A Post: Tax Burdens, Presidents, and Subsequent Economic Growth – A Few Pictures, Part
by Mike Kimel
A Post: Tax Burdens, Presidents, and Subsequent Economic Growth – A Few Pictures, Part 1
Last week I had a post looking at the relationship between the change in the tax burden in the first two years of a Presidential administration and the growth of real GDP during the remaining years of the administration. I’ve done variations of this exercise before. It turns out that the more an administration reduced the tax burden in its first two years, the slower the growth the in real GDP over the remainder of its term in office administration. Assuming the result is more than an artifact of the data (and it does seem to correspond with other results I’ve reported here over the years), it requires an explanation. While (I am not happy to report) an increased tax burden might in and of itself stimulate faster economic growth, I suspect a bigger effect is that a) the easiest way to move the tax burden is by increasing or decreasing tax regulation and b) there is a correlation between an administration’s views on tax regulations and its views on other regulations that are intended to prevent externalities. This theory is supported by the fact that the relationship between lower tax burdens and slower growth is strengthened by not including the administrations that served only four rather than eight years makes the relationship stronger.
As I keep noting, one doesn’t have to like the results. I personally would much prefer a world in which lower tax burdens do lead to faster economic growth. But the data doesn’t seem to show that. Still, every time I put up a post like this, I get a lot of flack. One thing people keep telling me is that the results are, at best, a coincidence. In their honor, in today’s post I’m going to describe a few more coincidences that the data shows in my next few posts. Some of these coincidences I expected to see, and some, to be frank, I did not. Today I’m going to stick with a few coincidences I expected.
So… let’s go with coincidence number one. The graph below shows the change in the tax burden from Year 0 (i.e., the last full year of the prior administration) to Year 2 on one axis, and the growth rate in the last full year of each administration. (Only eight year administrations are included.) As an example, for Ronald Reagan, we see the change in the tax burden from 1980 to 1982 along one axis and the percentage change from the 1987 real GDP to the 1988 real GDP.
Figure 1
Notice that the relationship between the tax burden in the first two years of each administration and the growth rate in its last year is extremely strong. That’s consistent with what I wrote in my last post (and so many times before): most administrations do not change their tax policy very much, but tax policy (and other policies that correlate with tax policy) can take a while to have an effect on the economy.
Before I go on, a few ground rules for those who want to comment or send me e-mail:
1. If you really believe that the growth rate in the last year of an administration is “causing” the change in the tax burden in the beginning of the administration, I encourage you to seek psychiatric help. I can’t do anything for you.
2. US’ participation in World War 2 prior to 1940 is best described as peripheral. Growth in 1940, or 1939, or 1934 for that matter, is not due to World War 2.
(If you find my constant repetition of these ground rules funny, hazard a guess as to what creeps into my inbox.)
Now, another coincidence… the next figure shows the the change in the tax burden from Year 0 (i.e., the last full year of the prior administration) to Year 2 on one axis, and the growth rate in the fourth year of each administration.
Figure 2
Again… the picture looks an awful like Figure 1. The fit isn’t as good (consistent with the idea that it takes a while for policy to have a a very strong effect. Kind of odd for a coincidence.
Now… you may be wondering… what about other years. I’ll tell you flat out, the fits in years 1 and 2 are awful… consistent with the idea that it takes a while for policy to have a very strong effect. As to the rest, that will wait for the next post.
To close, nominal and real GDP come from the Bureau of Economic Analysis. GDP was first computed in 1929, so the first complete administration for which we have data is FDR I. Data on the Federal government’s tax receipts comes from the Bureau of Economic Analysis’ NIPA Table 3.2.
As always, if you want my spreadsheets, drop me a line at my first name (mike) period my last name (kimel – with one m only) at gmail period com. I should also point out, you can find a lot more of this sort of analysis in Presimetrics, the book I wrote with Michael Kanell.
