Order this book
Vrije Universiteit Amsterdam
The notion of bounded rationality is unpacked into three programmatic claims. First, an individual's capacity to make the calculations necessary for forward-looking rationality is severely limited. Second, these constraints significantly affect decision-making. Third, the impact of these limitations depends on the difficulty of the task. So, the point is not that humans are irrational, but that constraints will increase when puzzles become more difficult. One way people reduce a complex problem to a cognitively manageable one is satisficing: they choose options which might not be optimal, but satisfy them enough. With only two options, this heuristic boils down to stick to an option when it delivered satisfactory solutions, and take up the alternative when the aspiration level was not reached. This strategy of "win-stay, lose-shift", labelled "Pavlov" by Nowak and Sigmund (1993), can outperform tit-for-tat in the repeated Prisoner's Dilemma game.
The extensive exploration of the impact of adaptive decision making of the electorate is the main innovative strength of this book. In this respect, it is fundamentally distinct from the work of Michael Laver (2005) (to which, surprisingly, is not referred to), who focuses on adaptation of parties in multi-party settings. Thus, this book provides an excellent complement to the forthcoming book by Laver and Sergenti (2011).
The authors address three components of elections: voter turnout, voter choice, and party competition. The consequences of the assumptions of the behavioural theory for each element are explored with a formal analytical approach in chapters 3, 4 and 5. Most readers of JASSS will probably find the integrated model presented in chapter 6, where computer simulations are introduced, the most challenging part of the book. We clearly deal with intractable interactions here: simultaneously, citizens decide whether to go to the polls or stay home, those who turn out choose which party to support, and parties compete by offering alternative platforms. Links to the models are made available by the authors at http://press.princeton.edu/titles/9352.html.
The simulations provide many interesting insights, of which I only select a few. First, they tackle the question why voters make costs to go voting in the first place. The solution to the voting paradox - an anomaly that has 'spectacularly falsified' (p.2) rational choice theory - is astonishing: breakout and sustenance of participation is caused by loser-driven mobilization. Suppose that initially only a handful of people bother to vote. The few who supported the winning party reap the benefits from the implementation of its policy platform, and are reinforced to do the same (vote again), but all free-riders that benefit from these policies as well (stay home). However, 'defecting' losers will often be unsatisfied with staying home and inclined to switch to participation. When supporters of the alternative party are increasingly mobilized due to losses in previous elections, and the challenger succeeds to win, then a similar counter-mobilization for the previous incumbent is generated.
The subsequent question is of course: when citizens turnout, which party is chosen? Interestingly, the authors do not assume that voters know the party platforms and choose the ideologically closest party. Instead, they use retrospective voting theory. Citizens do not directly compare the incumbent to the challenging party, but rely on their assessment of the incumbent's performance: has it delivered the policies (payoffs) that meet their aspiration levels? The system exhibits a restless dynamic and never settles down. This seems mainly due to two factors: random payoff disturbances and endogenously evolving voter aspirations. Thus, citizens learn 'what to hope for' and can acquire unrealistically high demands after prosperous periods. As a result, they can even mistakenly throw out of office an incumbent when it has implemented the median voter's ideal point. Another interesting outcome is that it may take some time for voters to switch their support from the incumbent to the challenger, even if the latter's alternative policy would make them better off. It yields a coordination problem: simultaneously, the challenger may unwittingly move away from an electorally superior platform when it does not succeed to beat the incumbent.
Rational choice scholars have difficulties with explaining non-divergence of platforms. In contrast, this model predicts that the winning policy does not become locked onto the central location. The more distinctiveness, the more accurate the electorate, i.e., the more people correctly vote for the party whose platform is closest to their ideals. Bigger spatial differences make it easier for adaptive voters to learn over time. Not surprisingly, voters are more often correct when policy platforms are fixed, instead of 'moving targets' that constantly change. Concerning the last component of elections, party behaviour, some readers may be disappointed by the fact that models only cover two-party systems (chapter 6) and three-party competition in plurality-rule (single-winner) voting systems (chapter 7). However, the advantage of this model, build up in an elegant, stepwise manner throughout the book, is that it is parsimonious and relatively easy to grasp. Extensions to include, for instance, proportional representation would make additional assumptions about coalition formations necessary (in order to define the government's payoff) and render interpretations much more difficult. Secondly, remember, that this is pioneering work. In fact, much remains to be explored.
As the authors admit themselves, a way to extent their work is the confrontation of simulations with real data. There is a rich, empirically inspired tradition of party positioning research (e.g., Budge 1994; Adams et al. 2005), but there is hardly a trace of these contributions. In contrast to studies of phenomena where data-gathering is time-consuming, costly, or virtually impossible, a natural question is to what extent predictions resemble patterns in for instance the US historical record of Republican and Democratic position-taking. Concerning the voter side, I did not fully understand the authors' claim that "our predictions are best understood as applying to populations or samples of elections, not individual cases" (p. 194). It seems an invitation to leave room for the notorious ecological fallacy. For example, the most convincing test of loser-driven mobilization explanations would be based on individual-level data, not on macro-level turnout figures.
In sum, this book offers plenty food for thought for both theoretical and empirical minded scholars and is a must read for anyone interested in understanding the dynamics of electoral competition.
BUDGE, I (1994) A New Spatial Theory of Party Competition. Uncertainty, Ideology and Policy Equilibria Viewed Comparatively and Temporally. British Journal of Political Science 24, pp. 443-467
LAVER, M (2005) Policy and the Dynamics of Political Competition. American Political Science Review, 99, pp. 263-281
LAVER, M and Sergenti, E (2011) Party Competition. An Agent-Based Model. Princeton, New Jersey: Princeton University Press
NOWAK, M. and Sigmund, K. (1993) A Stragegy of Win-Stay, Lose-Shift that Outperforms Tit-for-Tat in the Prisoner's Dilemma Game. Nature, 346, pp. 56-58
Return to Contents of this issue
© Copyright JASSS, 2011