Wednesday, April 20, 2011

In which Steven Landsburg utterly flips out

Steven Landsburg thinks he has come up with a proof that taxation of the "idle rich" is impossible:
[The extremely rich] Mr. Kendrick appears to do pretty much nothing but park and re-park his four cars all day long...Assuming the facts are as she states them, it is quite literally impossible to raise revenue by taxing the likes of Mr. Kendrick...

Here’s why it’s impossible: For the government to consume more goods and services, somebody else must consume fewer. But Mr. Kendrick...consumes almost no goods or services whatsoever. He just pushes cars around all day. His consumption can’t go much lower.

Ah...but there’s still that $84 million in the bank. Surely we can tax that, no?...[but] what happens if the government takes Mr. Kendrick’s $84 million away? Answer: A bunch of zeros and ones get shifted around on bank computers. Mr. Kendrick goes right on pushing his cars around. And nothing else has changed.

Unless, of course, the government decides to spend some of that $84 million. Now the government consumes more goods, Mr. Kendrick consumes no fewer, so someone else must consume less. Who is that someone else?...[T]he most likely answer is that when Mr. Kendrick withdraws $84 million from the bank to make his tax payment, the bank makes fewer loans, interest rates rise, and someone cancels a vacations, or postpones a car purchase, or abandons a half-built factory. Who bears the burden of the tax? The people who cancel their vacations and car purchases and factories, that’s who. Not Mr. Kendrick.

You can try to tax him, but any attempt to tax him turns into a tax-in-disguise on somebody else. And the reason for this is not ultimately to be found in the laws of economics; it’s to be found in the laws of arithmetic. You can’t drive a man’s consumption below zero.
OK, clever readers, your assignment is to reread the above blog post very carefully, and think of four reasons why Landburg's post is utter nonsense. That's right, four. I don't want to make it too easy for you.

All right, now here's the reasons I've come up with:

Reason 1: GDP does not equal consumption. This is a very, very basic Econ 101 fact. Like, the first thing you learn. GDP = Consumption + Investment + Govt. Purchases + Net Exports. Landsburg says: "For the government to consume more goods and services, somebody else must consume fewer." This is obviously false, since Government Purchases can rise if Investment or Net Exports goes down (or GDP goes up), with Consumption unchanged. There is no "conservation of consumption".

So, even before we go on to other, more sophisticated reasons why Landsburg is talking nonsense, we know from the most basic accounting identity in all of economics that he is, indeed, talking nonsense.

Reason 2: Revenue does not equal consumption. Landsburg writes: "It is quite literally impossible to raise revenue by taxing the likes of Mr. Kendrick." Wrong. Even if Landsburg's nonsensical idea about the "conservation of consumption" were right, this statement about revenue would be patently false, since a rise in Government Purchases accompanied by a fall in private Consumption would, if the deficit is unchanged, constitute a rise in revenue.

And yes, I know this is a fairly trivial and obvious reason that Landsburg is talking nonsense, but it bears pointing out before we move on to deeper and more interesting reasons.

Reason 3: Consumption today does not equal lifetime consumption. Maybe today Kendrick just reparks his cars, but tomorrow he might get tired of that, and might want to use some of his fortune to buy, say, a yacht. Confiscating his fortune takes away his ability to do this. Landsburg has decided that if Kendrick's current consumption is not reduced, it does not constitute "taxation" of Kendrick. By this logic, if I confiscate the entire contents of Steven Landsburg's bank account, it will not qualify as "robbery" if he did not plan on making an ATM withdrawal until tomorrow. 

Corollary to Reason 3: Local nonsatiation is an axiom of modern economics. This axiom states that, over his lifetime, Kendrick will consume his entire lifetime income (as Brad DeLong notes, this may include bequests to his descendants). Eventually he will do something with that $84 million, even if he just leaves it all to his kids untouched. So taking Kendrick's $84 million reduces the present value of Kendrick's lifetime consumption by...$84 million.

But wait, there's more.

Reason 4:  If Landsburg were right, aggregate savings could never change. Landsburg states that if Kendrick's fortune were withdrawn from his bank and spent on consumption goods, the bank would raise interest rates, and someone else's consumption would fall by an exactly countervailing amount. That is just wrong. Consider this: what if Kendrick himself decides to withdraw his $84 million and spend it on consumption goods (say, a few dozen more cars). Must consumption elsewhere in the economy fall by $84M? No? But in real terms, that is exactly the same as if the government confiscated Kendrick's $84M and bought the new cars for him - which, Landsburg insists, would reduce consumption elsewhere in the economy. 

