Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

September 21, 2010

Going Galt

A certain Professor Xxxx has received quite the internet thrashing for complaining that (cached version) — despite a household income (roughly $450K $400K ) nineeight times the median US household income — he’s really just an ordinary Joe, trying to make ends meet, and will suffer terribly if Congress allows the Bush tax cuts (on incomes over $250K) to expire.

I’m somewhat loath to pile on, as surely every cognitive flaw, exhibited in the post in question, has been more than amply exposed. But something struck me about one of his followup posts (cached version), which seems to speak to a more general bit of nonsense that one hears from those of Professor Xxxx’s ideological persuasion.

He says

So why doesn’t Prof. Krugman say that I’m a whiny loser because I’m complaining that the government needs 50% of the money I spend each month, on things like art camps? It may be that some of the decisions I made that led to these fixed costs were mistakes, but there will be a real impact from this increase in my tax burden. I’ve not seen any critic point out why my Polish-American house cleaner would be better off getting handouts from the government than earning her wage cleaning our house.

Now, arguing that your taxes shouldn’t be raised because, if they are, you’ll respond by employing fewer domestic servants, is probably not the winning sort of argument that a University of Chicago Law professor would teach his students to marshall. But let’s leave that aside. What I want to focus on is the fact that art camp, for apple-cheeked Amy is in peril. Two paragraphs later, he says

Fifth, lost in all of this is the impact of increased taxes on the work-leisure tradeoff. As marginal taxes rise, so does the disincentive to work. I’m asked with some frequency to write, consult, or testify, and when I do, I face the question of whether the effort and time is worth it. I can choose to watch the Steelers or help a hedge fund with a corporate law question. The higher my marginal taxes, the more likely I am to choose the former. This is a losing proposition from a social welfare perspective, no matter what you think of the quality of my advice or the role of hedge funds.

Let’s understand what’s being said here.

If his marginal tax rate is raised from the current 35% to 39% (the rate that prevailed during the Clinton Era), art camp for apple-cheeked Amy will become unaffordable. But, rather than take on that additional consulting gig (which, after taxes, would net him $122/hour, instead of $130/hour) to pay for Amy’s art camp, Professor Xxxx plans to sit on the couch watching football.

Is that really how Professor Xxxx plans to meet this (alleged) financial challenge? If so, he truly does inhabit a different psychological universe. Most of us, when cash-flow gets tight, try to find ways to increase rather than decrease our income.

But, really, who (except for the Xxxx’s) cares whether apple-cheeked Amy gets to go to art camp? What really matters to the rest of us is Professor Xxxx’s contributions to the overall economy. Is it really true that he (and others in his position) will park himself on the couch, watching football, instead of contributing productively to the economy?

Is it really true that raising the top marginal tax rate will hurt GDP, because people like Professor Xxxx will work less hard in response?

A simple test of this hypothesis is to plot top marginal tax rate versus real GDP growth, 1950-present. That’s 60 years worth of data, during which the top marginal tax rate varied between 92% (in 1952-1953) and 28% (in 1988-1989).

Surely, if Professor Xxxx is right, we should expect a negative correlation between real GDP growth and the top marginal tax rate.

scatter plot of real GDP growth vs top marginal tax rate.
Scatter Plot of Annual Real GDP Growth vs Top Marginal Tax Rate, 1950-2009

As you might expect, there’s a lot of scatter in the data, but there is no evidence of a negative correlation between the top tax rate and economic growth. If anything, a linear regression fit shows a positive correlation (albeit, with really crappy R2 of 0.06).

As Brad Delong points out, the CBO comes to the same conclusion about the ineffectiveness of upper-income tax cuts in stimulating aggregate demand.

On the other hand, that the Bush-era tax cuts are the major contributors to the projected deficits in the decade to come is pretty clearly established.

contributors to the near-term budget deficit
Contributions to the budget deficit over the next decade

Next to keeping Republicans out of office, letting the Bush tax cuts expire is the best thing we could do for the long-term budget outlook.

Posted by distler at September 21, 2010 10:44 AM

TrackBack URL for this Entry:   https://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/2279

20 Comments & 0 Trackbacks

Re: Going Galt

Hi,

Having been one of those who bugged you to post some (any) graphical representation to the cold, hard, math in this yarn of your blog, THANK YOU!

