Grumpy Old Men: reconsidering Rational Choice & Political Power

ABSTRACT

A new, expanded edition of Keith Dowding’s Rational Choice & Political Power (PC&PP) provides an opportunity to examine the uses of game theoretic models in political inquiry. Specifically, it raises questions about the apparent discontinuities between the way bargaining models are used in RC&PP and the way he defends formal models in more recent work. I prefer what I take to be his earlier view.

KEYWORDS:

• Explanation
• formal models
• game theory
• interpretation
• rationality

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. All parenthetical citations in the text and notes containing nothing but a page number refer to Dowding (2019).

2. Variations on the basic Nash bargaining solution (e.g. Kalai-Smorodinsky) differ from it largely insofar as they modify one or another of these axioms.

3. Rubinstein (1982) is arguably the founding contribution in non-cooperative bargaining theory. Binmore and Dasgupta (1987) collect it and other seminal early papers. Sutton (1986) gives an early, reasonably accessible overview. And Muthoo (2000) offers a more recent even more accessible introduction.

4. Please note the word ‘can’ here. Dowding (e.g. p. 181–3) arguably is at his best in differentiating resources, power and luck so I do not mean to imply that we can or should reduce power to resources.

5. Here one must clearly ascribe independent causal force to institutions and so depart from a too narrow commitment to methodological individualism. This implies not a structural or holist view but something like what Daniel Little (2016) terms ‘methodological localism.’ On this view we can ascribe causal efficacy to institutions (among other things) but still require that it operates by influencing the abilities of individual agents, the information available to them, the incentives they confront, and so on. This view seems to me to be broadly consistent with Dowding’s (p. 11–15) views on this matter.

6. Hausman (2000) offers a useful reminder that the refinements central to non-cooperative game theory make essential reference to the beliefs and preferences of the agents who populate its models.

7. Although it is not my focus here it is important to note that institutions are not directly observable. After all: ‘We cannot see, feel, touch, or even measure institutions’ (North 1990, p. 107). We can, however, model them.

8. See also Dowding and Miller (2019).

9. Dowding_old states: ‘a formal model is a set of statements related formally or analytically to generate testable hypotheses or predictions’ (2016, p. 88).

10. ‘A token is a specific example of a general class. A type is a general class which is made up of many token examples. Any given token may belong to many different classes, and any type may have many different token examples’ (p. 12).

11. Rubinstein repeatedly suggests that when we make game theoretic models we are constructing fables (e.g. Rubinstein 2012).

12. ‘In particular, I do not think that the bargaining models that I myself have studied have significant predictive value’ (Rubinstein 2012, p. 36).

13. This is the opening salvo in a contentious argument, offered as a prestigious address to the meetings of the Econometric Society. Importantly, however, Rubinstein does not reserve his views on the status of game theory for the initiated. He is entirely willing to puncture common interpretations of the enterprise in much more ‘lay’ settings (Rubinstein 2013).

14. ‘There is no need to go into the full complexity of Harsanyi’s model, for he is interested in providing an equation for the amount of power each individual has over any social outcome. For most individuals that power is miniscule. Further, as stated above, we cannot hope to place actually quantifiable values on the variables in Harsanyi’s model. Rather, the model’s utility comes from the clear thinking it engenders by its conceptualization of the elements that must be taken into consideration when studying the power of groups in society’ (p. 79).

15. I understand ‘rationality’ here in a basically Davidsonian way as consisting of concatenations of belief/expectation and preference (Elster 2007). It is something predicated to or withheld from action. Dowding_young (p. 24–5) endorses this conception of rationality.

16. ‘Measurement is not easy, but it is the stuff of social science’ (p. 200).

17. On all this see Johnson (2019).

18. Davidson was writing, in part, in response to an important and highly contentious set of debates in the philosophy of social science regarding the concept of rationality. These debates were prompted in large measure by studies in social and cultural anthropology. Wilson (1970) and Hollis and Lukes (1982) collect many of the relevant papers.

19. Lukes (1982, 2000) provides insightful guidance in this regard.

20. In this regard ‘behavioralism’ in political science resembles certain interpretations of the doctrine of revealed preference amongst economists. On those interpretations focusing on choice does not reveal the underlying preferences or beliefs that animate the action of agents who populate our models but purportedly allows us to dispense altogether with such occult entities (Hausman 2000).

21. This is perhaps a bit strong. Some suggest that rational choice models can jettison this sort of ‘folk psychology’ (Satz and Ferejohn 1994). And rational choice theorists can indeed depart from the canonical view. The price for doing so, however, is extremely high – namely, abandoning the Nash program in which various equilibrium refinements trade on the idea that in formulating her strategy a player must ascertain the credibility or otherwise of threats and offers other parties might issue. She must, in other words, have beliefs about what others say (Hausman 2000).

22. Here I follow Cartwright (1999, 2010) who not only depicts formal models as fables but, in so doing, rightly insists that we use them not to represent the world but to interpret concepts. See Johnson (2014, 2019, 2020).

Additional information

Notes on contributors

James Johnson

James Johnson teaches social and political theory at the University of Rochester and is a long-time instructor at the ICPSR Summer Program. His research traverses pragmatist political thought, philosophy of social science, democratic theory, and political economy.