Winner of the New Statesman SPERI Prize in Political Economy 2016

Thursday 24 January 2019

The key arguments for high top rates of income tax are political as well as pecuniary

When people complain that neoliberalism is a meaningless concept, I should point them to what has happened to the top rate of income tax since around 1980, not just in the US and UK but elsewhere. (Source HT Marcel Fratzscher)

Here is a chart of the United States top tax rate over the last century (HT Martin Sandbu). Eisenhower had top earners paying a 91 percent marginal rate.

No doubt there are complex reasons for these reductions, but key among them has to be a neoliberal belief that cutting top rates would lead to more dynamic CEOs who would produce more dynamic companies, and the benefits of this would trickle down to the economy as a whole. Low top tax rates would encourage entrepreneurs to take more risks that were socially beneficial and so on. The argument is so familiar, trotted out routinely by right wing think tanks, it hardly needs elaborating. It is a classic example of neoliberals using a bit of simple economics to justify policy that is advantageous to themselves or their paymasters.

Yet the evidence for such an effect is weak, at best. The intuition for why it should be weak is straightforward. Above a certain level of income, other incentives beyond the purely pecuniary become important. Top CEOs, like top footballers, want to be successful at what they do, and more successful than others. They will want success whatever the overall financial rewards of being successful.

Another bit of basic economics that neoliberals hardly ever mention is the diminishing marginal utility of consumption. This implies quite the opposite of low tax rates at the top. It is socially much more beneficial to tax those to whom one dollar is not worth the effort of picking off the sidewalk and transfer it to those who are poorer. A well known paper by Diamond and Saez found that, after allowing for disincentive and avoidance effects, the optimal top rate of income tax in the US should be 73%. [1] [2]

There are two reasons why even 73% might be an underestimate. Piketty, Saez and Stantcheva have argued that giving CEOs lots of money can have negative incentive effects. The CEOs start putting effort into increasing their salary rather than improving their firm. Part of your status comes from what you can afford. When all CEOs are taxed a lot at the margin the size of your salary has little impact on that, but when your salary is not taxed so much you can increase your salary and theefore status by extracting more from your own firm. To use some economics jargon, a low marginal tax rate on top incomes can be a good example of an incentive for rent extraction rather than an incentive for increasing social output. .

But while an extra dollar for a CEO is not going to incentivise them in a positive way very much, you could argue that it incentivises those with talent to aspire to be CEOs. CEOs are always going to be among the richest in society, because a lot of their income will be taxed at lower rates. A paper by Lockwood, Nathanson and Weyl turns that argument on its head. High salaries are associated with activities, like finance and law, that have what economists would call negative externalities, which means that they do much less good for society than the size of the salaries they pay might suggest. A lot of finance, for example, is about trying to take money off other people rather than growing the size of the overall pie. If high post-tax salaries incentivise talented people into those professions, that is negative for society, which would benefit if they worked in different jobs. You can reduce this misallocation of talent by having higher tax rates on top incomes.

Neoliberals have one last line of defence against raising top tax rates in a single country, and that is migration. The argument is that talent, which could be quite mobile, will move to where talent is most rewarded. There is clear evidence that this is true, to an extent. This concern does not mean we leave top tax rates where they currently are or even reduce them, but simply that we might not put them up as high as they should otherwise go while some countries that are attractive to talent continue to have low tax rates on top incomes. Sweden seems to do pretty well with a 70% effective top tax rate.

This danger of a race to the bottom with top tax rates makes it all the more important that the United States raises it’s top marginal tax rate, along the lines recently suggested by Democrat Alexandria Ocasio-Cortez. For various fairly obvious reasons the US does not need to worry too much about a talent drain if it raised top tax rates.

In my view the arguments for higher top tax rates are at least as much non-pecuniary, by which I mean they do not depend on the points raised so far. The evidence that social welfare is higher in more equal societies seems compelling to me. In other words we should increase top tax rates just because that helps produce a more equal society. I have seen a few attempts to debunk the evidence for this in The Spirit Level, but they are not convincing as a collective, while there is even more evidence to support the idea that people are happier in more equal societies..

