Marginal tax rates

One in four low-wage workers face marginal net tax rates above 70 percent, effectively locking them into poverty. Over half face remaining lifetime marginal net tax rates above 45 percent. The richest 1 percent also face a high median lifetime marginal tax rate – roughly 50 percent.

That is from a new NBER paper by David Altig, Alan J. Auerbach, Laurence J. Kotlikoff, Elias Ilin, and Victor Ye.


And now, with the CARES Act, my brother-in-law faces a marginal tax rate well over 100%. As long as the benefit continues, he will have to take a pay cut to go back to work.

True, and arguably they are trying to subsidize not taking risks to go out into the world. But the problem is it subsidizes non-productive behavior not covid-averse productive behavior.

The story is not true.

If you are referring to my story, he nets $400 a week from working and $600 from not working. If it weren't a violation of privacy laws and common decency, I'd post the pay stub and benefit check.

No! The article is wrong, the data is wrong.

In response to Marcus, a living wage is non-productive for whom?

> a living wage is non-productive for whom?

The person who must pay it IF the person they are paying isn't delivering the value.

Microsoft pulls in about $800K per employee. They pay employees about 20% of that. Overall, about $160K per employee.

Curiously, McDonalds has a labor cost that is around 20%. Same as Microsoft. In fact, across the board that rule holds.

In short, for every dollar and your you are paid, you need to deliver about $5 in revenue to your employer in order for them to find it valuable to have you around.

If you are making your employer $30/hour, then it's really hard for them to pay you $30/hour. Or even $15/hour.

If you want to earn $20/hour, then make your employer $100/hour and I promise you you'll earn $20/hour.

Actually no: the work force as a whole over time must deliver the needed productivity to make a wage viable.

> Actually no: the work force as a whole over time must deliver the needed productivity to make a wage viable.

Yes, but since McD's has ~300,000 US employees, and probably 500 highly paid execs, the productivity of the 299,500 non-highly paid people matters enormously. In other words, you cannot pay them all well above what they are delivering in terms of productivity. The math just doesn't work.

If you are workign at McDonald's making $12/hour and believe you should be making $20/hour, then build your own store and pay yourself whatever you want.

In response to Phinton: The Microsoft 42/160 (CEO to employee compensation ratio) suggests McD's 15/15 ratio should either be 4/15 or 15/57. The incentive of a 94% tax rate on McD's CEO $15 million compensation would perhaps improve the business model.

The employee to CEO ratio means nothing. Do you think a modern day McD's employee is 10X more valuable to the company than 50 years ago? Of course not. They are likely less valuable because employees of 50 years ago demonstrated the requisite hustle that the time required.

Any productivity gains the modern McD's employee delivers is from machinery purchased by the employer. Thus the employer should accrue the productivity benefits.

The CEO, on the other hand, has much more responsibility and influence compared to 50 years ago. For one, the company is global. Next, the net income of McD's today is 6X that of 25 years ago.

The McD's CEO earns about $20M/year. Divided up among the 1.7M global employees, that's a few bucks per employee. it's nothing.

In response to Phinton: Does Economics 101 cover which is more profitable, slavery or non-living wages? Was it up to taxpayers to keep slaves living? What is the point of McD’s average $20,000 non-living wage?

In response to Chuck, what I mean is that a subsidy to social distancing subsidizes social distancing activities across the board, some of which are non-productive like sitting at home doing nothing. Ideally, the subsidy should target industries where people can work at home to gain social distancing and not lose productivity, but that's an intensely tricky thing to do. Not sure if there's a good solution.

Why do you think reducing gdp is a good thing?

Or requiring lots of personal debt that is defaulted on as part of individuals borrowing and spending to prop up GDP when he returns to work?

It's not like driving down most household incomes to under $30,000 would be good for the US economy.

Economies are zero sum. Ie, you can't spend more than your income for long, so policies of driving down incomes drives down GDP.

mulp, you seem to not understand what zero sum means.

Please return when you've taken Microeconomics 101.

Does Micro 101 cover trickledown?

Micro 101 makes a very sharp distinction between gains from trade, which are not zero sum, and accounting identities which are.

You also can't buy more than you produce, so policies that encourage people to keep spending will lead to more inflation, more debt, and a bigger bubble. But, I guess that's someone else's problem; right, boomer?

