Theoretical models of voting behaviour

De Baripedia

It is no longer a question of explaining "why" people participate but "how", that is, in terms of voter turnout, what choice is made and what can explain an electoral choice. The theories that are supposed to explain the electoral choice also explain at the same time the electoral participation in particular with the sociological model.

Voting Explanatory Models[modifier | modifier le wikicode]

The distinction between the three main explanatory models of voting is often found. A distinction is made between the sociological model of voting from the Columbia School, which refers to the university where this model was developed. The reference work is The People’s Choice published in 1948 by Lazarsfeld, Berelson and Gaudet. The explanatory factors and aspects highlighted by these different models are always taken into account.[8][9]

The second very important model is the psycho-sociological model, also known as the partisan identification model or Michigan School model, developed by Campbell, Converse, Miller and Stokes in Campbell, Converse, Miller and Stokes, among others in The American Voter published in 1960.[10]

The third model is called the economic model of the vote or the Rochester School of Economics, developed by Downs in the book An Economic Theory of Democracy published in 1957.[11]

This model has given rise to the spatial theories of voting which are the dominant theories. Sometimes, indeed often, people combine the first two models incorporating the psycho-sociological model on the basis that the Michigan model is just an extension of the Columbia model that helps explain some things that the Columbia model cannot explain. Some people talk about membership voting for the first two theories and cognitive voting for the economic model of voting. We can talk about two major theories or two major models or even three models.

Sociological model[modifier | modifier le wikicode]

Proposals[modifier | modifier le wikicode]

This model emphasizes the role of integration into social groups. The term "group" can mean different things, which can be an ethnic group or a social class. Thus, the interpretation of differences in voting behaviour from one group to another is to be sought in the position of the group in society and in the way its relations with parties have developed.

For Lazarsfeld, "a person thinks politically as he or she is socially". In other words, social, spatial or group membership largely determines individual political actions. There is a direct link between social position and voting. Furthermore, "social characteristics determine political preferences". The political consciousness of individuals is based on social experiences and has little weight outside these experiences. In this model, importance is given to primary socialization.

Political Predisposition Index[modifier | modifier le wikicode]

As this is the first model that wanted to study empirically and test hypotheses on the basis of survey data, it was necessary to develop conceptual tools, in particular the political predisposition index, which focuses on three types of social affiliations that are fundamental in this perspective to explain electoral choices, namely social status, religion and place of residence. In other words, there is a social type variable, a cultural type variable and a spatial type variable. Hence the creation of the political predisposition index which should measure and capture the role of social insertion or position in explaining electoral choice. Today, when we see regression analyses of electoral choice, we will always find among the control variables social status variables, a religion variable and a variable related to place of residence.

Sociological model[modifier | modifier le wikicode]

It is possible to attribute some merits and some criticisms to this model at least in its initial formulation.

One of the merits, which can be found in Lazarsfeld's book entitled The People's Choice published in 1944 is that this model marks a turning point in the study of political behaviour. Linked to this, it is important to look at individual data empirically as well. Lazarsfeld was the first to study voting behaviour empirically with survey data, based on individual data, thus differentiating himself from early studies at the aggregate level of electoral geography. The sociological model at the theoretical level emphasizes something important that rationalist and economic theories have largely overlooked, namely, the importance of the role of social context, i.e., voters are all in social contexts and therefore not only family context but also a whole host of other social contexts. The sociological model is somewhat the model that wants to emphasize this aspect. Lazarsfeld's book created this research paradigm.

The sociological model obviously has a number of limitations like any voting model or any set of social science theories. These are some of the criticisms and limitations often made by proponents of other approaches. It has often been emphasized that this model and approach raises more questions than answers. It is a rather descriptive model, at least in its early stages. Its weak explanatory power has been criticized, and these are much more recent criticisms in the sense that we saw when we talked about class voting in particular, which from then on saw the emergence of a whole series of critics who said that all these variables of social position and anchoring in social contexts may have been explanatory of participation and voting at the time these theories emerged in the 1950s, but this may be much less true today in a phase or period of political misalignment. It is the idea of when does one or the other of these different theories provide a better explanation according to periods of political alignment or misalignment. The initial formation of this model was very deterministic in wanting to focus on the role of social inclusion while neglecting other aspects, even though today there is increasingly a kind of ecumenical attempt to have an explanation that takes into account different aspects. It is interesting to know that Lazarsfeld, when he began his studies with survey data, especially in an electoral district in New York State, was looking for something other than the role of social factors. He wanted to see the role of the media in particular and also the role of opinion leaders and therefore, the influences that certain people can have in the electoral choice. Lazarsfeld was interested in this and simply, empirically, he found that these other factors had less explanatory weight than the factors related to political predisposition and therefore to this social inking. He wanted to look for one thing and found something else. By finding something else, he shaped a dominant theory explaining the vote. This approach has often been criticized as a static approach since socio-economic or even socio-demographic characteristics do not change in the short term and yet the vote increasingly changes in the short term, what is called in electoral volatility, i.e. changes in voting behaviour from one election to the next. This electoral volatility, especially in a period of political misalignment, is becoming more and more important and is increasingly overshadowed by this type of explanation. Regarding the causal ambiguity, there are also critics who say that this approach is very strongly correlational in the sense that it looks for correlations between certain social variables and electoral choices, but the approach does not explain why this variable approach really has a role and therefore what are the causal mechanisms that lead from insertion, positions, social predispositions to electoral choice. The psycho-sociological model is intended as a development that wants to respond to this criticism.

Psycho-sociological model[modifier | modifier le wikicode]

Proposals[modifier | modifier le wikicode]

Often, in the literature, the sociological and psycho-sociological model fall into the same category, with a kind of binary distinction between the theories that emphasize social, belonging and identification on the one hand, and then the rationalist and economic theories of the vote, which are the economic theories of the vote that focus instead on the role of political issues, choices and cost-benefit calculations.

The psycho-sociological model has its roots in Campell's work entitled The American Voter publié en 1960. This approach emphasizes a central variable which is that of partisan identification, which is a particular political attitude towards a party. Often, in Anglo-Saxon literature, this model is referred to as the party identification model. From that point on, there has been the development of a whole body of literature on political psychology. We have to be careful, because when we talk about political psychology, we include that, but we also include the role of cognitions and rationality.

What is partisan identification? It can be defined as lasting feelings of attachment that individuals develop towards a certain party. The concept and this theory was developed in the United States by political scientists and sociologists and initially applied to the American political system with an attachment to the Democratic Party rather than the Republican Party.

This identification is seen as contributing to an individual's self-image. The image that an individual has of himself in this perspective is also the result of this identification. In other words, this identification is part of the self-image one can have of oneself. This identification with a party is inherited from the family emphasizing the role of primary socialization, it is reinforced over time including a reinforcement that is given by the very fact of voting for that party. This creates a concern for circularity of reasoning. Reinforcement over time since adult voters increasingly rely on this partisan identification to vote and to face the problems of information, namely partisan identification seen as a way of solving a problem that all voters have, which is how to form an idea and deal with the abundance and complexity of the information that comes to us from, for example, the media, political campaigns or others in relation to the political offer. In this approach, it is possible to say that the voter accepts the arguments of a certain party because he or she feels close to a party and not the opposite which would be what the economic model of the vote postulates, that is to say that we listen to what the party has to say and we will choose that party because we are convinced by what that party says. There is an opposite reasoning. The psychological and socio-economic model are strongly opposed, offering two explanations that are difficult to reconcile, even though there have been efforts to try to combine them. In this perspective, voting is essentially a question of attachment, identity and loyalty to a party, whereas in the rationalist approach it is mainly a question of interest, cognition and rational reading of one's own needs and the adequacy of different political offers to one's needs.

