As the eastern U.S. tries to dig out from under the devastation, I thought it might be useful to comment on the economic consequences that a storm like this could have.
One school of thought holds that our economic problems are entirely a deficiency of aggregate demand. According to this view, if we could just get consumers (or the government) to spend more, our problems would be solved. The classic extreme example of a policy that might boost GDP according to this perspective would be if the government were to pay workers to dig a hole and fill it back in.
That extreme case leads one naturally to ask how it could possibly be better to pay people to do dangerous and hard work digging a hole compared to just sending them checks to stay comfortably at home. The answer turns out to be that, if we pay for hole-digging, we not only have provided paychecks to workers, we also have produced a newly-filled-in hole, which the convention of our national income accounts treats as newly-created infrastructure. The economic value of that is taken to be whatever the government paid for it, and thereby adds directly to GDP.
In the present situation, there is a tremendous amount of real work that needs to be done– subways need to be drained, debris needs to be removed, fallen trees need to be cleared. Lots of hard work needed and jobs created in getting that done. More importantly, when the work is finished, we’ll have not a filled-in hole, but instead something really valuable– subways we can ride to work, roads that we can drive on.
Except, we had all those same good things two weeks ago.
In the mean time, trying to get by without those means people who can’t get to work, commissions on financial transactions that won’t be earned, sales that didn’t get made. The Wall Street Journal reports:
Along the East Coast, some restaurants, car dealers and other retailers are likely to see sales slip. Many workers aren’t being paid for lost hours. Shipments of goods through eastern seaports and airports are being delayed….
“It’s not catastrophic but it’s not trivial,” said Gregory Daco, an economist at IHS Global Insight. “You’re seeing major disruptions in trade flows, both on the import and the export side. You’re seeing disruptions in terms of construction, in terms of manufacturing, and also in terms of services—movie theaters closing, your local grocery store may not be open. It’s a number of combined shots that overall affect the U.S. economy.”
Those employed in the New York financial district had been constructively contributing to GDP last week, despite the weak economy. When people in good-paying jobs are prevented from getting to work, it has to be costly in any sensible economic model. This is particularly the case if it takes a long time to get the infrastructure repaired or if alternatives to New York brokers and transporters gain market share as a result of this episode.
One parallel to consider is the devastation from Hurricane Katrina in 2005. In addition to the short-run dislocations, this ended up causing lasting damage to offshore oil-producing infrastructure. An optimist might have thought this would create all kinds of new jobs trying to rebuild. The actual experience was not so cheerful.
 |
The Wall Street Journal reports that IHS estimates that Hurricane Sandy could reduce the 2012:Q4 U.S. real GDP growth rate by 0.6 percentage points at an annual rate. I’m not sure how one comes up with that kind of number.
But I am persuaded this was not a good thing for the U.S. economy.
The two measures – wealth and income – often conflict. It’s kind of stupid to say that paying to rebuild destroyed wealth is a good thing. But then look at how we measure performance generally: how many Wall Street firms tell you what their performance has been over x period? All of them. But they do that after they’ve lost a pile of money and then they show growth and profit – and pay bonuses based on profits made even while they’re still in net worth below where they were.
Can a place be better off if rebuilt after disaster? More likely the rebuilding won’t fix the actual problems. If the infrastructure added is an improvement, not just repairs, then it might stimulate further growth. But that doesn’t happen often. It never happens in less developed countries.
Ah! So, if instead of simply sending check to people, we also required them to write a little haiku on the back of the check in addition to their endorsement, from an national accounts point of view their poems would be worth whatever the government paid for them….
Thank you, Thank you, professor!!!
It happens every time there is a disaster. Self-appointed experts began to repeat the Broken Window Fallacy over and over. Just this morning as I was listening to Bloomberg on the radio I head a female “economist” talking about how this hurricane would stimulate the economy and how good it would be for GDP.
Professor, also thank you for this, “…if we pay for hole-digging, we not only have provided paychecks to workers, we also have produced a newly-filled-in hole, which the convention of our national income accounts treats as newly-created infrastructure.” It made me laugh right out loud!
