The top marginal tax rate

Taxing Top Incomes in a World of Ideas
Charles I. Jones∗
Stanford GSB and NBER September 26, 2018 — Version 0.5
Preliminary Abstract
This paper considers the taxation of top incomes when the following conditions apply: (i) new ideas drive economic growth, (ii) the reward for creating a successful innovation is a top income, and (iii) innovation cannot be perfectly targeted by a separate research subsidy — think about the business methods of Walmart, the creation of Uber, or the “idea” of These conditions lead to a new term in the Saez (2001) formula for the optimal top tax rate: by slowing the creation of the new ideas that drive aggregate GDP, top income taxation reduces everyone’s income, not just the income at the top. When the creation of ideas is the ultimate source of economic growth, this force sharply constrains both revenue-maximizing and welfare-maximizing top tax rates. For example, for extreme parameter values, maximizing the welfare of the middle class requires a negative top tax rate: the higher income that results from the subsidy to innovation more than makes up for the lost redistribution. More generally, the calibrated model suggests that incorporating ideas and economic growth cuts the optimal top marginal tax rate substantially relative to the basic Saez calculation.

Via Illya Novak.  Here are the slides, here is the paper.


Reminds me of the model from a few decades ago, referenced by one of TC's colleagues the other day in his blog, that the optimal tax rate for the top 1% is in fact zero. The idea being that it gives incentive to the middle class to reach the 1%. Fat chance of that happening politically however.

This paper seems to assume that the top 1% of the income distribution are innovators and not rent seekers or trust fund babies

Yeah, this. I’m sorry, model everything else all you want, but at least until your model accounts for innovation in rent-seeking and for exchangeability among actual innovators, reward looks to me self-evidently like a tenth-order binding constraint at best. Want more innovation? Eliminate the bottlenecks to securing startup capital. I’m looking at the gatekeepers.

@SC - glad to see you are for reforming patent laws, as I am, to make them stronger and fairer.

Actually, in the context of the post corporate law rather than patent law might be more relevant. Currently the corporation owns the intellectual output of its employees, even if all they provided was the paycheck, a desk and a computer similar to the one the employ owns. There might be some discussion with other employees driving the innovative though but again, that is not really something the corporation is uniquely providing.

Requiring some form of relative contribution share in the patent would certainly drive more of the revenue derived from innovative efforts towards the rank and file in society and out of corporate accounts, executive management of shareholders.

No one should have to pay more than 50% of their income in taxes, all taxes in total. Federal, State, local, property, sales, etc.

+1 Anything over 50, or even 40% is just theft.

So why is >40-50% theft, but 35% not theft?

I mean this is a theoretical paper... but also that is true?

68% of the UHNW population are self-made.

I don't think startup capital constraints are particularly high in a message where startups are rejecting VC capital (eg

Why is a tax rate of zero resulting in zero taxes paid better than a tax rate of 90% and tax payments of zero due the tax dodges of paying workers to gain tax dodges?

For example, Milton Friedman damned the limited liability tax dodge investments in drill baby drill that resulted in huge losses to investors, losses higher than the income invested, because this kept the US oil production growing, when fewer workers should have been paid, less oil produced in the US and more oil imported from regions that required more US military to protect US oil production assets in dictatorships.

He damned the high rates on profits that made building new factories, power plants, better machines, too cheap after taxes, so too many workers were paid, which caused both wage inflation plus too much consumer spending and consumer price inflation, rising income and living costs, and in the end, too small profits from too much competition. Thus less tax revenue than if there were less paid to workers and thus high monopoly profits.

Friedman was opposed to Keynes' view that:
"I feel sure that the demand for capital is strictly limited in the sense that it would not be difficult to increase the stock of capital up to a point where its marginal efficiency had fallen to a very low figure. This would not mean that the use of capital instruments would cost almost nothing, but only that the return from them would have to cover little more than their exhaustion by wastage and obsolescence together with some margin to cover risk and the exercise of skill and judgment. In short, the aggregate return from durable goods in the course of their life would, as in the case of short-lived goods, just cover their labour costs of production plus an allowance for risk and the costs of skill and supervision.

"Now, though this state of affairs would be quite compatible with some measure of individualism, yet it would mean the euthanasia of the rentier, and, consequently, the euthanasia of the cumulative oppressive power of the capitalist to exploit the scarcity-value of capital."

Friedman wanted high rent seeking, high monopoly profits, lower paid workers, lower consumption.

"...the point is that the existing literature has neglected an important consideration that appears to have a first-order effect on the calculation."