Most voters have a sense of allegiance to a party that is inherited through the family. We see the kinship of this model with the sociological model explaining that often they are put together. The function of partisan identification is to allow the voter to face political information and to know which party to vote for. Partisan identification becomes stronger over time. There may be a vote that is different from partisan identification, but in the medium to long term, partisan identification should strengthen. Voters who vote against the party with which they identify keep their partisan identification. Voters will vote for a party but that party is not necessarily the one with which they identify. In this approach, these voters keep their partisan identification and again in the medium or long term, they will go back on the electoral choice that is identified with the partisan identification, also called the homing tendency, which is a tendency to go back on the party with which one identifies. At the aggregate level, the distribution of partisan identification in the electorate makes it possible to calculate the normal vote. On the basis of this, we can know. In other words, if we know the partisan identification of voters, we can make a prediction about what the normal vote will be, which is a vote that is not or should not be influenced by other situational factors in a given electorate.

The diagram of the Michigan model[modifier | modifier le wikicode]

The psycho-sociological model, also known as the Michigan model, can be represented graphically or schematically. Here we see the key factors, namely electoral choice and, at the centre, the identification variable for a party, which depends on two types of factors, namely primary socialization and group membership. Inking and the role of socialization cause individuals to form a certain partisan identification that produces certain types of political attitudes. In short, it is an explanatory model that emphasizes the role of political attitudes.

Harrop, M. et W. L. Miller (1987). Elections and Voters. New York: New Amsterdam Books.[12]

There is a small degree of complexity because one can distinguish between attitudes towards the candidate or the party, attitudes towards the policies implemented by the different parties and attitudes about the benefits that one's own group may receive from voting for one party rather than another.

What we see here in relation to the sociological model and that these variables highlighted by the sociological model such as socialization, inking or social position play a role but only indirectly. The Lazarsfeld model would link membership and voting. It is in this sense that the party identification model provides an answer to this criticism that the sociological model does not highlight the mechanisms that make a certain social inking influence a certain electoral choice. The psycho-sociological model says that it is because this inking allows identification with a party which in turn influences political attitudes and therefore predispositions with regard to a given object, with regard to the candidate or the party, and this is what ultimately influences the vote. The psycho-sociological model can be seen in the light of an explanatory contribution to the idea that social inking is a determining factor in explaining the vote, or at least on a theoretical level.

Causal funnel[modifier | modifier le wikicode]

Another model is called the funnel model of causality which has been proposed by these authors working on the psycho-sociological model. The idea is to see what are all the factors that explain the electoral choice. This model shows that there is more than political identities, partisan identification and social inking. On the other hand, to explain the electoral choice, we must take into account factors that are very far from the vote theoretically, but we must also take into account the fact that there are factors that are no longer close to the electoral choice during a vote or an election. We must also, and above all, look at the links between types of factors. There are certain types of factors that influence other types of factors and that in turn influence other types of factors and that ultimately help explain the idea of the causal funnel of electoral choice.

Dalton, R. J. (1988). Citizen Politics in Western Democracies. Chatham, NJ: Chatham House.[13]

In this representation, there are factors related to the cleavages, but also other factors that relate to the economic, political or social structure of a country being factors that are far removed from the electoral choice but that still exert an important effect in an indirect way the effect they have on other variables afterwards. There are also intermediate variables that relate to loyalties to a certain group or sense of belonging. Value orientations refer to materialism as well as post-materialism, among other things, cleavages but no longer from a value perspective.

Partisan attachment is at the centre of the graph influencing opinions on certain issues being discussed or the attitudes of certain candidates. There are also external factors that also need to be considered, such as the actions of the government, for example, voters are influenced by what the government has done. There is a whole branch of the electoral literature that emphasizes government action as an essential factor in explaining the vote, and there is a contrast between a prospective vote, which is voting according to what the parties say they will do during the election campaign, and a retrospective vote, which is voting in relation to what has been done, particularly by the government, which has attributed the successes or failures of a policy. An important factor is the role of political campaigns in influencing the vote. Political conditions as well as the influence of the media play an important role, all the more so nowadays as more and more political campaigns and the role of the media overlap. The influence of friends refers to opinion leaders and circles of friends. The external factors would be the factors that, in the basic theory of the psycho-sociological approach, it would seem that this is what can do but if we have a certain partisan attachment to vote for another party because we are influenced by one or other of these factors but, basically, we keep our partisan attachment and the next time when these factors change, we return to the normal vote corresponding to the partisan attachment. All of these factors and their relationships have to be taken into account, but at the centre is always the partisan attachment.

Partisan Identification Index[modifier | modifier le wikicode]

How was that measured? For the sociological model we have talked about the index of political predisposition with the variables of socioeconomic, religious and spatial status. The psycho-sociological model also developed a measure called the partisan identification index, since this model wanted to be an empirical model with behaviourism and the idea of studying individual behaviours empirically with the development of national election studies and survey data to try to measure the partisan identification index.

The original measurement was very simple being based on two questions which are a scale with a question about leadership. The system in the United States is bipartisan and the question asked was "Do you consider yourself a Republican, Democrat or otherwise? ». Then a second question was supposed to measure the strength of that identification with the question "do you consider yourself a Republican, strong, weak or leaning towards the Democratic Party? ».

Four questions around partisan identification[modifier | modifier le wikicode]

Four questions can be asked in relation to this measure:

  • is partisan identification one-dimensional?
  • how does partisan identification develop?
  • does partisan identification work outside the United States?
  • has partisan identification weakened?

Is partisan identification one-dimensional?[modifier | modifier le wikicode]

For the first question, there are several studies on the fact that partisan identification is multi-dimensional and not just one-dimensional. It is multidimensional also in the bipartisan context of the United States because there are cleavages that cut across parties. There are other cleavages that cut across Republicans and Democrats that should be taken into account to explain the pattern. For example, a strongly conservative voter who votes Democratic may vote Republican because he or she feels more in tune with the party.

How does partisan identification develop?[modifier | modifier le wikicode]

With regard to the question of how partisan identification develops, the psycho-sociological model emphasizes the role of the family and thus of primary socialization, but several critics have shown that secondary socialization also plays a role. We must also take into account other socializing agents that can socialize us and make us develop a form of partisan identification. Several studies have shown that the very fact of voting for a party contributes to the development of a certain identification for that party. The cause-and-effect relationship is reversed, according to some who argue that this is a problem at the empirical level when we want to study the effect of partisan identification on electoral choice because there is a problem of endogeneity; we no longer know what explains what.

Does partisan identification work outside the United States?[modifier | modifier le wikicode]

Several studies show that the impact of partisan identification varies greatly from one context to another. While in the United States, several studies have shown that partisan identification is an important explanatory power on electoral choice, in other contexts this is less true. Studies have shown that, for example, outside the United States, a much larger proportion of voters who change their vote also change their partisan identification. In other words, the homing tendency that is the explanation that the model postulates is much less true outside the United States. We need to find identification measures adapted to the European context, which the researchers have done.

Has the partisan identification weakened?[modifier | modifier le wikicode]

Dalton, R. J. (1988). Citizen Politics in Western Democracies. Chatham, NJ: Chatham House.[13]

This is related to its variation in space and time. This table shows that for quite some time now there has been a strong decline in partisan identification. There have been several phases of misalignment. This is linked to a decrease in class voting and a loss of traditional cleavages. One can draw a kind of parallel with a loss of importance of the strength of partisan identification and also of the explanatory power of partisan identification. The idea that one identifies oneself, that one has an attitude, an attachment to a party was certainly true some forty years ago and has become less and less true and also the explanatory power of this variable is less important today even if there are significant effects. Even if there is still a significant effect of identification, there are other explanations and aspects to look for, particularly in terms of the issue vote and the assessments that different voters make of the issue vote.