You might want to send this to Paul Krugman.
The problem with using the Katrina aftermath as a basis for projecting is that the devastation was so great and the recovery so mismanaged after Katrina that it resulted in a substantial decline in the population of the whole state. The Louisiana population basically mirrors your graph over the same time period. There’s no reason to think the NYC region will see anything similar because while the destruction is significant, it won’t cause hundreds of thousands of people to move elsewhere. Also, oil production probably permanently declined because it wasn’t economically justified to rebuild the infrastructure needed to restore the flow of oil from older low-producing fields that were only producing because the infrastructure had been paid for years ago when the fields were new. There’s no reason to think financial markets or advertising or fashion or anything else that NYC does will be any less productive 6 months from now, so output is very likely to return to normal.
What about the idea that insurance money gets spent and then multiplied. Insurers sell investments (bonds) to pay claims. This money is spent on rebuilding (with what multiplier?). Insurance companies (or stock prices) are not effected because such disasters are actuarily priced in (as long as it is not something really out of the ordinary). Would this not have a positive affect on the GDP of the affected area?
Hey, it works for China.
http://www.theepochtimes.com/n2/china-news/frequent-bridge-collapses-help-boost-chinas-gdp-291060.html
One thing to keep in mind in the current case is that rebuilding will rely on excess capacity and be financed out of savings that have been earning near zero return. The jobs opened up to do the rebuilding will draw on a very large pool of unemployed workers.
This is just about the most advantageous circumstance imaginable for rebuilding after disaster.
@TF:
“What about the idea that insurance money gets spent and then multiplied.”
The implicit assumption in your statement is that the insurance companies have those insurance funds in mattresses, just sitting there, waiting to be paid out.
This is, of course, wrong. Insurance companies invest their funds, while they do not have any payouts. When they do, such payouts are retrieved from the insurance companies’ investment funds. Such payouts do not generate investments any more, and this is the opportunity cost of the insurance payouts.
For details, I refer you to Frederic Bastiat’s “That Which is Seen, and That Which is Unseen”. It is 160 years old, but still very, very modern.
‘The jobs opened up to do the rebuilding will draw on a very large pool of unemployed workers.’
I’d be very surprised if that turns out to be the case, given the power of the public sector unions in New York and New Jersey. It certainly wasn’t for the 2009 ‘stimulus’ spending. That money was just vacuumed up by the entrenched work force.
Mangred,
No, I am assuming insurance companies premiums are invested in fixed income securities and those are sold to pay claims. Of course this may cause somewhat higher rates for borrowers at the margin? Does this offset the growth from spending the money is my question.
Simple back of the envelope calculations help. First, if we take a production function approach Y=f(K,L) we have very little loss of life [and fewer routine deaths in traffic accidents…] and most of the destruction was to residential areas. So short-run disrutpions aside, there’s no impact whatsoever though if you’re living on the second floor of a water-damaged house you certainly see a shift to very substandard housing [which ought to show up as somewhat levels of imputed rent in the national accounts, albeit too many decimal points out to be noticeable].
From the demand side, some sales are put off (delays in buying a new car) without affecting levels — an early November sales turns into a late November one. Elected surgery gets elected later. So this shows up as a short-term disruption but with an offsetting rebound effect. Then there is the construction discussed by others, at low opportunity cost because of prior excess capacity. But note that it does add to productive assets, nature already dug that pothole for us, and someone is going from a house that may be impossible to live in to a repaired one providing housing services. Better that the wealth had not been destroyed, but restoring that wealth does represent a net addition.
Finally, to put numbers on it that a freshman Principles student or a journalist can follow:
if we lose 5 days output=final sales=value added in a region with 60 million people, that’s (rounding) 1/50 of a year and 1/5th of the population, or 1/250 = 0.4% of GDP. I chose numbers on the high side, and assumed away rebound effects, and assumed that all the carpenters and linemen would have worked just as many hours absent the storm. Make such corrections and you get under 0.1% of GDP. That’s a statement of the size of our economy, not a statement that the losses to those along the coast aren’t real.
