This week’s reading focused on changes in electoral systems of nations. The readings were divided between two different sub topics in the field on electoral system changes, the effects of potential changes in the electoral systems this area is covered by the Taagepera & Shugart, and Lijphart pieces, The other area the is examined in this weeks readings is the causes and situations that would lead a nation and legislature to adopt changes to their electoral systems. This aspect was examined by the pieces that were authored by Boix and Benoit. The selections were an interesting juxtaposition in the theoretical and more normative approach and the more real world approach. The more normative pieces talk about the effects on proportionality and representativeness that are associated with certain changes that can be made to the electoral system as well these pieces serve to prescribe ways to improve the electoral systems. The other two pieces approach changes in the electoral system in a more rational choice model. These pieces examine changes to the electoral system through the eyes of those that have the power to change the system; the incumbent legislatures. The litterature shows how these legislatures will only change when it is a rational choice for them to change in order to protect or increase their respective share of the legislature.
A very interesting aspect about changes in the electoral system that was noted by Taagepera & Shugart is that often it is beneficial to simplify the electoral system. The author states that unless there is a very egregious flaw in the electoral system it is often better to leave the system well enough alone rather than adding another layer of complexiety to the electoral system. The authors argue a very valid point that “complexity introduces an elitist inequity of its own, even if the purpose is increased “fairness”: the more complex a system becomes, the fewer people can comprehend it in order to make use its opportunities.” (228). The authors show that as an electoral system grows more complex it becomes increasingly harder to understand and therefore more difficult for the electorate and the parties to maximize their use of the rules of the electoral system so that might best represent their interests.
It is also shown that most of the time the same desire that it sought from adding an increased level of complexity to the electoral system can be found in a simpler fashion. The author uses several examples of this by suggesting that some problems can be fixed by increasing the district magnitude. The author’s note that unless the system in question regularly deviates from proportional representation by more than 10 percent then there is no need for another level of complexity to try and approximate the true level of proportionality. It is further shown that many of these extra levels of complexiety are designed to help a specific class or minority party and often this protection is no longer needed after around a decade or so, and that no electoral system should included rules to solve such a temporary problem.
The Boix and Benoit pieces show another perspective on how and why changes in electoral systems occur. They show that since the body that determines when and what electoral changes will occur in an electoral system are the exact same institutions that will be affected by these changes, it is only natural for the members of these institutions to try to protect their positions. Benoit states that “Electoral laws are quintessentially distributive institutions, improving the share of one group at the direct expense of another.”(366-367). This means that since the electoral systems are zero sum games any change in the system could result in a positive gain for one group and as a result another group must have a negative result. Those individuals that are in power would not seek to change the rules that got them elected into power unless they were at risk of loosing their power. The authors go on to explain this through rational choice models showing that when those that are in power fear that they might loose their power they change the rules of the electoral system so that might either maintain their control or at least minimize their losses.
Another point that was made by Benoit was that electoral systems freeze into a stable state around the same time that the party systems freeze. “As soon as the electoral arena became stable and the party system froze along certain cleavages, policymakers lost interest in modifying the regime.”(365). This shows that when a nation settles into a political culture the parties also learn to operate and succeed in this system and it is in this success and stability that the parties stabilize the electoral rules that brought the parties to power. When there is a reduced level of turmoil among the populace there are fewer voices calling for changes in the electoral system and the system stabilizes and changes are then few and far between. There is no need for the parties that are in power to change the electoral system that they work under since that might have potential negative impacts for the party. It is only when the party fears that it will loose under the current system exceeds their fear of the unknown consequences of a change in the electoral system that these changes occur.
A major question discussed in this week’s readings regards the role that established parties play in shaping electoral systems and the incentives for those parties to change the electoral system. Benoit (2003) provides a model for electoral system change in which parties are rational actors that seek to maximize benefits, their share of seats in the legislature. He argues that parties will seek to change institutional structures when changing these structures will “improve their expected seat share relative to the status quo” (374). A key variable in this model is the ability of a party or group of parties to enact a change in the electoral system. If enough parties in a multiparty system want to change the electoral system, then a change will occur. This ability or power is what Benoit calls “institutional fiat power,” which is the “voting weights held by each party for the purpose of changing the electoral law” (374). His model predicts that if the total fiat power exceeds more than what is need to change an electoral system and each party expects to gain seats under a new electoral system or rule change then the party system will change. In other words if a party holding seats perceive that a change in election law will gain their party more seats, and a coalition of parties agree that a change in election law will benefit them as well (the model does not state that the controlling party in government must be a member of this coalition), then an electoral system can change.
Boix (1999) articulates the above point in a somewhat different fashion. Boix (1999) also argues along the lines of strategic parties who shape electoral rules to their benefit (609). He extends his analysis back to prior to universal male suffrage and demonstrates that the entry of new voters and new parties into an electoral system has consequences for the electoral systems, two examples included are that the existing parties can either move to try and incorporate more votes at the expense of another established party or the existing parties will change the rules to a proportional representation system. In all of the cases Boix (1999) demonstrates hypothetically the existing parties anticipate the entry of new players into the electoral system. In empirical analysis Boix (1999) demonstrates that as the threat of socialist parties increase, the more likely the electoral system will switch to a PR system. Boix (1999) argues this is due to the winner-take all nature of plurality systems. The party systems changed due to the entry of new voters and parties and the reaction of the established parties to maximize their proportion of seats in parliament. The entry of the socialist parties into the electoral systems produced a lowering of the effective threshold because “Failure to reduce the electoral threshold would have led to an overwhelming victory of the socialist party” (609).
The two previous studies both argue that parties in electoral systems are strategic actors that want to maximize seats in the legislature. The final two readings for the week provide some insight and caution to electoral system changes from a different perspective, that of stability and representativeness. Both Lijphart (1994) and Taagerpera and Shugart (1989) advise against major changes in electoral stability.
