Originally published in Forbes.com.
I’ve been struggling with the question of how good the Tax Foundation’s dynamic model is. Although it does not support the notion that tax cuts pay for themselves, it does seem to lean in that direction. The Tax Foundation has a lot of credibility among tax bloggers and the rest of the media, because, well – What do we know? So if the Tax Foundation model is terrible, that is important. The TF analysis of the Ted Cruz plan is probably the best illustration.
The Postcard Program
According to the Tax Foundation, Ted Cruz’s tax plan borders on being a free lunch. Cruz calls for a single 10% rate, elimination of most deductions and credits, replacing payroll and corporate income taxes with a 16% business transfer tax and elimination of the estate tax – those are the high points anyway. On a static basis TF scores the plan as giving up $3.666 trillion over 10 years, but on a dynamic basis it only reduces revenue by $768 billion.
The Tax Foundation has provided a handy guide to its evaluation of the candidate tax plans. Everything is over 10 years and it gives us GDP growth ranging from 16.0% (Carson) to -9.5% (Sanders), capital investment growth from 48.9% (Rubio) to -18.6% (Sanders), wage rate growth from 12.5% (Rubio) to -4.3% (Sanders), added jobs from 5.3 million (Trump) to 5.9 million jobs lost (Sanders), static revenue increase from $13.574 trillion (Sanders) to $11.98 trillion static revenue decrease and dynamic revenue increase from $9.827 trillion (Sanders) to a decrease of $10.135 trillion (Trump).
The Tax Foundation, I think, shows its focus by putting the ten year GDP growth on top. Cruz’s plan is not at the top, but overall he ends up looking better when you consider dynamic revenue loss. His is the only tax cutting plan with change back from a trillion. Given the number of agencies he wants to eliminate, including the IRS, less than $80 billion a year in spending cuts should not be that hard to find. I think that we might find that his 16% business transfer tax might be harder to collect than people who promote such notions let on, so something like the IRS will still be required, even though we can put our returns on a post card. Still he gets even with the dynamic revenue loss just by getting rid of the Department of Education.
The other ]great thing about a Cruz presidency is another four years of birtherism, which I find pretty entertaining.
How Good Is The Model?
The Tax Foundation model results can be no better than the model that they are based on. So I have been trying to get a handle on that. The feedback I am getting seems to be that unsettled as the science is, assuming you can call it ]science, the [Tax Foundation results need to be taken with significant skepticism.
I’m Not An Economist
Laurence Kotlikoff, whom I cited in my last piece on this subject, gave me a pretty hard time when I spoke to him about it. Among the damning things he said about me, in a spirit of tough love he said, was that I am not an economists, have not studied the literature and really have no business writing about this and it is stupid thing to write about anyway since the Tax Foundation model is so bad. It is a “toy model” that has absolutely no scientific merit.
The one thing that I have to agree that I got wrong was my indication that there were no other dynamic models. He has one. Joint Committee on Taxation has one. Congressional Budget Office has one. Of course, my concern is that nobody, other than the Tax Foundation, is using their model to score the candidate tax plans.
Professor Kotlikoff believes that the Tax Foundation is getting paid to score the plans and that the people paying them like the answers that are coming out. I think that he is wrong about the Tax Foundation getting paid for scoring. At least directly. The reason I am pretty sure about that is that they scored a plan for me and all I had to do was ask.
Professor Kotlikoff insists that the TF model is very simplistic and that it ignores “income effects” which, in his view, swamp “incentive effects”. The incentive effect which the Tax Foundation model leans on heavily is the notion that if you get paid less for something – like your labor or your capital – you will provide less of it and a marginal tax rate change is like a change in your income. Since you are getting less for your labor, you might be inclined to work less. That is the incentive effect. On the other hand leisure is kind of a form of consumption, so if your income goes down, you might be inclined to work more. That is the income effect.
Other Critics
Taking some of Professor Kotlikoff’s tough love to heart, I spent some time poking around the Social Science Research Network looking for papers on dynamic scoring. In 2013 Bruce Bartlett wrote The Trouble With Dynamic Scoring (download). About the models that Congress will use for scoring we get:
Consequently, the push for dynamic scoring comes when economists are dubious that tax policy changes of any magnitude have much effect on the real economy, good or bad.
