A few days ago Greg Mankiw had an op ed piece in the NY Times talking about how even small increases in the marginal tax rate would keep him (and by extension, other talented folks like him) from working.

For giggles, I pulled data on the top marginal tax rate from the IRS and real GDP per capita from the BEA’s NIPA tables. Data on the latter goes back to 1929, and the tax rate info goes back further.

It turns out that the correlation between the tax rate in any given year and the growth rate in real GDP per capita from that year to the next is small but positive. That is, higher top marginal tax rates don’t seem to reduce real economic growth. Look at the tax rate and the annualized growth rate in real GDP per capita for two years, or three, or four, or five, or six (which is as far as I went) and ditto – the higher the marginal tax rate, the faster the economic growth over the next X years. The correlation is positive, if small.

Now, you may be saying to yourself – sure, but the world is very different now than in 1929 or 1942 or 1968. What about in recent times? So let’s start in 1981, which is more or less when a) the ideas that Mankiw endorses took hold and b) Mankiw’s career began. Here’s what that looks like:

Hmmm… a clear positive correlation between the top marginal tax rate and the growth in real GDP per capita over the next four years. Using three or five year lags decreases the correlation slightly but results are about the same. It would seem that discouraging Mr. Mankiw from working would be a very, very good thing for the economy. That shouldn’t come as a surprise to you if you’ve read my book and maybe I’ll write a bit more about this when I get a chance.

However, while it may seem like I’m being facetious about how discouraging Mankiw from working would be good for the economy, I really believe it. After all, Mankiw’s work consists in large part of advocating a position in his books, lectures, op eds, and as an advisor that, as the graph above shows, is consistent with slower economic growth, and he’s very good at what he does.

[...] Mike Kimmel at Presimetrics runs the numbers and determines there is “a clear positive correlation between the top [...]

“Hmmm… a clear positive correlation between the top marginal tax rate and the growth in real GDP per capita over the next four years.”

Except that in many of those four year periods, marginal tax rates declined or increased DURING the measured periods, making whatever the top rate was in year one irrelevant. For instance, a 70 percent top rate in 1981 would not be correlated to what occurred in 1982,1983, 1984 or 1985 because the top rate had been reduced to 50 percent. Likewise, a 50 percent rate in 1985 is not applicable to the years 1987, 1988 and 1989 because the top rate in those years was 28 percent. Alternatively, any 4-year period beginning in any of the years between 1987 and 1990 would include tax hikes made by Bush Sr. and Clinton.

There is no logical reason to suggest that a marginal tax rate would have any effect on any year other than the year in which it applies. How then does a 70 percent marginal tax rate in 1981 have anything to do with real GDP growth in 1984 or 1985? How does a 50 percent top rate influence what the economy did in 1988 or 1989 with a 28 percent top rate? How does one plausibly claim that a top rate of 28 percent in 1989 is evidence of weak growth over the ensuing four years when by 1993 the top rate is 39 percent?

The Numeraire,

Fine. That’s already mentioned in the post:

“It turns out that the correlation between the tax rate in any given year and the growth rate in real GDP per capita from that year to the next is small but positive.” That’s 19%.

And if you’re curious, during the 1981 onward period, its 6%.

Now, maybe you mean the tax rate in 1981 should apply to the growth rate from 1980 to 1981, etc. In that case, we’re talking figures of 20.7% (whole sample) and 13.6% (1981 on). Again, small but positive correlations between marginal tax rates and growth.

I assume you aren’t planning to argue that a small but positive correlation between taxes and growth somehow is evidence that lowering tax rates has been the way to generate economic growth. But are you willing to give up your views when they’re contradicted by evidence?

“I assume you aren’t planning to argue that a small but positive correlation between taxes and growth somehow is evidence that lowering tax rates has been the way to generate economic growth. But are you willing to give up your views when they’re contradicted by evidence?”

There is no contradictory evidence — it’s abundantly clear from the data that the best period for real per capita growth (both in terms of absolute growth and yearly consistency) was the period in which the top marginal rate declined from 70 percent to 28 percent. From 1990 onward, marginal tax rates are higher and per capita growth is inconsistent and especially weak during the period in which marginal rates climb rapidly.

It’s your viewpoint that is need of reexamination — the last sentence in which you suggest that Mankiw’s position is ” consistent with slower economic growth” is not borne by facts. Indeed, the reason the U.S. and virtually every other nation lowered marginal tax rates beginning in the 1980′s (see table 1 of link) was because they were at the time struggling with low growth.

http://www.econlib.org/library/Enc/MarginalTaxRates.html

I’ve noted before, you and I simply don’t share the same reality.

” it’s abundantly clear from the data that the best period for real per capita growth (both in terms of absolute growth and yearly consistency) was the period in which the top marginal rate declined from 70 percent to 28 percent.”

I’ll let you in on the numbers in my time stream – links are provided in the post. We’ll ignore the first year for which tax rates were cut (1981) because they were still close to 70%, and because growth in real GDP per capita sucked in Reagan’s first year. We’ll also ignore all but the first year (1988) for which tax rates were at 28% because in the subsequent years at that tax rate, growth rates were either below 1% or negative. So I’m cherry picking pretty good on your behalf, OK?

We’re left with the seven years between 1982 and 1989. I get a real GDP per capita increase of not quite 3.4% during that period. That is just shy of the the growth in real GDP per capita during the same number of years between 1960 and 1967, well shy of the 3.9% a year from 1961 to 1968, and still less than the 3.6% from 1962 to 1969, including Nixon’s first year.

Additionally, seven consecutive year periods beginning in 1932, 1933, 1934, 1935, 1936, 1937, 1938, 1939, or 1940 all (i.e., every single one) beat your cherry-picked years by at least 1.4% per annum. And of course, you’re aware that 1932 is the year that the top tax rate increased from 25% to 63%. It would increase again to 79% in 1936 and 81% in 1940.

In fact, the best seven year stretch for real GDP per capita growth came from 1938 to 1945 – when the top marginal rate rose from 79% to 81% to 88% to 94%.

All of which is to say, you seem to have written yet another statement that is in severe conflict with reality. I’m not sure what you get out of this, frankly, but its helped me crystallize my plans for my next project so I thank you for it.

As to why every other country lowered tax rates at the same time – the monetarists and supply siders were successful at selling their theory. They were less successful at producing results. Marketing beats product, every time.