@TF:
Even if insurance companies invest in fixed income securities, they give up return. Yes, affected people by the hurricane will get new stuff and there are people getting paid to produce that new stuff (builders, contractors, etc). Just like in Bastiat’s Broken Window story, where the glazier gets paid to set in a new pane of glass; but there is a loss to the economy, because that money could have been used for something else, even if it is fixed income securities – because those fixed income securities finance *something*, but since the insurance companies make payouts, that *something* that these fixed income securities finance, will not be financed anymore.
Again – read Bastiat.
I propose that we create a cabinet position in the executive branch and call it Dept. of Broken Windows (and GDP Dreams) and hire thousands of eCONomists to spend the next working lifetime devising theories and models to show how we can increase GDP by hiring the unemployed youth to, ahem, dislocate windows throughout their communities.
Beyond that, technical schools and universities around the country could create new disciplines to train for occupations such as window dislocation specialist, glass enhancement technician, window remediation design engineer, and personal rock-throwing trainer.
The labor division could be highly differentiated by those who use rocks, baseball bats, hammers, or Molotov cocktails to dislocate window glass.
Workers could be further differentiated within labor union classifications by the color of the masks covering their faces; that is, red, black, or some combination.
We could create multiples more employment opportunities by creating advanced degrees to train the trainers to train people to dislocate windows and replace them.
No doubt Fortune 25 firms could afford to create endowed chairs and scholarships at Ivy League schools to further the program and promote faster US uneconomic growth as part of the American Dream of Broken Windows.
Manfred,
The Broken Window Fallacy is so much a part of modern thinking that even college economics professors teach it. It is refreshing that Dr. Hamilton attempts to set the record straight, but then it is disheartening that so many here repeat versions of the fallacy. Sadly, Bastiat is not taught today because too many have gotten “too smart” to listen to an old theory.
The difference is that for Katrina, the U.S. was already near full employment. In fact, it was the peak of a housing construction bubble, so suddenly adding more housing demand would be expected to have very little impact on unemployment. You simple shift fully employed workers from one area of the country to another and increase their wages.
Stimulus effects are quite different during a recession when there is high unemployment.
Ricardo, nowhere did I see Prof. Hamilton endorse the “Broken Windows Fallacy” in this case, but he will have to speak for himself.
There are a several caveats to the Broken Window Fallacy:
If the replacement costs the same, but saves money, i.e. in the case of the window, the new one is more energy efficient; causing the owner to spend less on heating/cooling.
If the new window costs about the same, but is more appealing, increasing the value of the building.
Only in the case where replacement cost and benefits equal or are more than the original does the Fallacy apply.
Bastiat lived in the first half of the 19th century, while his observation is broadly valid, the concept of economic improvement by substitution wasn’t part of the paradigm at that time.
Mike Smitka First, if we take a production function approach Y=f(K,L)
Let’s take that a little further. If we assume a simple Solow growth model like one that you might find in a standard macro textbook like Romer’s, then the production function expressed as output on the vertical axis and investment per unit of effective labor intersects another curve that represents breakeven investment and is the sum of labor growth, knowledge growth and depreciation. And in the case of hurricanes and earthquakes the right way to think of these events is as exogenous depreciation shocks. There will be some temporary increase in investment to offset the depreciation shock.
What I don’t quite understand about JDH‘s position (at least as I understand it), is that he correctly argues that Hurricane Sandy reduced our wealth even if it increases GDP. And I think this gets to a larger point about the difference between gross domestic product and net domestic product. What we really value is net domestic product, so even though Sandy might increase the flow variable GDP our stock variable wealth will be lower as will the flow variable NDP. But strangely JDH does not always apply this seem reasoning when it comes to the extraction of natural resources, such as oil. Extracting oil increases GDP, but unless it is being extracted at an optimal control rate the extraction (which is really a special case of depreciation) might be reducing NDP. I’ll grant you that our NIPA accounts don’t really lend themselves to correctly evaluating the exhaustion of natural resources as depreciation, but I don’t think that makes the problem go away.
I don’t think the depreciation analogy works here. A fully depreciated asset is still productive. A destroyed asset is not.
