Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Politics. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 30 January 2023

The Economic Voter Decidesfree

The Economic Voter Decidesfree

  • Mary Stegmaier, Mary StegmaierDepartment of Public Affairs, University of Missouri
  • Michael S. Lewis-BeckMichael S. Lewis-BeckDepartment of Political Science, University of Iowa
  •  and Lincoln BrownLincoln BrownDepartment of Political Science, University of Missouri


In democracies, we elect our political leaders by choosing among a rival set of candidates or parties. What makes us pick one over all the others? Do we carefully weigh the platforms of all the candidates and then select the one closest to our personal desires? Or, do we select the candidate our friends and neighbors recommend? Perhaps, even, to save time, do we just vote for the same party we did last time? All of these are choice strategies, and there are many more. Here we focus on a well-known explanation of how voters decide, commonly called the Michigan Model, so named for the university where it was developed, in a path-breaking scholarly volume—The American Voter. The authors systematically gathered data, via scientific survey research, on individual voters in American presidential elections, measuring different traits, perceptions, and attitudes that they hypothesized might influence vote choice.

They arranged these different factors, or variables, into long-term forces and short-term forces that acted on the voter, and could be arrayed as if they were spread along a funnel of causality: from more remote, fixed variables, such as social class or party identification, to more proximate, fluid variables, such as issue preferences and candidate attributes. All these variables generally mattered, but those that concern us here deal with issues, in particular economic issues. How do voter evaluations of the economy help the voter decide what party to favor? Is it the national economy or the pocketbook that counts? How important are economic issues compared to other issues? What conditions make economic considerations more (or less) impactful? Does economic voting operate differently in different countries? These and other questions are addressed herein, with special attention to three leading democracies where economic voting has been heavily studied—the United States, Britain, and Germany. As demonstrated, economic considerations are pervasive and powerful elements in the democratic voter’s calculus.


  • Political Behavior
  • Political Economy
  • Public Opinion


Elections serve as a fundamental pillar of modern representative democracies. They provide a way for citizens to choose our leaders and call our representatives to account. Understanding how we, as citizens, form our political perspectives, evaluate government performance, and assign responsibility is of interest to anyone wanting to predict, change, or simply explain the outcomes of elections.

At first blush, such decisions might seem impossible to explain. As individuals, any decision we make is based on a combination of complex, often personal, factors. But when we study large samples of voters, predictable and steady patterns emerge. The foundational theories of how voters decide which candidate or party to support have been empirically tested over decades and in many countries.

The idea that the economy is a major influence on vote choice, government support, and electoral outcomes is a broadly supported proposition (Lewis-Beck & Stegmaier, 2000, 2007, 2013). Scholarship on the economic voter has a long lineage dating back to the 1950s and 1960s with classic contributions by Downs (1957), Campbell, Converse, Miller, and Stokes (1960), and Key (1966). In The American Voter, Campbell et al. (1960) included an entire chapter outlining how a person’s political choices were based on their evaluations of the economy. Downs (1957) proposed that voters were prospective and rational in their political decisions and made choices based on what a candidate could deliver in the future. Key (1966), on the other hand, argued that people reward or punish an incumbent based on past success or failure.

In the early 1970s, scholars began studying these relationships at the aggregate level using national measures to understand how the economy influenced presidential and government approval over time (Goodhart & Bhansali, 1970; Kramer, 1971; Mueller, 1973). But soon after, the research on how voters decide, using survey data, began to flourish. The expansion of national election surveys, like the American National Election Studies (ANES), that measure voter attitudes and intentions set the stage for a burst of international economic voting studies, which has resulted in an estimated 600 published articles (Lewis-Beck & Lobo, 2017).

This growing body of research has led to meaningful consensus on how and why the economy matters in vote choice. Stated simply, even when taking into account party identification and a host of other socio-demographic characteristics, the economy is a persistent and significant factor in voting decisions across democracies. Such economic evaluations can be measured in terms of the past or future national economy or one’s personal finances. In general, voters place greater emphasis on retrospective evaluations of the national economy when making their vote choice, while pocketbook evaluations and future expectations of the economy matter less (Lewis-Beck & Stegmaier, 2019). The impact of the economy on voting and election outcomes has been upheld in studies of parliamentary and presidential democracies, under single-party governments and coalitions, and during ordinary economic times and recessions, in nations around the world (Lewis-Beck & Lobo, 2017; Lewis-Beck & Whitten, 2013; Stegmaier, Lewis-Beck, & Park, 2017).