Taagerpera and Shugart (1989) argue that the familiarity of an electoral system reinforces stability of the electoral system, but change is not unwarranted in all situations. Specifically they argue that simplicity is better serves electoral system from a democratic standpoint. “[C]omplexity introduces an elitist inequality of its own, even if the purpose is increased ‘fairness’: the more complex a system becomes, the fewer people can comprehend it in order to make use of its opportunities” (228). The essence of their argument is that change in electoral systems should be few. Only when striking disproportionality is evident should change in electoral system occur, such as in the case of “overamplification” of the winning party in which the winning party wins a “narrow” plurality of the votes but gains “landside victories in term of seat shares” (221)
Lijphart (1994) like Taagerpera and Shugart (1989) also argues that electoral system change should be few and that stability is grown from familiarity with electoral systems. He makes several recommendations to increase proportionality of electoral system but none are major changes that would disrupt the established electoral systems. In his closing remarks to his book he says “Healthy partisan competition requires that the electoral system – the basic rules of the democratic election game- be broadly supported and not be changed too frequently” (151). The exception to his rule is in new democracies when rules must be chosen to “guide the new democracy’s elections for a long time” (152).
April 28, 2008
POL 628- Comparative Parties and Elections
Electoral System Changes
An electoral system can make a difference in which party wins, and how decisively it wins. Electoral systems can also influence which losing parties can stay around to compete again and which are eliminated for good. Electoral systems do not arise from a vacuum but from political debate and struggle…While they last, electoral systems, like constitutions and other institutionalized constraints, do shape politics. They can delay and exert pressure on policies and strategies, and this can make a difference” (Taagepera and Shugart 1989, 234).
As we discussed earlier in the semester, electoral rules are not wholly exogenous to a political system. While the institutional perspective maintains that electoral rules are key to shaping party competition and the behavior of parties and individual politicians, one must acknowledge that electoral rules are also endogenous to a system. That is, electoral rules are both shaped by and shape a political system. As political scientists, we place a great deal of emphasis on electoral systems, in part, because of their impact on other institutions. Electoral systems can have a variety of consequences, ranging from determining the number of parties that participate in government, to structuring incentives for individual politicians, and even to contributing to regime stability. As Duverger (1951) notes, electoral systems are important in terms of the mechanical and psychological effects they exert on voters and parties. This week’s readings examine the circumstances under which changes will be made to existing electoral systems and the shape such changes should take.
Before turning to an examination of factors contributing to electoral system change, it is first necessary to address the question (posed by several of this week’s authors) of why study electoral systems at all? Based on the preceding paragraph, weekly readings, and class discussions from earlier in the semester, one is left with the impression that electoral systems are extremely important in terms of their impact on other facets of the political system. However, both Lijphart (1994) and Taagepera and Shugart (1989) contend that the ability of an electoral system to exert any significant impact on other institutions is relatively small. It seems unfair to downplay the importance of electoral systems; a better argument might be that the impact of electoral systems is felt indirectly (which, in fairness, is something to which Taagepera and Shugart allude). Electoral systems determine which parties become part of a governing coalition, which, in turn, determines policy. Electoral systems, then, are important in terms of their indirect impact on policy, if for no other reason.
Taagerpera and Shugart (1989) argue that electoral systems should be simple. In their view, there are few circumstances under which changing electoral rules are appropriate. Even though changes in electoral rules may produce favorable results at the district level, for a small party, or relative to intraparty competition, often the change felt at the national level represents only a marginal improvement. Thus, these minor improvements are not sufficient for adding additional layers of complexity to an electoral system. Further, nations that already have complex rules really should not move to simplify them because the outcome would be marginally better; these countries would do well to stick with a system with which people are familiar. If, however, the decision to change the rules is made, leaders should aim for simplification, rather than adding another layer of complexity in an attempt to correct the problems that already exist. Taagepera and Shugart offer a normative argument regarding one aim for an electoral system (simplicity) and several suggestions for electoral engineering; they do not explain, however, the circumstances under which politicians will be motivated to change the rules or the situations in which those political actors not in power will succeed in lobbying for change.
Lijphart (1994) posits that achieving proportionality is the most important goal for an electoral system. He finds that, taken together, the effective threshold, electoral formula, assembly size, apparentement, and whether a country has a presidential or parliamentary system explains two-thirds of the variance in proportionality. While offering several suggestions for electoral system reform, Lijphart, like Taagepera and Shugart (1989), notes that changes to electoral rules should be infrequent in order to enhance stability and resist the whims of partisan interests.
By contrast, Boix (1999) argues that electoral systems change in response to the desires of politicians. The governing party (or parties) chooses electoral rules from which it will benefit. This party will not be motivated to change the rules unless something about the electoral arena changes that prevent it from continuing to enjoy the benefits of the original rules. While it is difficult to imagine that advanced democracies could ever experience anything as dramatic as the extension of suffrage or introduction of competitive elections, Boix’s analysis leaves room for changes in electoral rules in response to realignment or substantial changes in party organizations. Further, it would be interesting to extend his analysis to emerging democracies in order to note whether massive shocks to an electoral arena on the level of the advent of mass suffrage or the introduction of competitive elections lead to changes in electoral rules. Additionally, while Boix considers only alterations to the effective electoral threshold, it might be worthwhile to determine whether the government makes any other alterations to electoral rules in response to changes in the electoral environment.
In some instances, electoral systems are the outcomes of partisan goals; in other instances, they represent more general goals, such as representation or governability. Benoit (2003), in an argument similar to that of Boix (1999), notes that strategic political actors will attempt to change electoral rules when they no longer derive benefits from them. Political parties will choose the institutions that will allow them to maximize their seat share in the legislature. Change will continue so long as a party has the power to make changes and will gain more seats by doing so. Equilibrium is reached “when no party or coalition of parties with the power to adopt an alternative electoral system can gain more seats” by changing the electoral rules (374).
