February 14, 2005
The Infernal Machine of Imre Lakatos
Kevin Drum pins Tom Maguire down behind a stone wall and snipes at him:
The Washington Monthly: "POST HOC BULLSHIT....Tom Maguire... his theory that... it's quite possible to have both a low-growth economy (meaning that Social Security finances are in bad shape) and high stock market returns (meaning private accounts would be great).... Brad DeLong... concluded that....it's possible for the low growth/high return scenario to be true, but it's not very probable....
However, there's an aspect to this whole thing that bothers me... Tom's argument is... a random post-hoc effort to explain away a problem... part of the genus bullshit.... real stock market returns have closely followed GDP growth. GDP growth is projected to decline in the future, and common sense dictates that lower growth leads to lower corporate profits which in turn leads to lower stock market growth. What resulted was a bizarre series of Rube Goldberg inventions.... Maybe corporations will suddenly become far more profitable.... Maybe they'll start paying out enormous dividends. Maybe overseas investment will skyrocket.... they just have to sound plausible enough to create a cloud of FUD...
While Matthew Yglesias wheels out and discharges the Infernal Machine of Imre Lakatos:
Matthew Yglesias: Thinking Things Through: [Y]ou can make the numbers add up by means of some heroic assumptions about the future. But... it's a degenerative research program where you're making up odd empirical predictions in order to save the theoretical assumptions. The degenerative research program problem arises once again... the eagerness of privatizers to throw the Efficient Markets Hypothesis overboard... assert that the equity premium is... an actual outgrowth of systematic market failure.... [T]he consequences of the EMH being wrong are large and severe... John Quiggin... believes that the EMH is wrong and spells out what follows... virtually nothing Quiggin has to say is anything a reputable conservative would agree with.... Someone who was seriously contemplating the case against the EMH as part of a progressive research program would be coming to all sorts of socialistic conclusions....
So I've now gone on for a very long time about this, but I think it's important. Not just for Social Security but because it tells us a lot about the operations of the contemporary right.... [N]o one has run the numbers. It's a systemic breakdown throughout the movement...
First, let me agree with Kevin's and Matthew's big point: ever since Irving Kristol decided to push supply-side economics not because it was true but because it was politically convenient:
Irving Kristol and the Coming of Supply-Side Economics: rather cavalier attitude toward the budget deficit and other monetary or fiscal problems... [because] the task, as I saw it, was to create a new majority, which evidently would mean a conservative majority, which came to mean, in turn, a Republican majority--so political effectiveness was the priority, not the accounting deficiencies of government.
the quality of policy analysis--on economic policy, on social policy, on foreign policy--coming from Republicans has been on a steady decline. And the research programs--to the extent that there are research programs--are degenerative. Nearly all of the defenses of the SSA forecasts that we see--including those coming from the Council of Economic Advisers, which has shown no signs of even trying to think the issues through--are, as Kevin says, mud thrown against the wall in the hope that something will stick. That is a shame, and something about which people should be ashamed.
But make me also make two small and limited dissents:
First, let me provide covering fire for Tom Maguire, who as I understand it is engaged in a two-pronged project: (a) To get Dean Baker and Paul Krugman and me to admit that when we said "impossible" we meant "improbable and inconsistent according to standard neoclassical growth-model assumptions." Fine. I admit it. When we said "impossible," we meant "improbable and inconsistent according to standard neoclassical growth-model assumptions." (b) To try to figure out whether recent trends widening income inequality and raising the share of income going to capital could continue and so justify SSA's assumptions (even though SSA did not have any such process in mind when it originally made its forecasts by whatever process it makes its forecasts). Those are both not unreasonable objects to try to achieve.
Second, let me point out what economists do. We take a first cut at an issue by considering a very stripped-down and oversimplified model: one in which markets are efficient, expectations are accurate, and key macroeconomic ratios have had a chance to reach and stabilize at their steady-state values. Baker and Krugman do this standard constant capital-share steady-state growth path first cut. Economists then check the first cut by considering alternatives: suppose markets break down along dimension x; suppose expectations are flawed in aspect y; suppose disturbance z pushes the economy off its steady-state path; are the conclusions reached in the first-cut analysis still true?*
To a Lakatosian, this looks a lot like a degenerative research program: a lot of theoretically poorly-motivated tweaks are made to the basic framework for ad-hoc reasons. But I think it is something different: a process of incrementally checking the robustness of fundamental working assumptions and rules of thumb that we know are only crude approximations.
*Thus the major beef I have with the administration's mouthpieces is not that they perform the more complicated analysis, but that they do not perform the more complicated analysis--they simply make assertions they cannot know to be true about what the more complicated analysis would show if they were to do it. Assertions from the CEA that closed-econom calculations are seriously wrong because they "ignore global economic growth and investment in countries unaffected by the demographic slowdown" require that time paths for trade balances and overseas investment positions be calculated and their plausibility evaluated--which the CEA does not do.
Posted by DeLong at February 14, 2005 10:51 AM