Garett Jones  

The Folly of Obsessing over Marginal Tax Rates

PRINT
WSJ's "Roll Your Own" Deficit ... Go Fight Some Real Crime: Why ...
It's often wise to pay more attention to marginal tax rates than to average tax rates.  If you can make your first $100 tax free but the 101st dollar is taxed at a marginal rate of 99% you'll probably decide to earn $100 at most.  

But what is marginal?  When it comes to career choices or the state you'll live in or whether to have an extra child the marginal decision is very big, and a rational person will base that decision mostly on the average long run costs and benefits.  In cases like this, the official marginal tax rate won't matter nearly as much as the long run average tax rate.  

This isn't just a theory: It appears to be true for that most economistic of organizations, the multinational corporation.  When multinationals are deciding which country to invest in, they don't pay that much attention to marginal tax rates.  According to Glenn Hubbard

...investment location decisions are more closely related to average rather than marginal tax rates. 

When making the go/no-go decision, corporations care more about their long run tax bill.  That's because the marginal decision is the go/no-go decision.  

This matters for the current debate over taxes on high earners.  Speaker Boehner is open to raising taxes on the rich as long as it raises average rates not marginal rates.  But when people are deciding whether to become medical doctors or mere professors, lawyers or struggling novelists, entrepreneurs or Xbox champions, these go/no-go decisions will be shaped by the average tax rate that the rich will have to pay for decades to come. 

Why did the 90% marginal tax rates of yesteryear have so little apparent impact on labor supply?  I suspect part of the reason is because of the unlimited home interest deduction that used to be in the tax code: You knew that if you doubled your salary you'd move into a bigger house pretty soon, consuming much of your extra income tax free in the form of housing.  

That's just a case of applying Hubbard's idea: the marginal tax rate didn't matter when you chose your career, but the average tax rate--after deductions--mattered enormously.  

In the long run we should worry more about average tax rates.  

Comments and Sharing


CATEGORIES: Labor Market , Taxation



COMMENTS (19 to date)
MikeP writes:

In the long run we should worry more about average tax rates.

This is likely true, since total government burden is a very important metric of future economic health.

But is it not also true that, for the same average rate, we should prefer the lower marginal rate? Otherwise, we end up with choices people would not have made absent the government -- such as consuming more house than is efficient.

The president is looking for X dollars: he's already chosen the average. We would all like that average to be lower. But if it is not politically possible to collect less than X dollars, would it not be better to glean those dollars with fewer deductions and lower marginal rates?

If the marginal political effect is on marginal rates, it seems less like folly.

Hugh writes:
When multinationals are deciding which country to invest in, they don't pay that much attention to marginal tax rates

Most countries (but not all) do not have tax bands for corporate profit tax, but rather a flat rate. Sometimes you see lower rates for small or micro companies, but a multinational is unlikely to qualify for these straight from Day One.

The average vs. marginal discussion thus seems moot.


hhn1776 writes:

Pretty misleading article. You're confounding two types of decisions; those at the extensive margin such as go/no go, buy or don't buy versus those at the intensive margin. The extensive margin is about whether to work or not, or which profession to choose. The intensive margin is about how much to work.

Total and average benefits are those relevant at the extensive margin, so total and average tax rates are the ones relevant here. Yes, the average long run burden of taxes is what's relevant to choosing a lower income occupation relative to a higher one. The greater the tax burden on high nominal income occupations, the fewer we get of them. At the low end of the scale, the higher we tax jobs or the more we subsidize unemployment, the more unemployment we get.

Marginal taxes are those relevant at the extensive margin. No, it's not the nominal tax rate that matters, but the effective tax rate. The effect tax rate is the nominal tax rate adjust for exemptions such as the home mortgage deduction. To say that marginal tax rates were 90% in some edenic past is misleading. That may have been the nominal tax rate, but it was not the effective marginal tax rate.