There is a final argument for high tax rates at the top which seems particularly relevant to the US and UK at the moment. If you have a political system like the US where money easily buys political influence, you will find some of those who earn very high salaries trying to do exactly that (some references here). You can create the kind of plutocracy I discuss here. Because money can also help to buy votes, that plutocracy may also be able to continue with democratic elections without in any way threatening the plutocracy. Even when you have laws limiting the amount that can be spent on elections, the UK shows there are ways for the rich to get around that, particularly if they control large sections of the press.

This is the argument made in this excellent NYT op-ed by Emmanuel Saez and Gabriel Zucman. They write
“An extreme concentration of wealth means an extreme concentration of economic and political power. Although many policies can help address it, progressive income taxation is the fairest and most potent of them all, because it restrains all exorbitant incomes equally, whether they derive from exploiting monopoly power, new financial products, sheer luck or anything else.”

The economist Greg Mankiw, in a short response to this op-ed, says
“most rich people I know would have been happy to spend vast sums of money to keep Mr Trump out of the White House. And many tried. The Trump phenomenon is not an argument that the moneyed elites have too much influence on politics. If anything, it is an argument that they have too little.”

But this misunderstands (as some on the left do) the nature of the plutocracy that super incomes and wealth create. It does not, as I make clear here, create a kind of committee of the very rich that between them decide who rules. It is much more erratic than this. Instead it allows small groups among the very wealthy, who may be quite unrepresentative, to hijack a democratic system. Trump and Brexit are clear examples. Mankiw is right that one way to avoid that would be to create a more representative kind of plutocracy, but a far better way of avoiding disasters of this kind is to deal with the problem at its source, by reinstating high rates of tax on top incomes.

[1] Some intuition on this number. If disincentive and avoidance effects were so large that an increase in tax rates led to no additional income, then there is no pecuniary advantage from doing so. If these effects did not exist, the optimal marginal rate at the top would be 100%. The paper estimates these effects are somewhere between these two extremes, so the optimum tax rate at the top will be between zero and 100%.

[2] For a discussion of these issues in a UK context, see Note that HMRC calculations that the short-lived raising of the UK top tax rate to 50% did not raise any revenue involve debatable assumptions, and much of the avoidance only worked because the hike was temporary.    


  1. A lot of what seems to me contradictory statements in this article.

    In one paragraph, you say "Top CEOs, like top footballers, want to be successful at what they do, and more successful than others. They will want success whatever the overall financial rewards of being successful". In the next paragraph, you write "The CEOs start putting effort into increasing their salary rather than improving their firm". Likewise, you say "neoliberals hardly ever mention is the diminishing marginal utility of consumption" and then say "Part of your status comes from what you can afford.... when your salary is not taxed so much you can increase your salary and theefore status by extracting more from your own firm".

    High taxes discourage risky investments in startups. And the bigger problem is that a lot of good ideas sound bad which makes such investments appear even more risky and won't happen at all if the tax rate is high. Jones in "Taxing Top Incomes in a World of Ideas" arrives at about 30% as the optimal tax rate.

  2. As you've mentioned, a lot of work in finance is zero-sum and not beneficial to society. But I hardly see any cost-benefit analysis of administering taxes. There are so many accountants and people whose work is primarily determining or arguing about how much taxes are owed rather than something more productive. Isn't this a poor allocation of resources? Shouldn't we simplify taxation so that some of these people can be shifted to more productive work? This isn't an argument against high tax rates, but that certain types of taxation are easier to administer and even though, at first glance, this may seem unfair, we might still be better off.

    1. First, you do not understand what zero sum means.

      Second, the increasingly high cost of finance has greatly increased incomes of those in finance, and unless you can point to a hundred million starving Americans, its clear the high cost of finance in the food sector, eg commodity trading, gene patents, etc, which has paid lots of people working in finance, and hindering food production, and consumer spending on food, has not reduced the amount of food bought by American workers.

