Optimal Research on Optimal Taxation

November 28th, 2011 at 12:31 pm

After constructing an authoritative database on income shares covering many years and numerous countries, economist Emmanuel Saez, Thomas Piketty, and colleagues have been using these data to plumb important insights on the relationships between growth, policy measures (e.g., taxes, unemployment insurance), and inequality.

This paper examines the relationship between income shares at the top of the scale—the richest 1%–and different tax regimes.  The paper asks and answers two important questions: 1) do lower marginal tax rates on the rich result in higher income inequality, and 2) do higher taxes on the rich result in lower economic growth?

Answers: yes and no.

Panel A in the figure below shows a strong negative correlation between reductions in the top marginal rate and the increase in the share of income held by the top 1% of households.  Panel B, on the other hand, shows no correlation between such changes and the growth of GDP per capita.

Note that these income shares are pretax, so the correlation in the top panel is not simply induced by the rich keeping more or less of their earnings.  You lower tax rates and you just end up with a lot more concentration of the growth in earnings and assets.  And what do you get for that in terms of growth for everybody else?  According to the bottom panel, nothing.

I’ve long asserted that trickle-down tax policy—cut taxes on the rich and unchain the economic activity that will enrich everyone else—doesn’t lead to faster growth.  It just redistributes income upwards.  My views on this were strengthened by the very different outcomes in terms of growth, jobs, middle-class incomes, poverty reduction, and our fiscal health during the Clinton vs the GW Bush years.  Based on that natural experiment of supply-side trickle down, these findings shouldn’t surprise you at all.

They also complement Paul K’s oped today on why higher taxation on folks at the top of the income scale needs to be part of any plan for getting on a sustainable budget path.  Saez et al conclude that the optimal tax rate—the one that would raise the most revenue with the least economic distortions—for the highest bracket is 83%. 

Clearly, these guys aren’t running for office.  But you get the point.  Allowing the Bush highend cuts to sunset takes the top rate from 35% to 39.6%.  How about based on these results we nudge that up to 40%, shake hands and call it a day.


Source: Piketty, Saez, and Stancheva, 2011–see link in text.

Print Friendly

13 comments in reply to "Optimal Research on Optimal Taxation"

  1. Tyler says:

    Christina Romer, 2009: “The fundamental cause of [the recession of 1937-38] was an unfortunate, and largely inadvertent, switch to contractionary fiscal and monetary policy.”

  2. Tom says:

    We know that the tax rate for the highest bracket that raises the most revenue is greater than 70% from the natural experiment conducted in 1981-1982 (i.e., Reagan tax cuts.)

  3. Jake Lopata says:

    I can’t remember what blog I got it from, although I suspect that it might have been from your blog; but I read ‘The Case for a Progressive Tax:From Basic Research to Policy Recommendations’ by Peter Diamond and Emmanuel Saez, they also find that the top marginal tax rate should be in the ball park of 49-73% (depending on the methodology used); good findings.

    I know you know this Jared, but for your readers it might be important to note:

    “Correlation does not imply causation”

  4. jonathan says:

    The paper is complicated and seems to be written intentionally to avoid dealing with some of the issues it raises; it presents itself as a policy rather than an economics paper.

    They push 2 bits together and spend no meaningful time talking about 1 of them. The bits are tax rate effects, meaning income changes as tax rates change, and welfare utility for the rich. They assume away the latter and that leaves the paper open to criticism. They don’t, for example, include any numbers on what a range of utility values for the rich would generate as optimal tax rates. They note it would take a very high utility – something near .75 – to generate the current highest marginal rate given their other assumptions. That’s interesting but it needs context to be persuasive. It would also help if they talked about what it means to have certain utility values; I’ve already seen “conservative” commenters pointing out their number as a flaw.

    They also use what some would consider a low value for income sensitivity to tax rate change. They note that using a much higher rate would still generate a higher marginal rate than now exists but why not give a graph or table?

    The main point I took from the paper is that if you look at historical income and tax rate information and then apply a rational utility value, you get a range of tax rates that are higher than today’s and you only get today’s rates if you argue for radically skewed numbers.

  5. urban legend says:

    Better than one top 40% bracket: 40% on up to $1 million taxable (starting where top rate starts now), 45% on $1+ million to $10 million, 50% on $10 million +. Not unreasonable, probably politically popular — so what’s the revenue from that?

  6. Geoff Swenson says:

    I don’t think we need to have brackets any more. A simple, continuous equation will do just fine.

    rate = MAX(70 – ((12,000/Income)^0.4),0)

    This way any small increase in salary doesn’t put you in a worse tax bracket.

    $8,000 0%
    $16,000 0%
    $32,000 5%
    $64,000 21%
    $128,000 33%
    $256,000 42%
    $512,000 48%
    $1,024,000 54%
    $2,048,000 58%
    $4,096,000 61%
    $8,192,000 63%
    $16,384,000 65%

  7. Joe K says:

    How about we nudge it up to 40%, tax all income (including SSO) as ordinary income, add a 10% surtax on income over $1 million (until fiscal/economic situation stabilizes), and restore estate tax to a flat 50% $2 million exclusion. Still plenty to do to phase out tax expenditures and subsidies, with an offset at the low end by increasing the standard deduction. But this would be a real start toward fixing the economy if we could couple it with a serious direct jobs program (a la WPA).

  8. Mike Sax says:

    Hey Jared! There has been a pretty good debate about this paper on the econ blogs. At Angry Bear there’s this guy Mike Kimel who has been working on this same idea of the optimum tax rate. Of course Scott Sumner over at Money Illusion has been denoucing the idea.


    The idea that lowering top marigninal tax rates does nothing to growth but increases income inequality sounds about right from what empirical history tells us. So during the Bush years until the crisis hit we saw GDP get as high as 3.5 percent but we also saw inequality rise. So the Stephen Moores and Arthur Laffers have no basis to claim that Bush’s tax cuts gave us the high growth-or at best we could have found other ways to gain a similar level of growth with more income equality.

    Please try this


  9. Mike Sax says:

    I have come upt with my own tax policy. As I was born in Egland though I’m an American citizen it will have to be state policy where I picutre myself as governor of the state of NY. Please take a look. With all the tax proposals out there I add my own to the mix http://diaryofarepublicanhater.blogspot.com/2011/11/ok-lets-talk-taxes.html

Leave a Reply

Your email address will not be published.

Current month ye@r day *