This should have been an immediate suspicion of everyone to the 70% marginal rate finding. That so many people are willing to accept such an extreme and tenuous claim is really kind of sad. As much as I don't like to admit it, there is a good reason why economics has a less than stellar reputation.

This seems based on the idea that people are motivated by money rather than by relative status. I don’t that’s true. Especially at the top.

The better question is are the tax rates low enough to allow the successful to sufficiently signal their success.

Sure, *at the top* people may be motivated by status. People who are already there aren't the point. The point is incentivizing people to do the work to get to the top, and those people are absolutely incentivized by money.

Precisely why this paper is silly, as another commenter pointed out – people already at the top are surely rent-seeking, and not innovating.

People closer to the bottom are innovating (e.g. Walmart, Uber, Amazon founders) – that's why we have marginal taxes, you get to collect *all* that capital on the way to the top, but once you're there, the marginal dollar matters less.

Incorrect. The high top marginal rate will impact human capital decisions for those not yet at the top. They will choose to forgo investment in themselves, or investment in high-risk, high-reward ventures because the reward distribution is insufficiently long-tailed.

One argument in favor of low marginal tax rates is that money left in the hands of a founder who has successfully created a socially useful enterprise is likely to be used to create another socially useful enterprise. Think, for example, about Elon Musk creating Spacex after having created Tesla, or the existence of alphabet based on google’s wealth.



Maybe one argument against is making sure Elon Musk has access to more of the government subsidies that he relies on to fuel more "socially useful enterprises".

Yeah this is about risk-adjusted expected returns for a person weighing whether to become a founder or a safe/not-innovative desk job. It's true insofar as you believe that this actually happens (which is... probably true?).

I think you point is true for some but clearly not completely true -- those at the top don't really seem to say "I got all I need so I'll stop getting paid now. Give my money to someone else, or don't worry about a raise or a competitive package to keep me here.

I do agree some are driven by certain forms of status but a lot of that still remains tied to level of wealth they enjoy.

Now, there are some that are saying I'm putting my efforts into man kinds future and the status my success in that area will bring. Perhaps someone like Musk or Beitos fits that, or Gates or Buffet and Diamon (with a 3rd but forget who) with their focus on improving health or insurance costs. But for everyone one of those I suspect you get 100 Trumps who are self-centered and largely unconcerned about what bad impact their efforts have on others.

Wow, condition ii is doing a lot of lifting here. The individuals with successful innovations that make the world better for everyone else. Never mind that the most successful innovations tend to come from group efforts perhaps led by an individual entrepreneur. But if the gains were generated by a group, how best to allocate the ‘top income’ that results? And don’t forget that we’re assuming the innovator gets to capture the returns rather than lose to a second-mover or the innovator who builds on your innovation. It’s pretty easy to build a agent-based model that gets the desired kind of conclusion (don’t tax the makers!) but that doesn’t mean the simplifications used are appropriate.

'the simplifications used are appropriate'

Of course the simplifications are appropriate - they support what the author wants to support. This is the fundamental approach taken when talking about how those with wealth should be allowed to not only enjoy their wealth, but to use that wealth to ensure their rising wealth lifts all boats, or something to that effect, according to a Clinton appointee and 26 years long member of Goldman Sachs, a member of the board and co-chairman from 1990 to 1992.

'the reward for creating a successful innovation is a top income'

No, the Silicon Valley model of gaining immense wealth is not based on income, it is based on being one of those in a position to make billions from an IPO.

'top income taxation reduces everyone’s income, not just the income at the top'

And unicorn meat is delicious, if it weren't for taxation making unicorns unavailable.

How high would marginal tax rates have to be before you contributed something useful?

Warren Buffett became very rich when individual tax rates were as high as 70% and corporate tax rates over 50%, and the lower tax rates did not affect how he invested and ran his holding company, nor did lower tax rates do much to lower that taxes paid, until
Berkshire became so big that reinvestment to generate expensing and depreciation tax dodges became too hard due to scale.

Amazon took 25 years to get to the size that paying workers was not much higher than revenue so that depreciation and capital expensing didnt eliminate taxable income to below zero. Amazon is still net money losing over 25 years on a tax basis. But it has billions in zero tax basis capital assets.

In contrast to Apple which owns a small fraction of capital assets. China owns the majority of the capital Apple depends on: factories.

For Netflix, Amazon, Sony, Disney own the capital which they rent to Netflix, but Netflix is losing money paying workers billions to build capital: Amazon films and TV. When losing money, high tax rates give you greater access to cash to pay workers to build capital. Say an LLC that loses money developing something, all tax shelter for earned income, which becomes capital gains when sold to Netflix or Amazon with a lower tax rate. Back in the 60s, paying workers $100,000 and then selling the expensed/depreciated asset for $80,000 and a $60,000 capital gain converted $80,000 from 50-70% tax rate to $60,000 taxed at 20%. Today the rate difference and elimination of tax dodges making paying workers to build capital much less profitable than destroying old companies, like Sears, by stripping assets and then declaring bankruptcy to shed liabilities, and collect high management fees that are taxed at lower rates as incentives to kill jobs, etc.