In Personality traits and party identification over time published in 2014 by Bakker, Hopmann and Persson, the authors attempt to explain partisan identification.[14] They try to answer the question of how partisan identification is developing and how partisan identification has weakened because they look at the stability over time of partisan identification. First, they summarize the literature that has been interested in explaining why voters vary or differ in the stability or strength of their partisan identification. The main explanatory factors have been sought in socio-economic status and socio-demographic variables such as "age," "gender," and "education. They try to elaborate a bit and find out empirically how this happens. They find that partisan identification becomes more stable with age, so the older you get, the more partisan identification you have, so it's much easier to change when you're young. On the other hand, women tend to have less stable partisan identification, they change more often too. Finally, some studies show that high levels of education lead to weaker attachments to parties. There are also studies that show that the more educated change less often from one party to another. What is interesting is that they try to relate this to personality traits such as being open, conscientious, extroverted, pleasant and neurotic. These authors find with panel data that among their confirmed hypotheses that extroverted people tend to have a strong and stable partisan identification. They find that conscientious and neurotic people tend not to identify with a political party.

The psycho-sociological model: summary[modifier | modifier le wikicode]

As far as the psycho-sociological model is concerned, it has the merit of challenging the classical theory of democracy which puts the role on the rational actor. A rather subjective and almost sentimental citizen is placed at the centre of the analysis. The individual is subjectivity at the centre of the analysis. It is a model that is very close to data and practice and lends itself very easily to empirical testing through measures of partisan identification and different measures of socio-demographic factors among others. That is why there are many empirical analyses that are based on this model. Today, this may be less true, but until a certain point, there were relatively few empirical analyses based on the economic model of the vote. The psycho-sociological model initiated the national election studies and created a research paradigm that remains one of the two dominant research paradigms today and ultimately contributed to the creation of electoral psychology.

The limitations are the explanation of partisan identification, which is that the model has been criticized because it explains or does not explain too much about where partisan identification comes from except to say that it is the result of primary socialization. These criticisms and limitations are related to the original model. The relationship between partisan identification and voting is that the model postulates that partisan identification is the explanatory variable and that voting for the electoral choice is the explained variable. But there are studies that also show that the causal relationship goes in the other direction. The concept and measurement of partisan identification as conceived by these researchers as applying to the bipartite system and therefore needs to be adapted to fit the multiparty and European system. This model leaves little room for the ideology which is the idea that by putting so much emphasis on the emotional voter and feelings, it leaves little room for the ideology that is central to explaining the economic model of the vote. The role of the centrality of partisan identification has been criticized, especially today, because partisan identification plays a role that is still important but much less important than it used to be and may be much less important than some researchers within this paradigm have postulated.

Certain developments in the theory of the psycho-sociological model have in fact provided answers to these criticisms.

Economic model[modifier | modifier le wikicode]

We are going to talk about the economic model. The initial formulation of the model is based on the Downs theory in An Economic Theory of Democracy publié en 1957.[15] Then we'll look at the space theories of the vote.

In the literature, we often talk about the economic theory of voting. There are two slightly different connotations. We often talk about economic theory of the vote in the broadest sense in order to designate a rationalist theory based on rational choice theory and spatial theories of the vote. Today, in the literature, we talk about the economic vote in a narrower and slightly different sense, namely that the electoral choice is strongly determined by the economic situation and by the policies that the government puts in place in particular to deal with situations of economic difficulty. This is more related to the retrospective vote. Since the economic crisis, there has been an increasing focus on the economic crisis and economic conditions and how that can explain electoral volatility and electoral change.

Axioms[modifier | modifier le wikicode]

The basic assumptions of the economic model of the vote are threefold: selfishness, which is the fact that voters act according to their individual interests and not according to their sense of belonging to a group or their attachment to a party. The second criterion is subjectivity, which is that voters calculate the costs and benefits of voting subjectively, so they make an assessment of the costs and benefits. The third criterion is rationality, which is that based on the theory of rational choice, voters mobilize the limited means at their disposal to achieve their goals, so they will choose the alternative among the political offer that costs them the least and brings them the greatest possible benefit. In other words, there is the idea of utility maximization which is a key concept in rational choice theory, so the voter wants to maximize his utility and his utility is calculated according to the ratio between the cost and the benefit that can be obtained from the action, in this case going to vote (1) and going to vote for that party rather than this one (2).

Rational Voting[modifier | modifier le wikicode]

This theory presupposed that the voter recognizes his or her own interest, assesses alternative candidates, and on the basis of this assessment, will choose for the candidate or party that will be most favourably assessed in the sense of best serving his or her own political interests and interests.

Three elements should be noted. The vote is seen here as an instrument, that is to say, there is the idea of an instrumental vote and not an expressive one. Voting is an instrument that serves us to achieve an objective. On the other hand, the focus is on the political goals of the voters, whereas the psychological model puts a little more emphasis on the social use of the vote. Finally, there is an instrumental approach to information and voting.

There is the important opposition between an economic vote based on a choice, which is the idea that the voter makes a real choice based on a cost-benefit calculation, a choice that is rational in the end according to Weber's typology, while the psycho-sociological vote is rather based on a concept of loyalty that often makes the opposition between choice and loyalty.

Evaluation of costs[modifier | modifier le wikicode]

Voters calculate the cost of voting. There are different types of costs that this model considers and that need to be taken into account and in particular two types of costs which are the costs of going to vote (1) but above all, there are the costs of information (2) which are the costs of obtaining this information since in this model which postulates to choose a party on the basis of an evaluation of the different propositions of information which is available, given these basic postulates, the transparency of information and therefore the costs of information are crucial.

More specifically, the costs that the voter has to take into account according to the different parties and candidates must be evaluated, which is the partisan differential, i.e. the difference in the cost-benefit ratio that different parties give. We must assess the costs of going to the polls, of gathering the information needed to make a decision, but also the value of one's own participation, since the model is also supposed to explain voter turnout. Economic theories of voting explain both voter turnout (1) and electoral choice (2). One must assess the value of one's own participation and also assess the number of other citizens who will vote. This is the idea of collective action, since our own contribution to an election or vote changes with the number of other citizens who vote. This model relies heavily on the ability of voters to assess and calculate their own interests and all the costs associated with the action of going to the polls.

Voters assess the utility income of parties and candidates. Since the idea is to calculate the costs and benefits of voting for one party rather than the other, therefore, each party brings us some utility income. The voters have to make that assessment and then decide which one brings more income and which one we will vote for. On that basis, voters calculate the utility income of the different parties and then they look at and evaluate the partisan differential.

This is central to spatial theories of voting, that is, voters vote or will vote for the candidate or party that is closest to their own positions. This is called the proximity model. There are other models and economic theories of the vote, including directional theories that have a different perspective but remain within the framework of economic theories of the vote. When we talk about the Downs model, we also talk about the proximity model, which is the idea of a rational economic mode based on utility maximization. Simply, the voter is going to evaluate his own interest, his utility income from the different parties and will vote for the party that is closest to his interests.

Strategies to reduce information costs[modifier | modifier le wikicode]

Information is central to spatial theories, whereas in the psycho-sociological model, information is much less important. There are different strategies that are put in place by voters in a conscious or unconscious way to reduce these information costs, which are all the costs associated with the fact that in order to be able to evaluate the utility income given by one party rather than another, one has to go and see, listen, hear and understand what these parties are saying.

There are different strategies that are studied in the literature. One possible strategy to reduce costs is to base oneself on ideology. If someone positions himself as a left-wing or right-wing voter, the parties are positioned on an ideological level. Ideology can also be in relation to another dimension, for example between egalitarian and libertarian ideology. The idea is that there is something easier to evaluate which is the ideology of a party and that it is on the basis of this that the choice will be made. This means that we are not necessarily going to listen to all the specific arguments of the different parties.

Another strategy is the so-called "shortcut" that voters take within the rationalist framework of voting, since they are confronted with the problem of information and have to choose on the basis of this information. Another possible strategy is to rely on the judgment of others such as opinion leaders. For example, there is Lazarsfeld's theory with the idea that opinion leaders can be seen as people to whom we attribute a strong trust and maybe even an esteem in relation to the political judgment they may have and therefore, by discussing with these people, it is possible to form an electoral choice and therefore there is no need to go and pay these costs of gathering information. The idea is that it is in circles of interpersonal relations even if more modern theories of opinion leaders look at actors outside the personal circle. The role of the media and campaigns simplifies information by summarizing it.