Remember, depreciation is a non-cash item. In other words, a negative depreciation shock decreases my paper wealth. A shock that destroys my assets decreases my real wealth.
We are now into the third day of Sandy here in Princeton. The town, beyond the downtown and campus remain dark. The power lines in our street are compromised in at least four places. I doubt we will have power for several days, perhaps a week or more. Schools remain closed, and will almost certainly remain so for the balance of the week.
The primary shortage commodity is wifi. The town library is the refugee center. The staff were begging users to limit their wifi usage to a single device, as the wifi system at the library can handle only 500 users at a time. At least 2000 people were in attendance at the library, when on a normal day it might have a tenth as many. By 9:30 am, logging into the Internet was impossible. Power was also in comparatively short supply; I took my own power strip; I was charging four iPhones, an iPad, and a laptop simultaneously. Every outlet was otherwise taken.
Notwithstanding, someone walking in town would hardly have noticed a problem. The streets were lit, the restaurants and shops open, the streets packed with cars and pedestrians. The refugees were enjoying a day around town.
Great point about NDP vs. GDP. Similarly, extra depreciation from accidental destruction of fixed capital reduces national income, other things equal.
WRT insurance, P&C insurance output is measured as premiums less normal claims, averaged over some long earlier period. Difference between actual and normal claims is treated as a transfer to policyholders.
It is always soothing to read good logic and tempting to match the pair arguments as laid on this post.
K.Marx was saying that robbers were useful to society, as they would be security enhancers and would drive commensurate spending on security from the same society. There is as well a multiplier content in those expenditures.
On banks, it could be interesting to assume a bank underwriting loans and CDs and making promises only that means guaranties. When operating during working days it would be found that it makes less money than when it is closed since fixed expenses are largely decreased and loans losses the same.
Come to think about it, why are those sound principles not recommended to emerging countries.
Omitted to add that when banks are closed but still accruing commissions on loans and guaranties,booking less loans losses and consequently showing better performances.
The Laffer tax curve conclusion and theory becomes an irony
Joseph,
Dr. Hamilton’s reasoning supported the foolishness of the “Broken Window Fallacy.” Do you believe we should break windows to bring recovery?
New Jersey Power Restoration
The authorities are still trying assess damage and repairs, but we have some comps this morning.
ConEdison (NY) estimates 100,000 powerlines down. NJ is probably a similar order of magnitude. PSE&G, the NJ utility, has perhaps the equivalent of 500 2-man crews working. Figure that they can fix two power lines a day on average.
Do the math, and it’s a 100 days or so to put everything back in order.
Our house in Princeton is probably below the top 10% in priority, but probably in the top 33% (we’re essentially in town). So a better case scenario returns power to us in more than ten days (around next weekend) and probably less than three weeks.
We’ll know more in the coming days, but we could be offline for quite a while.
Our good Pr Hamilton has for the sake of pedagogy borrowed so many costums.One day he should be left alone, stranded in the sand bank of contradictions waiting for an helicopter
Sorry to hear that Steven. I hope things are back to normal for you as soon as possible.
tj I don’t think the depreciation analogy works here. A fully depreciated asset is still productive. A destroyed asset is not.
You are thinking like a business tax accountant advising a client, not an economist. Learn the difference. In economics a fully depreciated asset has zero economic value after disposal costs. Depreciation refers to the value of the capital asset that was consumed in use. If you just allowed an asset to depreciate without restoring any of its value (i.e., zero maintenance), then the total depreciation would equal the value of the asset, which would also equal the value the asset adds to the production process.
Bravo to Prof. Hamilton for the common sense observation that destroying wealth and rebuilding it is a zero sum game, and recognizing that GDP that goes to malinvestment is waste.
Further on the “broken window fallacy”
Maybe another way to convey this, especially to students:
“Foregone consumption is forever gone.”
Of course in the classroom I might then remind students of lifecycle / permanent income models, but return to note that in a disaster consumption smoothing isn’t complete, and real wealth losses have consumption analogs. If your window is broken, well, it’s November and with lots of windows broken, you probably aren’t first in line. And insurance may not cover it. Etc.