"Warren Buffett became very rich when individual tax rates were as high as 70% and corporate tax rates over 50% ...."

by avoiding paying taxes and keeping his money locked up in investments. Indeed, Buffet has paid far more money in taxes (as a percentage of wealth) at the lower rates 1983+ than he ever did in earlier years.

Yes, he was paying less taxes than his secretary... because the IRS was suing him for a billion in taxes he owed.

It strikes me that pretty much the same arguments can be made for protecting new ideas via patent and copyright laws. Both involve the same idea--guaranteeing a greater share of income of the inventor from his or her new idea. Yet, I get the impression libertarians are generally in favor of a lower top marginal tax rate and generally opposed to existing intellectual property law. What gives? All else equal, would strong intellectual property protection enable higher top marginal rates (and lower marginal rates weaker protection)?

The other thing is that income neednot be the only reward incentive for successful inventors and entrepreneurs. It is not necessary to impart status (one's social position relative to others) solely based on monetary reward. Perhaps we should institute some sort of knighthood for creators of new ideas. Also, rather than villifying the "rich", we should establish an honor roll and perhaps a national day of thanks for those paying the most into the tax coffers.

What gives? Taking money away from people with very high incomes results in an equal amount of money for other things (i.e. one or more people divide up that money.) Taking away IP results in millions of times as much for other things (i.e. everyone gets the IP for free.)

(That said, completely eliminating IP would in fact be a bad idea for the same reason that very high tax rates are bad.)

If you add a second-order term, which is that successful entrepreneurs are more likely to be successful again (because of better access to credit, better network,...). Then a bit higher marginal tax rate is more incentive. So the truth is somewhere between Jones and Saez. Not far from the actual situation. Which means that there is no room for improvement there and that economists are wasting our time discussing on this subject.

Income tax may not be optimal, fine.

A wealth tax would work better if rich people stopped pretending they're "middle-class".

Luxury taxes are actually pretty simple and could be effective. X% on cars and boats over $50k, $Y percent on homes over $1M.

And here's the amazing thing, if rich people have those things to signal their wealth, a luxury tax just reinforces the signal.

That dude has a Lamborghini, he must be rich!

"Luxury taxes are actually pretty simple and could be effective. "

That's a statement based upon shear historical ignorance.

"Ignorance, Stupidity or Connivance?

Let's look at what happened when Obama's predecessor George H.W. Bush signed the Omnibus Budget Reconciliation Act of 1990 ...Congress imposed a 10 percent luxury tax on yachts, private airplanes and expensive automobiles,


Within eight months after the change in the law took effect, Viking Yachts, the largest U.S. yacht manufacturer, laid off 1,140 of its 1,400 employees and closed one of its two manufacturing plants. Before it was all over, Viking Yachts was down to 68 employees. In the first year, one-third of U.S. yacht-building companies stopped production, and according to a report by the congressional Joint Economic Committee, the industry lost 7,600 jobs. When it was over, 25,000 workers had lost their jobs building yachts,
The U.S., which had been a net exporter of yachts, became a net importer as U.S. companies closed. Jobs shifted to companies in Europe and the Bahamas. The U.S. Treasury collected zero revenue from the sales driven overseas."
The net effect of the luxury tax was a loss of $7.6 million in fiscal 1991, which means Congress' projection was off by $38.6 million. The Joint Economic Committee concluded that the value of jobs lost in just the first six months of the luxury tax was $159.6 million.
Congress repealed the luxury tax in 1993 after realizing it was a job killer and raised little net revenue. "


Yacht owning has undergone a vast demographic change.

That and oops fiberglass last forever.

Chart the whole thing from the 70's fiberglass boom to the 90's "hey, no rot" bust.

"Chart the whole thing from the 70's fiberglass boom to the 90's "hey, no rot" bust."

You don't admit when your clearly wrong. The American Yacht industry experienced a 77% decline in sales within 1 year of the tax taking effect. Blaming it on fiberglass is both silly and a transparent refusal to admit your mistake.

I don't know if you found the cheese in the maze but today you are the king poster here.

From the Cal link below, the problem all these guys had:

"Although the brand has been out of production for three decades, the existing fleet is still substantially active in racing and cruising."

The timing of the tax may have sucked, but 10% wasn't the insurmountable problem, it was slips full of boats that weren't wearing out.

