Follow by Email

Thursday, January 18, 2018

Dani Rodrik: Telling interests and ideas apart [feedly]

Moderator: I find even Rodriks notes very interesting -- in addition to being very open and engaging as an  economist, he demonstrates great flexibility in seeking the truth of the multifaceted conflicts and ideologies in international relations and economics.

Telling interests and ideas apart

A long standing debate in the social sciences is whether behavior is driven by "interests" or "ideas." The debate is central in political science, where it plays out as an argument between realists and constructivists. It is less well articulated in economics, to the discipline's detriment. (See here for my thoughts on what economists would gain by taking the role of ideas seriously.)

As constructivists like to point out, interests are "congealed ideas." Or to put it differently, we don't have interests; we have ideas about what our interests are. Perhaps it's all about interests in the short run and it's all about ideas in the long-run.     

But if so, is there an analytically meaningful distinction between ideas and interests? And could we ever distinguish empirically between cases where outcomes are driven by interests as opposed to ideas?

I have never seen a good treatment of these questions. The two sides of the debate tend to talk across each other. In particular, both realists and constructivists tend to associate the interest-based perspective with rational-choice modeling, which is neither correct nor helpful.

A couple of years ago, I taught a seminar at HKS where the students and I discussed these questions in different domains – civil conflict, international trade, finance, human rights, and many others. It was a useful experience which helped my thinking along, though it also showed how tough it is to sort out the issues involved.

I plan to return to teaching and writing on this topic, but in the meantime here are a few ideas.

Let me begin my interest-based theories. Such theories are characterized by:

·        a parsimonious specification of agents' characteristics

o   based on economic (industry, occupation, etc.), social (class), or personal (dominant ethnic/identity marker) status;

·        a mapping from these characteristics to behavior through a payoff function;

·        usually, though not always, a game in which agents interact. 

Therefore we can say outcomes are "interest-based" when they are the direct result of agents' ex ante characteristics. More specifically,

·        these characteristics must be salient ex ante

·        there must be a tight mapping from these characteristics to perceived payoffs

·        the setting must not admit plausible alternative "causal models"

An important note: the payoffs need not be exclusively material/economic. Saying that behavior is driven by interests does not imply that individuals care exclusively or mostly about their incomes/consumption. These interests could also be defined in terms of cultural values or identities. A Catholic group that is lobbying against abortion is acting in its own interest. Yes, this is a result of some strongly held ideas (interests are indeed "congealed interests"). But individuals or groups favoring their material interest do so also because they think (they have the "idea" that) this is what they should be striving for.     

When are outcomes driven by ideas instead? When behavior cannot be directly predicted by interests as defined above, and when we can trace the impact of prevailing discourses/narratives on changing how interests are perceived. In particular,  

·        we must show ideas have independent traction, by delineating causal chain from ideas to how

o   they shape worldviews,

o   render salient identities, or

o   expand strategy (policy) space

·        importantly, we must also show influence of those ideas cannot be predicted from ex ante salient characteristics/markers of agents.

This way of thinking provides us with a way to testing interest-based arguments. We ask:

·        Are the individual characteristics that define preferences (and produce behavior in question) ex ante salient?

o   such as social, ethnic, or class identity, occupation, sector, weights on material versus other goals, etc. 

·        Is the strategy space ex ante determinate?

·        Are all the relevant options already on the table?

·        Is there unique, plausible model of the world?

o   or are there alternative models that could be plausibly considered?

Applying the framework

I provide two brief applications to show how this works. In the first case, the Reagan income tax cut of 1981, evidence suggests it was ideas that was dominant. In the second case, German support for austerity policies in the euro zone, I argue it was interest.

Reagan tax cut of 1981

This is typically viewed as the result of big business interest in low taxes. And it is true that business eventually became a supporter of low taxes on personal incomes. But as Monica Prasad among others has shown, business opposed those tax cuts in 1980. They were more concerned about cutting the deficit than about the provision of supply-side incentives. In Prasad's (2012) words, "the record could not be clearer that business groups opposed Kemp-Roth."

Ideas appear to have played a crucial role in changing perceptions of interests; it was all about selling a new model of how the world works. Here the policy entrepreneurship of Jack Kemp, Art Laffer, Jude Wanniski was particularly important. The Laffer curve may have been a gimmick, but it was effective in packaging, marketing and framing the tax cut proposal.

Most importantly, it convinced Reagan not only that personal tax cuts would improve incentives, but they would raise revenue. Business remained skeptical, and became a convert to the idea afterwards, once the tax cuts were passed. In other words, income tax cuts did become an interest for big business, but only eventually. It is difficult to attribute the actual reform to interests of big business or any other organized group. It was ideas that seem to have made the difference.  

German support of austerity policies in euro zone

This is typically presented as the result of peculiar German ideas on economics: "Americans are from Keynes; Germans are from Hayek." But one can present a counterargument that stresses the primacy of interests instead.

Note that Germany had strong ex ante salient characteristics that produced an "interest" in austerity:

•       Germany was structurally a strong country (with a current account surplus and full employment)

•        therefore was not in need of explicit stimulus unlike e other countries in the euro zone

•        in any case there were already strong counter-cyclical stabilizers in Germany, which did produce the intended fiscal expansion

•       euro-wide expansionary policies would mainly serve to help/bail out indebted countries

•       the hyperinflation experience had produced very high weight on inflation avoidance

•        yes, an idea, but one already embodied in ex ante preferences

•       Germany had no apparent desire for deepening political integration (which austerity policies would serve to undermine)

Therefore it was in Germany's interest to pursue austerity policies.


You can agree or disagree with the arguments in these specific cases. What I am more interested in is the analytical distinction between interest- and ideas-based outcomes. It seems to me that any meaningful, non-tautological distinction must rely on the empirical leverage provided by interest-based theories' reliance on a parsimonious set of attributes/characteristics of agents.

If we can predict outcomes based on these characteristics, by showing that they were salient ex ante and that they directly led to the behavior in question, we can argue that interests win the day. (Once again, these interests need not be material or selfish.) If we need to appeal instead to reconceptualization of objective functions or altered worldviews, and can show that specific ideas were responsible for that, then it is ideas that have the 

 -- via my feedly newsfeed

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.