Thursday, April 21, 2011

Landsburg on Taxes

Here's a piece from Steve Landsburg on taxation. DeLong and Krugman think Steve is batty, apparently, but it's actually an interesting puzzler. Let me add a bit to it so the setup is more precise. Landsburg imagines a rich guy, Robert (apparently a real person in the piece he refers to), who we'll say consumes zero and takes all his time as leisure. But, Robert has wealth. He carries around a bag of currency and he has an account with the Treasury and a Vanguard account. Robert will hold all this wealth and not spend it before he dies. When he dies, the currency is lost, Vanguard does not owe anyone anything, and neither does the US Treasury for what is in the accounts Robert owns. He has no heirs. This is very unrealistic I know, but these are assumptions. We're doing economics after all.

Now, what if the Treasury taxes Robert? What happens? Suppose the IRS person takes $100 in currency from Robert's bag, and spends it on good and services? Does that matter? Well, if the alternative was that the Treasury issues $100 in Treasury securities, and the Fed purchases the Treasury securities by issuing currency, then no, it can't matter at all, given any frictions we want to throw into this world - sticky prices, sticky wages, 10-foot gorillas, whatever.

What if the Treasury taxes Robert by deducting $100 from his Treasury account? But the Treasury has to finance the $100 in spending on goods and services somehow, so clearly this cannot matter at all. The balance that sits in Robert's Treasury account is irrelevant.

Finally, suppose now that the Treasury taxes Robert by taking $100 out of his Vanguard account, transfers it to the Treasury, and spends it on goods and services. Now, things get more complicated. There is an alternative, though, that makes this irrelevant too. Suppose that the alternative was that the Treasury issued $100 in Treasury securities, and the Fed bought them by issuing reserves. But, suppose that when the Fed taxed Robert and sold the Vanguard liability, the Fed bought it by issuing reserves. Now its equivalent.

Thus, we have a theorem. Taxing Robert, under certain conditions, is irrelevant. But DeLong says:
It is simply not true that "Robert Kendrick cannot be taxed." Not true at all. No, no, no, no, no.
Krugman says:
Discussions like this really disturb me; they indicate that there are a lot of people with Ph.D.s in economics who can throw around a lot of jargon, but when push comes to shove, have no coherent picture whatsoever of how the pieces fit together.
I'll let you be the judge of what is true and who has no coherent picture of how the macroeconomy fits together.


  1. Except that Landsburg claimed the man "can't be taxed," not that taxing him is irrelevant. And he stated it unconditionally, not "under certain conditions."

  2. Actually, it's even stronger than "he can't be taxed." In general, there is no way to extract any resources from the guy. That's under any conditions. I've shown you how it can actually be irrelevant for the allocation what financial assets you take away.

  3. Except he considers this not be an 'interesting puzzler' bur somehow relating to actual real-world policy making. Here is how Landsburg starts the debate:

    "Taxing people like Mr. Kendrick, she says, has to be part of any solution to America’s fiscal crisis. Here’s what Ms. Stevens misses..."


    Do you consider this sort of hypotheticals, under the patently false assumptions you outline, to be relevant to the policy debate?

    As for DeLong, this is more than a little unfair: Brad specifically talk about heirs, you assume no heirs, and then call him out for being wrong. Nice.

  4. DeLong says he's wrong, and he says it in his usual flamboyant way. I show you how he can be right. Now, the key question is whether the example has anything to say about anything practical. As someone pointed out to me, the standard assumption that is violated here is non-satiation. Essentially, Robert is satiated at zero consumption and leisure equal to his time endowment, which fits with the notion that he doesn't care about his descendants, if he has any. So, this is the crux of the problem. If we try to tax rich retirees in our current circumstances, do we extract no resources as a result? Not likely. The theory looks fine. But the empirical work is not so great.

  5. Actually, take some of the last comment back. Here is what is interesting about it, in a practical sense. Suppose Robert is satiated at some positive level of consumption, and he is such a creep that he is planning on (as per my assumptions) dying with positive wealth, in fact a huge quantity of it, which he is not giving to anyone. He is just the sort of person that Krugman, DeLong, and apparently our President, would love to tax. But so long as you don't tax him so much that he has to reduce his consumption, you extract nothing. The creepiest big rich jerk possible can't be taxed.

  6. Interesting post. Robert can't be taxed because all his wealth is either in the bag (i.e., in the Fed), in the Treasury, or in Vanguard. But the Treasury can tax Vanguard, and if we consider the Fed as a corporation separate from the Treasury, the Fed can be taxed too.