See, cold, hard math has this quality of being …well, how shall I put it? Cold and hard come to mind. So, when I am trying to convince some of my mathematically challenged acquaintances that they are voting against their own best interests, graphs like these really, Really, REALLY help. THANKS!!!

In turn, your post did prompt me to look up the ‘blog-wave Prof. Xxxx created, only to realize that it is the same line of “argument” Ben Stein used this Sunday (Sept. 19) on CBS’s “Sunday Morning Show”. So I did read up what Prof. Xxxx had so say, and find him saying things such as:

“The nanny is a must for two-income couples with kids.”

Well, if it is a MUST, than ALL two-income couples with kids should (afford to) have them, or else the system is not fair; plain and simple. Right? RIGHT?

Homework to your Readers:
(1) Estimate the total number of two(or more)-income couples with kids in the USA. (I’m not sure that the average family still has 2.3 kids, but that’s a good starting point.) Now, (2) estimate the number of two(or more)-income couples with kids in the USA who actually afford to hire a nanny. (3) Compare the two results and explain the difference.

Alas! Prof. Xxxx either truly believes his own mantra, or has never done math beyond what can be accomplished on fingers of one hand.

The moral to my post: folks, learn how to convince your mathematically challenged (or, innumerate, as John Allen Paulos would call them) friends, neighbors and acquaintances! For, if you don’t, they most likely will vote against their own (and most likely also your) interests.

Posted by: Tristan Hubsch on September 22, 2010 3:36 PM | Permalink | Reply to this

Re: Going Galt

Prof Henderson is changing the subject. His first post claimed he wasn’t really really rich. Then he objects to Krugman (and O’Hare and Delong) not discussing labor leisure tradeoffs when commenting on a post which didn’t discuss them either.

You are right to mention the ambiguous sign of the effect of taxes on labor supply. Oddly, advocates of low taxes typically ignore basic economic theory, when making their case.

In your post you don’t challenge the idea that income measures actual productive contribution to the economy. This assumption really is standard – it is assumed that income is equal to the marginal contribution to national income. But look at Prof. Henderson’s example – consulting with a hedge fund.

Most people would not consider that a productive activity and would rather hedge funds were a bit less active. The counterargument from Chicago (and not just from whiners in the law school) is “we assume that if you make money you are condtributing to national income.” The key argument is an assumption which no one believes to be simply true. The argument is that simplifying assumptions cna be useful. The full frank argument is that that clearly false assumption is useful to those who want to argue that their taxes should be lower.

Posted by: Robert Waldmann on September 23, 2010 6:14 AM | Permalink | Reply to this

Re: Going Galt

You may be right that, given everything we know, many of us might prefer that the good Professor would work less hard. But there’s no evidence that he will do so, in response to having his marginal tax rate raised.

Rather, his own description of the financial pinch he feels (and says he will feel even more acutely, if his taxes go up) makes it seem rather unlikely that he intends to cut back his income-generating activities, in response.

Still, I love that, even after explaining, at length, why he surely won’t, he threatens to “go Galt” anyway. It’s a charming verbal tick; one of those things that make libertarians so much fun at dinner parties.

Posted by: Jacques Distler on September 23, 2010 8:47 AM | Permalink | PGP Sig | Reply to this

Re: Going Galt

I really hate to defend this professor, but you misread his ‘art camp’ bit. He’s very concerned that apple-cheeked Amy is going to get to go to a government-funded art camp this year. The possibility that she might is obviously of great concern to a fellow like Henderson. He’ll probably need to buy box tickets to the next Stealers game just to cope with his well-justified outrage that his taxes are being increased by 4% to fund apple-cheeked Amy’s art camp.

He does almost make a good connection though. At least little Amy probably doesn’t need a nanny anymore because her dad probably doesn’t have a job anymore. Republinomics: Hey, at least we increased the number of stay-at-home dads.

Posted by: Elias on September 23, 2010 12:59 PM | Permalink | Reply to this

Re: Going Galt

I really hate to defend this professor, but you misread his ‘art camp’ bit.

No I didn’t (unless he’s a much worse writer of English than any Law Professor ought to be).

Posted by: Jacques Distler on September 25, 2010 1:03 AM | Permalink | PGP Sig | Reply to this

Re: Going Galt

I am genuinely concerned that an esteemed law professor seems so blithely unaware of what the government actually spends its money on. Between old-age pensions, keeping old people alive, and getting young people killed we’ve already spent over half the money.

