Mankiw challenges the CBO's scoring (here) on methodological grounds. Specifically, the CBO presumes constant rates of GDP growth. But Mankiw notes that the bill includes some tax increases on wealthy Americans that must, according to theory (and all else being equal), reduce work effort and, therefore, growth in GDP. This observation if correct, but does it merit the strong warning that Mankiw issues:
I have two concerns about Mankiw's heavy discounting of the CBO scoring.
(1) It takes a number to beat a number. Mankiw himself does not estimate, or cite any estimates of, (a) the GDP effects of the tax hikes in the health care legislation or (b) the effects of GDP losses from higher taxes on the net benefits (or costs) of the legislation. With respect to (b), he presumes that all else is equal, i.e., that there are no GDP-boosting elements in the health care bill that would offset (or more than offset) those tax hikes. For instance, David Cutler and Neeraj Sood claim (here) that health care reform could add between 250,000 and 400,000 jobs per year to the economy over the next 10 years. If their claim is correct, it would obviously go some way to offsetting the negative GDP effects of the legislation's tax increases.
(2) Mankiw still hasn't explained why the Clinton tax increases of the early 1990s did not curtail the high levels of GDP growth experienced throughout that decade. Perhaps raising taxes does not always have such deleterious effects on work effort and GDP, at least if there are other countervailing factors favoring high growth rates.
Having said all that, I am highly confident that Mankiw is correct that the CBO's scoring of the health care bill is erroneous. No surprise there. It is a truism that any ex ante economic estimate is going to be found to be in error on any ex post assessment. The goal, of course, is to minimize the range and amplitude of error in ex ante economic estimates. The CBO may not have done that in th case of its scoring of the health care bill; it may even be prevented from doing so by the rules under which it operates.
Even if assumed, however, that the CBO estimates are off to some significant extent, no one should be prepared to reach the opposite conclusion from the CBO. CBO estimates might be wrong and the health care reform bill could still provide net social benefits. If Mankiw wants to convince us that the health care bill is bad policy, he should give us a better, i.e., more convincing cost-benefit analysis, and not simply criticize the CBO, which, by the way, is something he only seems to do when it supports policies proffered by Democrats.
Ultimately, all Mankiw's blog has done is to provide political (not economic) cover for Republicans who find the CBO's scoring to be inconvenient.