Broadly, this week’s readings speak to both the importance of electoral systems and the circumstances under which those systems will change. Electoral systems are important because of their impact on other institutions and, indirectly, for their impact on public policy. Electoral institutions will change when it is in the interest of the governing parties to alter the rules. In a general sense, these conclusions emphasize the idea that electoral rules are both exogenous and endogenous to a political system.
 Unless greater complexity leads to a much better outcome.
 In PR systems in which voters choose among competing party lists, apparentement is the extent to which parties are able to formally connect their lists, “which means that their combined vote total will be used in the initial allocation of seats” (Lijphart 1994, 15).
In many of the articles and books focused on this type of research, electoral reforms are characterized as being encouraged by the self-interest of political actors. For scholars like Benoit (2004) or Boix (1999), parties are most importantly strategic actors trying to find electoral rules that would expand their power in general, and more specifically, their seat share. Parties look at what the resulting outcome would be of various electoral rules and choose the one that would give them the greatest political authority.
A main argument for majoritarian systems is the notion that when deciding between electoral systems, the most simple system is the best system. Taagepera and Shugart (1989: 236) subscribe to this principle but they seem try and avoid giving an opinion or choosing only one system by saying that they have “no emotional attachment to any electoral system”. First the authors say that electoral systems are best left alone. They contend that staying with what we know (however bad it may be) is probably better than going forth into the unknown. Yet, when speaking about a newly democratizing state, the authors hint at their being a better choice for small, multi-member constituencies, with some kind of proportional electoral formula. However, they still emphasize that it should be kept simple and there should not be too much sophistication with thresholds, adjustment seats, etc.
On the other hand, Lijphart (1994) does not seem to have any problems with particular electoral systems being too complex. He emphasizes the strengths of some of these features to include apparentement, two-tier districting, vote transferability, and national legal thresholds (Lijphart 1994: 145). Although Lijphart (1994: 151) does learn towards Taagepera and Shugart’s argument that in the case of existing electoral systems the preference should be for “incremental improvements, no revolutionary upheaval,” his advice for “electoral engineers in the new democracies” is to analyze all the options (1994: 152).
Following Lijphart and Taagepera and Shugart, Boix’s dependent variable is the effective electoral threshold, a measure that tries to capture the proportionality of the electoral system. Building upon Rokkan, Boix (1999) examines the choice of electoral systems in the set of countries that encountered at least a period of democratic government during the interwar years after the general introduction of adult suffrage. Change in electoral rules is composed as a strategic reaction by established elites to the changing political field. In conformity with this argument, older parties preferred the status quo until events persuaded them into thinking that current electoral laws permitted a new challenger to replace them. At this point they moved preemptively to endorse a reformed electoral system to minimize their political losses. When looking at it from this perspective, established parties operated strategically to maximize or at least to maintain their legislative strength in light of substantial changes to the distribution of preferences and number of political parties.
This line of reasoning is interesting in that it seems to point out the appropriate decision makers as well as to put forth a compatible and rational explanation of the variation one can observe in the choice of electoral rules. When established party leaders think that their grasp of majority support is decreasing, they wisely choose proportional representation. Or if they think their grasp of majority support is not decreasing, they keep with the single member district system.
Boix’s key contribution to studies of electoral systems is that instead of debating which type of electoral system is the best, we should just question how our view of electoral engineering alters when we admit that the those responsible for making changes are not benevolent planners. Boix essentially assumes that elites know and believe in Cox’s (1997) logic, and then asks what elites would do with this knowledge. However, this account of electoral choice as a predictable function of elite behavior depends on two implicit premises. It assumes that those elites had adequate information to be able to realistically estimate future voter preferences and the number of political parties. The function of uncertainty in the strategic calculations of party leaders is totally overlooked.
Rough Draft (very rough, very rough draft)
As a preface to the majority of this piece it is assumed that members of the electorate are rational creatures and as a result if given perfect information would be strategic voters. This paper tests the effects on the voter to make rational and strategic votes given the complexity of the electoral system in which they are a member of the electorate. It is noted that the strategic voter has not been proven in the masses by the field of political science but it is an accepted theory inside the academic field of political science. The author does acknowledge that this is a potential argument against this paper and its finding but with out this assumption no further work in the area could occur until that the assumption is proven true and that is very steep task that has yet to occur by the field but nor has it been disproved.
Different democratic nations utilize different styles of electoral systems and these systems vary greatly in their respective degree of complexity. To what extent is does the relative complexity of an electoral style affect the knowledge and understanding of the general electorate of what they are actually voting for and their understanding of the process of the election. As well how does the relative complexity of the electoral style affect the over trust and belief of the electorate that the elections are in fact fair and representative of the will of the people.
Nations and states vary greatly in their electoral systems from the most simple single member districts that operate in a pluralistic manner, in this system the voter votes for a single candidate and the candidate the receives the most votes on the first ballot wins even if a majority is not reached; to a proportional representation system with transferable votes that utilizes complex math equations to determine the vote quota required for a party or candidate to receive a seat in the legislature and then complex mathematical equations to divide up the remaining unoccupied seats and select winners for these seats. The voter in this model must select their multiple preferences for the legislature and rank them in order of their preferred preferences for whom they wish to be the electoral winners.
Does this increased level of complexity in the electoral system affect the ability of the voter to make informed and strategic decisions on how to best utilize their vote to better attempt to reach their legislative goals. Does the level of complexity that is imbedded into the electoral systems in an effort to increase the represenativeness of the legislature actually harm the represenativeness? Does this complexity harm the represenativeness of the results by lowering the ability of the electorate to make fully informed decisions for their legislature due to the face that they do not fully comprehend the electoral system and how the rules governing the selection of electoral winners work so they are unable to know how to best use their vote to impact their desired outcome.