Finally, the decisions at the extensive margin are made conditional on intensive marginal decisions. Thus, marginal tax rates that alter decisions and net benefits at the intensive margin do affect decisions at the extensive margin.

Pat writes:

Marginal rates matter to people already in a safe job who are thinking about taking a risk starting their own business.

Your path in life isn't neccessarily decided when in your late teens.

Potential novelists who have a day job factor marginal rates into their decision to focus on writing. High marginal rates change the risk reward.

MikeDC writes:

Is it just too common sense a notion to say that they both matter? Of course obsessing over one aspect of taxation to the exclusion of others is counterproductive.

Especially when there are so many obviously negative aspects to both our current system and the proposals I see.

Alex Godofsky writes:
Why did the 90% marginal tax rates of yesteryear have so little apparent impact on labor supply? I suspect part of the reason is because of the unlimited home interest deduction that used to be in the tax code: You knew that if you doubled your salary you'd move into a bigger house pretty soon, consuming much of your extra income tax free in the form of housing.

That's just another way of saying the effective MTR was much lower than 90%.

Philo writes:

“You knew that if you doubled your salary you'd move into a bigger house pretty soon, consuming much of your extra income tax free in the form of housing.” But you would move into a more expensive house than you really wanted, because of the tax incentive. So you wouldn’t get the full benefit of your extra income—it wouldn’t be translated fully into extra utility/happiness.

Still, you’re right: high marginal rates together with enough loopholes that most people can avoid the high rates will have less effect on career choices and lifestyle choices than would slightly lower marginal rates together with no loopholes. However, there are other objections to the multi-loophole scheme.

RPLong writes:

I totally agree with Dr. Jones here. The marginal rate definitely impacts the marginal dollar, but the average rate impacts the average dollar. I assume even the size of the marginal rate compared to the size of the average rate mattes. All kinds of things matter.

I would also add something less tangible: People will not attempt to amass great wealth if the perceived benefit of that wealth is low. Why break your back working 80-hour weeks at E&Y if the most it offers you is a slightly larger home and a Cadillac Escalade as opposed to a Ford Expedition. Big deal.

James Oswald writes:

On a long enough time horizon, everything marginal. When a firm invests in some capital, at the moment they buy the capital, they consider the expected tax rate they will face, which is the margin they are on. After the capital is purchased, that decision is no longer marginal, because it happened in the past. There is no cost which is purely "marginal" or "not marginal"; marginal is a statement about the time frame when a decision is made.

magilson writes:
Speaker Boehner is open to raising taxes on the rich as long as it raises average rates not marginal rates.

The way in which you stated this suggests you think it's a bad idea given the premise of your post. Maybe Boehner really believes marginal rates are the issue. Maybe he believes it because the people giving him money are entrenched business-folk for whom the marginal rate does matter most because they're not stepping out of their comfort zone any time soon. Twenty-somethings don't often catch the eye of a Senator in the middle of negotiations. So the reasonable concerns about average rates to a young person looking at their future don't enter the political dialogue quite as easily.

But at the same time I'm not sure what Boehner uses as justification will matter. If sufficient evidence amasses to convince a majority of the voting public that the taxes raised to avoid the fiscal cliff were a bad idea the Democrats will have a harder time pointing to Boehner as the problem as opposed to Republicans pointing to (former?) President Obama and the Democrats.

Daublin writes:

This is a very misleading article.

At a try, I believe a lot of the odd claims come from what you consider to be the "margin".

In the location choice, the relevant margin is the margin between the two locales. Yes, that's different than the margin between one dollar of income and the next dollar of income. It's still a marginal rate, though.

In the profession choice, you care about the marginal rates for the range in income in question. If a doctor makes a certain income, and a professor another, then what you care about is the marginal rate between those two levels of income. This time it really is the marginal dollar-over-dollar rate, so I really don't follow what you are trying to say.

The marginal rates seem extremely important. For income taxes, they matter for how people employ their labor.

