A program for cognitive economics

I’m visiting the American Economics Association conference in Philadelphia this weekend and looking forward to catching up with the latest in theoretical and empirical research. Behavioural economics has received another endorsement this year with Richard Thaler’s receipt of the Nobel Prize. The behavioural field still has only a small minority of the conference’s papers, but many more than a few years ago. It finally feels like an accepted part of the broader field.

Echoes of a new discipline have started to emerge. Miles Kimball published a detailed NBER working paper in 2015 that defined cognitive economics as “the economics of what is in people’s minds”. Before that, a book and conference in 2004 discussed the topic, and a few others (including Marco Novarese and myself, here and here) have discussed it in the meantime. It seems that the term has been invented more than once in parallel - the term is after all a natural counterpart to "behavioural" economics.

The field is related to behavioural economics but takes as its focus internal psychological processes and states, rather than external behaviour. It has links with happiness research, neuroeconomics and decision making theory.

In his paper Miles lays out some suggestions for what cognitive economics should include:
  • The economics of what is in people’s minds
  • The welfare or happiness they get from their state of mind
  • The use of survey data to find out what their state of mind is

I agree with the first two, and am happy with survey data as a methodology (indeed I use it myself all the time); but I wouldn’t go as far as Miles in implying that a reliance on survey data is part of the definition of the field. It’s enough to say that it’s likely to be a tool cognitive economists commonly use. I believe that theoretical approaches, reasoning from choices and market outcomes (once we have the theoretical scaffolding to support that), experimental economics, and neuroeconomic data are just as likely to be fundamental to the discipline. (I don’t think Miles and I would have a major disagreement here.)

I am fully on board with Miles’s observations that heterogeneity, welfare and happiness, and finite cognition are all important areas that cognitive economics can illuminate. To these I would add:
  • Fictional or symbolic utility – the pleasure we gain from our beliefs (independently of their accuracy or correlation to reality)
  • The structure of preferences – an understanding of cognition allows preferences to be endogenous to an economic model, rather than exogenously given
  • The real economic effects of marketing – aside from simply providing information, marketing persuades people – that is, it changes their beliefs and preferences

And depending on the theories we develop, the field could provide insight into:
  • Intertemporal tradeoffs, why they exist and how they are made
  • Social transmission of beliefs and preferences
  • Empathy for others and vicarious utility

I may be getting ahead of myself here. Before we can have such ambitions, I would suggest that some of the next steps in developing a successful theory of cognitive economics should be:
  1. Continuing to build alternative theories of individual decision making (DM), to supplement or replace utility maximisation. Miles talks about “new theoretical tools for dealing with finite cognition” and I agree this needs to be a core part of the program. Examples include Gabaix’s sparse maximisation models, and my own work on prospection. Much of the work in judgement and decision making would fit into this area, but a lot of it is oriented towards explaining individual decisions and not amenable to theoretical use in economic models. Of course we should avoid the trap of making assumptions about DM for the sake of modelling convenience rather than accuracy. However, it is definitely possible to make simplifying choices that retain more realism than rational agent theory, and provide more modelling power than empirical psychology.
  2. Using these DM models to explore important phenomena that are observable and have economic consequences. The models will only be used if they are useful; I believe we can show their value by explaining things like advertising, the evolution of preferences, and the significance of media consumption in our lives.
  3. Working out how markets operate under the constraints of finite cognition. The classical microeconomic models are beautiful but rely on utility maximisation to work. Many writers have observed ways in which markets depart from these ideals – which often have political implications, from Hayek’s “Road to Serfdom” on the right to Shiller and Akerlof’s “Phishing for Phools” on the left.
  4. After this, we might eventually look for macroeconomic – or macrosocial – principles of how whole economies and societies work on cognitive economic foundations. It is likely to be some time before the theoretical foundations are ready to support that kind of work.

Steps 1, 2 and 3 are likely to happen in parallel. When we attempt to apply our DM models to illuminate economic phenomena and markets, we will receive essential feedback on how good those models are.

I plan to organise a small conference or working group to discuss some of these topics and work out a research agenda. I would encourage anyone interested in participating, or discussing these subjects, to contact me (leigh@inon.com).


Popular posts from this blog

Is bad news for the Treasury good for the private sector?

What is the difference between cognitive economics and behavioural finance?

Dead rats and dopamine - a new publication