Landsburg's statement is therefore equivalent to the mathematical statement that economy-wide consumption cannot be raised by a decrease in a single individual's savings. And since Income = Savings + Consumption, Landsburg is therefore saying that any decrease in one person's savings must, mathematically, be accompanied by either a decrease in economy-wide income (GDP) or an increase in someone else's savings. If the former is the case - if reducing my savings by $X reduces GDP by $X - then increasing my savings by $X must increase GDP by $X. And that would mean GDP is maximized if people consume zero (i.e. everyone dies). This is clearly not the case. So, by process of elimination, Landsburg must be asserting that if my savings go down by $X, someone else's must go up by precisely $X, and hence aggregate saving is constant over all time.

Needless to say, that is total nonsense.

There, that's four (though you may notice that Reasons 1, 3, and 4 basically address the same fallacy). I'll leave it to you readers to come up with more (and I'm fairly sure there are some more).

But I'll end by pointing out that Alex Tabarrok of Marginal Revolution gives the following uncritical endorsement of Landsburg's nonsense:
Steve is quite right. The key is this sentence, “For the government to consume more goods and services, somebody else must consume fewer.”
He even repeats one of Landsburg's patently nonsensical assertions (see Reason 1 above)! And although he does acknowledges Brad DeLong's critique (i.e. my Reason 3) in an addendum, he never takes back the nonsensical assertion that present consumption is conserved.

WHY? "Economics by accounting identity" is dubious at the best of times, but "economics by patently false accounting identity" is just inexcusable. You can get economic theory to say damn near anything you want. But by Adam Smith and all that is holy, you just cannot get it to say that consumption is constant by definition! You just...can't!!!

Update: Paul Krugman explains why Landsburg is misunderstanding the nature of taxation, a point that Brad DeLong also noted. That's five... 

Update 2: Niklas Blanchard points out that the government doesn't just consume, it invests! The existence of public goods and government investment mean that Landsburg's statement about "government consumption", in addition to being nonsense, is also a bit irrelevant to real public policy issues.

Update 3: Landsburg, responding to Krugman, doubles down

1) A tax imposes a burden.
2) If a tax has no effect on a man’s lifestyle, then it imposes no burden on him.
3) Therefore, if a tax has no effect on a man’s lifestyle, then it must impose a burden on someone else.

This demonstrates a failure to understand Local Nonsatiation. If Kendrick's $84M is confiscated, either A) his present consumption is reduced, B) his future consumption is reduced, or C) his ability to bequeath money to his heirs is reduced. Either way, his choice set is reduced, so by Local Nonsatiation his utility is reduced, i.e. a burden has been imposed upon him, QED.


  1. Rebutting Landsburg is always exhausting work. Thanks for volunteering for shovel duty.

    p.s. I'd your take on this:

    It hits some topics you've discussed before but from a math/stat angle. I'd be interested to hear what an economist had to say.

  2. "WHY? "Economics by accounting identity" is dubious at the best of times, but "economics by patently false accounting identity" is just inexcusable. You can get economic theory to say damn near anything you want. But by Adam Smith and all that is holy, you just cannot get it to say that consumption is constant by definition! You just...can't!!!"

    In his example the man was at a constant consumption, and the rest of his money he was just putting in a massive pile. Lansberg then said, "IF this is true that the man isn't spending his money at all and taxes wouldn't change his behavior, then _____ is also true." What he said does follow.

  3. Reasons 1, 2, and 4 are just nit-picking about definitions. Obviously Landsburg is being a little careless with the words "consumption" and "revenue," but his substantive point is that the only limitation on the government is the availability of real resources. Ultimately, unless a tax results in someone's reducing their demands on real resources, it doesn't make resources available for the government.

    You reason 3 addresses this. You can "tax" Kendrick (in Landsburg's sense) by reducing his future consumption or that of his heirs. And yeah, if you take the intertemporal utility maximization stuff seriously, a tax (in the literal sense) today will ultimately reduce either his consumption or that of his heirs. But if you take the intertemporal utility maximization stuff seriously, you believe in Ricardian equivalence. Do you?

  4. @Doc Merlin: No.

    @Andy Harless:

    1. I think you are underestimating the significance of Reason 4. Consider what happens if Kendrick himself decides to take $84M out of his bank account and spend it on consumption goods. Does this necessarily reduce the present value of consumption in the "non-Kendrick economy" by exactly $84M? If you think it does, then you believe that wealth cannot be consumed. THAT is the point of Reason 4.

