David R. Henderson  

Selgin's Thermostat Analogy

Zwolinski's Weak Case for a Gu... Was Going to War with Japan a ...
Once, while a good friend was visiting me on a particularly cold winter's night, the temperature in the poorly uninsulated living room of my old Victorian house dropped to a distinctly chilly 62 degrees. "Can't you make it any warmer?" she asked? "I'm afraid I can't," I said; "the thermostat's already on 68." "Try setting it at 80," she replied.

I didn't indulge her (well, not that way). But I wonder whether those economists who have been calling for a higher inflation target would have.

This is from George Selgin, "New-Keynesian Thermodynamics," November 30.

I challenged George about the analogy by e-mail, writing:

I get the futility, obviously, about why setting the thermostat higher won't work. But what is the analog? Is it that the Fed has set an inflation target above the current inflation rate and so it doesn't make sense to set it even higher? If so, I get that. But is it as clear in the Fed case as in the home heating case that the Fed is actually setting a higher target rate? Couldn't it be that the Fed says it is, but it isn't? So, to go back to your home heating case, the analog would be you telling your guest that you already set it at 68, whereas you actually set it at the current 62.

George assured me, as did other monetary economists, that the Fed really means it when it says its inflation target is 2%.

Specifically, here's what Jeff Hummel wrote in response to my challenge to George:

You raised a good question: if the Fed really wants 2 percent inflation, why aren't they achieving it? After all, central banks traditionally could deliver any long-run rate of inflation they wanted as long as they had the will to do so. But I think that George and Jerry are right; 2 percent is their actual goal. So what explains the failure of technique? My answer would be Fed targeting of interest rates, yet again. As long as interest rates are really low (for reasons other than monetary policy), they think their policy is very loose and just not yet working. Couple that with the fact that paying interest on reserves and other policies designed to allocate credit are offsetting their expansion of the monetary base.

So, yes, George Selgin's thermostat analogy holds but it needs to be revised, and in a way that loses a lot of the analogy's power. It would be as if his guest asked him to set the thermostat higher than the 68 he thought he had set, but it turned out that he thought wrong. George had never set it. Instead, he assigned the task to someone else who assured him that it was set at 68 when actually the person kept it at 62. [The analog to this person's word is using interest rates as an indicator of tightness or looseness. The person he tasked lied. Interest rates "lie" as an indicator of tightness or looseness.]

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (17 to date)
Yancey Ward writes:

I suspect the Fed knows it can raise the inflation rate above 2% whenever it wants, but doubts its ability to control it afterwards. I view the people advocating the Fed to do more on the inflation front to be analogous to Selgin's friend advocating Selgin set the sofa ablaze.

Mike Hammock writes:

How about "He set the thermostat to 68, but the thermostat was in a small room in the middle of the house right next to a vent. The thermostat thought it was 68, but in the rest of the house it was 62. The thermostat was looking in the wrong place."

That's a pretty clumsy analogy, but I think it captures the intended idea.

David R. Henderson writes:

@Mike Hammock,
Yes, that’s a better analogy, I think.

Daublin writes:

I like yours, Yancey!

Arnold Kling has been asking just that heretical question, but without so good a metaphor. Yes, anyone can burn the house down if they want. The question is, just because you have learned not to play with matches, have you truly become a maestro that can control the temperature down to a tenth of a degree?

Roger McKinney writes:

This is really funny! Everyone assumes that the Fed can create as much inflation as it wants when it wants. Since it hasn't, some think it's being coy, while others think it's not doing enough.

Real economists should consider the facts and quit trying to be psychics. The fact is the Fed is trying to generate higher inflation but can't. Monetary policy is not as powerful as mainstream econ has assumed.

For QE to create inflation, the sellers of bonds to the Fed must spend the money on consumption. If they just buy another asset they won't create inflation.

If the Fed buys from a bank, the bank must lend the money to a consumer or a business. Banks aren't interested in lending for anything but housing and cars, and then only to borrowers with perfect credit.