In any case, I don’t think proponents of dynamic scoring are interested in accurate revenue estimates, but only in making it easier to cut taxes and harder to raise them
In short, dynamic scoring is about institutionalizing Republican ideology in the budget process, not about improving the quality of budgetary estimates
I asked Mr. Bartlett about the Tax Foundation specifically and he responded.
I think the Tax Foundation cherry picks its results–if there is a range of results they just assume that the most positive is the most likely.
Carl Davis of the Institute on Taxation and Economic Policy wrote me that the Tax Foundation’s assumption that capital is extremely sensitive to taxes is supported by some economists but not others making me sympathetic with Harry Truman’s famous desire for a “one handed economist”. He gave me several links, but the one I found most interesting was this article Capital Gains Tax Rates, Stock Markets, and Growth by Tony Kravitz and Leonard Burman. What is interesting about it is that it is empirical.
Capital gains rates display no contemporaneous correlation with real GDP growth during the last 50 years. Although the effect of capital gains on economic growth may occur with a lag, Burman (1999)1 tests lags of up to five years and finds no statistically significant effect. Moreover, any effect is likely small as capital gains realizations have averaged about 3 percent of GDP since 1960 and have never been more than 7.5 percent.
Some Empirical Data
Some economic models are based on theories about how people are inclined to behave. What tends to be hard is judging those theories against real world data, since it is hard to do controlled experiments. In 2014 the Congressional Research Service issued a paper by Jane Gravelle and Donald Marple – Tax Rates and Economic Growth, which did not find much in the way of correlation.
Historical data on labor participation rates and average hours worked compared to tax rates indicate little relationship with either top marginal rates or average marginal rates on labor income. Relationships between tax rates and savings appear positively correlated (that is, lower savings are consistent with lower, not higher, tax rates), although this relationship may not be causal. Similarly, during historical periods, slower growth periods have generally been associated with lower, not higher, tax rates.
Claims that the cost of tax reductions are significantly reduced by feedback effects do not appear to be justified by the evidence
I spoke with Jane Gravelle and the thing that she indicated that the Tax Foundation model appears to assume “infinite elasticity” in capital. Elasticity is a measure of how things respond to price changes. Apparently the TF model assumes that if we pay more for capital (by taxing it less), it will quickly appear, but the reality is that it is not that elastic.
On The Other Hand
Of course over at the Tax Foundation they have answers. When I wrote to Kyle Pomerleau of the Tax Foundation that some economists were telling me that the income effect tops the incentive effective, he responded:
That is a massively strong assumption that government spending is worthless on the margin. Remember, what the income effect is—when people receive less on the margin after tax, they work more to offset that. This of course runs opposite to the substitution effect that says when people get less for working an additional hour they would rather work a little less and spend a little more time with their family.
However, when taxes are collected, the government then provides services that people value. This offsets the income effect across the population because people will get an additional dollar in government spending after they give up an additional dollar in tax. What you are left with is only the substitution or incentive effect. Unless you are willing to make the argument that people do not value government spending at all.
Frankly I did not follow that logic too well, but here is a Tax Foundation source that explains why in tax policy, the “income effect” cancels itself out. You can also find a literature survey on the Tax Foundation site that indicates the relationship between capital and tax rates is empirically supported and a refutation of that survey on the site of the Center on Budget and Policy Priorities.
Where We Have Never Gone Before
The progressive income tax has been in force for over a century. The last time the top marginal rate was as low as 25% was in 1931. Yet the Tax Foundation is saying that the Cruz plan of a flat 10% income tax coupled with replacing payroll and corporate income taxes with something like a VAT will supercharge the economy. Even if they had uncontested data that when capital gains rates changed by x%, savings change by y% and boosted GDP by z%, which they apparently don’t have, there cannot be anything empirical on such a large change, since we have not had anything like that ever happen.
For whatever it is worth the Tax Policy Center scores Cruz’s plan as having an $8.6 trillion 10 year revenue loss more than double the Tax Foundations static score and more than 10 times its dynamic score.
I think the point that I was making above is better made by Len Burman who wrote:
The GOP primary tax cut competition is leading us into uncharted territory.
The Tax Foundation thinks that the uncharted territory contains the land flowing with milk and honey, but, methinks there mayhap be dragons there.
Trackbacks/Pingbacks