For a 77% decline in sales within 1 year? Bullshit.

I don't see a link to good data on that. With good data I think you will see a big decline say '85 to '95.

"77" does not appear in the link given above.

"This is one of the reasons these ruggedly built and surprisingly affordable boats remain popular nearly fifty years after they first appeared in the recreational marine marketplace."

The same tax hit airplanes and devastated the market for twin engine planes:

"Since planes that cost less than $250,000 and planes that were used 80 percent of the time for business (mainly jets) were exempt, the primary target of the tax -- wittingly or not -- was twin-engine propeller planes, like Beech's King Air. But for the first 18 months the tax was in effect, the IRS collected not a dime from the sale of a King Air, and Beech lost 34 King Air sales totaling at least $80 million."

I'm curious how you are going to rationalize this bit of data.


Local knowledge, lived history.

You might make a nuanced argument about precise levels of tax, but this is not it.

But I don't think you can really say sales taxes, of which this is an example, are impossible.

Not that a sales tax cannot be progressive.

"Boating is a major industry in America. As of 2016, there was an estimated 11.9 million registered boats in the U.S.—the vast majority being made of fiberglass. An average fiberglass boat has a lifespan of 30 to 40 years. Durable and relatively lightweight, production of fiberglass boats began to take-off in the 1960s as manufacturers were able to quickly and affordably produce them for a growing middle-class; a boom in sales was seen between 1960 to the late 1970s. "

Non sequitur, we're talking about yachts. There aren't 11.6 million yachts in the US. We're also not talking about the middle class.

Why is it so hard for anonymous posters here to just say once in a while "ok yeah that makes sense I missed that"

Maybe because I live not too from where the 1970s California yacht boom was located. A good example lifecycle is here:

I certainly call bs on the idea that the luxury tax caused that. The timing may have been painful, but the build out had happened.

Tell you what, show me that repeal of the luxury tax brought back California boatbuilding .. except oops, it did not.

One other reason for that is that the 90s were when many areas like so cal hit peak marina. That put a big crimp in mass boating, with slip fees climbing from a few dollars a foot to $10 or $15.

I just checked, available slips in Newport Beach go for $27 a foot, for 30 feet, a nice family size. That's $800 a month to park.

So bs on "tax explains it all.'

Note at the second link that boat owning has been in steady decline since 1995. Combine *that* with a product that didn't wear out.

Relevant paper from a few years ago argued we Can't All Be More Like Scandinavians.

I wonder if this line of thinking misses the endogenous agglomeration effect of top scientists and entrepreneurs. The US is lucky to be the place that the worlds' talent has settled; shouldn't they be taxing this fixed factor?

Well the argument seems to be labor is immobile and capital very mobile so is the talent like capital or labor? (I don't quite agree because it implies we cannot have an immigration problem...)

Still is that fixed because the conditions have supported staying or fixed independent of various policy options?

Add this paper to the bin of dumb ideas. Edison would never have invented the light bulb if there had been a top income tax rate of 70% because he would have been too discouraged to follow up on his idea? Gates would never have invented the PC? Page and Brin would never have invented a search engine? Zuckerberg would never have invented Facebook? Oops. Maybe we need a 100% top marginal rate to discourage the Zuckerbergs of the world.

The libertarian mind. Or is it? Krugman believes libertarian is just a smoke screen for cutting taxes on the wealthy and cutting government regulation (of drugs, food, etc.) so that good ideas (like selling untested drugs) can be brought to market without having to go through any oversight. Too cynical? It's been suggested that Cowen's idea of emphasizing the future over the present is just an excuse to cut taxes on the wealthy and to eliminate what some consider essential government services (i.e., services provided to the not-wealthy). What's inconvenient to those who would cut taxes on the wealthy is that productivity and economic growth were much faster back when the top marginal rates were 90% and 70%. But what about the 1970s? The oil embargo (supply) shock had consequences well beyond the immediate effect of long lines at the gas station, a shock that came on the heels of the catastrophic war in Vietnam. Supply shocks are consequential, except when they aren't? Libertarians suffer the same frustration as Trump because so many people won't agree with them. Trump pitches a temper tantrum. Libertarians, like Sisyphus, soldier on. [My view about Cowen is that he is a true scholar with libertarian leanings. And I agree with his idea that we focus more on the future and less on today. But I would not use that as an excuse to cut taxes or government oversight. Instead, I would implement a more robust government participation in activities that boost productivity. If that requires higher taxes on the wealthy, so be it; indeed, I would increase the top marginal rate, and I would complement the higher rate with a zero or near zero tax rate on investments in productive capital.]

