Andrew Coulson graciously responds to the slings and arrows I hurled at think tanks. I understand that EPI wishes to respond as well, so things should get lively around here.
I love Cato. I love school choice. I read their stuff all the time, and I think a lot of it is great. I cite it and use it.
But any movement is prone to groupthink. Yea, whenever two or more libertarians are gathered together, you have at least three opinions. But those opinions almost never extend to "You know what America needs? A single-payer national healthcare system." Likewise, I'm pretty sure the break room at EPI never hears the words "Right-to-work laws are awesome!" Groups extend to their own less scrutiny than they extend to those who disagree with them. They form their own domains of knowledge that tend to exclude sources of disconfirming data. Agreement on core principles like "Society should maximize individual liberty" means a lot of questions never get asked.
I don't think that think tanks fudge their numbers. I know Cato pretty well, so I know it's full of earnest, extremely smart people who genuinely believe what they write, and are scrupulous about doing high-caliber work. Most of them are smarter than me, and all of them are probably more likeable in person. But in any sort of policy debate, there's always the danger of asking yourself the question you want to answer.
Say you want to know whether Bush's tax cuts made the tax code more or less progressive. You can ask whether the gradient between brackets has gotten steeper, or you can ask whether the rich now pay a higher or lower percentage of the nation's tax bill than they did before. Those will give you different answers to the original question.
Hence the dueling factoids over whether Bush's tax cuts disproportionately benefitted the rich. The left likes to look at the average amount individuals got, which leads to the conclusion that the rich got a lot more. The right likes to look at who got a bigger share of the tax cuts, which leads to the conclusion that the poor and the middle class were the big winners. Neither of those ways to frame the question is obviously wrong.
It is easier to do this when everyone who works with you, and most of the people you socialize with, agree with you. They also influence who you consider reliable sources--the extreme version of this is Chomskyites, who reject any source that disputes The Great Man's
lies more fanciful interpretations of events. But everyone does it. Liberals like are fond of Card and Krueger. Conservatives like love Neumark, Wascher, and Murphy. The group acceptance of what are the "best" sources seriously influences work based on them.
I'm not saying academics are immune to this--indeed, the CK/NW divide is a good example of the tendency. But academics tend to ask narrower questions--not "Is the minimum wage a good idea" or "who benefits, rich or poor?" Instead they ask things like "what are the effects of the minimum wage on employment?" Now, often those figures get used as if they answered one of the other questions, either because the media needs a good lede, or because the professor has an axe to grind. But there's somewhat less room for choosing your data sources--and at least in economics, it will matter if your colleagues across the political aisle reject your approach. Cato loses little credibility with libertarians if CBPP publishes a withering critique of its work (I mean, it would if any such critique were possible.)