I’m skeptical of the possibility that the government will be helping little Amy out with art camp at all this year.

Posted by: Elias on September 23, 2010 1:09 PM | Permalink | Reply to this

Re: Going Galt

Didn’t Todd Henderson already admit he was mistaken by retracting his posts?

It seems that, after considering the critiques, he decided he was wrong and withdrew his original complaint.

Posted by: encernt on September 25, 2010 12:50 AM | Permalink | Reply to this

Re: Going Galt

My understanding is that he stands by his argument, but deleted his post(s) because his wife objected to his posting of personal financial information to the internets:

The posts that generated an unintended blogocane have been deleted. I stand by the posts, the facts in them, and the points they were making. The reason I took the very unusual step of deleting them is because my wife, who did not approve of my original post and disagrees vehemently with my opinion, did not consent to the publication of personal details about our family. In retrospect, it was a highly effective but incredibly stupid thing to do.

Posted by: Jacques Distler on September 25, 2010 1:02 AM | Permalink | PGP Sig | Reply to this

Re: Going Galt

It was stupid on many levels, but primarily because it totally undermines his argument that he is somehow “middle class” since, even in the richest areas of the U.S., median household income does not even come close to $400K:

http://en.wikipedia.org/wiki/File:US_county_household_median_income_2008.png

Posted by: Ernest Hua on September 19, 2011 2:25 PM | Permalink | Reply to this

Re: Going Galt

Over the past 97 years the correlation between income taxes and economic growth is positive since the top marginal bracket, the average or effective tax rate and the lowest tax bracket have each increased about ten-fold since 1913 when Amendent 16, personal income taxation, was added to the U.S. constitution.


Bob Bronson
Bronson Capital Markets Research

Posted by: Bob Bronson on September 25, 2010 7:58 PM | Permalink | Reply to this

Re: Going Galt

i happened upon your bog/posting….so you believe that the money being taxed belongs to the person earning that money and after that the government can tax….we can get into the wisdom of taxing the people who make the “whole thing go”…the entrepeneur…they put up the risk…put the name and abilities “out there” and many time risk everything…is it fair that reward be 30%, 40%, 50% of what they earned? in los angeleas the stimulus money provided my city produced 55 jobs at a cost od $2mil per job…government does not know how to create a job and should get out of the way…it should lay down the ground rules to play by and get out of the way…and get its hand out of everyones pocket, including yours.

Posted by: irs levy on October 6, 2010 12:26 PM | Permalink | Reply to this

Re: Going Galt

You write “Surely, if Professor Xxxx is right, we should expect a negative correlation between real GDP growth and the top marginal tax rate.”

This is precisely wrong. If Prof Xxxx is right, it does not follow that we will see a negative correlation between GDP growth and the top marginal tax rate. For example, suppose that Congress routinely cuts tax rates in response to recessions. That may create a positive correlation, even if Prof Xxxx is right.

I don’t mean to suggest that this is the right explanation. What I do mean to suggest is that this is a complicated issue. Getting it right will involve some delicate modeling, and it is certainly not amenable to the simplistic analysis you are offering. Your citation of the CBO report is, therefore, infinitely more convincing than your own analysis; even though I cannot see the original source, I presume the report was prepared by economists who would not make such simple errors.

If you think its good practice to make naive arguments about fields you vaguely understand solely on the basis of blog posts, then let me tell you about a theory of everything I’ve just discovered…..

Posted by: alex on November 4, 2010 4:46 PM | Permalink | Reply to this

Re: Going Galt

For example, suppose that Congress routinely cuts tax rates in response to recessions. That may create a positive correlation, even if Prof Xxxx is right.

It “may”, but it doesn’t.

There’s a simple way to test your hypothesis. Post-war recessions (as they are conventionally-defined, in terms of contracting GDP) have been quite short in duration. Only once (in 1974-75) has there been a period of two consecutive annual declines in real GDP. If your hypothesis is correct — that top marginal tax rates are cut, in response to recessions, hence (spuriously) coincide with contracting GDP — then introducing a relative shift of one year (I actually tried both 1 and 2 year shifts) would cure the problem, as the tax cut would then “coincide” with the post-recession rebound in GDP.