      And where people are starving, no one is profiting in food related finance. Eg, is North Korea's starving population caused by rent seeking by the global finance sector?

  3. On my reading of the Diamond-Saez paper, the 73% is actually optimal *if* you accept their priors on the social welfare function AND if you imagine a re-designed tax system where avoidance is eliminated/impossible. In fact, their own paper says that with the current U.S. tax code incorporating deductions etc, the revenue maximizing federal marginal income tax rate on top earners would be 48% (and combined with state, around 54%). That's certainly higher than the 37% today, but nothing like the 70% advocated by Ocasio-Cortez.

  4. I'd like to advocate for a replacement of the term "neoliberal" with the term "paleoconservative". When you think about it, there's nothing either new or liberal in what is being advocated by the "neoliberalism" you describe.

  5. perfect summary of the distortions created by low tax - thankyou Simon - suspect this will spread far and wide

  6. It's bizarre to say that Trump hijacked democracy when he had to win the GOP primary and then the general election and spent less money than his opponent each time. If that is hijacking, it doesn't mean the alternative was not also plutocracy and also caused by taxing the rich too little: Jeb Bush had $100 million from the start and Clinton spent over a billion. "Unrepresentative" of the rich, of course.

  7. Looking at the rates but not looking at the tax collected at those rates is not sensible.
    Punitive rates are 100% political.

  8. High marginal rates can be good for progressing society and the economy for many reasons. I completely agree with your thesis that high taxation rates motivate owners and executives in businesses to work for the benefit of the firm (an idea or an objective) rather than for personal gain. And diminishing net income after a certain threshold is still income.

    There are other benefits:
    1. High marginal taxes tend to make men and women participate in the economy, and taxation can definitely be designed to eliminate the marriage penalty.
    2. In the U.S., payroll taxes and state income taxes do have a net comparative effect on the taxes paid. Increase of teh marginal taxes should account for the cost of the tax burden from state income taxes, with a threshold set at a median top taxation level across all the states, and a fraction of the excess tax being deductible from federal taxes. Also, payroll taxes should be paid from the general fund with guaranteed funding, with part of the payments going towards funding future obligations as calculated from the demographics of the current population.
    3. One datapoint that offered a stark reminder of the level of inequality is the extent to which federal workers were put into hardship by the shutdown. This is a moral deficiency in US society, and must be remedied across the board. Businesses who pay workers less than minimum wage should be taxed higher to account for the social burden of offering top up living allowances to people who work. A national healthcare system and opportunities for higher education for students is another essential component.

  9. What low taxation has shown is pricing and the way pricing is set is false.whether in the real world or way economist calculate it! afterall price distributes money through the economy,and both the private sector and public sector is hostage to the price setters! in theory if price achieved what economist says it should create volume of (rather than slave labour)work! tax could be lower! the last 50 yrs as shown free markets are not good at setting good prices and the distribution within pricing is very poor even why it might be right!And the neoliberal deliberate paradox of low taxation,unfettered price administration (maximising profits)is one the world economies can never escape without breaking that paradox

  10. In the US, only in a few States is work/production taxed.

    The primary taxation is on net income, aka profit, aka money NOT paid to workers.

    The number one tax dodge for any business is labor costs.

    Milton Friedman circa 1970 argued high tax rates caused too many workers be paid too high wages and benefits because the "IRS pays half the costs".

    He argued for low tax rates and eliminating the tax dodges of quick expensing of labor costs, eg R&D expensing, and very fast capital depreciation expensing. He argued that unions had more power when tax rates were high because the corporate executive could solve a labor conflict by having the IRS pay over half the cost of union demands.

    And the high tax rates led to too much paid to workers who then consumed too much putting too much revenue into businesses that focused on revenue growth, not profit growth, business executives finding it easier to pay $2 more to workers because it only reduced profit growth by $1 to easkly grow the size of their business fiefdom.

