Steve Landsburg, Destroyer of Worlds
I stand in awe of Steve Landsburg’s mind, though sometimes I question the uses to which he deploys it. For example, I’m sure the math is correct (given his assumptions) behind his latest, counterintuitive result:
[E]ven in a thoroughly non-Keynesian world where markets work perfectly…and recessions cure themselves, we might still want [a destructive] hurricane.
Or, because we can’t always call forth hurricanes when we need them, we might want our government to simulate their effects by diverting funds from useful to destructive spending projects — or just occasionally showing up at people’s houses and trashing their furniture.
Here’s why: Hurricanes make us collectively poorer. When we’re poorer, we work more. When we work more, the government collects additional income tax revenue. But — taking total government spending as given — the government can’t continue to collect additional revenue forever; sooner or later it must lower tax rates. (This assumes we’re on the good side of the Laffer curve…) When tax rates fall, labor markets work more efficiently. So much so, in fact, that the efficiency gains can more than compensate for the initial destruction.
Now I know a bunch of you are going to go nuts over the premises in Steve’s argument above, most notably the idea that the government would hold total spending constant in the face of an influx of new revenues. But still, I like to play new games, and I’m willing to abide by Steve’s rules. Here’s what I think:
==> A previous, super-counterintuitive Landsburg result is that we should (he claims) encourage more people to sleep around, in order to reduce the transmission of HIV. But, assuming the spread of HIV is akin to a hurricane and stimulates more spending on health care treatments etc., I now conclude that Steve’s new result proves he was wrong about HIV, just like I said at the time.
==> Let’s drop the cuteness and focus on Steve’s new result. In conventional Pigovian terms, what’s going on here is that there is a “market failure.” When you’re deciding how much to work, you’re considering the after-tax return (and other advantages such as developing your human capital) in light of your forfeited leisure. But if the government indeed holds spending constant, such that more work on your part leads to lower marginal tax rates, then you are conferring a positive externality and thus aren’t working enough.
So, in order to correct this market failure, the government should either (a) subsidize you–with the money for the subsidy being raised ideally through a new lump-sum tax–for working more, on the margin, or (b) should tax you (on the margin) for not working.
I’m not going to check the math, but it would be neat if the Pigovian optimal solution turned out to exactly counterbalance the distortions of the income tax, so that we replaced it on net with a lump-sum tax that brought in the same revenues.
There is an even easier way to blow up this latest iteration of the age old consumptionist fallacy.
This premise: “When we’re poorer, we work more. When we work more, the government collects additional income tax revenue.”
is not only internally flawed, but ignoring the flaw, it is not a universal law. It could apply to someone in particular as an accurate description of their personal values, but it is not something that MUST apply to every actor, the way the law of marginal utility MUST apply. I for example work harder now than I did when I was poorer in the past. I know this because my desire for wealth has substantially increased, and with a higher desire for wealth, the more I want to work. It is not the case for me that being poorer makes me work more. So Landsburg’s premise is really just a reflection of his own personal values. HE might work harder if he were poorer, but that’s because his personal value scale is such that this is his behavior.
But the statement is internally flawed, so we don’t even have to consider it as possibly valid. It is flawed because it does not follow from “working more” that more net output is generated from which a government “collects more tax revenue.” Here Landsburg confuses labor with output. I could work “harder” but generate less output if my work is not efficient, e.g. I have less capital to work with, or I “work” with less emphasis on satisfying my customer’s demands, and hence get paid less. For suppose that a hurricane was so massive and destructive that it wiped out the entire capital stock. Output would collapse, and people might even end up “working harder” as per Landsburg’s premise, but does this mean the government can collect “additional tax revenues” If by “income” we mean dollars, then maybe, if the government inflates. But in real terms, not a chance.
I think your first objection is a bit off, MF, because presumably it’s an “other things equal” deal. E.g. back when you thought one way about wealth, at that moment you probably would have worked harder if you suddenly became a little poorer.
But your second one sounds good to me, I’m guessing Steve must be assuming a sufficiently small disaster–but I haven’t read his actual paper, just his blog post.
