Economists disagree on who pays if you tax „the rich“
Greg Mankiw points to an interesting disagreement among high-profile blogging economists about whether it is possible to tax rich people who don’t consume very much (or who consume the same before and after the tax), or more precisely who actually pays if you do.
As I had previously tweeted, I was quite surprised by Joseph Stiglitz‘ remark that in his opinion there are some convictions that all serious economists should share: „resources are limited, incentives matter“. The context was that he didn’t agree with what many other distinguished economists held should be supported by everybody in the profession. Now, the problem with this greatest common denominator of his is that it is barely more than what I remember from a textbook as the definition of the discipline as being concerned with scarcity.
Is the situation of economics really so dire that after decades of science dealing with numbers from complete (non-inferential, every psychologist’s dream) real-world datasets, basic public policy issues are still in the realm of ideology? It seems so. Now, before (as usual …) arguing for a constructivist turn, let me put some of the central irreconcilable statements side by side:
From the original blog post that sparked the debate, by Stephen E. Landsburg:
For the government to consume more goods and services, somebody else must consume fewer.
Now the government consumes more goods, Mr. Kendrick consumes no fewer, so someone else must consume less. Who is that someone else? The answer depends on the details of the transactions, but the 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.
Criticism by J. Bradford DeLong:
First, there is a full-employment assumption lurking underneath, which may or may not be valid.
Most important, however, is that there are actually not two parties to this situation–Robert Kendrick and the rest of us–but three: Robert Kendrick, the rest of us, and Robert Kendrick’s heirs. Looking over time, what happens when we tax Robert Kendrick and use the proceeds to boost government purchases (as opposed to paying down the debt) is:
- The government spends more on consumption and investment goods.
- Robert Kendrick leaves less to his heirs, whoever they happen to be.
- We today consume more government services (and benefit from additional government capital)
- We today consume less in our private capacities.
- We in the future consume more in our private capacities as Robert Kendrick’s heirs don’t have the money to consume in the future.
Another one by Paul Krugman:
So taxes are, first and foremost, about paying for what the government buys (duh). It’s true that they can also affect aggregate demand, and that may be something you want to do. But that really is a secondary issue.
And may I say that now is an especially peculiar time to think that taxes matter only if they reduce consumption. We have lots of excess capacity in the economy; the government can easily buy more goods and services without requiring that the private sector buy less. The only reason to raise taxes now, or promise future rises, is to address solvency concerns.
From a „Chris“ commenting on Landsburg’s blog:
If all there were in the world were currently produced goods and services, sure, you can’t have an excess supply. All markets clear, increase in someone’s consumption of currently produced good couldn’t happen except to be offset by reduction somewhere else. But if you have an excess demand for financial assets – which basically has to be true now given short term rates at zero there’s a market that’s not clearing – everyone wants to push consumption into the future and hold more safe financial assets – then you have an excess supply of currently produced goods and services.
No one else _has_ to consume any less. Essentially, the trick that you’ve pulled by taking his $84 million is to transfer it from those who want to hold cash to those who want to spend today, and there are those producers who have items that they can’t sell and will sell some of them.
Finally a reply from Landsburg, that seems to be aimed mostly at an argument like Chris‘:
Many are arguing in comments that increased government spending can lead to increased output. And so it can in many circumstances. But the increased spending yields the increased output whether or not Mr Kendrick gets taxed. The question here is: “What are the effects of trying to tax Mr Kendrick?” “What are the effects of government spending?” is a different question, no less interesting, but not the topic of this series of posts.
What fascinates me about these statements (and the debate as a whole) is that they all make perfect sense. Mankiw himself seems to see the problem in „people talking past each other“, which implies they just have different topics. But I disagree, because in the end they argue over the same practical question: Can I take money from the really rich and use it for the government? (As a sidenote: it is already an interesting feature of this scientific discourse that the question becomes one of „can“ instead of „should“ or „want“). So it comes down to different perspectives on the same problem.
Now, I am also tempted to hope that there is some way to determine the „right“ (meaning scientifically-empirically correct, not morally good) perspective. The issue seems at least partially due to inclusion or exclusion of different aspects — money vs. consumption vs. production vs. investment, the heirs (time perspective), employment, … I frankly don’t know enough about the subject to judge to what extent these decisions could be made in some sort of „evidence-based“ way. My suspicion is in any case that to a substantial degree they are a priori decisions, which ultimately can only be held accountable to the constructivist-pragmatist question: „How does this view affect my view of the world or my commitments to it?“ So, how can we bring a civilized way of talking about these values into economics (or politics, for that matter)?
PS: Re-reading the sources again I notice that both Mankiw and Landsburg make pretty blunt (and I’m wondering: conscious?) attempts at rescuing the argument by keeping out unwanted perspectives. Landsburg above who basically says „we are not talking about THAT right now, that has nothing to do with IT.“ And Mankiw who claims that it is enough to defend the much more general point that „the true incidence of a tax is not necessarily on the person who writes the check to pay the tax bill,“ ignoring the much more specific claim for government policy that was originally made.
Mittwoch, 4. Mai 2011 22:48
nice chat;-)
in this case I guess we have to be aware of the fact that those economists mainly want to play. Those blogs are read by students and they are written for students. I think the conversation is very good basis for introducing in tax incidence. And of course you can continue this game. Just immagine to consider the effect on utility. Mr. Kendrick looses assets, this might reduce his financial status within his peer group … You can also highlight the consequences of any international capital flows. Just immagine the neighbor of Mr. Kendrick wants to escape taxation and moves with all his assets to the Caymans causing a further drop in capital supply .… Finally, I think this is not a serious scientific debate but rather a nice game. convincing the majority of students by citing standard textbook thoughts in an appealing way means to win.
