I had the great pleasure of talking to a very sharp veteran of the federal bureaucracy yesterday, and he shared a very useful insight. When we think about the scoring conventions used by CBO and EIA and other agencies, it helps to keep in mind that it is very important to get the relative differences between legislative and regulatory initiatives right, e.g., it helps to know that A costs more than B and that B costs more than C. If the model then tells you that C costs more than A, something has gone wrong. Internal consistency is prized. But this means that you have to commit yourself to a number of arbitrary assumptions to make the numbers cohere. This approach can give us a very crude sense about how different policies compare, but it doesn’t give us a terribly accurate or precise portrait of how any discrete initiative will turn out.

I was thinking about this while reading David Roberts brief take on the EPA analysis of the American Power Act. In all of my interactions with Roberts, I’ve found that he is an exceptionally smart guy and a gentleman. And he also firmly believes that U.S. cap-and-trade legislation can stave off a profound climate catastrophe. Last night, I tweeted the following: “Do people really believe that the American Power Act will prevent climate change from happening?” The immediate response from some of my friends, predictably, was that “no one” was saying such a thing. And I’m well aware of the fact that the statement I made was extreme and implausible — but that’s how I (mistakenly) interpreted the following words from David:

The impact on U.S. consumers will be “modest.” Says the report: “Average household consumption is reduced, relative to the no-policy case, by 0.0-0.1% in 2015, by between 0.0-0.2% in 2020, by 0.2-0.5% in 2030, and by 0.9-1.1% in 2050.” Averaged over 2010-2050, households will pay an extra $79 to $146 a year. Not exactly a steep price to pay to avoid catastrophe. (Incidentally, overall household consumption will continue to rise, even with the mild constraints of the bill.) [Emphasis added.]

The notion that the American Power Act in itself would mean the difference between catastrophe and non-catastrophe struck me as wildly implausible. But Brad Plumer observed, correctly, that the EPA’s analysis strongly suggests that this is true.

That’s when I started reading the EPA analysis, and found a series of baffling assumptions, including highly implausible assessments of the international reaction to APA, the impact of international and domestic offsets, and how carbon prices would “generate” cost-saving technologies.

One can see how the EPA analysis might be useful when comparing the very broad impact of APA to other legislative proposals, but does it give us the foggiest idea of how it would work in practice?