    So if the IRS takes $100 from Robert's bag, the Treasury can redeem it at the Fed for a $100 T-bond. The Treasury now owes $100 less to the Fed. The Treasury is $100 richer, the Fed is $100 poorer, and Robert is unaffected.
    If the Treasury had issued a $100 bond which was purchased by the Fed, neither the Fed nor the Treasury would be richer or poorer.

    If the Treasury takes $100 from Robert's Vanguard account, then Vanguard is $100 poorer, the Treasury is $100 richer, and Robert and the Fed are unaffected. If the Treasury had issued a $100 bond, which was bought by the Fed, then neither the Fed nor the Treasury are richer or poorer. (You didn't mean to say the Fed taxed Robert did you?) But if the Treasury took $100 from the Vanguard account and sold it to the Fed, then the Treasury is $100 richer, Vanguard is $100 poorer, and Robert and the Fed are neither richer nor poorer.

  7. Yeah, I guess if a guy is a miser and the money he owns is not in circulation, than taking some of that money and spending it into circulation is no different than printing new money and spending it into circulation.

    Similarly, if there is one man with a bank account in the First Bank of Mars, and that bank contains only the deposits of that one man, and that man never spends or transfers any part of those funds into the terrestrial economy, and never even plans to spend or transfer them into the terrestrial economy, then those funds are not functionally included in the current terrestrial money supply. So whether the government legally directs the spending of those funds from the Mars Bank into the terrestrial money supply or simply prints the same amount of new money and spends it make no difference.

    Why is this a puzzle and why is it interesting?

    It is not the case, one should note that the guy "can't be taxed" - only that it probably makes no difference to the rest of us whether he is taxed or we acquire the money by printing it.

    And only if the miser is absolutely resolute in his determination never to spend the money does the taxation make no difference to him. Once the money is taken from his Mars bank account he loses purchasing power he once had, even if he wasn't very quick to exercise that purchasing power. If he was even considering that he might spend the money some day, he will experience that loss of power as a straight out loss.

  8. Dan:

    If you take $100 of the miser's money and spend it, you are $100 richer and the Fed is $100 poorer. If the Fed printed $100 and bought a $100 T-bond with it, nobody's wealth is affected---not you, not the Fed, and not the treasury.

  9. I mean, it looks like at first glance, the guy consumes nothing and never will, so if you take his cash (which he would never use anyway) it doesn't matter (unless he had heirs who would have used it).

    I guess if you never consume anything and never will, and have no real assets, the government can't tax you, but can tax your heirs if they would consume.

    How is this useful to know?

  10. You may not like Krugman's rant in the end, but perhaps he has a point worth elaborating on (even if to dismiss)? In his post the argument is that by taxing Robert the government can use the assets to cancel debts with others.

  11. Of course Landsburg is talking about tax burden. If the guy just has idle wealth (which is explicitly assumed by both Landsburg and the columnist he references), then his tax burden will be zero. Landsburg is talking about the incidence of the tax and in that sense he is correct.

  12. By the way Steve, I am greatly bothered by the contempt that some commenters on this blog seem to have for you.

  13. 1. No, Krugman's wrong. In the example, there is nothing about taking the assets away from Robert that allows you to relieve some debt burden you could not have been relieved by the government without taking the assets from him.
    2. As I said above, the key non-standard element of this is that Robert is satiated. In essentially all the economic theory we do (maybe someone knows about a counterexample), we assume non-satiation, which implies that everyone's budget constraint binds. If Robert's preferences satisfied non-satiation, then we would analyze this as we would any standard tax problem. The tax makes Robert worse off, it tightens his budget constraint, and it changes his behavior. Landsberg likes these non-standard things. He's got a mind like a mathematician's (he was one at one point), and he finds these things intriguing. The idea is useful, I think, but maybe not for exactly the reasons Landsburg constructed it (he's partly just trying to be provocative). Think about Bill Gates. He has a gazillion dollars in wealth, and he's satiated. If he buys more stuff, he'll have to look after it, or employ some more people to look after it for him. He thinks his children are well-enough-provided for, and seems worried that their lives will actually be worse if he allocates more wealth to them (we all know examples of people who inherit a large quantity of wealth and are bright and have a lot of human capital, but lead essentially wasted lives). So, what does he do with the wealth he does not want? He could let the government take it and allocate it, or he could set up a charitable foundation and do it himself. That decision actually matters, as Bill allocates the financial wealth in a different way than Congress will.

  14. Last anonymous,

    This is very mild, actually, relative to typical blogosphere talk. I'm not particularly bothered.