Let's Be Free writes:


An average rate is a the result of applying marginal rates across a range of income. Structure matters -- very much. Constant marginal rate, increasing marginal rate and decreasing marginal rate structures can be designed each to produce the same average.

What is marginal varies on what question is being asked or what decision is being made. For a location decision the margin is the rate difference between situating in location A versus location B, which may look like an average rate comparison, but which is a marginal rate with respect to that decision. If the decision is an output or how much should I work decision, the margin is a unit of output or work.

Mike writes:

Garrett, your spot on, but here's a simpler way of looking at the marginal vs average tax rate issue. Consider someone who has invested many years to reach the point where he is earning a "high" income, such as a doctor who has gone through many years of school and training or a business owner who has spent a decade (or more) building his business and mastering his craft. And suppose that working 40 hours per week, they net $500,000 (or $250,000 for 20 hours per week) and the tax rate is 0%. Now, the government comes along and decides that all income over $250,000 should be taxed at 50%. What is our high earned going to do, sit on his duff for 20 hours or continue to work 40 hours? Our high earner is essentially trapped - it's difficult to run most businesses part time or to maintain one's profession skills or referrals on a part time basis - and will continue to work the 40 hours (or more). But they sure won't be in a hurry to expand their business (because the margin does matter here) or to recommend a similar career path to their kids (this is where average tax rate applies).

Hi people! I`ve recently written a post , I may even call it a research about pretty the same topic . There is not much relevant information on that, but I was lucky to find some on mobile frequency jammers there .

UnlearningEcon writes:
This is likely true, since total government burden is a very important metric of future economic health.

Are there any peer reviewed studies to back this up?

MikeP writes:

Are there any peer reviewed studies to back this up?

Good question. Most studies look at marginal rates rather than average rates, since that is where the political margin is.

Also, the major effect of government burden on the economy is due to government decision making. So raw automatic transfers such as Social Security likely have much less of an effect and should be factored out of such analysis.

Peter H writes:

"Speaker Boehner is open to raising taxes on the rich as long as it raises average rates not marginal rates."

I don't know how I'm supposed to take this sentence seriously. To raise average rates is to raise marginal rates on at least some taxpayers. If you cap deductions instead of hiking nominal rates, then you raise marginal rates on people who take lots of deductions. Someone will be at the margin, and the more you want to raise average rates, the more someones will be at the margin.

The meaningful distinction is not marginal versus average, but rather nominal-marginal versus actual-marginal.

Suppose for example that I'm Mitt Romney, and I donate 10% of my income to the LDS church. This donation is fully deductible for purposes of my federal taxation, but even if it were not, for strongly held religious reasons it is my preferred disposal of .10 of my marginal dollars at any margin.

Say I make $10 million in a year, and am in the 35% marginal rate bracket. Well, the fact of my deduction for donating to my church means that my actual marginal rate is 31.5%. If I further pay 10% marginal state income tax on my money, and can deduct that income tax from my federal taxes as well, then my actual federal marginal rate drops to 28%.

Now say that my campaign proposal of limiting deductions to a certain dollar amount passes. That means that, at the margin, I lose my charitable and state income tax deductions. And my federal marginal rate shoots up to 35%. Even though the nominal rate is unchanged, limiting deductions has hiked my actual marginal rate by 7%. As economists, we should look at the actual margin, not the nominal one.

Steve Robinson writes:

The 90% rates of yesteryear had almost no effect because they applied to almost no income.

JEC Republican Staff Analysis Historical Tax Rates: Rhetoric vs Reality

http://www.jec.senate.gov/republicans/public/index.cfm?p=Studies&ContentRecord_id=e523a7a6-5387-4ee5-b6e4-932d29984e06&ContentType_id=11c26666-579d-432b-a65f-dab5849a0fce&Group_id=74fc1c2f-4d37-4b97-b74f-5ee72cff4880

mike writes:

I wonder if average rates are taken as approximations of marginal rates in making some choices.

Comments for this entry have been closed
Return to top