Even in simple electoral systems such as the pluralistic majoritarian single member district as is found in the United States there is great confusion as to the impact of the electorates vote. This was very clearly evident in the 2000 Presidential election where many people where upset and confused by the set up of the Electoral College. Many in the electorate believed that they voted directly for the president and the man that won the majority of the vote was the winner, as had occurred in nearly all of the American presidential elections. These members were unaware of the true rules governing the election that instead of voting for the president directly they were voting for the slate of electors that their respective legislature would select to vote in the Electoral College. This small difference in what the electorate was actually voting for allowed for Albert Gore to win the popular vote but loose the White House to George W. Bush because Bush garnished more votes in the Electoral College and thus became the President. In a system that is not as easily laid out as the American system what effect does the complexity of the electoral system on the ability of the electorate to fully understand the electoral system and its rules, and be able to make strategies accordingly.
As the level of complexity that is associated with the electoral system increases so to does it become harder for the electorate to fully understand the electoral system and as a result we will see greater confusion on the part of the electorate. With an increase in the level of complexity we will see a reduced level of knowledge of how the electoral system operates and therefore a reduction in the level of strategic voting by members of the electorate will be expected. As one of the side effects of the complexity of the electoral systems we will expect to see that the nations must spend extra efforts to educate the electorate on the rules that govern the electoral system. These extended efforts to educate the electorate would take the form of public service announcements and an increase in the amount of time that is allocated to teaching civics in the education system of the nation. Lastly as the level of complexity that is imbedded into the electoral system increases the electorate will have a reduced understanding and as a result the proposed benefit of increasing the represenativeness does not occur by increasing the complexity of the electoral system.
This experiment needs to be survey based and will start with a survey of twenty five democratic nations that use a variety of different electoral systems from the majoritarian systems to several different proportional systems. Inside of every each nation a random sample of one thousand likely voters who have also voted in at least three of the last five elections will be chosen to make up the sample. This will guarantee that the respondents in the survey are the individuals that are active voters and can be expected to continue to vote. These repeat voters are also the voters that you would expect would be the most strategic in their voting and have the greatest understanding of the electoral system. This is to be expected because of their repeated use of the system and the fact that the voter is willing to sacrifice the effort need to be a regular voter.
There might be arguments with this style of selection because it discounts the general public and instead focuses on the more active voters in the nation; this has the likely potential of skewing the understanding of the electoral system by the respondent in an upward manner. But by having these selection restrictions in finding the respondents also reduces the chances of a selection bias by including those individuals that are not interested in the political system of the nation and therefore lack a strong working knowledge of the electoral system and also choose not to participate in the elections that occur in the country.
The survey would take place inside of an election year for the respective country; it should occur in a year when the lower house is to be elected if the nation is a bicameral system. Also the survey should be conducted in the year that holds the highest voter turnout in the nation’s election cycle for example the survey would be conducted in the United States in a year in which the president is to be elected. This is because the entire U.S House of Representatives is elected that year and this election draws far greater attention from the media and the election has much stronger voter participation than does the midterm elections. This will assure that the survey is conducted in a time when the electoral system is drawing the most attention from the electorate and the electoral rules and how to utilize their vote in a method that will best affect their respective interests are at the front of the respondents mind.
Two potential problems that might arise out of this section of the design are the fact that there might be an upward bias in the knowledge of the electoral system due to the increased media coverage given to the upcoming election, and that it might be physically challenging to conduct the survey in a timely manner under certain electoral systems. The increased level of awareness by the general public might bias the survey in an upward fashion, this is not a terrible problem due to the fact that it is the most prime situation to study the effects of the understanding of the electoral system based on the rules is when the electorate is most aware of the system and how it operates. It might be a challenge to conduct the study in an election year in certain nation due to the fact that some of the countries due not have regularly scheduled elections. This makes preparations to conduct the survey harder as well it might limit the amount of time in which the survey might take place due to period of time that is allotted from the time an election is called until when the election occurs. This could also bias the responses in a downward fashion in these nations because of the limited amount of time that the electorate has to be exposed to a coming election, as opposed to nations that have a regularly scheduled election and there is a gradual build up of attention is given to the election in the period prior to its scheduled date. There really is no way around this problem so its affects are acknowledged but can not be corrected for.
The survey would consist first of a simple civics examination, this exam would center on the electoral rules of that nation. This exam would be a multiple choice exam and the percentage of correct answers would be recorded to be compared cross-nationally. Then an additional section would be a discussion with the respondent about whom they intended to vote for in the upcoming election. This discussion would be recorded and would contain questions that pry into the motives of the individual to vote for that candidate and look for any voting strategy that might be behind the respondent decisions and motives. These responses would be reviewed and scored by a pool of three scorers trained to look for signs of strategic voting and the average of their scores would be used as the score for the respondent. There are several potential problems that could arise out this design. First having to adjust the civics exam for each nation might potentially bias the results but it is unavoidable because of the differences in electoral system would make the constant exam useless. Another potential problem that arises out of this design is the potential measurement error that could occur by the process of recording the respondent’s level of strategic voting.
Then the amount of energy that is devoted by the nation to educating the electorate about the electoral rules would be another area of the study. To do this the amount of election related public service announcements that occurred in the year prior to an election would be recorded and compared cross nationally. Also the amount of time that is dedicated to studying civics in the education system can be compared. This would be for the education of all students from the start of school to the end of secondary education but not inclusive of the university and professional and trade schools. This comparison would help to show the amount of energy that the state feels is necessary for it populace to become informed citizens and able to fully participate in the electoral system.
