Her er en semesteroppgave jeg skrev i faget Core Concepts of Political Economy i fjor, om public choice-teori og velgeratferd. For spesielt interesserte (hvem nå det er).
To what extent can voters be considered as ‘consumers’ that make rational choices between alternative policy packages offered by parties?
Public choice theory, drawing on micro-economic assumptions, view voters much like political consumers: rational actors that shop politics as an act of personal value-maximization. In this essay I will present some of the assumptions implicit in public choice theory and discuss their implications to studies of voter behaviour. To challenge the theoretical outset I will argue that if man is in fact by nature a political animal, then studying the political life of homo economicus is bound to have its' limits.
Public choice theory builds on the conventional micro-economic concept of rationality. This implies that actors (1) have a given set of preferences, and (2) pursue actions perceived to maximize the personal benefits ascribed in these preferences (Shepsle and Bonchek 1997: 16-18). At a glance public choice theory seems straightforward and intuitive – an actor knows what she wants and acts to attain these goals.
Applied to an election in a two-party system, this means that a citizen will vote for the party in power only if her perceived benefit of being governed by the adversary party is less than her benefits gained during the reign of the governing party (Downs 1957: 38-40 refers to this as the voter's current party differential). In other words, she votes to keep the incumbent party in power if she believes the adversary party will provide her less benefit.
Similarly, if voters are viewed as consumers, the two parties should be expected to compete for voters in much the same way as two gasoline stations compete for customers, and set up shop side by side. Black's median voter theorem predicts that both parties will try to establish themselves in the middle of the political spectrum. Of course, each voter is still expected to vote for the party that is situated most closely to her political preferences – her ideal point (Shepsle and Bonchek 1997: 104-115).
These examples show the intuitive and straightforward appearance of public choice theory. The assumption of rationality, however, rests on some vital simplifications that need to be addressed. As Downs notes, the purely rational voter inhabits a "world of complete and costless information – the same world in which dwell the rational consumer and the rational producer of traditional economic theory" (Downs 1957: 45). Although he discusses the difficulties arising from these assumptions, Downs concludes that theories should be judged not on the improbability of their assumptions but on the accuracy of their predictions (Downs 1957, quoted in Dunleavy 1991: 104).
Although Downs may be right in this respect, a closer inspection of the assumptions implicit in public choice theory will uncover some important practical consequences. First of all, in the real world political information is neither costless nor complete. Although Downs does adapt his theory to include imperfect information such as errors, false information and ignorance, he also excludes deliberate lies from the model altogether (Downs 1957: 46). In the real world of politics scandals and hypocrisy arguably play important roles, and a blatant exclusion of lies from political science seems ill-placed.
The cost of information, furthermore, has consequences highlighted by the paradox of voter participation: faced with virtually no chance of affecting the election outcome, why should the rational voter even bother to vote? As only the single tie-breaking vote will actually decide the outcome of an election, no voter should be expected to be willing to bear the costs of gathering and sorting information on politics, yet alone walk to the election office (Dunleavy 1991: 80).
Many attempts have been made to overcome this problem; perhaps voters derive satisfaction from the act of voting, or want to show off their patriotic motivations, or perhaps they vote in order to build a reputation for themselves as civil-minded and trustworthy. Downs explains voter participation by stating that if nobody voted, the party in power would have incentives to "shirk on the provision of political services, and everybody would be worse off" (cited in Laver 1997: 93). This explanation hardly seems convincing, and goes against the point forcefully demonstrated by basic game theory of how value-maximizing actors will encounter problems of collective action and produce a sub-optimal, yet fully rational, outcome (Laver 1997: 94-97).
Furthermore, the most simplified public choice models assume unidimensional party competition; parties are generally thought to be situated along a single left-right socioeconomic dimension (Laver 1997: 90). Crucially, when applying the median voter theorem to a given electorate our predictions of each party's stance are shaped by whatever distribution curves we ascribe to the ideal points of the voters. The two parties wooing an electorate with an assumed normal distribution along the left-right dimension will differ significantly from those of a system with a right-leaning, left-leaning or multi-peaked distribution curve (Dunleavy 1991: 90-8).
To be fair, multidimensional models exist and may present a more nuanced view of the breadth of the political debate beyond the simple left-right dimension. However, the shape of the distribution curves within each of these added dimensions will still have crucial consequences for our predicted placement of the median voter. According to McKelvey's Chaos Theorem the probability for normal distributions along several dimensions is diminishingly low, and under these conditions conventional public choice analysis will probably not be fitted to make predictions (Shepsle and Bonchek 1997: 97-101).
Finally, public choice theory treats voter preferences as exogenously given: that "every citizen has a fixed conception of the good society and has already related it to his knowledge of party politics in a consistent manner" (Downs 1957: 47). If, however, political campaigning is viewed as something more than a mere presentation of each party's views – a struggle to shape the voters' perceptions and preferences, perhaps – ascribing fixed interests to voters will fit poorly with reality. Critics of public choice theory have asserted that electoral choice is in fact a complex product of voters' preferences, empirical analysis and moral judgements, constructed within the context of a campaign (Dunleavy 1991: 98). In short, voters' preferences can be viewed as social constructs.
This has spurred great debate, and some defenders of public choice theory have claimed that abandoning the assumption of exogenously given preferences would lead to "a highly elitist interpretation, contrary to normal justifications of representative democracy, in which parties shape elector's preferences" (Budge and Farlie 1977, quoted in Dunleavy 1991: 100). This critique hardly seems convincing, first of all because party politicians should not be viewed as the only actor or force affecting voters' preferences. The press, other voters, external events and other forces will probably have some degree of influence on each voter's perceived preferences. Indeed, if choice and options are constructed within a social context, no actor can possibly control the political discourse singlehandedly.
More formally, changes in voters' preferences can be described by different mechanisms: a spatial shift of the voter's ideal point, a change in the weighting of relevant political issues, or a shift in what the voter regards as politically feasible (Dunleavy 1991: 108-110). Changes in individual voter's preferences will produce a certain degree of change in the distribution curve of the total electorate.
Thus, empirical analyses to test the median voter theorem, such as Robertson's (1976, presented in Dunleavy 1991) study of British party competition during the 1924-1966 period, should be critically assessed. The findings of this study, e.g. that Labour moved towards the ideological centre to compete for the median voter in the 1945 and 1950 elections, are only feasible if the distribution curve of voter preferences did not change during the time span of the research (Dunleavy 1991: 105-107). In other words, Robertson's model does not tell us whether Labour moved towards the ideological centre, or the median voter moved to the left. The conclusion rests on the slightly dubious assumption that the views of the British electorate stayed basically unchanged throughout a forty year period.
In this essay I have outlined the practical consequences that follow from viewing voters as political consumers. I have argued that the assumptions implicit in public choice theory regarding the quality and availability of information, the existence of several political dimensions, the distribution of voters' ideal points, and the possibility of voter preferences changing during the choice process, are bound to limit which conclusions we can possibly draw. Although public choice theory can provide useful explanations and predictions of voter behaviour under certain conditions, the underlying theoretical assumptions should lead us to question the conclusions drawn.
Downs, Anthony. 1957. An Economic Theory of Democracy. New York, Harper & Row.
Dunleavy, Patrick. 1991. Democracy, Bureaucracy and Public Choice. London, Prentice Hall.
Laver, Michael. 1997. Private Desires, Political Action. London, Sage Publications.
Shepsle, Kenneth A. and Mark S. Bonchek. 1997. Analyzing Politics. New York, W.W. Norton & Co.