    2. Ricardian Equivalence does not automatically follow from the Income-Expenditure Identity, because production is involved. The Kendrick example deals only with consumption.

  5. Anonymous1:25 PM

    I'm no economist. But doesn't the manufacturing capacity rise to meet demand? And why on earth would one person spending money they already have (GIVING it to someone else) decrease consumption elsewhere since its not like "oh no! there's only one ford focus left in the world and if I buy it, you aren't going to buy any other car at all."

    This is just a rich guy saying "If you tax me, YOU will suffer." Sick, greedy SOBs.

  6. 1. Landsburg seems to be assuming a closed economy and using a definitional simplification in which he considers investment to be a form of consumption. In that case, aggregate savings cannot change because it is always zero. You could expand the model for an open economy and make the distinction between investment and consumption, and this will change the way he expresses his argument, but it won't change the substance of it.

    2. My point about Ricardian equivalence was in response to your "corollary to reason 3," which is a behavioral point, not an accounting point. If you really think that Kendrick will consume his entire lifetime income because of local non-satiation, then you must believe that Ricardian equivalence applies, unless you think that Kendrick is somehow deceived about his future taxes. (Or maybe he has enough political power to guarantee that the taxes will fall on someone else, but in that case, why are we even discussing increasing his taxes?)

  7. Ryan P2:16 PM

    Landsburg essentially says, "Stevens claims Kendrick does not and will not ever consume all of his wealth (and ergo we should tax Kendrick). Under that assumption, is it possible to tax Kendrick?" and then goes on to say no. If your response includes anything like the phrase "suppose Kendrick consumes", your response is nonresponsive.

    So your point 3 isn't disproving his claim so much as saying it's not worth saying anything about statements by journalists that there exist people who have assets that they have no intention of consuming or even of bequeathing to someone who will consume. Which is of course fine.

  8. @Andy Harless:

    1. If you include investment in "government consumption" then GDP is not constant, and hence Landsburg's statement breaks down for the reasons Niklas Blachard gives (see Update 2).

    2. Your point about Ricardian Equivalence isn't right. Even if everyone consumes their lifetime income (including bequests as consumption), that doesn't mean Reicardian Equivalence has to hold, again for the reason that public goods exist (again, see Niklas Blanchard at Modeled Behavior).

  9. Oh, and Andy, for another explanation of why Ricardian Equivalence fails in a world with public goods,! :)

  10. Krugman is 100% wrong about this. Taxes don't exist "to ensure government solvency." Solvency is an accounting condition that, in the case of the federal government, is ensured by the ability to print money. Taxes exist to ensure government control over real resources -- which they accomplish by persuading the private sector to relinquish control (usually by reducing consumption). Inflation, which Krugman cites as a limit on seignorage, will happen only if there is a shortage of real resources. If there's a surplus of real resources, then there is (at the margin) no need for taxation, because there is no limit on seignorage.

  11. I guess Local nonsatiation is key to the argument; could that be violated though, if say Kendrick were to withdraw his $84M as green pieces of paper and set fire to the lot of them? Of course that's unrealistic, but so is the premise for this whole discussion (that Kendrick will never spend any of his money). But in that crazy scenario, would Landsburg be correct?

  12. I agree with Andy Harless. Normally I start a post after reading something like Krugman's posts with "With friends like this you don't need enemies". Taxes according to MMT serve two purposes: They give value to a FIAT currency. And they drain nominal demand. By draining nominal demand government can (1) prevent nominal demand crossing the inflation barrier and (2) force the private sector to relinquish control over real stuff. Taxes fund nothing. The US government can buy whatever it wants as long as it is denominated in US$ by simply crediting bank accounts. If it should do so is another question.

  13. Ryan P8:43 PM

    Sumeet, precisely. The whole premise is exactly that: if you suppose Kendrick won't spend his money, ever, then what? Claiming local nonsatiation is just to refuse to answer the question and claim you've disproven it.

    Trim off the flamboyant language, and Landsburg's point becomes a plain vanilla tax incidence question, AFAICT

  14. Butter2:13 AM

    My God, what are you talking about in this post? Not Landsburg's post. So much confusion in the blogosphere over a simple (and theoretically correct) point by Mr. Lansburg, which is: "You can't tax idle resources because it's eqivilent to printing money. Confiscating money in a mattress doesn't create any real resources because it doesn't induce people to forgo consumption. The burden of such a tax would be exactly equivilent to the burden of inflation." That's my best shot at reframing the argument. Hope this helps.