For low interest rates to work, businesses need to borrow and invest, but they aren't.

One would think economists would have learned more from Japan's attempts at goosing the money supply over the past two decades.

Roger McKinney writes:

PS,in other times the government has done the borrowing and spending to create inflation, but the US is limited by already high debt.

Yancey Ward writes:


You basically beg the question when you assert that the Fed is trying to increase the CPI and failing.

Hamood writes:

Hi All,
I like your comments, but I personally think that the Fed is some how powerless nowadays in its monetary policy. I don't think it needs to keep the interest rate that low, it is time to increase a bit, and try to manage the inflation rate more effectively!!
what do all think about that?keeping in mind that the 3rd quarter GDP has increased to 3.8%!!!

Benjamin Cole writes:

The Fed says it has a 2 percent target, but acts like it has a 1.5 percent ceiling,,,it does not help that various FOMC board members rant in public about the perils of inflation and the glories of deflation.

I think the market has correctly pegged the Fed as not serious about the 2 percent inflation target...otherwise, would not the Fed now target 3 percent to make up for the years at much less than 2 percent?

The sad thing is, with an effective 1.5 percent inflation ceiling, will the economy ever recover..or will we hit another recession while still close to the zero bound?

Sam Thomsen writes:

Similar to Mike Hammocks point about the thermostat being in the wrong place: The thermometer component of the thermostat could be miscalibrated or broken.

Roger McKinney. Thank you for these insights, new to me:

"For QE to create inflation, the sellers of bonds to the Fed must spend the money on consumption. If they just buy another asset they won't create inflation.
"For low interest rates to work, businesses need to borrow and invest, but they aren't."

Roger McKinney writes:


You basically beg the question when you assert that the Fed is trying to increase the CPI and failing.

How so? The Fed claims it wants higher inflation. Is it getting it?

Richard, Thanks! Mainstream economists forget that the economy isn't a hydraulic system; we're dealing with people. The Fed has to persuade people to spend cash on consumer goods, or businesses to invest, else nothing happens.

Mike Hammock writes:

For what it's worth, we had Larry White at MTSU last week. He said that there are primarily two reason why all that new money the Fed has printed hasn't resulted in much of an increase in money supply.
1) The fed is paying interest on reserves, so rather than starting the money creation process by lending out the new dollars, banks just keep the money on deposit at the Fed.
2) Regulators and the Fed are actively discouraging lending, through regulation or threats of regulation.

None of which really answers the question of why the Fed is pursuing this apparently contradictory policy.

Andrew_FL writes:

Well, CPI less food and energy is pretty close to 2%. But, correct me if I'm wrong, the Fed prefers the Personal Consumption Expenditures Index, yes? (Also less food and energy? or not?) That's, at least presently, significantly lower.

Perhaps we are dealing with, in part, a measurement problem. There are prices that are rising, but not necessarily in the mix that the Fed is looking for. There then isn't necessarily then a lack of the amount of inflation they are looking for, it's just occurring in the wrong prices.

Chris F writes:

@Mike Hammock

Maybe the Fed recognizes that the chances of spillover are a lot more likely than they want to admit? So they're trying to get as much traction out of their saying they want a target yet can't credibly commit.

Patrick writes:

Thanks for clarifying the analogy at the end. It makes a lot more sense when you when bring to light that the heat was never truly increased and the man was lied to much like trying to use interest rates as an indicator of tightness or looseness.

This type of analysis or misleading information can easily lead the Feds to misinterpreting that data and and coming up with inaccurate goals or estimates.

prometheefeu writes:

I would have indulged my guest. I have a single thermostat for my apartment. If my thermostat is set at 68, it may very well be 62 in one room and 68 in another and setting the thermostat at 80 may bring up the temperature in the cold room. I will let seeing the parallels to the Fed and the economy as an exercise to the reader.

Comments for this entry have been closed
Return to top