An obvious potential problem with this system is the possibility that different cultures place a different value on education and having a well informed citizenry. A nation that values this higher than another would be expected to invest a far greater amount of time and energy into the education of its citizens and might not necessarily be a function of the complexity of their respective electoral system.
Party system stability is crucial for the consolidation of democracy. “Where parties fail, it will hardly be a matter how efficient other institutions of state may have become. The new system will lack legitimacy and be vulnerable to takeover” (Innes, 2002). Tavits (2005) adds to Innes (2002) arguing, “when the existing party system is not strong or stable the potential for voter to be attracted to populist parties and demagogic leaders is considerably higher.” Parties are the vehicles that link the electorate to those who make decisions (Tavits, 2005; Roberts and Wibbels, 1999; Muller and Katz, 1997). Muller and Katz (1997) explores the party as a “linkage” arguing one way of conceptualizing a party is that parties provide a way for citizens to influence decisions, and provides for elites a way to communicate with the public.
When parties in government fail to produce this linkage, voters will turn their support and votes to other parties (Tavits, 2005). This shifting of votes is known as electoral volatility and is defined as: “the change in vote shares for individual parties across consecutive elections” (Tavits, 2005). The post-communist countries of Eastern Europe, compared to established western democracies, have experienced high levels electoral volatility since their transitions to democracy in the early 1990’s (Bielasiak, 2002). This high rate of volatility has lead to weak and unstable party systems in these countries. This essay explores electoral volatility in 12 post-communist Eastern European countries as determined by economic conditions. The question then is: What is the relationship between economic voting on electoral volatility in post-communist Eastern Europe?
The theory of economic voting is based on the assumption that voters cast their ballot with economic interests in mind. It employs the assumption of a rational voter as well. A rational voter is one who uses all the information available to the voter to make a decision and casts his preference at the ballot box. If the voter is using economic interests to cast his or her ballot, are they evaluating past performance of the economy and the decisions the incumbent government made, the retrospective voter; or are they evaluating the past performance of the economy but also with perceptions of how the economy will change in the future, and how does this affect volatility?
Theories of Economic Voting
The theory that economic conditions affect the voting behavior of the electorate is abundant throughout both comparative studies and American studies of voting behavior (Conover, Feldman and Knight, 1986; Pacek, 1994; Bohrer, Pacek and Radcliff, 2000; Duch, Palmer and Anderson, 2000; Duch, 2001; Tavits, 2005; also see Kramer 1971). This general theory of economic voting can be divided into two competing but related perspectives about how the electorate responds to the economy. The retrospective theory of economic voting posits that voters hold accountable incumbents for the past performance in managing economic conditions in a country (Conover, Feldman and Knight, 1986; Duch, 2001;; Kelly, 2003). This theory is based on the Downsian model of voting in which “rational citizens base their choices mainly on assessments of parties’ past performance, from which they predict parties future performance” (Weyland, 1998). They punish the incumbent through the ballot box either voting for the status quo, a vote for the incumbent, or against the incumbent, in a show of essentially no confidence. Kelly (2001) accurately describes this perspective in her assessment of economic and political accountability of Peru: “The retrospective link between economic and political assessment evokes the traditional understanding of accountability, wherein incumbents are held responsible for past actions and policies.” In multi-party systems, a number of parties can benefit from an increase in vote share, which comes in part from the decrease of the incumbents vote share (Tavits, 2005).
A second perspective of economic voting, what has been termed prospective economic voting, posits that the electorate holds the incumbent accountable for performance of the economy and also for past decisions that affect future economic performance (MacKuen, Erikson and Stimson, 1992; Duch, Palmer and Anderson 2000; Duch, 2001). From this perspective voters incorporate information from the past performance of the economy and current conditions that affect the future of the economy. “People develop economic expectations, which are formulated using retrospections as well as all other available information such as economic forecasts and economic indicators” (Kelly, 2003). The members of the electorate make calculated guesses about the future outcomes based on a plan for the future. “ Changes, such as political reforms and economic regulation or deregulation, may be incorporated because such change is viewed as incremental adjustments or refinements and do not upend the character of the political-economic system” (Cohen, 2004). Cohen (2004) argues that when uncertainty exists about the future voters are less likely to use prospective evaluations to judge government performance.
Previous Studies of Electoral Volatility
Electoral volatility in Eastern European party systems is especially common compared to the democratization periods of Western industrialized party systems, Latin American as well experiences higher rates of volatility (Roberts and Wibbels, 1999; Bohrer, Pacek Radcliff, 2000; Bielasiak, 2002). Bielasiak (2002) finds that “an open electoral marketplace has persisted” for post-communist regions. He argues voters’ favor the opposition parties because of economic transformations and social dislocation, though this is beyond the scope of his study of institutionalized party systems.
Tavits (2005) empirically tests economic factors as a cause for electoral volatility, specifically Gross Domestic Product (GDP) and the rate of inflation, along with institutional and sociological variables. She finds that the two factors exert a significant effect on electoral volatility with GDP causing a one percent decrease in volatility for every increase in GDP. Inflation does not exert as significant effect as GDP but it still makes a substantial contribution to the degree of volatility. She does not test for unemployment or an economic reform variable, which I believe, has a substantial impact on volatility. Roberts and Wibbels (1999) also find that GDP and inflation are significant indicators of electoral volatility, again not test for unemployment or economic reforms. A seminal work by Pacek (1994) tests economic conditions and voter turnout rates for four countries just after the fall of communism in Eastern Europe. He finds that higher rates of unemployment cause voters to switch their vote away from the incumbent. His sample size though is small and limited to elections that are not long after transition, and can therefore be expected to have high volatility as citizens. All the elections Pacek (1994) looked at occurred during and before 1992, Bielasiak (2002) argues that changes in electoral formula were confined to this period and founding elections took place in all four of these countries between 1991 and 1993 in the countries that Pacek (1994) included in his analysis.