Hmm, is the TFA source code posted somewhere? Only fair that it gets a good shakedown given what was found with the Imperial College model.

How high are the GRE scores of the paper's authors?

It is working as designed. Keep them stupid and poor.

Or as Pelosi said: “As you hear from these stories, this is a liberation,” Pelosi said at a Capitol Hill news conference Thursday.

“This is what our founders had in mind--ever expanding opportunity for people. You want to be a photographer or a writer or a musician, whatever -- an artist, ..."


Around the same time that Nancy Pelosi made that statement, Mitt Romney disparaged beneficiaries of the Earned Income Tax Credit, Child Tax Credit and various tax reforms that lessened the tax burden on low income people by describing them as members of the irresponsible 47%.

Conservatives mostly defended Romney's comments at the time and some even explicitly said they want to see higher taxes on low income people.

"Conservatives mostly defended Romney's comments at the time"


It was one of the dumbest comments ever made during a campaign, and effectively ended Romney's chances in an election that should have been easy for him. I was horrified. He has never been a good campaigner, but I think he would have made a much better president than Obama.

I will defend him to this extent: we would be much better governed if everyone had some stake in paying the cost of government. When an election is a contest between net tax payers and net tax recipients, guess who wins? That is why the Founders limited the franchise to property owners, not because they owned property but because they paid property taxes, which were the primary source of government revenue at the time.

"and major differences in marginal and average net taxation across states"

Anyone have a copy of the paper and want to tell us which states to move to?

Varies by cohort. For example for the 30-39 age cohort in the lowest income quintile (from a federal perspective), "median rates [vary] between a low of 38.8 percent in South Carolina and a high of 55.0 percent in Connecticut."

See you at Waffle House!

Note that if you're switching states you're no longer making a marginal move but instead are making a .... whatever move: discrete, quantum, qualitative, long run, global. So the relevant comparison is the average tax rates in each state, not the marginal tax rates.

No, the relevant comparison is the total tax take for the person making the move. The average in the state has nothing to do with the decision.

Alas, the paper is behind a paywall for those of us who don't meet the various criteria. Maybe the US now meets the 'transition economy' point.

Non-paywall version:

Doing God's work, Bob.

Net tax rate, marginal net tax rate, median lifetime marginal tax rate, so many undefined terms. This is economics? This may not be fraud but it certainly is obfuscation. Shame!

Every single one of these terms is not only well-defined, but the definition follows in a straight-forward way from the term itself. You seem to be (once again) mistaking your own intellectual deficiencies as insights.

Not rayward, but I honestly don't know what "Net Tax Rate" means, and my searches unfortunately produce results for "Net of Tax" accounting terminology, nor do I think it follows immediately from the term itself (net of what?). Care to explain what it means? I understand the terminology otherwise.

Net Tax Rate = effective tax rate, "net" of ALL taxes (income, FICA, sales, etc) and transfer programs (social security, SNAP, section 8, etc)

Marginal net tax rate = same as above but marginal instead of effective tax rate.

Median lifetime marginal tax rate = Calculate the marginal net tax rate for each year of life and take the median. Note that this also refers to "net" taxes even though it's not in the name (I guess you'd call that "nuthin but net")

How do taxes you paid in the past affect your future actions?

Ie, you had a great job that paid hundreds of thousands requiring you pay tens of thousands in taxes, then things change and you lose that job and can't get it back so now you pay very little in taxes while sinking to being homeless and hungry.

A skilled person earning high incomes loses the use of his hands, eg, baseball pitcher, surgeon, jeweler/watchmaker, etc, would happily pay 50% marginal tax rates to return to their high income job.

After all, everyone of them started out earning nothing at all and then were lucky to be required to pay 50% marginal tax rates.

I've gotten a number of huge tax cuts I never wanted, and I wish I had had to keep paying what you consider crushing taxes.

And I refused jobs that required I pay very little in taxes - I'd much rather have my free time and live off my saving than get paid less than it costs to work and thus spend my savings faster. (The biggest marginal cost of work is transportation for low income jobs.)

tl:dr... someone please cull out the driving assumption that catapults their numbers to where they wanted them to be.

I did note one important caveat: it matters what state you live in. Exactly. For some reason, people keep choosing to live in "high tax" states. Must be something else going on in the decision making besides income taxes.

