Unsurprisingly, the Tax Foundation objected to my critique of their pumped-up growth results from evaluating the House GOP tax cut. Their response makes a bad point and a good point (so, on net, they got nothin’!).
Their bad point is blowing smoke around net and gross numbers. This is just obfuscation, and not worth arguing about. The TF themselves prominently report the following as a “Key Finding” from their analysis of the House GOP tax cut (it’s their bullet #3):
“The plan would reduce federal revenue by $2.4 trillion over the first decade on a static basis. However, due to the larger economy and the broader tax base, the plan would reduce revenue by $191 billion over the first decade.”
Like me, they don’t get into net or gross either here. What matters, and what I and anyone else who worries about smart fiscal policy must question, is their implausible claim that the growth effects from the plan would turn a $2.4 trillion revenue loss into one that’s 92 percent smaller.
Their second point is much better, and warrants a response. As have many other critics of their model, I pointed out that by assuming away “crowd-out”—the idea that budget deficits compete with private borrowing and thus lead to higher interest rates, which in turn slow investment and growth—the TF model gets results that are far, far more favorable to tax cuts than the official scorekeepers.
Here’s what I wrote:
“In fact, in the Congressional Budget Office’s and JCT’s analyses, these types of proposals don’t just generate growth plus-ups based on lower tax rates or capital costs. They also, when they’re not paid for, generate larger budget deficits that crowd out private borrowing, raise interest rates and thus hurt investment. As my Center on Budget and Policy Priorities colleague Chye-Ching Huang notes, “JCT’s analysis of permanent bonus depreciation concluded that in the second and third decade, due to deficits and crowding out, it’s impossible to tell whether counting macroeconomic effects even reduces or increases the cost of the proposal.” The TF model, in effect, assumes away such potential negative impacts.”
That’s unquestionably true, and if TF disagrees, it would be useful to hear about it.
What I didn’t get into is my own views on the crowd-out question, and TF correctly and relevantly point out that in other writings I’ve raised questions about the validity of the crowd-out assumption. In recent years, interest rates have stayed low across many parts of the globe, often irrespective of government finances.
Interestingly, this dynamic is likely related to what economist Larry Summers has dubbed “secular stagnation,” a condition driven by the failure of even low interest rates to generate the investment growth needed to achieve full employment. The TF model ignores this (as do most other such models), yet another reason I consider their results too sunny.
Their accusation is that I’m being inconsistent on crowd-out. In fact, the point of this aspect of my critique was simply to show that they’re making very different assumptions than the official scoring agencies, and that these assumptions are partially driving their big offsetting results. To the extent that my own views on crowding out matter here, I suspect there’s probably less near-term crowding out than CBO/JCT assumes and more than TF assumes.
What would really be inconsistent would be for me to argue, as does TF, that because of this diminished near-term correlation between deficits and interest rates, tax cuts generate enough offsetting revenue to significantly lower their costs. To the contrary, I’ve long argued against the supply-side growth effects that undergird the TF’s favorable scores for big tax cuts.
And that’s what really matters here: how confident can we be that the House GOP plan will pay for 92 percent of its (net) cost? As I wrote in my critique, I suspect we’ll see a lot more of this sort of scoring from the TF, possibly next re a new Trump tax plan.
In this regard, crowding out is but one of the model’s problems. What’s more problematic about TF’s model, as with all hyper-aggressive supply-side models, is their implausible tax-cut responses from “factor inputs,” like labor and capital, a point Josh Barro made in a highly critical recent NYT piece on the TF model. After consulting with top tax economists, he concludes, as I do, that the “Tax Foundation assumptions…take us farther away from accuracy, and make unsupportable promises of tax cuts paying for themselves.”
Consider: as noted re low interest rates, the cost of capital is already very low, and we’ve consistently been tweaking the tax code in recent years with bonus depreciation on investments. Yet we’ve seen nothing like the response to capital investment that TF assumes.
It’s this empirical evidence that I find most compelling and most damaging to the TF case. I’ve looked under many rocks for actual correlations (a lower bar than causation) between supply side tax cuts and growth. I come up with nothing.
That, instead of modelling assumptions, is what the TF folks, who are capable and serious analysts, need to show. If you want us to take your model more seriously, show us the evidence of past supply-side tax cuts coming anywhere close to paying for themselves, and conversely, the negative growth and revenue impacts from past tax increases.
Evidence first, assumptions later.