It is necessary to distinguish between two types of voters and to make a distinction between a literature that has become increasingly important in recent years on opinion formation in an election or voting context. A distinction is often made between two types of voters and votes between the :

  • systematic voting, i.e. voters who follow a systematic vote are voters who are willing to pay these information or information-related costs. They are voters who make the effort to inform themselves, to look at the proposals of the different parties and try to evaluate the different political offers. There are a whole bunch of individual characteristics related to the fact that one is more of a systematic voter of something else.
  • and voters who choose to use euristic shortcuts to solve the information problem. In other words, they are voters who are not prepared to pay all these costs and therefore want to reduce or improve the cost-benefit ratio which is the basis of this electoral choice by reducing the costs and the benefit will remain unchanged.

There are these two types and a whole literature on the different types of euristics that can be set up. It is a very detailed literature today. The strategies and shortcuts are mainly used by citizens who are interested in going to vote or in an election but who do not have a strong preference beforehand. Voters who rely on strong partisan identification do not need to go and do systematic voting or take one of the shortcuts. If we look at it a little more broadly, partisan identification can be seen as a kind of shortcut.

Four types of voters[modifier | modifier le wikicode]

On this basis, four types of voters can be identified in a simplified manner:

  • those who inquire: they are willing to pay these costs. These are voters who proceed by systematic voting. In general, they are politically more sophisticated and better educated;
  • those who rely on the opinion of the media and opinion leaders;
  • those based on ideological differences;
  • those based on partisan identification.

It is possible to start from the assumption that the characteristics of these different voters are very different. In other words, when we are interested in trying to explain the vote, we must already know what type of voter we are talking about. There is a kind of heterogeneity of voters. There are different types of individuals who take different kinds of shortcuts or not, who vote systematically or not, and so on. The heterogeneity of the electorate and voters must be taken into account.

One important element of this model must be highlighted in relation to the others. This economic theory of the vote, this rationalist theory, has a great advantage over the other models, which is that it does not only focus on voters, that is to say, it does not only focus on political demand, but it also looks at supply and especially at the interaction between supply and demand. Moreover, there are analogies that are made even explicitly with the idea of the market. We talk about the electoral market in the media or the electoral supply. This jargon comes from this type of explanation. We are not ignoring the psychological model, which focuses on the identification people have with parties without looking at the parties. We are looking at the interaction. In the spatial theories of the vote, we see the strategic link between a party's supply and a demand from voters or electors. There is an idea of interdependence between political supply and demand, between parties and voters, which is completely removed from other types of explanations.

Economic model: summary[modifier | modifier le wikicode]

The theory of the economic model of the vote is also a model that allows predictions to be made about party behaviour. There is the idea of the interaction between a political demand and a political offer proposed by the different candidates during an election or a vote. From the point of view of parties and candidates, the economic model and in particular the model that was proposed by Downs in 1957 and which predicts a convergence of a party position towards the centre. This is the median voter theory. The ideological space can be defined as a left-right ideological space but can also be defined more precisely in relation to certain issues. There is in fact the idea that the choices and preferences of voters in the centre will cause the parties, since they are aiming in this model, to try to maximize their electoral support.

In summary, it can be said that in the economic model of voting, the political preferences of voters on different issues, are clearly perceived by the voters themselves which is the idea that the voter must assess his own interest, he must clearly perceive what are the political preferences of voters. On the other hand, the political preferences are exogenous to the political process which is the fact that when the voter goes to vote which is the moment when he or she starts to think about this election, he or she already arrives with certain fixed or prefixed political preferences. In other words, the voters' political preferences on different issues, in other words, in this type of theorizing, they know very well what they want, and what is more, these positions are very fixed and present when the voter is going to have to vote. The economic model makes predictions and tries to explain both the participation but also, and above all, the direction of the vote, which is the electoral choice.

Voters vote for the candidate or party closest to their own position which is the proximity model. Proximity means the closeness of the voter's interests to the political proposals that are made with the parties. It is a paradigm that does not only explain from the macro-political point of view an electoral choice, but there is the other side of the coin which is to explain the choice that the parties make. The strategic choices made by parties can also be explained by this model since, since this model postulates an interdependence between supply and demand, we address the demand but we can also address the supply. We speak of cognitive preference between one's political preferences and the positions of the parties. A distinction must be made between the affective vote of the psycho-sociological model and the cognitive vote of the theories of the economic model.

From this point of view, parties adopt political positions that maximize their electoral support, what Downs calls the median voters and the idea that parties would maximize their electoral support around the center of the political spectrum. What we are interested in is on the demand side, how can we explain voters' electoral choice.

The economic model has put the rational and free citizen back at the centre of attention and reflection, whereas if we push the sociological model a bit to the extreme, it puts in second place this freedom and this free will that voters can make since the psycho-sociological model tells us that voting is determined by social position, it is not really an electoral choice that we make in the end but it is simply the result of our social insertion or our attachment to a party. The economic model of the vote puts the notion of electoral choice back at the centre. It is a theory that is made in the interaction between supply and demand, that is, between parties offering something and voters asking for something. It is a theory that makes it possible to explain both the voting behaviour of voters and the organisational behaviour of political parties. The theory of partisan competition was completely eliminated by the other types of explanations. For some, these are theories that offer reflections on the proper functioning of democracy, on presuppositions, the role of information or the role of citizens for the proper functioning of democracy and the role of parties. This has created a research paradigm which is perhaps the dominant paradigm today. Much of the work in electoral behaviour draws on this thinking. Today, there is an attempt to combine the different explanations trying to take into account, both sociological determinants but also the emotional and affective component as well as the component related to choice and calculation.

With regard to the limits, methodological individualism has often been evoked, saying that it is an exclusively micro-sociological perspective that neglects the effect of social structure. There is little room for context even though there are more recent developments that try to put the voter's freedom of choice in context. Some have criticized this model saying that it puts forward the one-dimensional image of the human being and politics, that is, that it is purely rational, hypercognitive in a way without taking into account sociological but also psychological elements. Others have criticized this analogy between the economic market and the political market as being a bit simplistic, saying that, basically, the consequences of buying a consumer product have a certain number of consequences, but they are much more limited compared to what buying a vote can have in terms of choosing a party. This approach would be elitist, this assumption that voters have the ability to know what is going on which is the idea of information and this ability that voters have to look at that information and process it. For some, this model overestimates the capabilities that voters have. There is a whole literature on opinion formation, quite consensually, that says that citizens have a limited capacity to process information. So there is an overestimation in this model with respect to capacity. One of the answers within spatial theories is based on this criticism that voters are not these cognitively strong beings as the original Downs theory presupposes. There has also been the criticism of abstention as the result of rational calculation. It was this model that proposed that abstention can be the result of a purely rational calculation. There was a whole series of critics who said that if it's something rational, there's a problem with the way democracy works.

Spatial theories of voting[modifier | modifier le wikicode]

Spatial theories of voting are nothing other than what we have seen so far with regard to the economic model of voting. Nevertheless, some of these spatial theories depart from this initial formulation.

The presupposition for spatial theories of voting has already been mentioned, namely the stake vote. Psychological theories are based on a type of explanation that does not focus on the issues discussed during a political campaign, for example. Even more plausibly, election campaigns are built around several issues. There may be one that is at the centre, but there are also others that are discussed. In Switzerland, the idea of an issue is particularly important because there is direct democracy, which is something that by definition is based on issues. But more generally, when there is a campaign, the issues are discussed. In this theory, we vote for specific issues that may be more or less concrete, more or less general, and which form the basis for explaining electoral behaviour. This idea of an issue was not invented by the proponents of the economic model of voting but was already present in the psycho-sociological model. In the Michigan model, the idea of stakes was already present but was somewhat underdeveloped, and this perspective on the role of stakes in the psychosocial model lent itself to both theoretical and empirical criticism from proponents of rationalist models. The theoretical criticism consists in saying that in this psychosocial approach or in this vision that the psychosocial model has of the role of political issues, the evaluation of these issues is determined by political attitudes and partisan identification. In the retrospective model, some researchers have proposed an alternative way of viewing partisan identification as being determined by the position voters take on issues.