  15. But is there a substantive disagreement with DeLong? Behind the rhetoric, his point seems to be that even if Robert is satiated, taxing him means taxing his heirs.

    You assume he has no heirs, which in real-world terms means that his property escheats to the government when he dies. In your specific assumptions, the Treasury account literally escheats to the Treasury. The cash is burned, which reduces the Fed's liabilities. The Vanguard account is more problematic: if the stocks and bonds are inherited by Vanguard, that's the taxable heir; if they are cancelled, the issuing company's shareholders are. So I believe you need them to escheat to the government too for the theorem to work.

    Thus it seems you and DeLong agree that if the government is certain to inherit Robert's untouched wealth then it cannot tax him. That's because by assumption it is already the effective owner of his wealth. If with any positive probability Robert might consume or bequeath any part of his wealth, then he (or his potential heirs) can be taxed.

    Your very well-taken final point also feels in full agreement with DeLong. Bill Gates can be taxed: the tax transfers the power to spend money from him to the government. True, the marginal tax on Bill Gates is likely to fall overwhelmingly on the Gates Foundation rather than on his household's current or future personal consumption. This matters for assessing whether taxing Bill Gates is desirable, but confirms that it's possible.

  16. This so reminds me of Scrooge. In the dream, IIRC, he dies a rich man, but is castigated for having helped nobody.

    But presumably, he must thereby have helped his heirs. They get to consume all the resources he didn't.

    (Dickens' story only makes sense if you assume Scrooge lived during a time of deficient-demand unemployment that the government was unable and unwilling to fix, and died at a time when the economy returns to full-employment.)

  17. Giacomo--I think DeLong is correct but off-topic. If Kendrick plans on leaving his fortune to an heir, then the burden of taxing Kendrick's fortune today obviously falls on his heir, assuming that heir would have spent the money. But if the heir was going to spend the money, he clearly wasn't idle rich. So taxing Kendrick isn't actually taxing the idle rich; it's taxing the normal rich. That may or may not be desirable, but it's unrelated to the original discussion.

  18. Yes, DeLong is missing the point. He wants to be argumentative and call people names, rather than settling down to learn something. You could carry the example on, as Ryan says, through multiple generations if you want. There could be an infinite chain of descendants, all of whom are satiated and pass the excess wealth on to their descendants, a la Barro, and get the same thing.

  19. I'm a little confused why
    "Mr. Kendrick can't be taxed"
    "There is a set of assumptions, practically certain to be violated in reality, under which a tax on a fictional character does not actually impose a burden on him".
    are supposed to be equivalent statements.

    Landsburg's piece is based on a condescending and very uncharitable reading of the NYTs article (the main point of which is a very high estate tax!) - there is absolutely no reason why everything he writes should be interpreted in the most charitable way possible.

  20. To me, he looks like a creative thinker, and you can learn something about taxation and government financing from what he wrote. I certainly have, so for me it was a productive piece of writing. You are not capitulating on some political point by thinking about what he wrote.

  21. Thanks for replying, but I think you're misunderstanding which political point I'm debating. I'm fine with Landsburg being a libertarian, against taxing rich people etc.
    What I mind is that he is one of the prime examples of what I like to call "econ 101ism" - people take the assumptions of an introductory econ class, apply them to real world issues and then claim that they have provided an economic analysis.

    I'm fine with your example above as a thought excercise, but Landsburg, without being clear about the assumptions, talks about real people and real economics, not about a stylized world. He attacks the writer of the underlying article for not understanding economics. He claims to show that taxing the idle rich is impossible as a solution. And in that context what he writes is just ludicrous.

    People like DeLong and Krugman (or Milton Friedman or Ken Rogoff so that this isn't about ideology) are such good economists because they have an incredibly good economic intuition - they don't just see simple models, they have both an intuitive grasp and a deep understanding for how the economy works. Landsburg doesn't (he does this type of thing all the time) - so he resorts to cookie-cutter models. The problem is that he thinks they apply 1:1 to the real world.

  22. I don't have a firm opinion about Landsburg. He has been writing these pieces for a long time. Some used to show up on Slate, and I read a few. He's not the same type of economist as DeLong and Krugman, who actually have serious academic credentials. If you have been reading my stuff, you probably know that I don't agree with you about DeLong and Krugman being "incredibly good economists." They are two people whose agenda is primarily political, and they really are not up on modern macroeconomics. They put their faith in some overly simplistic and old-fashioned theory, and make you want to think that everything they tell you is incontrovertible, which is not true.