Suspect that it’s all about the net.
ken,
Yeah. That one comes up often too. But its a second tier “explanation.” Time travel speculations (in general – growth that comes later explains tax burdens that come before, and in specific – the rapid growth that began in 1934 is due to World War 2 (a mistake even Krugman seems to repeat in his posts)) are particularly popular.
“1. If you really believe that the growth rate in the last year of an administration is “causing” the change in the tax burden in the beginning of the administration, I encourage you to seek psychiatric help. I can’t do anything for you.”
The increase/decrease in the level of tax receipts (or tax burden) is highly dependent on whether or not the economy is in recession during the change in adminstration.
Consider: in 2001, the U.S. economy went into recession. In years prior, both the federal government and the state of California had budget surpluses and projected surpluses beyond 2000-01. In the first two years of a new administration, the level of federal receipts fell faster than GDP, which you argue is caused by specific tax collection policy. Yet, the level of tax receipts in California fell just as sharply despite no change in state leadership or change in tax policy. The culprit in both cases, was the recession and the nature of the recession; sharp declines in personal and corporate income taxes as business profits turned to losses and capital gains vaporized. Obviously along those lines, the decline in income of Californians would affect the level of federal tax receipts.
Why did the tax burden decline during G.H. Bush’s single term in office; was it because he was wish-washy on tax enforcement (despite hiking tax rates) or was it because a recession was causing tax collections to dwindle?
In the Clinton adminstration, it took until 1995 for tax receipts/GDP to reach the level of 1987-89 despite a higher top marginal tax rate (39% vs. 28%). Incidentally, personal income tax as a percentage of GDP was lower during the the first Clinton adminstration than during the the late ’80’s cyclical peak despite the higher marginal rates. The first Clinton adminstration had higher corporate tax collections to offset the difference.
Did Herbert Hoover hike tax rates across the board in 1932 but offset this attempt to raise the tax burden by becoming lazy and oblivious towards tax enforcement such that tax receipts declined even further in 1932 or did the depth of the decline in 1932 reduce taxable income and GDP and cause the tax burden to shrink?
In 1933-34, the levels of receipts began to grow because the economy began to grow again. There was no major change in taxation rates (those occurred in 1932). You seem to have the chain of causality backwards — arguing that the increase in the level of tax receipts caused the economy to expand ,or at the least, that the increase in tax revenue did not deter growth
There is no denying that the level of tax receipts (tax burden) will always be relatively highest at the peak of an expansion and relatively lowest coming out of a recession. At a cyclical peak business profits are at their highest levels, income tax collections are high due to growing incomes and low unemployment, payroll tax collection is high due to the level of employment etc. The position within an economic cycle is far more pertinent to the level of tax collection than tax policy. Tax policy perhaps explains some of the nature and duration of these cyclical expansions (there are other factors; monetary policy is paramount among other factors) but that is an entirely separate issue.
The Numeraire,
“In the Clinton adminstration, it took until 1995 for tax receipts/GDP to reach the level of 1987-89 despite a higher top marginal tax rate (39% vs. 28%). “
1. I keep pointing out that you can cut the tax burden while increasing marginal tax rates or raise the tax burden while decreasing marginal rates. Its called enforcement. Tax/GDP rose every year of the Clinton administration. Do you really think that the economy improved every single year of the administration, or that Clinton raised the marginal tax rate every year?
2. Yes, the business cycle can make a difference. It can decrease ttax/GDP. But every Republican administration in our sample reduced taxt/GDP. Are you trying to imply that this is because every Republican administration since 1929 has tanked the economy? That’s where you want to stand your ground? You want to go with “Republican administrations always make the economy worse than it was when they took office, but its OK because that isn’t caused by taxes”?
3. So… it took thre years from 1989 to 1992 for tax/GDP to drop from 18.9% to 18.1%, and then another three years from 1992 to 1995 for it to rise back to 19%. And that is evidence of what exactly?
4. You are aware that the economy was still shrinking in 1933… and the tax burden went up. My little theory explains how. Does yours?