“I think your first objection is a bit off, MF, because presumably it’s an “other things equal” deal. E.g. back when you thought one way about wealth, at that moment you probably would have worked harder if you suddenly became a little poorer.”
Well things are never equal in this respect. It is a model that is supposed to apply to real people who live in a temporal world, right?
But let’s suppose “all else equal” does apply. Are we saying that there is no way, no possibility, that right now, a wealthy, hard working man who had all his wealth taken away, say from some freak natural disaster, that “all else equal” he may become jaded/depressed and become some loafer relative to his former hard working self?
I mean, the way Landsberg states his case as a universal law, we’re supposed to treat that as impossible. But really it’s a value judgment. Landsburg is treading into objectifying subjective preferences.
“But your second one sounds good to me, I’m guessing Steve must be assuming a sufficiently small disaster–but I haven’t read his actual paper, just his blog post.”
I was thinking that too, but I used the extreme to make the same principle clearer. Even if it was a “small” hurricane, that destroys real wealth, namely capital wealth, the productivity of labor will decline. The government could only collect more taxes here if it increased the tax rate.
I don’t see how more output could come out of less capital even in a context of some folks choosing to work harder because they’re poorer (while others work the same or less hard).
Suppose I used to work with a steam shovel, but a hurricane destroyed it, and all I have left is a regular hand shovel. Landsberg might consider me to end up “working harder”, because I develop more blood sweat and tears and so forth, but my output won’t be as high.
I think what needs to be emphasized here is that no matter how wealthy we get, our minds are constituted in such a way that we continually act to further our interests regardless.
I think Landsberg might be thinking that people becoming wealthier naturally makes them lazier, but lazier in such a way that their real output declines, which is implied by “becoming poorer makes us work harder” where of course “harder” somehow implies more real output to be taxed.
“I mean, the way Landsberg states his case as a universal law…”
But he stated no such thing. You just inserted that yourself.
I realize he stated no such thing. That’s why I said “the way” he states his case.
“When we’re poorer, we work more.”
There are no explicit caveats to this, no contingencies, no exclusions to a rule. It says that if we are poorer, then we work harder.
That to me sounds like a universal law, and I think anyone who isn’t more concerned with disagreeing with me for the sake of disagreeing than they are with honest analysis, would interpret it the same way.
Whether intended as a universal law or not SL was clearly applying it as a wider principle to this hypothetical case, not arguing that in this particular case it will happen to be true.
Just to play along: If the positive externality of working more is a lower tax rate, or increased productivity via a lower rate, wouldn’t the government negate this eternality with its tax-funded subsidy? Or are we to assume the new tax would be levied on something other than productivity–something less distortionary than an income tax–thus providing a net gain?
Right Tony, I said the subsidy should be paid out of a new lump sum tax.
My apologies, Bob. My comment was the product of a thought carried over from my original reading of Landsburg’s post. Should have read your commentary more carefully.
Now all we have to do is find a real world example of a lump-sum tax that is minimally distortionary yet sufficiently robust to account for the revenue lost through the income tax cut. Or we can simply capture and harness the majestic powers of a unicorn.
It would have to be a horn tax on the unicorn. Otherwise we would discourage it from running.
If working more and having higher productivity actually led to a lower tax rate, then we’d see a historical trend where tax rates asymptotically approach zero as productivity skyrockets. That’s… the opposite of what’s happened.
Government is not going to reduce tax rates simply because it finds increased revenues are coming in anyway…
Ahem: “taking total government spending as given”
My point was that is an entirely unrealistic assumption. It’s like saying “taking it as given that we live on the sun”…
Why even introduce goofy assumptions about government spending to get more efficient labour markets? Why not just say, “When people are more desperate, they try to find more ingenious and efficient solutions. Therefore, conditions that make labourers desperate (like trying to rebuild their lives after a hurricane or mob visit) will result in more efficient labour markets”?
Where there’s a will there’s a way, and where there’s a whip there’s a will.
With apologies to that hobbit guy, or other.
