Economists as experts: Overconfidence in theory and practice

Abstract

Drawing on research in the psychology of judgment and decision making, I argue that individual economists acting as experts in matters of public policy are likely to be victims of significant overconfidence. The case is based on the pervasiveness of the phenomenon, the nature of the task facing economists‐as‐experts, and the character of the institutional constraints under which they operate. Moreover, I argue that economist overconfidence can have dramatic consequences. Finally, I explore how the negative consequences of overconfidence can be mitigated, and how the phenomenon can be reduced or eliminated. As a case study, I discuss the involvement of Western experts in post‐communist Russian economic reforms.

Keywords:

• overconfidence
• calibration
• economists
• experts
• public policy
• Russian reforms

ACKNOWLEDGEMENTS

Earlier versions of this paper were presented at the 2001 St Louis Philosophy of Social Science Roundtable, and at a session organized by the International Network for Economic Method at the 2002 Meeting of the Allied Social Science Associations. I am grateful to the participants at those meetings, especially to Ross Emmett, and to Wändi Bruine de Bruin, Baruch Fischhoff, Balázs Gyenis, Kevin Hoover, Rikard Levin, Thomas Mayer, Gualtiero Piccinini, Daniel Steel, and anonymous referees for thoughtful comments. This research was supported by a fellowship from Jan Wallander and Tom Hedelius' Foundation, for which I am most grateful. Errors remain my own.

Notes

1. Although I use the case of Åslund as an illustration, my case does not hinge on his being overconfident. The main argument in favor of my thesis is based on evidence from the psychology of judgment and decision making, along with some general facts e.g. about the nature of the judgment task facing the economist‐as‐expert.

2. In some contexts, the term ‘expert’ is defined so as to imply that the person in question is calibrated. Under this definition, my title should read ‘Economists as purported experts’. However, I will use ‘P is an expert in domain D’ to mean ‘P professes to be exceptionally knowledgeable in domain D’. Thus, I allow for the possibility that an expert may not be very knowledgeable at all.

3. For useful overviews, see Lichtenstein et al. (1982), Yates (1990: ch. 4), Griffin and Tversky (1992) and Plous (1993: ch. 19).

4. In many of these studies, unlike the ones discussed above, confidence is assessed by asking subjects how their performance compares to that of other people. As Don A. Moore and Daylian M. Cain (2005) have shown, the relative nature of judgments in such tasks leads to additional complications that sometimes lead subjects to exhibit underconfidence. Note that this research does not undermine my thesis, since the judgments in which I am interested do not typically have this relative nature.

5. See Gigerenzer et al. (1991), Dawes and Mulford (1996), Griffin and Tversky (1992), and Brenner et al. (1996) for a fuller discussion of the criticism of overconfidence research.

6. All translations are my own.

7. Hence, my argument does not presuppose that economists deliberately exaggerate their degree of confidence. While they may do so, the research reported in the previous section indicates that overconfidence stems in large part from cognitive factors, which affect also the most honest, forthright and well intentioned.

8. Unless otherwise noted, all statistics come from the World Development Indicators Online (World Bank 2004).

9. Åslund et al. (1996) do attempt to define more clearly what they mean by ‘radical economic reform’. In order to attenuate the effects of the confirmation and hindsight biases, however, the definition needs to be provided at the time of the prediction.

10. There is no reason to think, as far as I can tell, to think that these experts deliberately exaggerated their confidence.

11. It is unclear on what basis Åslund postulates ‘existential shock’ as the specific cause of Russians' early deaths.

12. See chapter 13 of Luce and Raiffa's Games and Decisions (1957) for a classic discussion of principles of decision making under uncertainty.

13. I thank the editor and referees of this Journal for pushing this point.