Variables and Case Selection
To test the effects of the economy on electoral volatility a set of variables that capture retrospective voting and prospective voting and the economy as a whole are needed as well as a series of control variables to control for effects of other factors such as sociological forces and institutional effects. The dependent variable, electoral volatility, is calculated by the Pedersen index of electoral volatility. This measure measures the net change in votes from one election to the next(1). The independent variables I use are the rate of GDP growth rate, rate of inflation, unemployment, and the number of economic reforms the government has implemented during office. The GDP growth rate captures the rate at which the economy grew (or declined) during a political parties time in office. It will not be a measure of GDP growth rate at the time of the election but rather a change in growth rate while the incumbent party is in office prior to the election. Tavits (2001) voices some concern over economic variables and causality, the model I propose captures only the change during the tenure of the incumbent party. Therefore the causality is one way in that economic conditions will cause volatility. The rate of inflation also captures the economy by documenting the rate at which the price of goods goes up in a country. Again I propose only capturing the change during the tenure of the incumbent party. Unemployment provides a relationship to the economy in that it is the percentage of unemployed in the country. The number of economic reforms will capture the party in government’s attempts to reform the economy in a country. Control variables will include the effective number of parties in a system district magnitude. The countries in which the elections will be chosen from those that are classified as “partly free” or “free” in the Freedom House Freedom in the World Report. The report classifies countries as “free”, “partly free”, or “not free,” based on political rights and civil liberties. The data for elections and countries will come form the International Labor Organization, The University of Essex Database on Post Communist Eastern European elections as well as election results for the various countries from their respective election commissions. The countries to be included are: Armenia, Bulgaria, the Czech Republic, Estonia, Georgia, Hungary, Lativia, Lithuania, Moldova, Poland, Romania, Russia, Slovakia, Slovenia, Ukraine. I will look at the parliamentary elections for proportional representation systems, in the case of mixed systems; I will only look at those parties in the PR segment of the system.
Hypotheses and Methods
Given the nature of economic voting and volatility that we have discussed above several hypotheses can be made concerning electoral behavior in post-communist countries:
- 1. Countries with a higher GDP will have lower electoral volatility than those with a lower GDP.
- 2. As the performance of the economy increases, that is a GDP increase, electoral volatility will decrease.
- 3. If economic reforms are implemented volatility will be decreased, this is because the government in power is making strides to improve economic conditions. It is also a symptom of prospective voting, implying a sophisticated voter.
- 4. As unemployment decreases volatility will also decrease.
- 5. As the number of electoral reforms increases, electoral volatility will decrease.
(1)The Pedersen Index for electoral volatility is V= ½ S|vp,t-vp, t-1|. vp,t is the percentage of the vote obtained by a party at election t, vp,t-l is the percentage of the vote obtained by the party at the previous election. To calculate electoral volatility take absolute value of the difference between a party’s percentage of the vote at election t and percentage of the vote at the previous election t-1. Sum the absolute differences and divide by half. An alternative proposed by Sara Birch and quoted in Tavits is to divide by the “sum of the fractional shares of the total vote at each election of the parties which are included in the calculus” (Birch, 2001, 4 quoted in Tavits, 2005) I would calculate both and use the one that gave the most efficient results.
Bielasiak, Jack. 2002. “The Institutionalization of Electoral and Party Systems in Postcommunist States.” Comparative Politics 34:189-210.
Bohrer, Robert E. II, Alexander C. Pacek, and Benjamin Radcliff. 2000. “Electoral Participation, Ideology, and Party Politics in Post-Communist Europe.” The Journal of Politics 62:1161-1172
Cohen, Jeffrey E. 2004. “Economic Perceptions and Executive Approval in Comparative Perspective.” Political Behavior 26: 27-43.
Conover, Pamela Johnson, Stanley Feldman, and Kathleen Knight. 1986. “Judging Inflation and Unemployment: The Origins of Retrospective Evaluations.” The Journal of Poltics 48: 565-588.
Duch, Raymond M., Harvey D. Palmer, and Christopher J. Anderson. 2000. “Heterogeneity of Perceptions of National Economic Conditions.” American Journal of Political Science 44: 635-652
Duch, Raymond M. 2001. “A Developmental Model of Heterogeneous Economic Voting in New Democracies.” The American Political Science Review 95: 895-910.
Kramer, Gerald H. 1971. “Short-Term Fluctuations in U.S. Voting Behavior, 1896-1964.” The American Political Science Review 65:131-143
Innes, Abby. 2002. “Party Competition in Postcommunist Europe: The Great Electoral Lottery.” Comparative Politics 35: 85-104.
Kelly, Jana Morgan. 2003. “Counting on the Past or Investing in the Future? Economic and Political Accountability in Fujimori’s Peru.” The Journal of Politics 65: 864-880.
MacKuen, Michael B. Robert S. Erikson, and James A. Stimson. 1992. “Peasants or Bankers? The American Electorate and the U.S. Economy.” The American Political Science Review 86: 597-611.
Muller, Wolfgang C., Richard S. Katz. 1997. “Nominations and Reflections: Party as linkage.” European Journal of Political Science 31: 169-178.
Pacek, Alexander C. 1994. “Macroeconomic Conditions and Electoral Politics in East Central Europe.” American Journal of Political Science 38: 723-744
Roberts, Kenneth M. and Erik Wibbels. 1999. “Party Systems and Electoral Volatility in Latin America: A Test of Economic, Institutional, and Structural Exlanations.” The American Political Science Review 93: 575-590.
Travits, Margit. 2005. “The Development of Stable Party Support: Electoral Dynamics in Post-Communist Europe.” American Journal of Political Science 49: 283-298.