Great point. By analogy, it certainly seems like Landsburg would welcome slavery, or at least contemplate it, for its efficiency gains.
I find it difficult to invest thought into a what-if question where the premise goes against all experience and political nature. Here, the premise is that a government will freeze spending and then lower tax rates in an attempt to avoid having too much revenue.
How can one begin to deal with that question, with no experience of any such government or any such people to guide the thought? What is the purpose of working out what the world would be like if governments and people were different from who they are now, other than sci-fi entertainment?
Math is fun, and one can pose systems to work out. But, the results from unreal premises tell us nothing about reality, even if the quantities are labelled with words taken from reality.
“Here, the premise is that a government will freeze spending and then lower tax rates in an attempt to avoid having too much revenue.”
No, no, no: he simply said “taking total government spending as given.”
He is not saying anything at all about how governments really behave.
Those are the same things worded differently
To Gene Callahan,
Landsburg says this in his post: “When we work more, the government collects additional income tax revenue. But — taking total government spending as given — the government can’t continue to collect additional revenue forever; sooner or later it must lower tax rates. (This assumes we’re on the good side of the Laffer curve, where the way to collect less revenue is to lower rates, not raise them.)”
Why would the government have to lower tax rates to reduce its income? Only if its spending is “given”, that is set at a given amount. If the government can spend the additional income, as it always does in reality, then there is no reason to lower tax rates to reduce its income.
You repeat my point. Landsburg assumes a government we have never seen. A government has sometimes reduced tax rates to promote growth, gain votes, or avoid killing its economy, but never merely to limit its income. If a policy of intentional destruction actually raised government revenue, then governments would be tempted to do it.
When one starts with an assumption that does not represent reality, one can derive lots of things that also don’t represent reality.
EasyOpinions
Slightly OT: Call me crazy, but the more sex = safer sex argument never made sense to me, even with the favorable assumptions, and for (very simple!) reasons that I’ve never seen pointed out, AFAIR.
The simple flaw is that adding non-infecteds to the pool can only increase the number of infected/non-infected pairings, which is the key parameter. Landsburg’s argument seems to be that adding non-infecteds increases the probability that anyone (non-infecteds especially) will pair with non-infected, which is true, but irrelevant.
You might as well extend his argument to prove quarantines don’t work as a disease containment measure. After all, letting people out of the quarantine will decrease the probability that one of the not-yet-infecteds will come into contact with an infected! Woo-hoo!
The problem there (as, it seems, with STDs) is that you massively increase contact between infecteds and non-infecteds! It’s irrelevant that the not-yet-infected in the quarantine w
Even in the canonical example of switching from “every man pairs with his wife and Francine” to “every man pairs with his wife and (Francine or one of the other wives)”, you’ve still strictly increase the non-infected/infected pairings.
So … what gives?
Sorry, 3rd-to-last para should end “… in the quarantine would have fewer of their encounters be with those infected”.
I posted a whole class of counter examples, but Steve ignored it. Here’s a simple one: two complete subgraphs, disjoint, one fully infected.
Steve’s point is that under some conditions by increasing the number of couplings you can increase the likelihood that the infected screw the infected, and the clean screw the clean. Comapred to what would happen otherwise you can reduce spread that way. Even if it holds for a large class of special cases it won’t hold true for long if there is even a very small bit of crossover. This is just Hardy-Weinberg in an unusual guise I think.
Ken B: Your summary of “Steve’s point” is 1000% incorrect. A far more important point is that it’s better to have the virus passed to people who are relatively unlikely to pass it on. And a far more important point yet is that this is a model of maximizing behavior which rules out your digraph example. If population B is fully infected, then members of population A won’t choose to mate with them. So you’ve found an extreme example in which the effect goes to zero, not an example where the effect is reversed. You also seem to have missed the point that the counterintuitive effect on *incidence of disease* depends on parameter values, while the counterintuitive effect on *welfare* is independent of parameter values. It’s the latter that’s both most striking and most interesting,. Read the book!
