Tax Rates v. Real GDP Growth Rates, a Scatter Plot
by Mike Kimel
Tax Rates v. Real GDP Growth Rates, a Scatter PlotCross posted at the Presimetrics blog.
This post was submitted by Kaleberg.
In this post, I will look at the relationship between top marginal income tax rates and real GDP growth using a scatter plot.
I am inordinately fond of scatter plots. The nice thing about a scatter plot is that you can present a lot of data in a fairly small space, so rather than just comparing tax rates at time period t against real GDP growth rates from period t to t+1, I can also show real GDP growth rates from period t to t+2, from t to t+3, and from t to t+4. (I.e., the scatter plot shows tax rates at any given time, and the growth rates over one year, two years, three years, and four years.)
The vertical axis is the GDP growth rate, the geometric average for multiple years. The horizontal average is the top marginal tax rate. The one year comparison is shown in dark blue, and each subsequent year is shown with a paler color and a smaller marker.
Figure 1
Data is for the period from 1929 to 2009 (i.e., all the years available from the BEA.)
Lower top marginal tax rates seem to limit economic growth with a rate of about 60% seeming to divide the restricted growth phase from the unrestricted growth phase. There might be a little falloff when the tax rate passes 90%, or there might not. There are lackluster growth rates associated with higher and lower top marginal tax rates. Mediocre growth is not all that hard to achieve. Finally, if high top marginal tax rates had a multi-year effect, we’d see a distinctive pattern in paler blue in this chart, but we don’t. The paler blue, longer term comparisons seem bounded by the single year effect.
—
The data used in this scatterplot is the same data used to build the bar chart in this this post.
—
Note from Mike Kimel – as always, if you want the spreadsheet (I have a copy of Kaleberg’s work), send me an e-mail. I’m at my first name dot my last name at gmail.com, and my first name is “mike.” My last name has only one “m.”
A coupe of things strike me:
1) lots more scatter in the above 60% tax region.
2) The GDP dip at 70%
It’s awfully hard to believe that the relationship would be bimodal – so what else is in play for the 70% years? I don’t think tax rates can be the only things that effect GDP.
Is this a reasonable hypothesis: As the tax rate gets beyond, say 60%, the tax rate itselt becomes less determinant, and other factors take prominence.
Nice piece of work!
Cheers!
JzB
You divided off pre-1933 years in one earlier analysis. Perhaps you should be dividing of pre-1980 years as well. The range of years at which you could put such a divider is pretty wide.
Potential causes of a change in that range:
OPEC became strong enough to make a difference.
Nixon’s resignation caused a massive loss of confidence.
Medicare.
I entered the workforce.
Reaction to stagflation.
Something changed the character of the variation in growth; the same thing would likely also effect average growth. Why should it be tax rates rather than some combination of meaningful events?
Mike:
About all the scatter Plot shows is the data appears to be non-normal. beyond that I believe you would have more analysis to do to determine distribution and a relationship. It certainly points a direction.
Should try graphing the effective top tax rate against GDP growth rate. I wonder what the GDP growth rate would be without the deficit spending. I think because of low tax rate that the rich save so much money that they severely depress demand so much that the economy goes in a tail spin. Without the federal government deficit spending the decline would be worst. Over the next twenty years or so I expect more tax cuts for the rich with much less spending by the government. I think our economy will fall apart and that will result is more tax cuts and more budget cuts. We will quickly become a third world economic.
I would have to agree with you because I didnt actually understand anything about this graph. what image did you see on the inkblot?
I also love this scatterplot display method.
Mike, thanks for sending me the spreadsheet. A few findings:
Correlations between marginal rate and GDP growth, t+1 through t+4
0.23 0.23 0.23 0.21
I dropped in Real GDP/Capital in place of Real GDP. Of course the growth rates are slightly lower — the population was growing.
Here are the correlations with marginal tax rate — also lower, but darn close:
0.23 0.23 0.23 0.21
Short story, there is a statistically significant correlation between marginal tax rates in year X and GDP/GDP-per-capita growth over ensuing years.
It’s pretty small, but consistently positive — a higher marginal tax rate in year X correlates with faster growth over the ensuing four years.
I also love this scatterplot display method.
Mike, thanks for sending me the spreadsheet. A few findings:
Correlations between marginal rate and GDP growth, t+1 through t+4
0.27 0.28 0.28 0.27
I dropped in Real GDP/Capital in place of Real GDP. Of course the growth rates are slightly lower — the population was growing.
Here are the correlations with marginal tax rate — also lower, but darn close:
0.23 0.23 0.23 0.21
Very consistent.
Short story, there is a statistically significant correlation between marginal tax rates in year X and both GDP and GDP-per-capita growth over ensuing years.
It’s pretty small, but consistently positive — a higher marginal tax rate in year X correlates with faster growth over the ensuing four years.
This is a pretty weak analysis. First, you need to be looking at 5-10 year moving averages of the tax rate for the 75-90th percentiles in income. The top tax rate in the 1920s applied to incomes over $1 million and in the 80s it applied to incomes below $200,000. The relative brackets aren’t at all comparable. Second, the growth effects are lagged by 5-10 years or more since the effect we are trying to measure is entrepreneurship and new business formation. The theory is at confiscatory tax rates no one is going to spend 2-3 years funding a business out of debt on the hopes that the business will generate a windfall in later years – generally 5-10 or more – particularly if your path to making money is IPO.
Andy, show us yours. We’re always looking for good analyses of these data.
Get on it.