I’m writing a longer post on economists and the TPP for the Monkey Cage, and perhaps another post about the idea of “mood affiliation” for here (short version – I don’t think it at all does what its advocates want it to do) but in the meantime, a specific response to this post by Tyler Cowen. Tyler, partly perhaps as a result of an argument I had with him and Noah Smith on Twitter, argues that:
I’m familiar with studies showing estimated economic gains from TPP in the neighborhood of $1.9 trillion (pdf). Given the past performance of trade models, I am willing to believe that might be an overestimate. So let’s cut those gains roughly in half to say a trillion. (That said, if I understand the Peterson document correctly, they are not even trying to incorporate gains from reallocation on the production side, as might result from comparative advantage or dynamic specialization; in this sense $1 trillion may be a considerable underestimate of the upside.) That is still a sizable sum of economic gain. What would convince me to oppose TPP if is somebody did a study showing the following: when you use a better trade model, use better data, and/or add in the neglected costs of TPP (which are real), those gains go away and indeed become negative.
This is fundamentally the wrong way to think about these models. If as Tyler accepts, thinking about TPP primarily in Ricardian terms is likely to lead one ‘substantially astray,’ then starting from a Ricardian model and stipulating that you’ll lower the expected benefits by half to give your opponents a bit of a leg up, is fallacious. Specifically, it’s a weaker version of the Iraq war fallacy that Daniel identified in his 1 minute MBA.
Fibbers’ forecasts are worthless. Case after miserable case after bloody case we went through, I tell you, all of which had this moral. Not only that people who want a project will tend to make inaccurate projections about the possible outcomes of that project, but about the futility of attempts to “shade” downward a fundamentally dishonest set of predictions. If you have doubts about the integrity of a forecaster, you can’t use their forecasts at all. Not even as a “starting point”.
This isn’t to say that the Peterson Institute model that Tyler is working from is “fundamentally dishonest.” Although Alan Beattie notes that Peterson used to be “notorious for claiming trade deals would create thousands of jobs, cure scrofula and turn base metals into gold,” he accepts that they’ve gotten better under Adam Posen, (also, in fairness to the pre-Posen regime, this). But it’s highly problematic as a starting point for debate. As Jared Bernstein describes the results of the Peterson model in email conversation: “Only in DC-style econ would a number like 0.4% by 2025, derived from a model of a 29-chapter trade agreement that the modelers never saw, be taken seriously (Remember, we can’t accurately forecast monthly jobs numbers–yet we can somehow tell you to count on a miniscule change to GDP 10 years hence).”
I’ll have more to say about these issues in a follow up post. For now, just this. If you’re trying to build theoretical argument, you can reasonably ask someone to provide you with a better theory before you abandon your own. However, if you’re trying to advocate for a policy measure, and you accept that your preferred model is likely to lead people substantially astray, you don’t have any very good warrant for suggesting that this model should still anchor policy debate in lieu of someone coming up with a better one. Far better to admit ignorance (while berating the ignorance of your opponents as you like) and to accept that everyone’s views on the policy (including your own) are likely more the product of political values than dispositive evidence.