There has been the whole emergence of the rational actor, which is the vote in relation to issues, which is not something that comes simply from our affective identification with a party, but there is a whole reflection that the voter makes in terms of cost-benefit calculations. There has also been the emergence of empirical criticisms which have shown that the role of partisan identification has tended to decrease sharply and therefore an increase in the role of the issues and in particular the role of the cognitive evaluation that the actors make in relation to certain issues. There has been a lot of criticism that has allowed the idea of issue voting to develop in a rationalist context and models.

There are a whole host of typologies in relation to issues, and we distinguish different types of issues such as position issues and issues that are more or less emotional. There is also a literature on whether certain parties have certain issues, which voters believe are the parties that are better able to deal with a certain issue. What interests us is that the idea of issue voting is fundamental to spatial theories of voting. If we do not accept the idea that actors will vote according to their assessment of certain issues, to be more precise, according to their assessment of the position that the various parties have on certain issues, if we do not understand that, we cannot understand the spatial theories of voting either.

Voters try to maximize the usefulness of the vote, that is, they try to vote for the party that makes them more satisfied. Voters try to maximize their individual utility.

These theories are called spatial theories of the vote because they are projected. We project voters' preferences and political positions, that is, the positions that parties have on certain issues and for the preferences that voters have on certain issues. Theoretically, it is possible to have as many dimensions as there are issues being discussed in an election campaign.

There are two important issues in relation to the spatial theory of voting.

The first question is how to assess the position of the different parties and candidates, since we start from the idea of projecting voters' political preferences and party projections onto a map. If we accept this premise, how will we position ourselves? A set of theories has given some answers. These theories are the retrospective voting theories and the theories of ideological space.

The second question is according to which criteria to determine the individual utility of voters. The answer to this second question will allow us to differentiate between proximity models and directional models because these two subsets of the spatial theories of voting give diametrically opposite answers to this question. Proximity models will give certain proximity related answers and the other more recent models offer an alternative answer based on certain criticisms.

How to assess the position of different parties and candidates[modifier | modifier le wikicode]

The problem of information is crucial in the spatial theories of voting and who would need an answer to fully understand these different theories. In the psychological approach, the information problem is circumvented by the idea of the development of partisan identification, which is an emotional shortcut that voters operate. On the other hand, in rationalist approaches, shortcuts are cognitive shortcuts.

Retrospective voting: Fiorina[modifier | modifier le wikicode]

Fiorina's theory of retrospective voting is very simple. In the study of electoral behaviour, there is a simple distinction between what is called prospective voting and retrospective voting. Prospective voting is the one that has been postulated by Downs and by all other researchers who work in proximity models but also in two-way models. Fiorina proposed the question of how to evaluate the position of different parties and candidates: how can voters know what the position of different parties is during an election campaign? Prospective voting says that voters will listen to what candidates and parties have to say. This is called prospective voting because voters will listen to what the parties have to say and evaluate on the basis of that, that is, looking ahead. The voters choose the candidate whose positions will match their preferences. That is what is called the proximity vote, that is, having a preference over a policy.

There is a particular requirement, which is that this way of explaining the voting behaviour of the electoral choice is very demanding in terms of the knowledge that voters may have about different positions, especially in a context where there are several parties and where the context of the political system and in particular the electoral system must be taken into account, because it may be easier for voters to know their positions when there are two parties, two candidates, than when there are, as in the Swiss context, many parties running. The idea of prospective voting is very demanding. Voting requires voters to know the candidates' positions on issues, but when there are several candidates or several parties, it is not very easy for some voters in particular. The degree of political sophistication, political knowledge, interest in politics varies from voter to voter.

Fiorina proposed an alternative way to explain why voters vote for one party rather than another, or a different answer to how the position of different candidate parties can be assessed. This is called retrospective voting, which means that we are not looking at what the parties said in their platforms, but rather at what the parties did before. It is easier to look at what someone has done than to evaluate the promises they made. Prospective voting is based on election promises and retrospective voting is based on past performance. There is also the economic vote, which is the role of the economy.

According to Fiorina, retrospective voting is that citizens' preferences depend not only on how close they are to the political position of a party or candidate, but also on their retrospective assessment of the performance of the ruling party or candidate. Voters are more interested in political results than in political programmes, and the choice is also made from this perspective. A corollary to this theory is that voters react more to the government than to the opposition because performance is evaluated and a certain state of the economy, for example, can be attributed to the performance of a government. Thus, voters find it easier to assess performance than declared plans during an election campaign.

This is an alternative way which is another answer to the question of how to evaluate the position of different parties and candidates. Prospective voting says that the evaluation is based on what the parties and candidates are going to say. For Fiorina the voter does not do that, he will rather look at what has happened, he will also look at the state of affairs in a country, hence the importance of the economic vote in the narrower sense of the word. In other words, in this retrospective assessment, the economic situation of the country plays a crucial role. It is the state of the economy that will decide who will win the election or not. The government is blamed for the poor state of the economy.

The ideological space: Hinich and Munger[modifier | modifier le wikicode]

The premise of prospective voting is too demanding for most voters. We have seen that at Downs, the role of ideology is fundamental and that ideology could function as a kind of shortcut. Moreover, retrospective voting can also be seen as a shortcut. Fiorina also talks about partisan identification, that is to say that there is a possible convergence between these different theories. Among these bridges, one of the first bridges between the psycho-sociological voting theory and the rationalist theories was made by Fiorina because he considers partisan identification to be an important element in explaining electoral choice. However, he conceives the origin and function of partisan identification in a different way from what we have seen before. The Michigan model was based on the idea of socialization and partisan identification as a long-term attachment to a party that is the result of primary socialization in particular, and therefore as insertion into a given social context. Fiorina reverses the question, in fact, partisan identification can result from something else and it also produces electoral choices. Voting for a party and continuing to vote for such a party repeatedly makes it possible to develop an identification with that party which, in a way, then reinforces the electoral choice. There is a small bridge that is made between these two theories with Fiorina on the one hand and the Michigan model of another party that puts the concept of partisan identification at the centre and that conceives of this concept in a very different way, especially with regard to its origin. According to Fiorina, identification with a party is not necessarily the result of a long phase of socialization, but it is also the result of evaluations of a certain party, it is the fact of voting for that party that makes it possible to develop a partisan identification. It is quite interesting to see the bridges that can be built between theories that may seem different.

Downs already put ideology at the centre of his explanation. In the sociological and psycho-sociological model, there was no place for ideology, that's another thing that counts, on the other hand, in economic theories, spatial theories and Downs' theory of the economic vote, ideology is important. Ideology is to be understood as a way of simplifying our world in relation to the problem of information. According to Downs, based on the prospective assessment that voters make of the position that voters have and their position on various issues, voters arrive at and operate this shortcut by situating and bringing parties back to an ideological dimension that may be a left-right dimension but may also be another one. So, voters evaluate the positions of the parties and from these positions, this party is a left-wing party and this party is a right-wing party. Then they evaluate their own position in relation to the issues and they do the same operation positioning themselves on this left-right axis. Positioning on a left-right scale is related to this type of theory.

Hinich and Munger take up the Downs idea but turn it around a bit. The basic idea is somewhat the same, namely that it is a way that voters have at their disposal, a euristic and cognitive shortcut that voters have at their disposal to deal with the problem of complex information. The basic assumption is that voters decide primarily on the basis of ideologies and not on the basis of specific positions on issues.

In their view, ideology is a means of predicting political positions on a significant number of issues and also a basis for credible and consistent engagement by the party or candidate that follows it. Ideology is a means of predicting and inferring political positions during an election campaign.