    Note, FDR paid for social insurance by a flat tax on work/consumption.

    And in the rest off the world, work and profit and consumption are all taxed by a single tax rate, the VAT aka GST aka sales tax, in the place where the work and profit are paid for, at the point of sale.

    Something that either greaatly confuses Trump and his advisors, or frustrates them because they can't cut the taxes Europeans pay on US goods below the taxes paid on European goods. Ie, not taxing labor costs in the US and taxing profits at less than 21% does not prevent US goods from being taxed at a typical 25% VAT. But EU produced goods are not taxed when sold in the US, except for any profit recorded in the US, other than tariffs. Trump sees the rest of the world cheating by having high taxes on consumption from all over the world, stealing tax revenue from the US, and then taxing US businesses unfairly when the US does not tax US businesses.

    But the thing the GOP does not want taxed is NOT PAYING WORKERS. Paying fewer workers, paying workers less, should be rewarded with lower tax payments to the IRS.

    As for individuals, limited partnerships were used widely to start businesses in order to get investment losses to shelter more income than actually invested in cash. The expenses were ultimately labor costs. This meant US oil production increased until tax laws were changed making paying workers to drill baby drill unprofitable, even as the real price of oil was falling.

    The real reduction in tax rates and real elimination of tax dodges happened in 1986, but now tax tax deductions did not significantly cut taxes due, ie, paying workers $100 would only reduce taxes by $35 max, tax credits were added which would reduce taxes due $50 for every $100 paid to workers, up to $100 in reduced tax bills for every $100 paid to workers. For example, paying workers to reduce energy consumption in your home, or when you drive to work earns you tax credits for some of the money paid to workers. Given many workers pay too little in taxes, the tax credits a working class homeowner earns get bundled and traded in limited partnerships to those paying the "crushing" Obama ~38% income marginal rate.

    Oh, and I grew up when economksts held as an axiom that profit was a sign of economic inefficiency - the economy was not delivering what customers demanded, so instead of increasing production, prices were simply increased far above cost and the money pocketed by those not increasing production to meet demand. Keynes argued for tax and other policies to punish not paying more workers to increase production and drive prices together with labor costs, ie prices down and labor costs up.

  11. Simon, I agree with you that top tax rates should be increased, and I agree that the political influence of the hyper-wealthy is problematic. That said, I think there are a couple of arguments here which are problematic:

    First, you've graphed statutory marginal tax rates in the U.S., but effective marginal rates were much lower than effective during the post-war period. (And, FWIW, to whatever extent effective rates did fall, "neoliberal" is far broader and far fuzzier than a description of marginal tax rates; it remains an ill-defined cuss word.)

    Second, the Spirit Level is terrible empirical work consisting largely of underspecified and unidentified ecological regressions. One of the major problems with it, and some of the literature on which it's based, is it confounds poverty and inequality. There is no evidence that destroying the incomes of relatively rich people without increasing the incomes of people at the bottom does any good; probably the opposite. The evidence from more sophisticated empirical work suggests that most of the ecological association between inequality and bad outcomes results from concave relationships at the individual level: if e.g., i's health, h, as a function of her income, y, is h(y_i), and h() is concave, then mechanically associations such as those in the Spirit Level obtain, but at the same reducing inequality does harm if (1) it fails to sufficiently lift the bottom or (2) sufficiently reduces mean income.


  12. Hi Simon,

    Interesting and informative article.

    If possible, could you advise where one can find data on the history of British Tax Rates at all levels for the past 150 years or so (have tried finding this myself but been unable to find any relvant sources)?

    Also, is there any 'standard' work that outlines the associations between tax rates and levels of GDP growth?


  13. Do you know that how many people pay taxes every day in the USA? You will be amazed by the figures issued by the IRS. You can check them here IRS Figures and read in detail about it.


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.