“There is no denying that the level of tax receipts (tax burden) will always be relatively highest at the peak of an expansion and relatively lowest coming out of a recession. “
Live and learn. And here I thought there was a recession in 1980, and yet the tax burden was rising that year, and continued rising until 1982. I’m too tired to check the data for other examples, but I think the data is denying what you claim there is no denying. I’ve noticed this pattern with your comments in the past – blanket statements that contradict the data. Its one thing to disagree what it all means, but if you’re going to assume the data is wrong the onus is on you to explain why. And you should start off by stating precisely that, otherwise it seems as if the statements you are making are not in opposition to the data but actually supported by the data.
“Do you really think that the economy improved every single year of the administration, or that Clinton raised the marginal tax rate every year?”
It’s well known that the economy accelerated in growth during Clinton’second term and improved every year. If you dissect the data its obvious that the biggest driver in enhanced revenue collection is from increased capital gains tax collections and related one-time windfalls from stock options on personal tax returns. If you want to play up the enforcement angle, the least you could do is present some analysis on the enforcement techniques used and estimates of how enforcement enhanced revenue.
If enforcement is the prime determinant of the level of tax collection, then the Dubya adminstration must of had an enormous change of heart from 2003-2007 when revenues grew faster than GDP. I haven’t seen any evidence of that — in fact folks like David Cay Johnston make the argument that enforcement was lax during this time period.
“2. Yes, the business cycle can make a difference. It can decrease ttax/GDP. But every Republican administration in our sample reduced taxt/GDP. Are you trying to imply that this is because every Republican administration since 1929 has tanked the economy? That’s where you want to stand your ground? You want to go with “Republican administrations always make the economy worse than it was when they took office, but its OK because that isn’t caused by taxes”? “
In your sample there are instances when the economy was already in recession or rapidly deccelarating. 2001 is a prime example, and the Nixon adminstration is as well. During the period in which the Johnson surtax was in effect from mid 1968 thru 1970, economic growth was far slower than in the period before and after the surtax was in place. The surtax may have raised the level of tax receipts but slower growth in GDP offset much of this. When the surtax ended, the tax burden returned to its pre 1968 level but growth increased.
3. So… it took thre years from 1989 to 1992 for tax/GDP to drop from 18.9% to 18.1%, and then another three years from 1992 to 1995 for it to rise back to 19%. And that is evidence of what exactly?
Evidence of the impact of a recession as it applies to the level of tax receipts. How can I be sympathtetic to your alternate view when you provide no evidence of lax enforcement of tax law from 1990 to 1992?
4. You are aware that the economy was still shrinking in 1933… and the tax burden went up. My little theory explains how. Does yours?
Technically the economy did decline in 1933, but it all occurred in the first quarter. Statistics on industrial production and merchandise trade confirm this. Furthermore and most imporantly, you are looking at real GDP, but taxation is based nominal GDP. The GDP deflator turned positive in 1933 after being sharply negative in the three preceeding years.
Live and learn. And here I thought there was a recession in 1980, and yet the tax burden was rising that year, and continued rising until 1982.
Guess you still have a great deal of learning to do. Nominal growth is the basis for tax collection not real growth Nominal GDP growth in 1980 was nearly 10 percent despite the recession — in the 1981 recovery nominal GDP was nearly 15 percent. The tax burden peaked in 1981 and began declining in 1982. The biggest contributor to tax collection growth in […]
Numeraire:
You do not believe Greenspan had an impact during the Clinton years? Also, I do see Mike Kimels data supporting his beliefs. As a LSS practitioner, the data certainly supports his beliefs and points in the direction he believes. I do not see much in the way of statistics to support your contention. Unsupported opinion?
It’s well known that the economy accelerated in growth during Clinton’second term and improved every year.