  15. Anonymous9:56 AM

    This is all very silly. The argument Landsburg made, even if true, is a pointless and irrelevent one.

    The point is that we have idle resources -- or capital or wealth, however you wish to categorize it -- in our economy that should be put to use in order to increase the economic multiplier of money transactions.

    Therefore, taxing (or shifting, if you prefer) idle resources into a more productive use has a net economic benefit.

    Call it the "use it, or lose it" theorem. Taxation is a means by which the government can - by compulsion -- shift idle or non-productive resources into a more productive use.

    In a situation like now, where you have a lot of slack capacity and idle resources, it makes no sense for the Gvt to continue to expand the money supply to prop up economic activity. It is better to use the power of taxation to sop up the idle resources and put those resources to more productive use.

  16. OK, so taking money from Kendrick's account doesn't change anything because it just reduces the money available to lend for something else. And we know the amount of money is equivalent because Landsburg said so in bold letters which makes it true.

    So I want to play this game too. Everything the government does, it fails at. There, it's in bold. So it must be true. And as a result of this obvious truth, I think we should get rid of fire departments because they're a waste of resources. You see, even if you call the fire department, it won't be able to help you put out the fire. (Remember, government fails at everything. I put it in bold, so it's true.) So therefore, it doesn't make any sense to have fire departments.

    Gee, that was a fun game. Now what does any of this have to do with what sort of taxation and economic policies should be implemented?

  17. Landsburg of the "We need herd immunity, everyone should take home more random strangers for sex" theory?

    People take this guy seriously?

  18. "...either A) his present consumption is reduced, B) his future consumption is reduced, or C) his ability to bequeath money to his heirs is reduced. Either way, his choice set is reduced, so by Local Nonsatiation his utility is reduced, i.e. a burden has been imposed upon him, QED"

    This seems like rather a pyrrhic intellectual victory. First of all, the marginal utility of his wealth may be very low, and we can guess that it is low from the observation that he isn't consuming it. OK, he might really love his heirs a lot, but I have to say, that's not my impression of the guy in Landsburg's example. So, yeah, technically there's local nonsatiation, and you can tax him, but you can only tax him a tiny bit.

    And in any case his heirs almost certainly get significantly more utility from the bequest than he does. So the point, "You thought you were taxing him, but (for the most part) you're really taxing someone else" pretty much stands.

    Moreover, the richer you are, the more marginally altruistic you're likely to be, since the marginal utility of individual consumption goes down. So the richer someone is, the more the incidence of his taxes will be on the recipients of his charity. Technically, you can tax him, at least a little bit, no matter how rich he is, but in practice, you're mostly taxing third world orphans and such.

    And anyhow, to the extent that Landsburg is technically wrong, and you really are taxing the rich by taxing the (non-consuming) rich, that actually makes the case against taxing them stronger. If you tax a middle class person with a high marginal propensity to consume, the tax only hits once; it's more or less just a transfer from the taxpayer to the government (except for consumer surplus, but we all know that Harberger triangles are small). If you tax a rich person with a low marginal propensity to consume, the government still only gets the revenue once, but the utility hit on the paying side happens twice, once to the taxpayer himself, who is deprived of the opportunity to make a gift, and once to the would-be recipient of the gift. So one might say, "Hmmm, Steve Landsburg almost convinced me that taxing the rich doesn't do any harm, but now that I've read Noah Smith, I realize it's a really bad thing to do. I guess I will change my vote on Obama's tax bill from 'present' to 'no'"

  19. There are two reasons Landsburg cites to prove taxing Mr. Kendrick is not possible, only one of which is correct. The first is that his money is presumably invested somewhere, meaning that someone is using his money to create wealth. If the government takes the money, it is really taking it from the people currently enjoying its use. Mr. Landsburg claims that when the government takes the money, nothing changes, and that it is only when government spends the money that someone gets hurt. This is incorrect; the government takes wealth only from Kendrick, and he is the one who is harmed.

    To understand the true nature of this process, it is helpful to realize that “money” is itself a form of wealth. It is a financial asset like any other, and Mr. Kendrick has loaned all of his assets to other people (if he holds it in a bank, stocks, mutual funds, etc). Because he doesn’t actually hold the wealth, but has instead loaned it to others so that it may find productive use, the government theft transfers ownership of those loans to the government. So far, the only loser is Mr. Kendrick.