Steve
If A and B are complete disjoint graphs then “more sex” — by which you mean more partners per person not more coition per person — must entail a member from A and a member from B. That is the only way you can get “more sex”. In that case the contagion spreads from B to A. So this is a counter example.
I also — it was long ago — gave some conditions on rerlative rate of infection per coition. I gave the example of an undetectable incurable disease with a 100% gestation rate and long incubation period, and situations where that can lead to the opposite of your prediction.. (And I noted wickedly that there is commonly believed to be such a disease: sin.)
Let me elaborate a more telling counterexample on the same lines *that confroms to Kremer’s recommendation*. We have a lot of small groups of 2,3 or 4 persons who have sex within the group. So married couples, faithful small group polygamy, couples where one cheats with a faithful corespondent, etc. These folks are all free of the contagion. And everyone else is part of a bath-house culture, very promiscuous. And very high rates of infection. Now Kremer specifially wants the memebrs of the *first* category to step out more. Well if they step out with the second category, and transmission rates are high, and the second has a high infection rate, then Steve and Kremer are wrong. This is a counter example.
It should be clear that the result depends on the rate of transmission, the rates of infection of the different groups, and the detectability of the infection. Some will amke Steve right, some wrong. My guess is most will make him wrong over the long term, because the conditions where he is right will erode.
It’s not like I’m the only one who isn’t convinced http://gilkalai.wordpress.com/2008/05/29/is-more-sex-safe-a-book-review/ This guy is a combinatorics expert of some note.
“. You also seem to have missed the point that… the counterintuitive effect on *welfare* is independent of parameter values.”
I was addressing the claim made, which was *safer*. It is quite clear that is the point Silas was making.
Ken: I was addressing the claim made…
The claims made were that a small increase in the number of partners on the part of the relatively cautious a) increases welfare regardless of parameter values, b) reduces the spread of disease for some parameter values and c) reduces the spread of disease for realistic parameter values. Which of those claims did you think you were refuting?
Let’s start with this one:
“this is a model of maximizing behavior which rules out your digraph example. If population B is fully infected, then members of population A won’t choose to mate with them.”
In order to have what is meant by more sex they MUST. That is why I made the second, more interesting example of a large population of naerly monogamous uninfecteds. That’s a not implausible model when we are talking about a new disease.
Note that I have not ever said there are no parameter values where your effect won’t happen. But your characterization HERE does not match what you said on your blog many moons ago where you asserted the claim for ALL realistic parameter values.
I also note that the conditions where you are right can and I think generally will deteriorate into conditions where you are wrong. This is easy to see in my examples. Over time the bath houses get higher infection rates.
Once tha bath house population is fully infected more sex can only increase the spread of the disease of course. That is why I say the conditions deteriorate.
Ken B:
what you said on your blog many moons ago where you asserted the claim for ALL realistic parameter values.
If by “the claim” you mean the assertion that the infection rate goes down, then your memory is playing tricks on you. This isn’t the place for an extended discussion (witness the extreme narrowing of the columns), but in any event I’m not interested in defending a misremembered assertion that I never made.
Steve
Oops. I checked the post I was thinking of, about the python simulation, and I have *completely* misremembered it.
My apologies.
This all assumes money is government fiat with no inherent value. I can’t think of a better way to revere the state apparatus, than to canonise its tools, deficits and taxation. One MMT’er said to me that without a currency monopolist, society would resort to barter. He was serious!
“This all assumes money is government fiat with no inherent value.”
So, gold has “inherent value”? Mises would be very surprised to hear about this!
It is rather unfortunate that laymen confuse preexisting market value with “inherent” value.
Well, yes and no. Some goods gradually become more traded and thus BECOME the money.
This video explains it well:
Smashing Myths and Restoring Sound Money | Thomas E. Woods, Jr.
http://www.youtube.com/watch?v=HAzExlEsIKk
This guy is nuts. “When we’re poorer, we work more.” Is he serious? Has he never heard of the welfare system? How about food stamps? WIC? All these programs allow the poor to work less. And besides, people in the upper middle class are some of the hardest working folks out there from what I’ve seen.