Unfortunately, that’s not what happens. The (weak) positive correlation between top marginal tax rates and real GDP growth remains.

You are right, that having a model is better than a stupid linear regression. In particular, a model would suggest a causal explanation, rather than merely signs of a correlation (or lack thereof).

But you need a model that fits the historical data. And, wishful thinking aside, I don’t believe that anyone has a model that backs up Professor Xxxx.

You’re welcome to suggest one, though ….

…let me tell you about a theory of everything I’ve just discovered….

Let’s hope it’s more convincing.

Posted by: Jacques Distler on November 4, 2010 8:37 PM | Permalink | PGP Sig | Reply to this

Re: Going Galt

Dear Prof. Distler - your reply misunderstands my point. My words were carefully chosen: “If Prof Xxxx is right, it does not follow that we will see a negative correlation between GDP growth and the top marginal tax rate. For example, suppose that… This does not mean that I believe that what follows after “for example” is the truth. I am merely listing one way in which your claim “If Prof Xxxx is right, then…” can fail.

You respond by outlining a more sophisticated analysis, by that only proves my point - that a more sophisticated analysis is necessary! If you really want to argue your point, you also need to anticipate other ways your analysis can fail, other mechanisms in which the choice to cut/increase taxes is intertwined with economic growth. This can turn out to be a quite sophisticated production, but fortunately there is a hundred years of economic thought on the subject. There is no need for you or me to engage in “amateur economics.”

“I don’t believe that anyone has a model that backs up Professor Xxxx.”

I believe this is not right. What about

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=225613

Note that I am not saying I agree with the conclusions of this paper. I am merely trying to give a counterexample to your statement that no one has a model that backs up Professor Xxxx.

But I am incredulous that those other effects would be so precisely-tuned, as to completely wipe out the effect of the marginal tax rate

You shouldn’t be incredulous. My understanding is that this is quite common in the social sciences: a phenomenon which is absent at first glance suddenly appears once a key variable is controlled for.

Posted by: alex on November 7, 2010 4:35 PM | Permalink | Reply to this

Re: Going Galt

I believe this is not right. What about

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=225613

Note that I am not saying I agree with the conclusions of this paper. I am merely trying to give a counterexample to your statement that no one has a model that backs up Professor Xxxx.

Actually, I would say that paper (whose full text is available here) backs up my contention quite nicely. They conclude that a 5% across-the-board tax cut would yield a 0.22% increase in GDP growth.

Given the statistical (and systematic) uncertainties in each of the steps of their analysis, any rational person would call that effect “statistically indistinguishable from zero”.

And that’s for a 5% across-the-board cut. Presumably, a 3% change in just the top marginal tax rate (which is the subject of Professor Xxxx’s lament) would have an even smaller effect (i.e., would really be indistinguishable from zero).

Note that I am not saying I agree with the conclusions of this paper.

Well, I don’t believe it, either … for the simple reason that they make no attempt whatsoever to estimate the errors in their analysis.

But it really would be more interesting, were you to cite a paper that you did believe.

You respond by outlining a more sophisticated analysis, by that only proves my point - that a more sophisticated analysis is necessary! If you really want to argue your point, you also need to anticipate other ways your analysis can fail …

I studied lots of variants of the above. In addition to the 1- and 2-year shifts in the GDP growth time series (mentioned above), I also looked at 2- 3-, and 4-year trailing averages.

My motivation, for doing so, is obvious from looking at the graph above (and even moreso, from looking at the actual time-series). There’s a lot of vertical scatter in the GDP growth rates, at any given tax rate. The reason is simple: the tax rates tend to stay constant, often for many consecutive years, even as the economy itself goes through the usual business-cycle fluctuations in GDP-growth. If you’re trying to isolate the effect of tax rates, perhaps it makes sense to average over those business-cycle fluctuations.

Sure enough, employing trailing averages does produce a less-noisy graph, with a marginally better R 2R^2 for the linear fit. Which, I hasten to add, is still a slight positive correlation. But, like the graph above, the only honest conclusion is that the result is consistent with there being no correlation at all.

The reason why I didn’t present these latter graphs is that they suffer from an obvious objection: while there are long periods when the tax rate remained constant, there are other periods when it changed multiple times over the space of a few years. It obviously makes little sense to study a trailing average GDP growth as a function of top marginal tax rate in year nn, when the top marginal rates in the out-years were very different from the rate in year nn.