Yeah, it's odd when people are fixated so much on state taxes. State taxes are much less for most people than federal taxes. The lowest federal income tax bracket is 10% and that's only for the first $9,700 of income. Almost no states have any income tax brackets above 10% (or even above 6-7%). And if you really don't like the tax burden, it's pretty easy to move between states, much, much more so than it is to move between countries. Cutting taxes should be all about federal taxes; taxes and services should be decentralized to as local a level as possible where people can vote with their feet. Small enough local governments (like a township) where you could easily live in the next town over without any sacrifices to your work or even social life should basically be considered private entities with taxes equivalent to a service fee.

"The U.S. has a plethora of federal and state tax and benefit programs, each with its own work incentives and disincentives."

Someone was just telling me that centralization is bad, but I think it's kind of necessary, if you want to avoid one program stepping on another's toes.

If you want one consistent curve, smooth from assistance and transfers for the poor through to higher taxes for higher earners, doesn't there have to be one agency applying one curve?

(possibly there could be a national formula where you plug in some local numbers)

Are we in favor of simpler government or just "smaller" government with a crazy complex framework?

It strikes me that "simpler" would be smaller in many important respects.

You may be correct, but it is unclear to me as a matter of first impression that we should seek to avoid programs "stepping on each other's toes". New Dealers like Rex Tugwell made similar arguments about how competition between businesses led to waste and inefficiency, but we now know the exact opposite is true. It may be the case that allowing local governments to "compete" by offering a variety of tax curves incentivizes them to legislate effectively. At the very least, it seems fair to assume that transaction costs between neighboring municipal governments are low, so it's likely they would be able to build a cooperative tax structure without needing default terms imposed from above.

What the covid unemployment pay rate bump is telling us is that ordinary wages are too damn low, and that too many people are too damn broke.

The perverse system here is not the unemployment pay that is more money than their ordinary pay. The perversity is that their ordinary pay is too low to start with.

Somebody noticed the elephant in the room.

How much do you pay your employees?

Are they measuring poverty using the supplemental poverty measure? People who work for a living don’t set their wages, they take what they can get. Somehow I doubt that cutting benefits would transform below poverty line jobs into good jobs.

This is unfortunate, but the vast majority of workers face much lower marginal tax rates. My wife and I, for example, earn roughly $270k per year and face a 36.9% (my wife's marginal fed, state, local and 7.65% fica) and 30.7% marginal tax rate (my fed, state and local combined rate, with 1.65% fica). Without fica, which is a bit different because it leads to increased SS upon retirment, our total marginal rates are in the high 20s, which is hardly crippling. Our early 20s, single daughter earns high 40ks and faces a marginal rate in the low 20%s, including fica.

That seems astonishingly low considering the marginal Federal tax rate alone is 24% (Married: $171,051 to $326,600).

We live in Pennsylvania, which has a flat 3% state marginal rate and about a 2% local marginal rate (wages only, not investments), which adds up to 29% marginal rate before fica. My salary is over the SS max, so on the margin I only pay the 1.45% medicare tax.

the margin I only pay the 1.45% medicare tax."

No, that's incorrect. The marginal Medicare tax rate is actually 2.9%.

(This is an economics blog, we all know that the employer pays portion is a fiction to make people feel better and that contracted employees get paid enough to cover the whole amount).

But still a 32% marginal rate isn't bad all things considered.

Your wife of course is paying more like 44% with the full FICA charge. It's nice to live in a medium-low tax state.

Yes, I am an economist and fully recognize the employer paid issue. But counting the 1.45% employer payment opens a can of worms in which the numerator (salary) then also has to rise by 7.65% Similarly, I am paid an additional 9% of my base in retirement contributions. These are pretax for the federal rate, but zero percent margin for both my and my employer's fica. So my combined marginal rate is thus (.24+.05+.029)/(1+.0145+.09) =28.9%. Right?

Whoops, should proofread. changes the _denominator_ by _1.45%_. Calculation not affected by those, so:
But counting the 1.45% employer payment opens a can of worms in which the denominator (salary) then also has to rise by 1.45% Similarly, I am paid an additional 9% of my base in retirement contributions. These are pretax for the federal rate, but zero percent margin for both my and my employer's fica. So my combined marginal rate is thus (.24+.05+.029)/(1+.0145+.09) =28.9%. Right?

"(.24+.05+.029)/(1+.0145+.09) =28.9%. Right?"

