David R. Henderson  

Politi-Obfuscate

PRINT
Shocking data on wealth inequa... Designer Babies Are Nothing to...

Bill Gates stated yesterday:

The highest economic growth decade was the 1960s. Income tax rates were 90 percent.

He stated this on CNN's Fareed Zakaria GPS.

Politifact's Punditfact decided to check the truth of Gates's statement. To their credit, they did point out that for the last 6 years of the 1960s, the top tax rate was not 90 percent. Instead it ranged between 70 percent and 77 percent.

Also, to their credit, Punditfact briefly covers the controversy about the connection between tax rates and economic growth. Here's the skeptical quote about tax cuts:

A more sober analysis comes from William Gale at the Brookings Institution and Andrew Samwick at Dartmouth College and the National Bureau of Economic Analysis. In a 2014 article, the two economists found that "U.S. historical data show huge shifts in taxes with virtually no observable shift in growth rates."

Punditfact also gives the "other side:"
But there are studies that reach other conclusions. Two economists at the University of California-Berkeley, Christina Romer and Dave Romer, broke down tax changes based on the driving purpose behind them. For example, it is very different if taxes were cut because the economy hit the skids, rather than wanting to shrink the size of government by curtailing revenues. Romer and Romer -- they're married -- figured that separating the reasons for tax changes allowed them to factor in a number of outside conditions that determined the impact of the policy shifts.

Their key finding?

"Our estimates suggest that a tax increase of 1 percent of GDP reduces output over the next three years by nearly three percent," they wrote. (Christina Romer is former chair of President Barack Obama's Council of Economic Advisers.)


Punditfact's conclusion, which they label "Our Ruling":
Gates said that in the 1960s, high taxes, 90 percent, and high economic growth came at the same time. The underlying reasons are complicated, but the numbers largely bear out. On average, annual growth was about 4.3 percent for the decade, higher than any other post-World War II period. Gates was a bit off in talking about a 90 percent tax rate. During the 1960s, the marginal rate fell for a time to 70 percent.

Gates did not go so far as to say that higher taxes bring higher growth. And for good reason. The connection between taxes and growth is quite tricky.

We rate this claim Mostly True.

Punditfact bothered to check average growth rates of GDP by decade. Punditfact also checked top tax rates during the 1960s and found the split noted above. They found Gates's claim about tax rates being 90 percent in the 1960s to be false (if I were to put in in their terms, I would say "Mostly False" because tax rates were below 90 percent for over half the decade.)

But then that raises an obvious point that inquiring minds would want to know. Given that top tax rates were substantially lower in the last 6 years of the 1960s than in the first 4 years, was there a difference in the growth rate of real GDP?

There was.

Here are the growth rates of real GDP from 1960 to 1963, the 4 years in the 1960s when the top tax rate was 91 percent:
1960: 2.5
1961: 2.3
1962: 6.1
1963: 4.4

Average growth rate (arithmetic, not geometric): 3.825 percent.
Average growth rate (geometric): 3.8 percent.

1964: 5.8
1965: 6.4
1966: 6.5
1967: 2.5
1968: 4.8
1969: 3.1

Average growth rate (arithmetic, not geometric): 4.8 percent.
Average growth rate (geometric): 4.8 percent.

In other words, the annual growth rate averaged a whopping one percentage point higher during the six years of the decade that had top tax rates substantially below 91 percent.

But you read that here, not at Punditfact.


Comments and Sharing






COMMENTS (20 to date)
Brad writes:

"The connection between taxes and growth is quite tricky." Unquestionably true.

Here's the thing - if Bill Gates thought confiscatory tax rates really benefited the economy and improved the economic well-being of the middle class, why is he giving away all his wealth to charity? Why not write Uncle Sam a big, fat check? Same for Warrent Buffet. Talk is cheap.

E. Harding writes:

It's not the top marginal income tax rate that matters, but actual taxes paid and the taxes' incentive effects.
Also, I'm pretty sure Bill Gates is not a huge fan of fossil fuels:
http://research.stlouisfed.org/fred2/graph/?g=Vgm
Also, the Vietnam War.

love actuary writes:

Dr H, while you don't draw explicit conclusions, are you reasoning from a (price) change here?