Weyland, Kurt. 1998. “Peasants or Bankers in Venezuela? Presidential Popularity and Economic Reform Approval, 1989-1993.” Political Research Quarterly 51: 341-362.
The notion of representation is central to democratic theory. Elected officials are expected to represent the interests of their constituents; the evidence of this representation includes the types policies the government enacts. From a normative standpoint, democracy sounds like a great idea. In practice, however, democracy is a terribly inefficient form of government.
Political parties are a convenient means to streamline the system of representation and make it more efficient (Holler 1987). Parties decrease voters’ information costs by providing them with short cuts or cues that allow for the quick evaluation and classification of political information. They organize and aggregate collective interests, consolidate the lines of political conflict, serve as mechanisms of representation, and directly connect the voters to the government through the process of elections, thereby increasing government accountability to the citizenry (Desposato 2006). Parties also increase the stability of democracies. Research demonstrates that democratic systems with weakly institutionalized parties are less stable and have lower rates of survival (Gosolov 2003). Just as parties are important for democracy, they are also important for politicians. Parties provide elected officials with a brand name, campaign workers and funds, and, once elected, afford officials with access to pork (Mayhew 1974).
Given the necessity of parties to democratic regimes, it is important to understand how individuals acquire their partisan attachments and how these individual-level attachments shape aggregate-level behavior. While much of our understanding of the development of party affiliations and impact of partisanship on mass behavior comes from the American literature, it may be that these explanations do not always translate well to other countries with multiparty systems. This paper examines the explanations for how partisanship develops in new democracies with multiparty systems. In particular, it highlights some the reasoning behind the argument that the explanations for the development of partisanship and its subsequent impacts on behavior in the
Parties, Partisanship, and Democracy
The concept of partisan affiliation has at least two implications for democracy. Since they represent the aggregation of the electorate’s interests, they are responsible for encouraging the electorate to participate. It was once thought that those who identified themselves as strong partisans were those most likely to vote and participate in party politics. However, it seems that the current trend is toward increased numbers of apartisans, individuals who are politically active but lack partisan attachments (
In spite of the decline of parties thesis, most political issues are still presented in partisan terms and individuals are able to use the partisan cue as a short cut for evaluating and classifying information (Ray 2003; MacDonald, Listhaug, and Rabinowitz 1991). In order to partisanship to be helpful in orienting an individual to the political world, parties must represent distinct and competing sets of issues (Desposato 2006).
Parties serve to connect the government to the electorate (Desposato 2006; Rueda 2003). They provide a mechanism for citizens to choose people who represent their interests to create the national policies by which they will live. When elected officials switch parties for their own political gain, they seriously undermine this relationship. According to democratic principles, political parties were not intended to be instruments of reelection but the apparatus of representation.
However, the meanings associated with partisan affiliation and the role of parties in the political system are not the same across all democracies. Too often, the tendency is for researchers to compare the dynamics of other democracies to those of American democracy. If this area of comparative politics research is to progress, it will be important for researchers to recognize that partisanship comes in a variety of forms, from identification to general political orientation. Also, the ways by which parties influence the public, either in terms of behavior or mass opinion, may be different in a multiparty context than a two-party system. Different contexts will produce not only different types of partisanship but different impacts by the parties.
Competing Explanations for the Development of Partisanship
Early explanations asserted than an individual’s partisanship could be understood as some type of psychological attachment to a political party. Within this framework, there are several explanations for how partisanship is acquired. First, Converse (1969) offers a temporal explanation in which partisanship results from socialization and experience. In terms of socialization, children are expected, at least initially, to mirror the partisan affiliations of their parents. This suggests that if an individual’s parents were Democrats, for example, that said individual would also be a Democrat and convincing that person to support the Republican Party would be no small task. Further, individuals “practice” being partisans when they vote for a particular party over the course of several successive elections or when they use the partisan cue to orient themselves to the political world. Using the partisan cue in this way causes partisanship to take on a sort of historical significance for individuals, implying that partisanship is temporally stable. Thus, partisanship is inherited from one’s parents, strengthened through practice, and stabilized over time (see also Green and Palmquist 1990). While the Converse model, like most models of partisanship, was developed in the context of the
An alternative to the temporal model, generational theories suggest that young adults will be very sensitive to the political events taking place when they first become politically aware. The way an individual perceives reacts to, and experiences these events will help determine his partisan affiliation. Over time, the impact of issues diminishes, individual partisanship stabilizes, and switching to a different partisan affiliation in response to new political issues is less likely (Niemi and Jennings 1991; Luskin, McIver, and Carmines 1989; Budge, Crewe, and Farlie 1976). In this sense, an individual’s reaction to his political environment determines his partisanship.
Taking a slightly different approach, Billingsley and Tucker (1987) assert that partisanship is the product of operant conditioning. Operant conditioning is a process through which behavior is shaped via a series of reinforcements or punishments. If an individual responds to some stimulus and is subsequently rewarded for his response, the likelihood of that response being repeated when confronted with that same stimulus increases. Similarly, if an individual is punished for his response, there is a decreased likelihood that he will respond in that way in the future. For Billingsley and Tucker, partisanship is formed during early adulthood, at which time individuals first have the opportunity to participate in politics. Also, for the first time, individuals are confronted with the relationship between political decisions and the performance of the economy. Partisanship develops through a series of responses to political events and the economic rewards or punishments that an individual experiences based on those responses. Individuals use partisanship learned in this way to make political decisions over the course of a lifetime.
[Explanation of partisanship as a running tally goes here—basically, Fiorina (1978) argues that rather than having a stable psychological attachment to one of the major parties, voters are constantly updating their partisan preferences in response to candidates, issues, and the changing political environment.]