    When the government trades those loan obligations (taken from Mr. Kendrick), they do not destroy the loans. They only transfer the ownership of the loans to a road builder or submarine manufacturer. The people who were originally putting Kendrick’s assets to good use creating more wealth continue to do so, oblivious to the fate of Mr. Kendrick. Landsburg’s attempt to elaborate on from whom the assets are taken is unnecessary; the wealth is taken from Kendrick alone.

    The bank has invested it in myriad ways, but to keep it simple, let’s think about it this way: Mr. Kendrick owns 84 ounces of gold. He has loaned them to a shoe manufacturing company in exchange for an IOU from the company, stating “Bob’s shoe company owes the bearer of this note 84 ounces of gold and 5% interest until the gold is returned to the owner of this note.” This is a simple way to think about the fate of the money Kendrick places with the bank.

    When the government taxes Mr. Kendrick, it does not take the gold from the shoe company. It takes the note issued by the shoe maker, and when it spends the note the government trades it to the submarine company for a new warship. The shoe maker remains unaffected; it will repay the 84 ounces of gold to the submarine company instead, and continues to make use of the gold for its own shoe-making purposes. It is only Mr. Kendrick and his heirs that are penalized by the tax. Taking the money does not destroy wealth, does not cancel the loans Kendrick made, and does not tax others inadvertently. The tax plainly and clearly taxes Mr. Kendrick and Mr. Kendrick alone.

    Refuting the suggestion that it is good for society to take Kendrick’s money is the same argument used to refute any type of central planning, which involves rearranging and directing the wealth of a nation, and has never proven successful. The reasons it is harmful to take Mr. Kendrick’s money are the same reasons it is harmful to take anyone’s money, or force anyone to do anything at the point of a gun. In the end, we are all happier, richer, and live with more freedom when government taxes us as little as possible. Telling Mr. Kendrick what to do with his wealth is no different from telling any citizen what to do with his or her wealth. If I choose to keep keep 84 pennies in a shoebox and never spend them, that is my personal choice. The situation does not change when if I choose to keep $84 million in my bank account, and rearrange my cars until I die. Removing the the freedom of individuals to do as they please with the fruits of their labor is called central planning.

  20. Money is Wealth:

  21. Anonymous7:22 PM

    "In the end, we are all happier, richer, and live with more freedom when government taxes us as little as possible. "

    Indeed. On the one hand, Somalia. On the other, Denmark.

    How ridiculous can some people get?

    The best-functioning societies on earth, ever, by any widely accepted measure, are places like the Nordic countries and Switzerland. Their taxation rates are pretty high.

  22. This comment has been removed by the author.

  23. It is no big nbews that Steven Landsburg is a know-nothing of macroeconomics. With his record, I think a nice post would show him saying something smart.

  24. justanothereconomist1:34 PM

    Steven Landsburg is wrong.

    The government can raise revenue by taxing Kendrick. The government takes Kendrick's wealth and pays off savings bonds. Bondholders buy Kendrick's former wealth with the cash they received from the government. It's a triangle. The government raised revenue, and Kendrick was taxed- he has less stuff. Is this really so hard?

  25. Anonymous11:38 PM

    You guys simply don't know what you are dealing with. Landsburg is a first rate mathematical genius of the highest order. Hence you have all completely misunderstood his arguments.

  26. Steven Landsburg is an idiot. Anyone who has ever read his works on Scrooge realizes that. He manages to bifurcate theoretical logical reasoning from common sense, and opts for the former and completely expunges the latter.

    Let this be a lesson to never do that.

  27. Isn't Landsburg the guy who, earlier this week, said that having unconsensual sex with an unconsicous person shouldn't be a crime because it doesn't cause harm?

  28. Anonymous11:06 AM

    GDP calculations are a joke passed on by hilarious government pawns who want to legitimize theft and actually claim it is beneficial. hilarious... I can't believe we are seriously have a conversation about how stealing from people doesn't hurt them and actuall makes everyone else better off... talk about rape...

  29. "Moreover, the richer you are, the more marginally altruistic you're likely to be, since the marginal utility of individual consumption goes down."

    I believe this is empirically false.

  30. I think this is a case of person not making the most charitable reading. Do really think Steven Landsburg is ignorant? I do not, and I think his post would give most voters something to think about that would help their understanding.