R Richard Schweitzer writes:

Not stated: What is the "measure" of "economic growth?"

If it was the usual GDP, which includes the substantial segment of expenditures via (all)governments, was that "economic **growth**" or expansion of economic activities (transactional "growth')?

During those periods, what was the rate of increase in the proportion of GDP represented by governmental expenditures?

During those periods did not inflation (the consumption or depreciation of savings and capital) increase about 500%?

What immediately followed those periods? Declining inflation or acceleration?

Taxation creates burdens on economic activities. We have seen what happens when they have been excessive (disproportionate). The subsequent avoidance of taxation by fiscal borrowing shifts those burdens to an uncertain but conceivable set of consequences.

Perhaps we should look at the issue not as the RATE of taxation, but the proportionate extractions from economic activity (including those occurring by regulatory impacts).

Phil writes:

Brad - Whether or not Gates or Buffet believe the federal government could be both efficient and effective at redistributing their wealth, they still would not do what you suggest because they would lose control over where those funds go. Giving to government puts the money in the general fund to be allocated through a political process. Giving directly targets the impact the policy areas most important to them. They can both donate and believe in high tax rates without being hypocritical.

The Original CC writes:
love actuary writes:

Dr H, while you don't draw explicit conclusions, are you reasoning from a (price) change here?

I think his point was that the traditional analysis reasons from a price change. Romer & Romer explicitly avoided doing that and found that lower tax rates improve economic growth.

Or were you suggesting that DH reasoned from a price change in a different part of the post that I missed? (Seriously! No sarcasm intended!)

David R. Henderson writes:

@love actuary,
Dr H, while you don't draw explicit conclusions, are you reasoning from a (price) change here?
If a tax is a price, then yes. And there’s nothing necessarily wrong with doing so. The problem with not understanding all the reasoning behind a rule such as Scott Sumner’s is that you can sometimes misapply the rule.
If someone asks, “What happens to cigarette prices when the cigarette tax is raised by a dollar a price,” an economist, if he knows the elasticities of supply and demand, can answer that question.

Floccina writes:
A more sober analysis comes from William Gale at the Brookings Institution and Andrew Samwick at Dartmouth College and the National Bureau of Economic Analysis. In a 2014 article, the two economists found that "U.S. historical data show huge shifts in taxes with virtually no observable shift in growth rates."

So if low tax rates do not decrease growth then we should keep them low. Or at least keep taxes just high enough to fund the Government functions that show really good results.

mickey writes:

I'm interested in what he thinks a ~90% kind of rate should start at bracket wise. According to the Tax Foundation, in 2013 dollars, such a high rate had a married filing separately or individual bracket starting at above $1.5 million. Though marginal rates increase significantly even for those making much less.

Zeke writes:

Rates miss (by and large) the point. What we really should care about is the ETR and the METR. It also ignores that a lot of taxable income today was not taxable income back then. Basically, Gates' statement is a nice correlation, but lacks the nuance needed to make a strong argument.

Mark V Anderson writes:
Brad - Whether or not Gates or Buffet believe the federal government could be both efficient and effective at redistributing their wealth, they still would not do what you suggest because they would lose control over where those funds go. Giving to government puts the money in the general fund to be allocated through a political process. Giving directly targets the impact the policy areas most important to them. They can both donate and believe in high tax rates without being hypocritical.

No Phil, Brad is right and you are wrong. Gates and Buffet have both argued that taxes aren't high enough. They didn't say everyone should give more to charities. They are arguing that everyone should give up their control of their money. So when they don't give up their own control, they are being hypocritical.

I have often argued this very point, that people that want to help others should give more to charity, not government, because with charity they can give the money to where they believe it would be the most valuable. Those that disagree seem to think that individuals will make worse choices than governments. Gates and Buffet seem to be of that belief, except with their own money.

Zack writes:

Zeke's point is important. Politifact doesn't even mention the 1986 Tax Reform. Those allegedly high rates in the 50's and 60's existed under an entirely different tax code with a much narrower definition of taxable income. Comparing rates before and after '86 is apples and oranges, isn't it?

