Herding and the quest for credit

Abstract

The system for awarding credit in science – the ‘priority rule’ – functions, I have proposed elsewhere, to bring about something close to a socially optimal distribution of scientists among scientific research programs. If all goes well, then, potentially fruitful new ideas will be explored, unpromising ideas will be ignored and fashionable but oversubscribed ideas will be deprived of further resources. Against this optimistic background, the present paper investigates the ways in which things might not go so well, that is, ways in which the priority rule might fail to realize its full potential as an incentive for scientists to work on the right things. Several possible causes of ‘herding’ – an outcome in which a single research program ends up with a number of researchers well in excess of the optimum – are considered.

Keywords:

• priority rule
• social structure of science
• reward system in science
• herding
• economics of science

Acknowledgements

I would like to thank the participants at the Workshop on Methodology, Systemic Risk, and the Economics Profession for many helpful comments and conversations over the course of that most fruitful weekend in December 2012.

Notes

1. From a Bayesian perspective, the Superintendent will not have a single ‘best guess’ but a probability distribution over possible returns functions. In what follows, talk of a ‘best guess’ stands in for all such other implementations of the Superintendent, on the details of which nothing in this paper turns.

2. I exaggerate to some degree; for example, the NSF (National Science Foundation) and other grant agencies sponsor initiatives to draw scientists to work in areas considered undersubscribed, and grant money determines where there are postdoctoral positions to pursue in the first place. The power of the NSF to determine the distribution of cognitive labor is nevertheless minuscule compared to that of the Superintendent.

3. In other cases, additivity fails less completely; for example, two cures for the same disease may, if the cases where they do some good overlap but not entirely, provide more aggregate social good than either cure in isolation, but not so much as the sum of the goods provided by each in isolation.

4. Some connections to the recent literature: Kitcher (1990), without discussing the priority rule explicitly, argues that self-interested scientists may distribute themselves among research programs in a way that is more advantageous to society as a whole than the distribution achieved by high-minded scientists. Dasgupta and David (1994) propose other functions for the priority rule to play in science, such as the provision of motivation to publish as soon as possible. For some reservations about economic approaches to thinking about the social structure of science in general, see Hands (1997).

5. Even if the braking effect damps the bias, herding happens to a certain degree, since the braking cannot undo the fact that a greater proportion of scientists join the predictive program than is warranted by its intrinsic merits.