I’ve been reading through Rick Perry’s official analysis of his tax plan (scored by John Dunham and Associates), and hoo boy, it’s a doozy. Difficult to know where to begin.

The intellectual dishonesty is breath-taking. The basic conservative argument that tax cuts increase economic growth goes like this: tax cuts for the rich results in more capital formation, which fuels greater productivity and higher economic growth. In other words, supply-side growth with demand catching up. It’s a pretty ridiculous model in the face of huge demand shortages, but hey, at least it’s a model.

But as Matt Rognlie points out, the model that Rick Perry used is very different, and in fact has a lot more in common with traditional Keynesian demand-side economic models than conservative supply-side models. In fact, this economic model (called IMPLAN) is usually used to justify government public works projects, such as sports stadiums, because this model shows that the economy can be significantly improved through government spending.

Wait, what?  Rick Perry’s using an economic model that shows that government spending helps the economy? That’s right, partner. In fact, the IMPLAN model (which we’re decently familiar with at EPI) will actually show that spending cuts reduce demand more than tax cuts boost it (because a portion of the tax cuts are saved), so the net effect is likely negative. And to be clear, JDA did assume that the plan would cut spending by the same amount as the revenue loss (page 7, footnote 7).

So how did the analysis show a positive economic impact? Simple: they assumed the spending cuts, but didn’t model them, despite the fact that it’s logically inconsistent to claim that tax cuts have an effect on demand but spending cuts don’t. The only reason they got a positive economic impact was because they did not model the entire plan. Had they, it’s pretty clear that the analysis would have shown that Perry’s overall budget plan—tax cuts paired with offsetting spending cuts—would, on net, hurt jobs and economic growth.