Ryan Avent and Matt Yglesias ponder whether the degree of disagreement among economists is exaggerated in public debate. The classic statement of this argument, of course, is Alan Blinder’s dictum in Hard Heads, Soft Hearts that:
Economists have the least influence on policy where they know the most and are most agreed; they have the most influence on policy where they know the least and disagree most.
But ever since reading this argument, I’ve wondered whether it was quite right. Blinder’s observation helps explain a readily observable empirical correlation between (a) disagreement among economists and (b) apparent prominence of economists’ arguments in public debate. But prominence is not the same thing as influence – and I can’t help wondering whether the causation goes the other way, so that economists are only middling influential at most when they disagree. Consider the following model (for extremely casual senses of the term ‘model’).
(I) Many important political actors have stable preferences over economic policy outcomes, regardless of whether they make economic sense or not. Perhaps e.g. a party or a well-placed interest group would like to see regulations that redistributed benefits to their members or to their backers, even if this led to lower growth. There will typically be other political actors with opposed distributional interests (who would oppose such a regulation, perhaps push for their own alternative regulation which favored them or their backers etc). These opposed interests lead to bitter political debates.
(II) Other political actors, or, to the extent it pays any attention at all, the public, have weaker preferences over policies, and can potentially be persuaded to support policies that are validated by experts (economists in this instance).
(III) Economists – the validating experts – adhere to a common set of methodologies or modeling techniques that suggest some general conclusions but that can also with a little jiggery-pokery (your common or garden partial equilibrium model, or folk theorem result) can be made to support (PDF) any policy prescription you might care to be having yourself. Furthermore, some economists (perhaps because of ideological commitments, perhaps external pecuniary motivations, perhaps some combination) can be swayed to find theoretical support for arguments that support external political actors in their own sets of debates.
(IV) Political actors with stable preferences have means (either intellectual – weekend seminars, perhaps in pleasant parts of the world, or directly pecuniary) that they can employ to sway those economists that are swayable, so as in turn to influence political actors with weaker or undefined preferences.
Now, I don’t at all fully buy into this drafter’s sketch of a toy model. There is surely a greater degree of coherence to economic thinking than it suggests, so that some positions will be more controversial than others. Nor are ideas mere ideological justifications for preconceived political agendas. Even so, the model’s plausible outcomes map onto real world outcomes in some interesting ways. One might reasonably predict the following.
(1) Where economic debates are not salient to political debates, economists will be left undisturbed. There will be many propositions in economics that are not particularly relevant to political actors. Some propositions may simply not have any very great distributional impact for politicians or their clients. Others may be politically irrelevant, because they are probably repugnant to all political actors. Here, there will be no external pressures causing economists to disagree, allowing them (to the extent that their natural vexatiousness permits) to reach a happy concordance.
(2) Where debates among economists are politically salient and economists have a genuine consensus among themselves, economists’ influence will be at its peak. They will not influence actors whose interests would lead them to oppose this consensus. But they may expect their arguments to be championed by political actors whose interests accord with the consensus among economists, and to be listened to by those who are open to influence. Note, however, that this happy (for economists) state of affairs will surely be unstable. Those political actors who are disadvantaged by the given consensus will have strong incentive to undermine it, by sponsoring or encouraging potentially dissident economists to make alternative claims, flying them on all expenses paid trips to week long boondoggles where they listen to counter-arguments etc. Precisely to the extent that economists are politically influential, disadvantaged political actors will have incentive to try to open up debate among them so that they can turn to their ‘own’ economists for validation in public argument. Hence, state (2) will sooner or later collapse into (3)
(3) Where debates among economists are salient, and where there is disagreement among economists, economists will only have moderate influence on political outcomes. Each side in political argument will have its own economists to validate its claims, provide helpful models, econometric papers etc. This state of affairs (we might dub it the ‘John Lott’ equilibrium) will, unlike (2), be quite stable, except under extraordinary circumstances.
So what does this predict? Like Blinder’s aphorism, it suggests that we will observe a broad empirical correlation between (a) the extent of disagreement among economists, and (b) the involvement of economists in political disputes. ‘Eat your greens’ propositions that are popular among economists, but more or less equally uncongenial to all political actors in a given system will, as in Blinder’s formulation, be systematically ignored. But economists’ influence will not be particularly high when they disagree with each other, since different economists arguing for different sides of the political debate will at least partially cancel each other out. It will be far higher on those rare and fleeting occasions when economists unite in favor of the one or the other side actively participating in a political debate.
I note as a postscript that this toy model contains a simple public choice explanation for the emergence of public choice. Drawing this out is left as an exercise for the reader.