They say the top rate in the 80's was "50 then 38.5 then 28," when in fact that 50% rate didn't even go into effect until 1982. They also don't adjust for differences in the rate of population growth over time. They ignore the capital gains tax too. This reminds me of when Elizabeth Warren recently claimed the typical American family has a lower income today than in 1980, and somehow they rated it "mostly true" if I remember correctly.

Dan writes:

"@love actuary,
Dr H, while you don't draw explicit conclusions, are you reasoning from a (price) change here?
If a tax is a price, then yes. And there’s nothing necessarily wrong with doing so. The problem with not understanding all the reasoning behind a rule such as Scott Sumner’s is that you can sometimes misapply the rule.
If someone asks, “What happens to cigarette prices when the cigarette tax is raised by a dollar a price,” an economist, if he knows the elasticities of supply and demand, can answer that question."


This is not correct. We don't know the counter-factual. You should read Sumner's oil example, he makes a similar point.

Tax policy does not happen in a vacuum. Suppose the government pursues a counter-cyclical policy and cuts taxes after a negative shock that will last couple of years. You'll get a negative correlation between tax cuts and growth. It's easy come up with other scenarios in which you'll have zero or positive correlation.

You can make things more complicated by thinking about future taxes. Clinton, for instance, raised taxes but reduced the deficit (hence future expected taxes). Who knows what the net effect is.

Looking at the correlation between taxes and growth (which is close to zero) will tell you absolutely nothing about the true underlying economic relationship. So, it is like arguing from a price change.


David R. Henderson writes:

@Dan,
Good point. And, as I’m sure you noticed, the Romers take that into account and so they can’t be criticized on this in a way that Gale and Samwick can.
Here’s why I think I’m fairly safe in my reasoning: I’m not singling out a particular year. We have 4 years versus 6 years, which helps wash out the effect you’re talking about. It’s not perfect.
Remember also the main point I’m making: Punditfact should have pointed out what I pointed out.

Erich Boldt writes:

What no one seems to be pointing out is that Gates is comparing top individual tax rates in the sixties with the current top corporate tax rate. According to the IRS, the top corporate tax rate in the sixties started at 52%, then dropped to 50% in 1964, then to 48% in 1965, and finally back up to 52.8% in 1968. That's not exactly low, but it isn't 90%.

Using the annual growth rates from this post, for the years when the top tax rate was over 50% the growth rate averaged below 4%. The growth rate for the rest of the decade averaged over 5%.

It's also interesting that the growth rate for 1963 (the year the tax rate cut would have been passed) was higher than the average for 1960-63, even with the odd 6.1% growth in 1962. Similarly, the growth rate for 1967 (the year the tax rate hike to the highest level for the decade would have been passed) was well below the growth rate for the previous years with the same tax rate.

Perhaps incentives do matter.

Dan writes:

"@Dan,
Good point. And, as I’m sure you noticed, the Romers take that into account and so they can’t be criticized on this in a way that Gale and Samwick can.
Here’s why I think I’m fairly safe in my reasoning: I’m not singling out a particular year. We have 4 years versus 6 years, which helps wash out the effect you’re talking about. It’s not perfect.
Remember also the main point I’m making: Punditfact should have pointed out what I pointed out."


Unfortunately, averaging would not wash out the endogeneity problem. By that argument, you should look at an even a longer time series instead of two points in history. For the US, you'll find a significant positive correlation between top tax rates and growth. This i also true at the local and state level. If you just look at the historical record, Bill Gates is correct.


The Romer study ATTEMPTS to resolve the endogeneity problem, but short of a controlled experiment this is an impossible task. Even if you could identify exogenous tax changes (which I don't think you can), expectations about future changes in taxes are impossible to quantify. e.g. how do you know that a "permanent" tax decrease won't be changed by the next administration.

I think that Punditfact got it right. Historical record is supportive of the Gates' claim that higher taxes are associated with higher growth rates, but this does not imply that higher taxes cause high growth rates.

Jim Glass writes:

There's another big logical fallacy in the whole "90% tax rate" argument that few people (who aren't tax professionals) notice: It assumes the rich actually paid that rate when it was on the books. And that, thus, the richest paid more in taxes than anyone else.

