold on to addition cash is pretty freaky.

Doc Merlin writes:

Just because monetary policy can't do much short term good (in your view), it doesn't follow that it can't do short term harm.

Bill Woolsey writes:


I think at MIT they teach them that monetary policy _means_ the Fed buying T-bills. Buying milk isn't monetary policy by definition.

Bob Murphy writes:


How do the MIT guys classify buying mortgage-backed securities?

(Myself, I call it theft.)

Doc Merlin writes:

Minor quibble, but the fed doesn't just buy T-bills anymore it buys all sorts of stuff.

Scott Sumner writes:

Maybe I am unusual, but when I go to the store with a wad of bills in my pocket, it doesn't much matter to me whether I have 10 fives of 5 tens. But if I had some T-bills in my pocket, I might have a bit more trouble exchanging them for goods.

My post was mostly directed at normal times, when T-bill yields are around 4%. I say this because Arnold Kling's hypothesis didn't seem to merely be limited to liquidity traps. But in normal times the price of a T-bill is very different from its face value, hence making it not very useful in transations. And the other great source of demand for cash is hoarding to escape taxes. Haven't anonymous T-bills been eliminated?

woupiestek's first point is very important and often overlooked--the Fed didn't need to change expectations, it needed to maintain them. That is far easier.

Comments for this entry have been closed
Return to top