Yes, that sounds correct. Thanks for the response.

I didn't read the paper, so not sure if this is applicable, but you'd also need to add in property, sales and gas taxes. These could easily add 10% to the total.

I have a marginal tax rate of maybe 50% in Australia if I want to make it look high. But it's not really that amount on account of how I respond to incentives. So I presume the richest 1% in the US don't have a 50% lifetime marginal rate in reality. Not unless they are really dumb or the US tax system is a lot more efficient than I think it is.

I do respond significantly to incentives, hitting the post-50 year-old maximums for 401k investments, using maximum health savings account contributions and using "backdoor" Roth IRAs to move funds to tax free investments. Our marginal rates are essentially unchanged, but our _average_ tax rate is 19% (or 25% with fica). Given that both our income and net worth are at about the 95th percentile for the USA, that does not seem crippling to me.

Marginal tax rates have to be graduated across partitions.

How many tax brackets can we keep stable before tax evasion and income volatility fouls the brackets up? I doubt that we can even classify income to more than ten brackets or so.

How do these brackets count? Well, the poor get grouped into one bracket as there are more of them. The wealthy are spread out, they consumer more brackets for the simple reason they pay more taxes. Government wants to maximize tax gain from the upper brackets and spend more bracket space at that level.

So, due to efficiency, the poor always have the larger jump to the next bracket up, mainly because of the costs of managing more granular tax brackets for a low tax generators.

This graph from the paper is less exciting: "Across all age groups, the median lifetime marginal net tax rate is 46.6 percent for those in the lowest resource quintile. This exceeds the median lifetime marginal net
tax rates for the next three quintiles of 41.4 percent, 41.1 percent, and 42.6 percent, respectively. But it’s lower than the 50.2 percent rate of the top quintile."

The richest 1 percent also face a high median lifetime marginal tax rate – roughly 50 percent.

That perverse incentive effect must explain why their income hasn't risen at all in recent history!

Because the richest 1% get a substantial part of their true income from capital gains on stocks, real estate, and other assets, which are not taxed at all until they decide to cash out and spend part of their gains, their effective rate seems unlikely to be anything like 50%.

How about we give everyone a restricted pair of shoes and then increase the price of the shoes a person owns by $100,000 per year.

The restrictions being they can only be worn, eg, productive capital, but never sold. Though they can be abused and discarded, like a lot of factories in the US making toilet paper and PPEs.

If we did that, everyone could have a very high income, not pay any taxes, and not cost employers or society anything.

"One in four low-wage workers face marginal net tax rates above 70 percent, effectively locking them into poverty. Over half face remaining lifetime marginal net tax rates above 45 percent."

This is an argument in favor of a UBI.

And it most be Universal (for all workers) for it to not have a built in marginal rate tax hike built in. If there's any kind of income cut off point, then you effectively have placed a high marginal rate at that point.

I'm not convinced that a 70% marginal tax rate would lock one into poverty, especially when much of it is in kind benefits that recipient values much lower than the cost.

+1, I agree that's a silly presumptive statement. And as you said, the reason for the marginal tax rate is because a previous benefit is phasing out.

A better phrase would be a high marginal tax rates incentivizes low income workers to keep their declared wages low.

The paper simply highlights the fact that thousands of chefs have worked on the tax recipe. Of course there are going to be weird cliffs and plateaus all over the place. It will take another thousand to tweak those out, but that will yield another set.

The simplest recipe for taxes is (A-B)*C, where A is your earnings, B is a deduction, and C is your tax on all funds earned above the deduction.

I've never understood why we don't offer B and C as public numbers required to set the budget at breakeven every year (pay for the government you get). And if we decided we want a new benefit, such as "government healthcare", then the folks that want the benefit offer new numbers for B and and C. And based on that, everyone can vote. Do you want to lower the age of retirement from 65 to 64? If so, here are the numbers, America. Your deduction will drop from $33K to $30K and you're tax rate will increase from 0.22 to 0.235. So, a $40K earner will see their taxes go from $1540 to $2350. Simple and easy for everyone to understand.

Of course, those controller the levers see all this confusion and obfuscation as a feature, not a bug....

"Of course, those controller the levers see all this confusion and obfuscation as a feature, not a bug...."


At iPayTotal, we make it our mission to ensure that your business can process credit card payments with a high-risk merchant account.

Comments for this post are closed