While Downs said that there are parties that take positions on issues, the voter has difficulty with this inferring a position on a left-right axis. Hinich and Munger say the opposite, saying that on the basis of their idea of the left-right positioning of the parties, they somehow deduce what will be or what is the position of these parties on the different issues. Voters have knowledge of the ideological positions of parties or candidates on one or more ideological dimensions and they use this knowledge to assess the political positions of these parties or candidates on specific issues.

What are the criteria for determining the individual usefulness of voters?[modifier | modifier le wikicode]

The spatial theory of the vote postulates that the electoral choice is made in the maximization of individual utility.

A distinction can be made between the simple proximity model, which is the Downs model, and the proximity model with Grofman discounting. It is a variant of the simple proximity model which remains in the idea of proximity but which adds an element which makes it possible to explain certain voting behaviours which would not be explainable by other models. These two proximity models are opposed to two other models that are called directional models with Matthews' simple directional model but especially Rabinowitz's directional model with intensity. Curiously, the intensity directional model that adds an element to the simple directional model chronologically precedes the simple directional model.

Simple proximity model: Downs[modifier | modifier le wikicode]

The basic idea is the representation of a point that is an ideal point for each voter in a hypothetical space. The idea is that each voter can be represented by a point in a hypothetical space and this space can be a space with N dimensions and each dimension represents an election campaign issue, so that this point reflects his or her ideal set of policies, i.e. preferences and positions.

The political position of each candidate is represented in the same space, it is the interaction between supply and demand and the voter will choose the party or candidate that is closest to the voter. Maximizing utility is done in proximity to certain issues.

Simple proximity model utility function.

The utility function of the simple proximity model appears, i.e. the maximum utility is reached at the line level. When the voter is in the same position, i.e. maximum proximity, as the party, his or her utility increases, and when the voter moves away from the party, his or her utility decreases. Otherwise, our usefulness as voters decreases as a party moves away, i.e. as a party's position moves away from our political preferences.

It is also often referred to as a point of indifference because there are places where the voter cannot decide. This is especially important when applying this type of reasoning empirically. This model explains for Downs why we abstain. It is because we are rational, and if we are rational, rationality means maximizing our usefulness on the basis of the closeness we can have with a party. If that is true, then if there are two parties that are equally close to our preferences, then we cannot decide. That is called the point of indifference.

Proximity model with update: Grofman[modifier | modifier le wikicode]

Grofman introduces a central element which is the position of the status quo which is not necessarily the neutral point but the current policy. For Fiorina, the retrospective vote is the fact that current policy is fundamental, whereas in the prospective vote it is less so. In prospective voting, Grofman said that the position of current policy is also important because the prospective assessment that one can make as a voter of the parties' political platforms also depends on current policy. The distance must be assessed on the basis of what the current policy is.

Grofman's idea is to say that the voter discounts what the candidates say (discounting) based on the difference between current policy and what the party says it will do or promise. Discounting is saying that the voter does not fully believe what the parties say. This is a very common and shared notion. That discounting depends on where the policy is right now in relation to what the party is promising, and that is the directional element. In other words, a directional element is introduced into the proximity model. It is a small bridge between different explanations.

Proximity model utility function with update.

The curve instead of the simple proximity model, or obviously the maximization from the parties' point of view of electoral support, lies in the precise proximity between voters' preferences and the parties' political programs on certain issues, in this case this remains true but with a lag that is determined by discounting from a given status quo.

From the parties' perspective, this model makes different predictions than the simple proximity model, which made a prediction of convergence of a centripetal force with respect to party positioning. This model predicts a convergence of party program positions around two distinct positions, there are two types of convergence.

Simple directional model: Matthews[modifier | modifier le wikicode]

In the literature, spatial theories of voting are often seen as one of the main developments of the last thirty years which has been precisely the development of directional models since the proximity model dates back to the 1950s. Beginning in the late 1980s and early 1990s, there has been a strong development of directional models.

The idea of the directional model, and this applies to both the simple directional model and the intensity directional model, is that voters basically cannot clearly perceive the different positions of political parties or candidates on a specific issue. What voters perceive are directional signals, that is, voters perceive that some parties are going in one direction and other parties are going in another direction on certain issues. This is the basic motivation for the development of these directional models. Candidate choices are made towards parties or candidates who are going in the same direction as the voter, this being understood as the voters' political preferences on a given issue.

Utility function of the simple directional model.

The simple proximity model is that the voter will vote for the party or parties that are in the same direction. What determines direction? There are two variations. It is possible to determine direction based on the "neutral point" which is the point in the middle, or it is also possible to determine direction from the "status quo". For most theories, and in particular Matthews' Simple Directional Model theory, the neutral point determines direction.

All parties that are in the same direction of the voter maximize the individual utility of that voter. In this case, there may be other factors that can contribute to the voter choice; and all parties that are on the other side of the neutral point minimize the voter's utility, so the voter will not vote for that party all other things being equal.

Directional model with intensity: Rabinowitz[modifier | modifier le wikicode]

The intensity directional model adds an element that is related to the intensity with which candidates and political parties defend certain positions. This is a fairly reasonable development, as is the discounting model, whose proximity was something reasonable and which makes the model more consistent with reality. The same can be said of the directional model with intensity.

The choice of candidates is made both according to direction but also according to the intensity of positions on a given issue. Thus, voters will vote for candidates who are in the direction (1) and who are going in that direction in the most intense way (2), that is, who propose policies going in that direction in the strongest and most intense way.

There are several reasons that the authors of these directional models cite to explain this choice of direction with intensity rather than a choice of proximity as proposed by Downs. The aspect is based on the idea that there is an information problem that represents a difficulty and costs that voters must pay to gather information and to become informed about an election. The idea is that voters are not really able to really evaluate in a forward-looking way the different positions of the parties. Therefore, they cannot really situate where the different parties stand. On the other hand, this is true for the directional model; they manage to perceive a policy direction. As for the intensity model, they manage to perceive something more, that is to say, not only a direction but an intensity through which a political party defends certain positions and goes in certain political directions.

This model of directional proximity with intensity illustrates what is called symbolic politics which is related to the problem of information. Symbolic politics says that what is important in politics are not necessarily the rationally perceived positions or the political positions of the parties but what the political symbols evoke in relation to certain issues. The importance of symbolic politics is especially capitalized on by the intensity directional models. The importance of symbols lies in what arouses emotions. Symbols evoke emotions. The idea of intensity can also be seen as the idea that there are certain issues, that there are certain political positions that put forward symbols and some of these symbols evoke making these two issues more visible to voters but in the sense of making voters say that this particular party is going in that direction and with a high intensity.

At the basis of the reflection of directional models, and in particular of directional models with intensity, there is what is called symbolic politics. A symbol is evaluated on the basis of two parameters, namely direction (1), a symbol gives a certain direction in the policy and in addition a certain intensity (2) which is to what extent is one favourable or unfavourable to a certain policy.

Utility function of the directional model with intensity.

The advantage of the intensity directional model is that it goes in a more intense direction, i.e. it takes a political position that evokes the idea of symbolic politics in a more salient way.

The utility function of this model is modified compared to the simple model, i.e. the further a party moves in the same direction as the voter, the more likely it is to be chosen by that voter. The further a party moves in the other direction, the less likely the voter will choose it because the utility function gradually decreases.

In this model, there is a region of acceptability of positional extremism which is a region outside of which the intensity of the positions or the direction shown by a party cannot go because if it goes beyond that region, the voter will no longer choose that party.

As far as the proximity model with discounting is concerned, there is a concern when we are going to apply it empirically: we need to be able to determine what the degree of discounting is, how much the voter is going to discount. This is something that remains difficult in theory, we don't know how much the voter will discount.

Summary of the four models[modifier | modifier le wikicode]

Merrill, S. III et B. Grofman (1999). A Unified Theory of Voting. Cambridge: Cambridge University Press.[16]

Four possible answers to the question of how voters decide to vote[modifier | modifier le wikicode]

To summarize these approaches, there are four possible answers to the question of how voters decide to vote. We want to know how and why a voter will vote for a certain party.

