A modest proposal part II: the debate over US tax reform

Dear colleagues:

We appreciate that in response to our questions you clarified a number of points in your letter to Treasury Secretary Steven Mnuchin and, in particular, that you are backing off from the statement in your original letter that “the gain in the long-run level of GDP would be just over 3 per cent, or 0.3 per cent per year for a decade”. As you state in your response to us, “we did not offer claims about the speed of adjustment to a long-run result”.

The only three studies you explicitly called out in your original letter do, however, provide specific estimates of the speed of adjustment that would imply that a 3 per cent increase in long-run output would increase the annual growth rate by 0.1-0.2 percentage points per year for the next decade — rates of growth that would not come close to paying for the cost of the tax cut.

Even these growth rates, however, are likely to be too high. We have honest differences with you on how the economy operates, including how responsive behaviour is to tax changes. But these are not the source of the differences here. Instead, much of the difference appears to be that you continue to mis-cite your sources while failing to consider the actual Tax Cut and Jobs Act (TCJA) rather than hypothetical theoretical proposals. In particular:

First, your use of the OECD study is flatly erroneous and we request that you publish an explicit correction. The OECD study translates the magnitude of corporate tax cuts into predicted output increases. You inputted a corporate tax cut costing $3.3tn over 10 years into its model in order to generate the prediction that the long-run level of output rises by 1.8 per cent. There is no way to justify the claim that the corporate tax cut costs that much. Moreover, even this $3.3tn corporate tax cut would only generate an increase of growth of 0.05 percentage points per year in the first decade, according to the OECD study.

Second, your letter to Secretary Mnuchin reported only one of the three models the Treasury used to assess the growth impact of the Bush tax reform commission’s Growth and Investment Tax Plan, specifically the one that reported long-run output effects more than twice as high as the other two Treasury models. Your new letter asserts this was because you “believed [this model] most accurately reflects likely saving responses and thus capital accumulation”. Do you have any basis for this belief?

Your original letter emphasised the importance of analysis that reflects the fact that “the United States operates in an international capital market” while the Treasury model you chose to present “do[es] not account for international trade or capital flows” because it treated the US as a closed economy. You did not cite Treasury’s Overlapping Generations (OLG) model that actually did incorporate international capital flows, in fact the same model that Doug Holtz-Eakin has been repeatedly featuring as the basis for the dynamic scoring of the American Action Forum Tax Plan. Frankly it is hard for us to escape the conclusion that you cherrypicked the highest reported estimate in the Treasury report instead of making a considered economic judgment.

Third, your letter admits that you did not model the specific provisions of the TCJA, including measures that could potentially reduce growth. For example, your modelling did not include a number of significant aspects of the bill, including the negative growth effects from two of the quantitatively largest offsets, a provision that would raise the cost of engaging in research and development and another that would raise the cost of risk-taking.

Finally, you did not address one specific question we asked: if your estimates are correct would the tax bill result in a large increase in the trade deficit? We assume you agree that it does — consistent with textbook economics. Let us know if that assumption is wrong.

Our sense that you are implying substantially larger growth effects than are warranted is reinforced by the fact that your conclusions appear to be at odds with the fact that only one of the 42 economists surveyed by the Chicago Booth IGM Panel agreed that if the tax bill passed, the result would be that GDP would be “substantially higher a decade from now than under the status quo”.

Given that we are not likely to reconcile our differences any time soon, all of us should instead send a clear and united message that encourages Congress to rely on estimates from the expert and non-partisan Joint Committee on Taxation.

Sincerely,

Jason Furman, professor of practice, Harvard Kennedy School

Lawrence Summers, Charles W Eliot University Professor and president emeritus at Harvard University

UPDATE – DECEMBER 1:

Since we sent this letter, the Joint Committee on Taxation (JCT) has come out with its dynamic score of the Senate bill. The score shows that the tax bill would cost more than $1tn over the next decade, would add less than 0.1 percentage points to the annual GDP growth rate, and the long-run output effects could be smaller or even negative.

We are sure any of us could debate any particular assumption, but we see the value in an impartial, expert umpire that is actually modeling the specific legislation before the Senate. We would be happy to continue the discussion with you but also note that you do not take a stance on how long it takes to get to your 3 per cent of GDP effect so, in fact, your estimates may even be consistent with the JCT’s first decade estimates.