There are ways around this, and I thought about doing a yet-more-sophisticated analysis. But I concluded that would be a waste of time. There’s no hint of a correlation in the time-series, and further attempts to coax a correlation, from it, is like hoping to see an image of Elvis emerge from the “snow” of late-night TV (an antiquated reference that will make no sense to this generation, raised on Cable, and 24-hour TV programming).

By the way, the fact that the time-series of US GDP-growth versus tax rates (whichever rate you choose for the latter) holds no evidence for a negative correlation between tax rates and GDP growth, is also the conclusion that Engen and Skinner come to. They turn to cross-country comparisons, to try to tease out an effect, with the “desired” sign. That introduces its own set of methodological problems, which make my analysis, crude as it was, seem positively respectable.

Now, of course, you are right that there are a gazillion papers on this subject. Doubtless many of them come to stronger-sounding conclusions than Engen-Skinner. After all, it is something of an article of faith that lower tax rates “should” lead to faster GDP growth (well, OK, not in the Solow-Swan model, where they have no long-term effect, whatsoever).

So I’ll retract my “nobody has a model.” Doubtless, somebody does. Instead, I’ll content myself with the statement that (to use your formulation) “‘a hundred years of economic thought on the subject’ has produced little-to-no evidence to back up Professor Xxxx.”

Posted by: Jacques Distler on November 7, 2010 10:34 PM | Permalink | PGP Sig | Reply to this

Re: Going Galt

Let me put it another way.

I am willing to believe that you have a model in which lowering the top marginal tax rate has a significant positive impact on GDP growth. Of course, other effects will impact GDP growth, too. But I am incredulous that those other effects would be so precisely-tuned, as to completely wipe out the effect of the marginal tax rate, leaving no correlation whatsoever.

Or, more pithily, “correlation ≠ causation” but the absence of correlation…

Posted by: Jacques Distler on November 4, 2010 9:24 PM | Permalink | PGP Sig | Reply to this

Re: Going Galt

What Professor Xxxx and other supply-side kool-ade drinkers conveniently forget is that Reagan followed his huge 1981 tax cut with two large tax increases. In fact, no peacetime president has raised taxes so much on so many people. This is not a criticism: the tale of those increases tells you a lot about what was right with Reagan’s leadership, and what was wrong with the leadership of Duh-bya.

Posted by: Chuck Beretz on March 1, 2011 12:05 PM | Permalink | Reply to this

Re: Going Galt

ugh, math.
but! you had me til the math started.
Professor Xxxx sounds like a douche.

Posted by: rebecca on March 28, 2011 12:28 PM | Permalink | Reply to this

Re: Going Galt

Your scatter plot is idiotic. A two-dimensional scatter on something as complex as the point you’re trying to make (or assail) is dumber than either the pro or con position on this issue. People just love to wield ill-conceived math as if it’s some irrefutable dogma. Scattered plot is more apt, methinks.

Posted by: Tripp Knightly on May 29, 2011 11:19 AM | Permalink | Reply to this

Re: Going Galt

I came across your post while researching for a tax article, there is alot of debate at the moment over here in the UK about taxes.

If the loveable Prof was in the UK (casting aside the fact that he would not have that he probably would not have that salary in the first place) he would be liable to pay 50% tax on anything he earned over about $150,000.

Of course this is all a bit of a farce, even people earning over that amount would rarely pay it. They would use things called accountants and form a company.

All the money would then be paid to the company rather than the person and then paid from the company to the person in the form of dividends which would not attract income tax at 50% but corporation tax at 20%. Nice.

They would then pay themselves a token salary in which attract a tiny rate of income tax if any at all.

But those tactics, totally legal, are for the underachievers. The people earning serious money form their companies off shore or are registered as non-domiciles (they do not pay UK tax).

Lovely. So in summary in the UK we have a multi-tier tax system where the very wealthy pay very little and in some cases, is actually a tax haven (even though the UK is trying to jump all over other tax havens). Those that get caught by the 50% tax rate are tiny and could get out of it by re-arranging their affairs.

I presume their are similar things going on in the US, each state has their own laws, some don’t charge income tax other do?

Posted by: James on September 24, 2011 6:19 AM | Permalink | Reply to this

Post a New Comment