The first answer is that basically, they vote according to their position, according to their social characteristics or according to their socialization, which refers to the sociological model. There is no real electoral choice in this type of explanation, but it is based on our insertion in a social context. For Lazarsfeld, we think politically how we are socially, there is not really the idea of electoral choice. The concept of electoral choice does not belong to the sociological model but rather to rationalist theories. Rationalist theories and spatial models of the vote have had the very beneficial relationship of putting precisely the free choice of voters at the centre of analyses.

A second possible answer is that they will vote for the candidate who belongs to the party with which they identify. This refers to the Michigan model, the psycho-sociological model.

A third possible answer is that they will vote for the candidate whose political ideas are closest to their own. This is the proximity model. Proximity can be calculated on the basis of the programmes and actual positions declared by the parties or on the basis of a discount factor, a perception factor or a difference factor according to the discount model.

Finally, they can vote for the candidate who is most likely in the voters' perception to change things in a way or in a way that leaves them the most satisfied.

These are models that should make us attentive to the different motivations that voters may or may not have to make in making an electoral choice. There are other variants or models that try to accommodate this complexity. As part of spatial theories of the vote, some theories consider the characteristics of candidates. There are other theories that highlight the impact of economic conditions and how voters compare different election results in their electoral choices, which refers to economic voting in the strict sense of the term. There are other models that try to relate the multiplicity of issues to an underlying ideological space, i.e., instead of looking at specific issues, everything is brought back to a left-right dimension as a shortcut, for example, and there are other theories that consider the degree of ambiguity and clarity of the candidates' positions. In directional models with intensity, there are models that try to show how the salience of different issues changes from one group to another, from one social group to another, or from one candidate and one party to another.

Unified Voting Model: Merrill and Grofman[modifier | modifier le wikicode]

Other researchers have tried to propose combined models that combine different explanations. Merrill and Grofman have proposed unified models that want to get out of this hyper-simplification with respect to spatial theories where one either makes a choice of possibilities or a choice of direction but evacuates any other element such as partisan identification, socialization, social inclusion, economic conditions as well as the role of opinion leaders as seen in the funnel model of Michigan theory.

According to Merril and Grofman, one cannot determine whether one pure model is superior to another because there are methodological and data limitations. Of course, there have been attempts to assess the explanatory power of directional models, but according to these researchers, these spatial models were designed to be purely theoretical in order to highlight on a purely theoretical level what motivations voters may have for their electoral choice.

However, this is empirically incorrect. According to them, it is necessary to combine different types of explanations and in particular, in the electoral choice, the components related to proximity, leadership, and also the rather "intensity" leadership, all of which play a significant role in the positioning of candidates and parties. In other words, they propose something quite ecumenical that combines directional and proximity models. Numbers abound, since we have seen that, in the end, both models systematically have a significant effect.

They are both proximity choices and directional choices with intensity, since there are voters who may choose intensity and others who may choose direction. One must take into account the heterogeneity of the electorate and how different voters may have different motivations for choosing which party or candidate to vote for. The extent to which the usefulness of voters' choices varies from candidate to candidate, but also from voter to voter. It is also possible to add that the weight of partisan identification varies from one voter to another.

On the other hand, preferences for candidates in power are best explained by the proximity model and the simple directional model. On the other hand, the intensity directional model better explains the electoral choices of candidates who are not currently in power.

Reviews of the proximity model[modifier | modifier le wikicode]

Iversena proposed a way of classifying the different explanatory theories of voting that allow to add a very important element that has been neglected until now. Many researchers have criticised the Downs proximity model in particular. A first criticism that has been made is that the simple proximity model gives us a misrepresentation of the psychology of voting. This is the idea that gave rise to the development of directional models, which is that, according to Downs and those who have followed him, because there is transparency of information, voters can very well see what the political platforms of the parties or candidates are.

The second criticism is the lack of an adequate theory of preference formation. These spatial theories start from the assumption that there is a voter or voters who have political preferences with respect to certain issues, but completely discard the explanation of how these preferences are formed. This theory is not about the formation of political preferences, they start from the idea that there are voters with certain political preferences and then these voters will look at what the offer is and will choose according to that offer. The choice can be made according to different criteria, but they start from the assumption that there are these voters who arrive in an electoral process that refers to the idea of the hexogeneity of voters' preferences.

A third criticism of the simple proximity model is the idea of the median voter, which is the idea that all voters group around the centre, so parties, based on this observation, will maximize their electoral support at the centre, and therefore if they are rational, parties will tend to be located more at the centre. However, we see that this is not always true and that there are parties that propose more extreme policies that receive considerable electoral support. So there is this empirical anomaly where there is a theory that presupposes and tries to explain the electoral choices but also the positions of the parties in a logic of proximity to the centre of the political spectrum, but on the other hand there is the empirical observation that is the opposite and that sees parties and voters located elsewhere. The directional model also provides some answers to this criticism.

Responses to criticisms of the proximity model[modifier | modifier le wikicode]

There are several responses to criticisms of the proximity model. These are possible answers more to justify and account for this anomaly. The anomaly is that there is a majority of the electorate around the centre, but there are parties at the extremes that can even capture a large part of the preferences of the electorate. There have been attempts to address this anomaly. Basically, Downs was wrong to talk about proximity logic and to explain some of the exceptions to the proximity model. There are three possible answers:

  • that of the law of curvilinear disparity proposed by May;
  • the directional model of Rabinowitz and Matthews;
  • Przeworski and Sprague's mobilization of the electorate.

Law of Curvilinear Disparity: May[modifier | modifier le wikicode]

May's Law of Curvilinear Disparity is an answer that tries to stay within the logic of the proximity model and to account for this empirical anomaly, but with the idea that it is distance and proximity that count. It is an answer that remains faithful to the postulates of Downs' theory and the proximity model.

Hirschman wanted to explain what happens in organizations when they enter a situation of crisis or decline. The idea was that there were two possible responses that are put in place by members of that organization: one of "exit", to withdraw, to go to another organization. The organization is in crisis and no longer reflects our own needs. Applied to the electorate, this means no longer voting for one party and going to vote for another party. Hirschman contrasts the "exit" strategy with the "voice" strategy, which is based on what he calls "loyalty", which is that one can choose not to leave but to make the organization change, to restore the balance between one's own aspirations and what the organization can offer. The idea is that you stay loyal and you do "voice", that is, act to make things change.

The law of curvilinear disparity takes up this distinction. There are three actors at play in this theory: there are voters, candidates, and an intermediate group represented by activists who are in fact voters who become activists going to exercise "voice".

The starting point is that there is a congruence of attitudes between party leaders and voters due to the possibility of exit for voters when the party no longer represents them (exit). If voters, who prefer more extreme options, no longer find these options within the party they voted for, then they will look elsewhere and vote for another party. This ensures congruence and proximity between the party and the electorate. On the other hand, ideologically extreme voters try to influence party policies through party activism (voice).

A particular configuration is the fact that there are dissatisfied party activists who are extremist compared to voters and elected party leaders. In other words, party activists tend to be more extreme in their political attitudes than voters or party leaders.

If certain conditions are present, such as good democratic functioning within the party, activists will have the opportunity to exercise "voice" and influence positions. The idea is that the extremist attitudes of those former voters who become party activists push strategic positioning in a direction that takes them away from their constituents. We end up with a configuration where there is an electorate that is at the centre, there are party activists who are exercising the "voice" and who have access to the extreme, and there are party leaderships that are in between. There is this curvilinear disparity because the three actors position themselves differently.

It is by this configuration that May tries to explain this anomaly which is due to the fact that there is a group of voters who become activists within the party and who succeed in shifting the party's positioning towards the extremes.