Both beliefs are false. In those olden days the nose-bleed rates were accompanied by a massive web of tax preferences (exemptions, deductions, credits, business tax subsidies of all sorts, etc.) which made those top rates much more decoration than reality -- plus a massive tax shelter industry that the rich with their lawyers and CPAs used to game all those tax preferences better than anyone else.

The result is reported e.g. in the dead-tree IRS Statistics of Income volume for 1965 sitting on my shelf. With the top tax bracket rate at 70%, effective tax rates by income level were (tax actually paid/reported income, 1965 dollars)…

Over $1 million: 30.8%
$500k to $1,000k: 32.8%
$100k to $500k: 36.4%
$50k to 100k: 31.2%

So the people who had income of $50k to $100k paid a higher effective tax rate those who had 20x more income. The richest paid a lower effective rate than people who had only 5% of their income.

The naive always dream that if politicians enact a tax people will actually pay it! But the deadweight cost of taxes rises not by the increase in the tax rate but by the square of the increase in the tax rate. So say you up a tax rate from 30% to 90%, "triple the size of the tax, the deadweight loss increases nine-fold". And this creates massive incentives for politicians to sell tax breaks and rich interest groups to buy them, gutting the tax. And of course the richest are better at playing this game than anyone else.

This was the world until the 1980s. The richest didn’t start paying the *highest* effective rate until after the Tax Reform Act of ‘86 lowered the top rate to 28% while also killing off the tax shelter industry and eliminating the favorable rate for capital gains.

The progressive left signed on to and co-sponsored that bi-partisan reduction of the top rate to 28% because they were so outraged by the tax shelter industry, and by the richest using it to legally game their rates down so *low* (in many highly publicized cases to $0) so often under the 50%-70% rate era's rules.

How quickly they forget how the world works -- and their own history.

BTW, this is the same thing we see today with the estate tax, which has comparably high rates and plentiful opportunities to avoid them if you can afford the lawyers and CPAs. The very largest taxable estates pay effective tax rates lower than the smaller ones.

E.g., Buffett and Gates in particular, while publicly saying others should pay more taxes, have both set up things so neither will ever pay a dollar of estate tax, while assuring their descendants will carry on with more tax-favored inherited wealth than they will be able to spend in generations.

Footnote: Anyone old enough to remember Art Linkletter may enjoy an entertaining talk he once gave about how when he as a broadcast celebrity found himself pushed toward the 90% tax bracket he sure as heck wasn't going to pay it, so he set up a personal tax-shelter conglomerate that engaged in businesses ranging from drilling oil wells to making the hoola hoop famous, ran his broadcasting income through it and got it back all squeaky-clean and low-tax. It used to be on the Web, though I haven't checked on it recently.

Mr. Econotarian writes:

If you look at the numbers, Federal Income tax revenue went UP significantly the year after the "JFK Tax Cut".

This implies that a lot of income was being "hidden" from the 90% top marginal rate, which is in agreement with a lot of anecdotal evidence of massive tax avoidance during that time (for example, see the 1960 book "Swiss tax shelter opportunities for U.S. business" by Bianchi & Walter, or the 1963 book "Encyclopedia of tax shelter practices" by Prentice-Hall).

Lowering the top marginal rate allowed more income to become "visible" and be taxed.

Imagine the challenge of the IRS trying to find rich people's income in a time before wide spread computers and when the politics did not encourage the IRS to get money from rich people anyway.

Regardless, 2014 US total government revenue (federal, state & local) was 34.5% of GDP, 1965 US total government spending was 26.5% of GDP. Maybe governments (especially state governments) need to spend less to achieve the higher growth rates of the 1960's?

Several years ago I put together a list of studies and their conclusions of the impact of taxes on economic growth:

Taxes and Economic Growth

Zack writes:

Great comments from Jim Glass and Mr. Econotarian. Looking at taxes actually paid as a percent of GDP, it's pretty difficult to argue that the 50's and 60's were a particularly high tax era. I'll also add that according the Tax Policy Center (not exactly a conservative group) the average effective rate on capital gains in those years typically ranged from about 14-16%.

I could also point out the fact that government spending as a share of GDP was generally lower in those decades, a higher share of the budget went to the military, no Alternative Minimum Tax, and environmental regulation was virtually nonexistent compared to today. Almost sounds like a Republican's dream.

Comments for this entry have been closed
Return to top