"What's inconvenient to those who would cut taxes on the wealthy is that productivity and economic growth were much faster back when the top marginal rates were 90% and 70%. "
--a meaningless correlation (and meaningless for lots of reasons). Libertarians have so many good ideas, but seem to ignore the benefits of cooperation (when that cooperation comes at the government level).

When the top marginal rate was 90% (91% actually), the effective tax rate was 45%.

"People at the top are motivated by status" is not such a killer argument for high-marginal rates of income taxation.

Would Bill Gates have the status he has were he not the richest man in world (at least for a long time)? If the most successful startup founders were as well-off as a suburban orthodontist (which is a typical member of the top 1% of incomes, so nothing to scoff at), would Silicon Valley hold anything close to its cultural cachet?

If tax policy makes it impossible for any startup founders to reach the very top, so that the list of richest people was always dominated by inherited wealth (see Sweden), then startup-innovation would likely have lower status while "schmoozing the wealthy to try to grab a marriage proposal" would have higher status.

Don't you just love papers that discuss tax level in an abstract sense, without respect to deficit or debt?

Here's a shocker, all tax levels should be as low as possible, covering government expenditure.

In fact we would be a hell of a lot better off if tax level was automatically set that way, indexed off of previous year expenditure.

Then politicians could not play this game of voting for expenditure and then being shocked shocked that tax necessary.

You can prove any thing with assumptions :-)

The liberals I know won't be convinced by this argument, because they don't believe in the incentive effects of tax rates. That Steve Jobs, Bill Gates, et al would have invented their creations even at 90% rates.

Both those guys benefited from capital gains, and used family trusts in tax planning.

If anything the better argument is that a high marginal rate on simple income would miss them entirely.

(In fact Steve Jobs famously took a salary of $1 per year.)

I hope everyone who used to talk a good game about fiscal responsibility during the Obama years is staring at their shoes right now.

Staring at my shoes, nope. Against the deficits now as well as back then. It's BS to run a deficit now, even though most everything else he's done has been pretty good. But likewise, those who said nothing during those years shouldn't be running their mouths now.

ITT: People confusing high wealth individuals with high income individuals.

Two functions of taxation are simply the funding government one and the one related to income redistribution.

I wonder if the goal of achieving a less skewed wage/working compensation distribution could not be address both more efficiently and more directly than leveraging tax policy.

Some will reject the thought out of hand but would be interested in considering an approach along the lines of top-bottom pay rations within the public corporation. We know that measuring productivity within large, joint production settings is problematic. We know there is a wage premium for firm size. We also know that investment and valuation of a firm is not directly related to the internal distribution of revenue surplus among the members of the firm.

Rather than attacking the distribution indirectly via tax and government distribution would a rule saying if a company wants to enjoy the benefits of limited liability (and other benefits governments grant) then it's highest paid employees cannot make more than X times the lowest.

I do not think the internal compensation process in large corporation is that much like economists talk about the labor markets. It is more like political economists talk about distribution within economies. Large firms are only partly market participants. They are also markets in their own rights; and markets lacking a clean price mechanism.

And in his continuing drive to turn right-wing economic dogma into a full-blown religion, TC invents heteroousian trickle down theory.

From the conclusions:

Lots of unanswered questions ... Why is evidence on growth and taxes so murky?

i.e. the actual data doesn't fit this theory at all, but I'm having trouble coming up with an explanation why reality is wrong.

By all means let us dispense with income and payroll taxes

Let us go to national property, sales, pollution, Pigou taxes, and heavy tariffs.

Income taxes have become far too complex and sometimes uncollectible.

That latter characteristic is probably why there is no movement towards a non-income tax based federal tax system.

Your assumptions are wrong, which is why the rest of the paper is worthless. High marginal tax rates do not penalize innovation. High marginal tax rates penalize a person who works by-the-task from working extra hours to make more money. For example, a surgeon who is good at a specific procedure is penalized for Spending more time to do more of the same procedures, because the value of his time declines the more time he spends (and the more procedures he does).

The optimal solution under high marginal taxes is to Keep working until the marginal income of time worked is equal to your marginal utility of taking time off. High marginal tax rates would incentivize the hypothetical surgeon to reduce the number of low-value procedures performed (allowing others to perform them to fulfill the demand) and concentrate on innovating procedures and methods to maximize value and efficiency.

A person who spends all his time in innovation will never be the highest compensated person because most ideas are bad ones, and he will be compensated based on the average value of his innovations (the mode and median value of his innovations will be 0) to. This is obvious from the fact that an orthopedic surgeon makes far more money than a research scientist who develops new ideas.

Comments for this post are closed