Freitag, 6. Mai 2011 17:03
Hello my dear! Thanks for the de-escalating re-framing of the discussion I agree that one should keep in mind that professors blog partly for their students, and that blogging most of the time has a certain playfulness to it. I think that causes the above statements to be stronger and less ambiguous than what the authors would say or write in a more formal academic setting.
Still, I don’t think the disagreement can be fully explained away like that, and I want to remind you that at least Krugman’s statement appeared on a popular newspaper‘ website (NYTimes) and clearly aims beyond students at the general public, and addresses people in their role as political decision-makers (voters).
Bottom line: I do think there is a strong serious core to the disagreement, granted that there is some playful spice. And I would be interested if and how you think this debate can be scientifically resolved?
Montag, 9. Mai 2011 18:23
… it can be resolved by something called applied general equilibrium models. These are general equilibrium models that take all „relevant“ aspects into account (heterogenous agents, international capital flows, saving decisions, …). In fact such models might include all the effects we are talking about. The incidence depends finally on parameters. If capital markets e.g. are highly deregulated you can assume that taxing the rich does not effect capital supply as people from somewhere else are going to invest in German companies… However, if the elasticity of tax evasion is high as well, this could lead to a decreasing tax base in some years. So, economics is mostly about highlighting single effects that are theoretically sound. To communicate the net effect of a complex adjustment process that includes several effects is rarely possible in a public debate. Just compare it with the question of nuclear power. The optimization is on such a complex level that even fully informed people are not able to solve it with out any focussing bias. But still economist got the job to separate valid arguments from nonsense.
Dienstag, 10. Mai 2011 17:53
Ich hatte die Blogs zur Zeit gelesen und jetzt nicht mehr reingeschaut, ich hoffe ich erinnere mich richtig. Ich habe das Gefuehl, dass da grosse Konfusion herrschte. Erstmal zu zwei Fakten, die darin vorkommen, aber nicht unbedingt so klar herausgestellt werden:
1. Steuerinzidenz. Es ist nicht immer der, der die Steuerlast traegt, der sie an den Fiskus abdrueckt. Zum Beispiel wenn ein Gut besteuert wird, der Einfachheit halber Rotwein (angenommen, der Verkaeufer muss die Steuer bezahlen), dann teilen sich gewoehnlich Verkaeufer und Konsument die Steuerlast — der Marktpreis steigt, aber nicht um die komplette Steuer. Wer wieviel der Steuerlast traegt kommt vor allem auf die Elastizitaeten an.
2. Nicht jeder erleidet gleich viel „Nutzenverlust“, wenn er einen bestimmten Steuerbetrag bezahlt. Fuer manche ist es ganz schlimm (sei es, weil sie sowieso schon wenig haben oder weil ihnen Geld wichtig ist) fuer manche nicht.
In diesem idiotisch kuenstlichen Beispiel aus dem Blog erleidet der Reiche, der besteuert wird keinerlei „Nutzenverlust“, weil er sich kein Stueck fuer Geld interessiert. Daraus schliesst der Autor, dass man ihn nicht besteuern kann (und daraus koennten/sollen dann andere folgern, dass man Reiche nicht besteuern kann und es deshalb am besten auch gar nicht probiert).
Darueber regen sich viele linkere Oekonoblogger auf und stellen klar, dass der Sinn von Steuern ist, fuer verschiedene staatliche Leistungen zu bezahlen und das geht sehr wohl vom Geld der Reichen (und falls es die Reichen gar nicht stoert, wenn man ihnen Geld wegnimmt, wie in diesem idiotischen Beispiel, dann ist das ein extra Grund es genau von ihnen zu nehmen).
Mankiw interpretiert dieses Beispiel sehr viel wohlwollender und sagt, alles was der Autor wollte ist auf die Fakten 1 und 2 oben (Mankiw schreibt es ein bisschen anders, aber ich glaube vom Sinn her ist es das, was er meint) hin zu weisen.
Ich persoenlich sympathisiere sehr viel mehr mit der Meinung von Krugman/DeLong/Yglesias (wenn ich mich recht erinnere) und sage, wer ein idiotisch Beispiel erfindet um voellig falsche Dinge zu suggeriern, der verdient, dass man seinen Artikel kraeftig auseinandernimmt und nicht all zu viele guten Haare daran laesst.
Dienstag, 10. Mai 2011 17:56
Lieber Matze, vielen Dank für die ausführliche und erhellende Antwort, es scheint Du hast die Debatte tatsächlich noch ganz gut im Kopf. Interessant finde ich, dass sowohl der ursprüngliche Blogger als auch Mankiw Deine Punkte 1. und 2. zu vermischen scheinen, bzw. sogar als identisch sehen, was mir recht absurd erscheint. Ich bin auch froh, dass Du die realpolitische Implikation ebenfalls siehst, die wie Du sagst aus einem absurden Beispiel heraus impliziert wird: Wir können (ergo: sollten) Reiche nicht besteuern.
Dienstag, 10. Mai 2011 18:04
Dear Joe!
Thank you, this is what I was looking for. But does this imply that the only scientifically rigorous solution to a question as (seemingly) simple and as (definitely) practically important as the possibility of taxing „the rich“ can only be determined by running computer models/simulations? Which are impossible for outsiders to evaluate and in turn dependent on the choice of variables to include and the (empirically?!) chosen parameter values?
This would still leave me preferring the „constructivist“ perspective, with its strong call for democratic foundation and open debate of values. Especially after just now reading and thinking about an article by Krugman on this computational research culture.