David R. Henderson  

In Defense of Dudley

PRINT
Minogue on intellectuals and p... Macro events (in big economies...

Looking for Bastiat's Broken Window Fallacy Where There Isn't One.

A libertarian friend on Facebook linked to this two-minute interview of William Dudley, president of the New York Federal Reserve Bank, and claimed that Dudley was subscribing to Frederic Bastiat's famous broken window fallacy.

He's not. There is one statement I wish Dudley had made to clarify his point, but the statement I wish he had made is implicit in one word he uses.

Let's quickly review the broken window fallacy. Bastiat tells of the boy who breaks a window at a bakery. People in the crowd say that's good. Why? Because then it will give more work to glaziers. Bastiat points out that what's unseen is that then the baker will have less money to buy shoes for his kids. Whereas Bastiat focuses on the fact that the glazier has more work while the shoemaker has less work, I, when I teach this article (and I teach it in every course I teach), focus on the effect on wealth. Net result: same number of windows, but one less (fewer?) pair of shoes. Wealth is destroyed.

Dudley makes two points: (1) an obvious one that libertarians will agree with, and (2) a somewhat less obvious one that libertarians should agree with.

The first point is that the economy will recover. While these wealth losses are awful, the long-run effect is probably fairly small. Or, in his words, "Those effects tend to be pretty transitory,"

His second point is that, in his words, "The long-run effect of these disasters unfortunately is it actually lifts economic activity because you have to rebuild all the things that have been damaged by the storms." That's roughly right. I would have said that that's the medium-run effect but Dudley and I might have the same amount of time in mind and, to a central banker, a year or two is the long run. Bastiat doesn't address this point. Implicit in Bastiat's analysis is a full-employment model in which the number of jobs and the number of work hours is the same but the number of increased hours worked by the glazier is exactly offset by the number of decreased hours worked by the shoemaker. That works fine to make Bastiat's important point. But Dudley is addressing a different point: when people's wealth falls, they will temporarily step up their work effort. The result: real GDP may well by slightly higher than otherwise.

This doesn't mean, as my libertarian friend and some of his commenters said, that Hurricane Irma is good for the economy. Hurricane Irma is bad for the economy. And Dudley agrees that it's bad.

What did Dudley say that should have keyed my libertarian friend into the idea that Dudley thinks Hurricane Irma is bad? One key word: "unfortunately."

Look back at his sentence: "The long-run effect of these disasters unfortunately is it actually lifts economic activity because you have to rebuild all the things that have been damaged by the storms." He wouldn't have used the word "unfortunately" if he had thought Hurricane Irma is good for the economy. When your wealth takes a hit, you work harder to offset the wealth loss, even if you never completely offset it. Your output goes up, but your leisure goes down. You are better off by working harder than by not working harder, but you're worse off than if your wealth hadn't fallen in the first place.

So I wish Dudley had clearly made the distinction between wealth and income (GDP). That distinction was implicit in what he said, but he would have done well to make it explicit.

But claiming that he could have said it better--couldn't we all?--is different from claiming that he said it wrong. He didn't.

UPDATE:
Eric Boehm at Reason gets Dudley wrong.


Comments and Sharing






COMMENTS (16 to date)
George J. Georganas writes:

There is, also, a difference between an act that humans can actually do something about, such as punishing or preventing vandalism and a natural disaster that humans cannot do much about.

John Hall writes:

Dudley's a smart guy who knows his stuff. However, most people hear economic activity will be slightly higher and turn that into we will be better off. But that's not true. We would only be better off if wealth would be higher in the years after the natural disaster than it would be without out.

Jon Murphy writes:

I'm glad you posted this. I had missed the subtlety in his comment

Jerry Brown writes:

I think you explain this very well, and I am also glad you posted it. But what if the economy happened to be in a recession at the moment the window was broken, and what if the baker was insured for his loss by say, the government? Not saying that is optimal, but does the economy really take a hit if otherwise idle resources were employed to repair the window? Could it be said that it is at least 'less bad' under those circumstances?

I'm also not saying that right now the US has lots of unemployed resources- Irma is going to be bad no doubts about it. Probably good for carpenters in Florida- assuming they aren't spending all their time fixing their own houses though.

Jerry Brown writes:

On the topic of the broken window- I know it just doesn't work that you can pick some unemployed person off the street and hope they know how to fix a window. But what if you got lucky and happened to get an economist? Then everyone is happy- the baker can assume the pane of glass, the economist gets paid, the economy booms, and no real resources get used up!

john hare writes:

I do construction in Polk county which is in Irmas' path. We got hit in 2004 by three storms that were cat 1 or cat 2 hurricanes when they passed through. At that time there were far more qualified people doing construction in this area. As much trouble as it is getting work done before the storm, I expect it to compound the problems when it comes through tomorrow.