Please inform the BEA that it is well known that their figures for real GDP are incorrect. (http://www.bea.gov/national/xls/gdpchg.xls)
You can tell them it is well known that despite the fact that they show real GDP growth rates to be lower in 95 and 96 than in 94, and that real GDP growth also fell in 98 and 2000, the economy actually Improved every year while Clinton was in office.
I’m sure they’ll give your assertion the respect it deserves and throw out the official GDP figures they’ve been compiling
“If enforcement is the prime determinant of the level of tax collection, then the Dubya adminstration must of had an enormous change of heart from 2003-2007 when revenues grew faster than GDP. “
Oy, gevalt. I keep saying, enforcement is one factor, but over the long haul, more important than the marginal rate if you want to determine tax collections / gdp.
And bear in mind… by 2007, the tax/GDP was still 1.8% below its 2000 levels. (And below its 1995 levels.)
“In your sample there are instances when the economy was already in recession or rapidly deccelarating.”
Its called statistics because there are other things happening. Not every point fits on a straight line.
“During the period in which the Johnson surtax was in effect from mid 1968 thru 1970, economic growth was far slower than in the period before and after the surtax was in place…. When the surtax ended, the tax burden returned to its pre 1968 level but growth increased. “
Um… technically, yes. If you skip ahead to 1973 and ignore the intervening years, yes. But what about those intervening years?
I’m tired of the details, let me skip ahead…
“Guess you still have a great deal of learning to do. Nominal growth is the basis for tax collection not real growth Nominal GDP growth in 1980 was nearly 10 percent despite the recession — in the 1981 recovery nominal GDP was nearly 15 percent. “
Yes? Remember, we’re looking at nominal tax collection / nominal GDP, or real tax collection / real GDP, which are the same thing. You are arguing that collections / GDP rose because nominal GDP rose. But that only is true if collections rose faster than GDP.
“You are presenting the idea that the level of tax receipts (the tax burden) is something that a government can actively target and effectively manage (and providing no evidence of how this is performed) when in fact government is subordinate to the business cycle.”
You are correct that I haven’t (yet) presented evidence that its enforcement. Enforcement is my hypothesis, the only one I have that actually fits the data that the BEA presents.
You can tell them it is well known that despite the fact that they show real GDP growth rates to be lower in 95 and 96 than in 94, and that real GDP growth also fell in 98 and 2000, the economy actually Improved every year while Clinton was in office.
You misinterpret what I wrote (I’ll admit that my original sentence was poorly worded). When I said “accelerated during Clinton’s second term” I meant the total 4-year growth from 1997 thru 2000 measured against the first term. You read it as meaning accelerating rate of growth year after year.
And bear in mind… by 2007, the tax/GDP was still 1.8% below its 2000 levels. (And below its 1995 levels.)
Correct, but not due to enforcement but due to changes in tax rates. Especially, the deadweight loss portion of the tax cuts — namely doubling the kiddie tax credit and making it refundable and the cut to 10% from 15% of the lowest bracket. Neither of these tax cuts create any marginal supply-side incentive to supply labor or declare additional income during a tax year.
Yes? Remember, we’re looking at nominal tax collection / nominal GDP, or real tax collection / real GDP, which are the same thing. You are arguing that collections / GDP rose because nominal GDP rose. But that only is true if collections rose faster than GDP.
I am arguing that tax collections rose faster than nominal GDP in the years of heavy inflation because that is what happened. I even hypothesized the likely explanation (bracket creep).
In the 1980-81 period I was discussing, only personal income and social insurances taxes rose faster than nominal GDP. Social insurance taxes rose rapidly from 1980 to 1981 because of a huge spike in employee and employer contributions (more than 40% annualized) during Q1/1981.
The NIPA tables on personal income show disposable income growing slower than nominal personal income during 1981 and shows the opposite during 1982. Of course, income tax and social insurance tax measures against GDP (tax burden) show the exact same thing. I’m showing my work to prove my points, why don’t you show yours to prove your much broader point on the importance of tax enforcement?
You are a very clever person!
adjustable bed reviews