“Has he never heard of the welfare system? How about food stamps? WIC?”
Have you never heard of ceteris paribus, Neil?
Global warming will surely thwart the coming crisis then.
OT: Mike Norman isn’t happy…..
http://www.youtube.com/watch?v=CuIWUArl2g4
That guy is creepy.
The main problem is his baseless assumption that the supposed “efficiency gains” will more than compensate for the initial wealth destruction. If a hurricane wipes out $1 trillion of wealth/capital, and the the economy becomes 0.1% more efficient, is it really worth it? What metric is he using to compare these costs/benefits?
Also wouldn’t the efficiency cancel out once we return to baseline? If people truly work harder when their poorer, then it follows that they work less when they’re richer. So after months/years of working harder, tax cuts, etc…, people will get richer and want to take a break. Then the government will collect less revenue, have to raise taxes, muddle up the labor market, and the whole thing’s a wash.
An interesting thought experiment but seems to me like he didn’t fully think this one through
Since we can’t call forth hurricanes at will, and since the whole point of the thing is that poorer people work harder, couldn’t we simulate hurricanes by raising taxes? Raising taxes make people poorer than they otherwise would have been just like a hurricane.
Here we get to the glorious conclusion. Raising taxes, by making people poorer cause then to work harder, thus lowering taxes.
With that, I leave the field.
What if the hurricane was destructive enough and just happened to only destroy all government buildings and employees. Tax rates would really drop off then, and we could still work really hard. Or not, we could just soak up the leisure good and enjoy our long sought after freedom.
Oh, Collin. Didn’t you know? You work so as to keep working, not to “improve your living standard” or anything.
Bob, there is a very serious problem here that deserves your attention: “I stand in awe of Steve Landsburg’s mind, though sometimes I question the uses to which he deploys it…”
This, I suggest, is a seriously flawed picture of the mind: it is like a hammer, and I can deploy it to strike nails or heads. I say the mind is not like a hammer, but like hammerING: it is nothing at all but its activities. In that Landsburg is thinking incoherently, his mind is nothing to be in awe of!
How very Heideggerian of you
I’ve actually only read a few pages of Heidegger! But a number of other thinkers have made similar points.
Fichte especially.
Ah but that mind must lead the body to making the objectionable blog post. So perhaps Bob should be concerned with the the the use to which his mind deploys Steve Landsburg!
If the mind is like hammering, then all thinking is incoherent since the thinking would not be purposeful.
If a computer appears to think, it’s only because we know that someone programmed a response to an input. So, while the computer, itself is just doing what it’s programmed to do, there exists a “someone” that is responsible for the output – the only reason the output of a computer is meaningful is because of this.
Thinking isn’t just a reaction; We can make choices. And when we reason in our own minds, our conclusions can become the basis of our actions; This means that we are agents of first causes, at least as far as our bodies are concerned.
The Krugman/Landsburg economic solution:
Once per week, The Federal government will randomly select a city in the United States and completely destroy it. As a result of rebuilding, we will create renewed consumer demand, deploy idle resources, people will work harder, earning more money, paying more taxes reducing the deficient, and in a few short years we will all be rich.
Gee this economics stuff is easy.
Unless on a given week a city (e.g. Detroit) should happen to step forward and destroy itself.
Since we at “destruction”. Thanks Tel for mentioning “The Wages of Destruction” from Adam Tooze some time ago. I am currently reading it and it really is a fantastic book!
And Ken is right here. Schacht really is a strange dark and shady figure…
http://consultingbyrpm.com/blog/2012/07/updated-yield-curve-chart.html#comment-41782
*Since we are at “destruction”*…
Since I live near Detroit, I really am.
What do you make of his arguments (and evidence) re the bombing? Will not be popular on this blog!
I have always argued the bombing was very effective. Just look at how many 88s were not at Stalingrad or Kursk. But it was nice to see the supposed statistical analysis against the bombing’s effectiveness dismantled.
Overy has a good analysis of it in one of his books too.