Many broken windows are going to stay broken for much longer this time leaving many people with a much lower quality of life for many months. Jerry is right about picking up some unemployed person not working, as most of us have been trying to hire for quite some time.

More money helps move some people back to the trades for a while. More money does not instantly create tradesmen. It also doesn't convince many people that they should work hard in the sun or learn a trade to begin with.

Scott Sumner writes:

Disasters also reduce the capital stock. So the net effect on GDP is ambiguous, but in my view it's usually negative

Tim Worstall writes:

When I try to explain this I may well be wrong in detail but I think I manage to get the point across.

GDP is value added the economy. Repairing things adds value, so, repairs increase GDP.

NDP is net (or NNI, etc). This takes account of the destruction as well as addition.

More normally we think of N as taking into account depreciation etc but it works just as well to think of it as damage, pollution etc.

Our problem is that the N takes a year or two to work out, thus we're rather stuck with the G as a management tool.

Julian writes:

If a number of windows are suddenly broken then the the increase in the business for firms in the “window business” could lead them to hire and train additional people to do the job. This additional training, in turn, could have positive externalities.

For example, learning how to repair a window increases you human capital which would be beneficial for a firm that hires you in the future (say, because it has to pay less to train you on the new job). Now, depending on the size of the externality, a broken window could be a net gain on welfare.

Of course, there are more efficient ways of reaping the positive externalities of knowing how to repair windows (one way is for the government to subsidize learning window repairing). Also, given that what is beneficial is not the act of breaking the window but knowing how to repair them one could ask what prevented the market from reaching some private arrangement (i.e. why did the Coase theorem did not apply).

R Schadler writes:

I think the confusion is that "increased economic activity" entails the assumption that this is a good thing, despite the insertion of "unfortunately."
Increased job opportunities for the unemployed sounds like a good thing in itself. But maybe not if a plague or genocide killed half the population. Feldstein in yesterday's WSJ addresses a somewhat similar point, regarding the accuracy and relevance of economic tabulations in a vacuum.

Tim G writes:

I've often wondered if one of the laws of thermodynamics applies to the laws of economic activity. Energy cannot be created or destroyed. It only heads toward increasing entropy. Money is fictionally created, is it not? Yet it is not usually destroyed. It creates value in a new form to another owner. I enjoyed reading the book on the natural economy where the author compares the falling leaves (businesses destroyed, windows broken) to creating life in a new returning state.
Where businesses fall apart, others can pick up the pieces of value (equipment) for a much cheaper price. The value is not totally lost. There is usually a new beneficiary from whatever has cost the broken windows.

JFA writes:

"I would expect that by the time we get to the end of the year and early 2018, the transitory negative effects of this storm I think will be over and we actually will start to see some of the benefits of the rebuilding efforts in terms of boosting the economy." This Dudley quote does muddy your analysis just a bit. Yes, he did use "unfortunately" in your preferred quote, and it is correct to suggest that there will be a measured increase in economic activity. But the quote that I (and Boehm) provide (while not a smoking gun) does appear to suggest that he sees "benefits" in the rebuilding. If the lift in "economic activity" were "unfortunate", I'm not sure "benefits" would be the correct word.

All that being said I don't think it is clear from Dudley statements whether or not he is committing the broken window fallacy.

Yaakov Schatz writes:

David, I think you are missing an important point. Dudley was not talking to students in an advanced economic seminar but was talking to a newspaper in a manner which is hard to interpret other than purposely giving an ambiguous answer meant to confuse the public.

This happens too often, and economists should not be explained away in such circumstances.

David R Henderson writes:

@Yaakov Schatz,
Dudley was not talking to students in an advanced economic seminar but was talking to a newspaper in a manner which is hard to interpret other than purposely giving an ambiguous answer meant to confuse the public.
Please identify, for other readers and me, where you think the ambiguity is.

David R Henderson writes:

@JFA,
Yes, he did use "unfortunately" in your preferred quote, and it is correct to suggest that there will be a measured increase in economic activity. But the quote that I (and Boehm) provide (while not a smoking gun) does appear to suggest that he sees "benefits" in the rebuilding. If the lift in "economic activity" were "unfortunate", I'm not sure "benefits" would be the correct word.
Unlike you, I think “benefits” is the correct word. When your wealth falls and I observe you voluntarily working harder to build it up again, I say that you are benefiting from that work.
It’s important to recall the broken window fallacy. The fallacy is in seeing the broken window as being good. If Dudley had even hinted that the hurricane was good, he would be committing the broken window fallacy. He didn’t even hint that. He’s saying that people’s reactions are good.

Bob Murphy writes:

FYI for those interested, I tried to clarify each move in the debate's progression, starting with Bastiat and ending with Scott Sumner's caveat above.

Comments for this entry have been closed
Return to top