Studying the Development of Partisanship in Multiparty Systems
It is unusual for a modern democracy to have a two-party system similar to that of the
In most cases, how individuals acquire particular partisan affiliations (or not) is studied using cross-sectional survey data. In particular, researchers at the
In general, there are two basic questions asked during the course of the NES that are designed to tap a respondent’ partisan attachments. First, the proverbial party identification question asks a respondent whether he thinks of himself as a Republican, a Democrat, an independent, or as an affiliate with some party. A second question is used to assess the strength of that individual’s partisanship. The respondent is asked to place himself along a seven-point scale ranging from “strong Democrat” to “strong Republican”.
Surveys examining political attitudes in other democracies contain questions aimed at measuring partisanship that mirror the partisanship questions on the American NES. For example, a typical party identification question asks survey respondents to choose, from a list of parties, the one with which they identify. “Independent” is always among the possible options. The party closeness question then asks respondents “which political party do you feel closest to?” After giving the name of the party, respondents are asked to indicate whether they feel “not very close,” “fairly close,” or “very close” to that party
Barnes et al. (1988) contend that the party closeness question is much more appropriate measure for cross-national studies of multiparty systems that the traditional party identification question. The authors argue that many people who classify themselves as “independents” when asked the party identification question only identify as independents when presented with that option. When those same respondents are presented with the party closeness question, they indicate that they feel close to one of the parties. It may be that some individuals feel “close” to one of the parties but do not possess the cognitive attachments to that party such that they self-identify as partisans. Additionally, the party closeness question is a better measure of partisanship because it asks for a spontaneous response, which enables researchers to separate the apoliticals from those who simply do not consider themselves partisans. Further, in multiparty systems, the party closeness question is much easier to administer than they party identification question. This is an example of one occasion where a measure used commonly in the study of American electoral behavior does not translate well to other democracies.
In Western European democracies, partisanship tends to be most stable among followers of old cleavage parties, like labor parties, social democrats, Christian democrats, and liberals.
Barnes, McDonough, and Pina (1985) argue in favor of a broader understanding of partisanship that goes beyond simple party identification and includes measures for attachment to political symbols that provide electoral guidance. This approach would be more appropriate for multiparty systems as it would be able to capture the individual partisan tendencies that would likely be overlooked by focusing on partisan identification only. They note that, in
While a great deal of research has been conducted on partisanship in the
Partisanship in Emerging Democracies
In this section, I will examine various explanations for the emergence of partisan attachments in new democracies. It is important to note that the explanations offered below are applicable only to a single country or a small group of countries.
1. Moser (1999)- Russia- Partisanship develops first at the elite level as a tool for winning elections and then is transmitted to the electorate as the number of elites identifying with a party grows over the course of several elections.
3. Brader and Tucker (2001)-
4. Powell (1976)-
Theory and Hypothesis
While there are a variety of explanations for the development of partisan attachments, the dominant account (at least in the context of the
The account of partisanship as a running tally should better explain the development of partisan attachments in new democracies. For example, Miller et al. (2000) argue that partisanship need not be a byproduct of childhood socialization (487). With the increasing availability of information, voters in former Soviet states are able to educate themselves about party platforms and identify the party which best represents their interests. As a result, they are beginning to develop stable partisan attachments. As Miller et al. demonstrate, partisan loyalties can develop in a relatively short span of time, an explanation that is more in keeping with the running tally model than the psychological attachment model. The running tally explanation is particularly attractive because it is likely generalizable to a broad range of countries. Earlier research regarding the development of partisanship in emerging democracies is limited in that there has not been a systematic attempt to apply a general theory to multiple countries.
Data and Methods
As of yet, I have been unable to find suitable data for testing the hypothesis that the running tally explanation for the development of partisanship accounts for partisan attachments in emerging democracies. As noted earlier, most earlier studies that examine this type of question focus on a single country. While a number of Western European democracies conduct national election studies and while the Eurobarometer contains questions that attempt to tap partisanship, new democracies are not included in these surveys. The data necessary to test the hypothesis would be composed of national election studies of former Soviet states and third wave democracies in
If the running tally explanation of partisan attachment holds for emerging democracies, this would suggest that party loyalties could develop rather quickly. Further, we would not expect aggregate partisanship to be stable, at least during the early years of democratic governance. If, however, aggregate partisanship is stable in emerging democracies, this might suggest that partisan attachment is based upon affect for political symbols that have endured over generations or social cleavages, although this paper does not explicitly test these claims.
 Viewing the party label as a “brand” with which voters are familiar is generally thought to be important in the context of campaigns and campaign strategy.
 Barnes, McDonough, and Pina (1985) contend that partisanship should not be defined is such limited terms. They argue that partisanship can also be understood in terms of tendance (a general partisan orientation that may exist independently of some continuing attachment to one political party) or even an individual’s affect toward political leaders (insofar as political leaders are viewed independently of their respective parties).
 Particularly their fathers.
 The categories from which respondents may choose are “strong Democrat,” “weak Democrat,” lean Democrat,” “independent,” “lean Republican,” “weak Republican,” and “strong Republican”.
 While similar, the two measures are not the same. However, unless one is interested in studying the group of individuals who self-identify as independents, the two measures can be treated as interchangeable.
 Green and Palmquist (1990) would disagree with this approach, arguing that only the party identification question is capable of capturing an underlying enduring trait.
Baker, Andy, Barry Ames, and Lucio R. Renno. 2006. “Social Context and Campaign Volatility in New Democracies: Networks and Neighborhoods in
Barnes, Samuel H., M. Kent Jennings, Ronald Inglehart, and Barbara Farah. 1988. “Party Identification and Party Closeness in Comparative Perspective.” Political Behavior 10: 215-231.
Barnes, Samuel H., Peter McDonough, and Antonio Lopez Pina. 1985. “The Development of Partisanship in New Democracies: The Case of