Partisan Competition Theory: Przeworski and Sprague[modifier | modifier le wikicode]

The second explanation refers to the directional model, i.e. it is an element of direction and not an element of distance or proximity that counts. It is an explanation that is completely outside the logic of proximity and the spatial logic of voting. It is a third explanation given by Przeworski and Sprague in their theory of partisan competition, also known as the theory of mobilization of the electorate. The presupposition is that voter preferences are not exogenous but are endogenous - they change within the framework of an electoral process. We leave behind the idea of spatial theories that preferences are exogenous, that they are pre-existing and almost fixed. Here, preferences are endogenous and they can change.

Political parties can make choices that are not choices to maximize the electorate, unlike spatial theories, where parties seek to maximize their short-term electoral support in an election. For Przeworski and Sprague, there may be another logic that is not one of maximizing the electorate in the short term but one of mobilizing the electorate in the medium and long term. The idea is that a party is ready to lose an election in order to give itself the means to win it later by giving itself time to form an electorate. So, we are going to the extremes precisely because we are trying to mobilize an electorate. The assumption is that mobilizing an electorate is done by taking clear positions and not a centrist position. The idea is to create a party that forges ideologies and partisan identities.

In this way, parties can offer relatively extreme political platforms that are not optimal in the short term, but that generate higher levels of support in the medium and long term. Parties do not try to maximize the vote, but create images of society, forge identities, mobilize commitments for the future. Some parties have short-term strategies for maximizing voting and others have long-term strategies for social mobilization. In order to explain this anomaly, another explanation beside the curvilinear explanation beside the directional theories of the vote, a third possibility to explain this would be to say that there are some parties that abandon the idea of maximizing the vote or electoral support in order to mobilize this electorate and for this we have to go to extremes.

Four voting models: Iversen[modifier | modifier le wikicode]

Iversen, T. (1994). “The Logics of Electoral Politics: Spatial, Directional, and Mobilizational Effects”. Comparative Political Studies 27: 155-189.[17]

If we take into account Przeworski and Sprague's idea that preferences are exogenous and not endogenous, it is possible to create a typology as Iversen did. It is possible to create a typology that distinguishes between four approaches crossing two important and crucial elements: "is voting spatial? "i.e., if it is proximity, it is 'yes', otherwise it is 'no' and therefore directional; 'are the preferences of the actors exogenous? "The answer is "yes", as postulated by spatial theories, or "no", as stated by Przeworski and Sprague, for example.

The idea is that this table is the Downs-Hirschman model that would have been made in order to summarize the different responses to the anomaly we have been talking about. So there are four main ways. From the perspective of the issue vote, there are four main ways to explain how and why voters are going to vote a certain way and why parties are going to position themselves. In the Downs-Hirschman model, the vote is spatial in the sense of proximity and preferences are exogenous; on the other hand, in the directional theories of Rabinovirz and Macdonal in particular, we remain in the idea of the exogeneity of preferences but the vote is not spatial in the sense of proximity. If we take into account Przeworski and Sprague's idea that there can be a mobilization of the electorate in a logic of endogenous preference and non-maximization of the utility of voters.

For Iversen, distance is also important. Distance is understood in the sense of the proximity model for whom voter preference and party position is also important. Distance must be taken into account and the idea of mobilizing the electorate must be taken into account. His conclusion is that the vote is explained both by elements of leadership, partly by an element of proximity and distance, but also, for some parties, it must also be taken into account that there are parties that act according to a mobilization of the electorate according to the approach of Przeworski and Sprague. So all these elements help to explain the vote and must be taken into account in order to explain the vote.

The process of misalignment[modifier | modifier le wikicode]

Some have another way of talking about convergences and showing how the theories explaining the vote can be reconciled with the process of political misalignment. These authors have tried to say that the different explanatory theories of the vote can be more or less explanatory in the sense of having more or less importance of explanatory power depending on the phases in which one is in a process of alignment and misalignment. Three notions must be distinguished: a phase of political alignment (1), which is when there is a strengthening of partisan loyalties, i.e. emotional ties between voters and parties; a phase of political misalignment (2), which may be the one we are currently in in Europe since the economic crisis, which is a weakening of partisan loyalties resulting in increased electoral volatility, i.e. it is easier to change parties from one election to the next; a phase of realignment (3), which consists of creating new partisan loyalties.

Harrop, M. et W. L. Miller (1987). Elections and Voters. New York: New Amsterdam Books.[18]

This diagram shows the process of misalignment with changes in the generational structure and changes in the social structure that create political misalignment. Misalignment creates greater electoral volatility that creates a change in the party system that can have a feedback on the process of alignment, misalignment or realignment.

Relationship between voting explanatory models and realignment cycle[modifier | modifier le wikicode]

These authors proposed to say that there would be a relationship between the explanatory models of the vote and the cycle of alignment, realignment, misalignment in the sense that the sociological model would be better able to explain the vote in phases of political realignment. It is a moment when social cleavages directly influence the vote in this approach and therefore the sociological model, perhaps, at that moment, better explains the vote. In a phase of alignment, this would be the psycho-sociological model, i.e. party loyalties are freed from their social base and thus these party identifications are formed and crystallized. Finally, in a phase of misalignment, this would be the economic model, since there is a loss of these partisan loyalties, so these voters become more and more reactive to political events and therefore may be more rational in their decision-making process. They may rely less on their partisan loyalties, so their vote may be explained less by their social base and more by their choice among an offer that is the economic model.

Apart from the combined models, it can be thought that different models may explain differently according to historical moments and phases of a process of political alignment and misalignment just as models may better explain certain types of candidates or according to the profile and type of voters.

Annexes[modifier | modifier le wikicode]

References[modifier | modifier le wikicode]

  1. Marco Giugni - UNIGE
  2. Marco Giugni - Google Scholar
  3. Marco Giugni -
  4. Marco Giugni -
  5. Marco Giugni - Protest Survey
  6. Marco Giugni - EPFL Press
  7. Marco Giugni - Bibliothèque Nationale de France
  8. LAZARSFELD, PAUL F. PEOPLE'S CHOICE : how the voter makes up his mind in a presidential campaign. S.l: COLUMBIA UNIV PRESS, 2020. Print.
  9. McClung Lee, A. (1949). LAZARSFELD, PAUL F., BERNARD BERELSON, and HAZEL GAUDET. The People’s Choice: How the Voter Makes Up His Mind in a Presidential Campaign. (Second edition.) Pp. xxxiii, 178. New York: Columbia University Press, 1948. $2.75. The ANNALS of the American Academy of Political and Social Science, 261(1), 194–194.
  10. Campbell, Angus. The American voter. Chicago: University of Chicago Press, 1976. Print.
  11. Downs, Anthony. “An Economic Theory of Political Action in a Democracy.” Journal of Political Economy, vol. 65, no. 2, 1957, pp. 135–150. JSTOR,
  12. Harrop, Martin, and William L. Miller. Elections and voters : a comparative introduction. Houndmills, Basingstoke, Hampshire: Macmillan Education, 1987. Print.
  13. 13,0 et 13,1 Dalton, Russell J. Citizen politics in Western democracies : public opinion and political parties in the United States, Great Britain, Germany, and France. Place of publication not identified: Chatham House, 1988. Print.
  14. Bakker, B. N., Hopmann, D. N., & Persson, M. (2014). Personality traits and party identification over time. European Journal of Political Research, 54(2), 197–215.
  15. Downs, Anthony. “An Economic Theory of Political Action in a Democracy.” Journal of Political Economy, vol. 65, no. 2, 1957, pp. 135–150. JSTOR,
  16. Merrill, Samuel, and Bernard Grofman. A unified theory of voting : directional and proximity spatial models. Cambridge New York: Cambridge University Press, 1999. Print.
  17. IVERSEN, T. (1994). The Logics of Electoral Politics. Comparative Political Studies, 27(2), 155–189.
  18. Harrop, Martin, and William L. Miller. Elections and voters : a comparative introduction. Houndmills, Basingstoke, Hampshire: Macmillan Education, 1987. Print.