In this article, The American Voter's (Campbell et al., 1960) “funnel of causality” serves as the guiding theoretical framework to understand vote choice. Economic evaluations fit neatly within this framework and have a strong and predictable impact on vote choice, even when controlling for other theoretically relevant factors. The next section describes the “funnel of causality” and then moves on to look at contemporary voting behavior scholarship, with an emphasis on economic voting in the United States, Britain, and Germany. These three established democracies offer the opportunity to test the funnel of causality and economic voting in three quite different institutional contexts, allowing us to assess the generalizability of vote choice theory.

The Funnel of Causality

The field of modern election studies was launched with the publication of The American Voter. In it, Angus Campbell et al. (1960), working at the University of Michigan, put forth the funnel of causality “merely as a metaphor” to provide structure for the difficult task of explaining how voters decide (p. 24). (For a heuristic sketch of the funnel, see Lewis-Beck, Jacoby, Norpoth, & Weisberg, 2008, ch. 2) Nearly 60 years on, that metaphor has become a comprehensive model, sometimes referred to as the “Michigan Model,” for understanding the causal antecedents for vote choice in democracies around the world.

The funnel of causality, as structured originally by Campbell et al. (1960), established a causal flow from more distant to more immediate factors that shape the vote. The more remote factors tend to be stable characteristics and attachments of voters. They shape citizens’ perspectives and gain relevance after being translated though more short-term, election-specific factors.

The most distant set of factors, the voter’s sociodemographic characteristics, are located at the mouth of the funnel. These typically include gender, race, education, religion, and income. Next, is party identification—a person’s long-term psychological attachment to a political party. Moving closer to the tip of the funnel, and ultimately the vote choice, are issue positions and then finally evaluations of candidates. To think about this in practice, imagine a person who reports identifying as an Evangelical Christian. This is a socio-demographic characteristic that flows through the “funnel” to impact party identification; party identification shapes attitudes on issues such as abortion laws; issue positions influence how the person evaluates the candidates running for office; and these candidate assessments ultimately lead to the voter’s candidate choice.

Scholars using the “Michigan Model” to analyze individual-level vote choice in a particular election will include measures capturing all or some of these parts of the funnel. In some countries, because political parties have developed to represent certain groups of people, including party identification captures much of the effect of sociodemographic characteristics. For example, often parties on the left represent the working class and union members, while parties on the right represent the interests of owners and managers. Likewise, some parties specifically represent the interests of religious or ethnic groups.

According to Campbell et al. (1960), party identification is first acquired through socialization—what we learn from family, education, peers, and religion. Thus, socio-demographics play an important role in how people develop their party identification. And, while people sometimes change their party identification at critical points in their lives, for many, party identification remains quite stable.

The concept of party identification works well in the American political context, but in many other democracies, citizens do not think of parties in the same way. Instead, political ideology, whether someone adheres to a liberal or conservative perspective, is commonly included with or in lieu of party identification as a long-term political anchor in a vote choice model.

The most proximate factors to the vote choice concern political issues and the candidates who are running for office. Voters will evaluate these from the perspective of their identity—their sociodemographics and their partisanship. The types of issues included in a voting model depend on what issues are salient in politics at that time. For example, there are times that attitudes about immigration policy would be important to include because the candidates have repeatedly highlighted their differences on the issue, while in other elections, the topic might not be discussed among the candidates, and thus it would not help us explain vote choice.

This article will primarily focus on the economy as a perennial issue which is regularly included in voting models and looms large in its impact on the vote choice near the tip of the funnel. Campbell et al. (1960) originally recognized the importance of the economy for vote choice, writing “[T]he crucial questions are: (1) Is a person’s economic outlook associated with his partisan choice between candidates? (2) Is a person’s economic outlook associated with his level of political participation? The answer to both of these important questions is substantially affirmative. . . . [I]t would seem that economic outlook must be placed, under our theoretical scheme, antecedent both to political behavior and partisan attitudes, but between them and most of the more remote factors” (pp. 397, 399).

Even though The American Voter laid the theoretical groundwork for the various dimensions in which the economy enters into a citizen’s vote choice, this went largely unrecognized for decades (Lewis-Beck & Stegmaier, 2009). Since the 1980s, the most commonly used measures of economic evaluations ask survey respondents to evaluate the national economy (sociotropic) and their personal financial situation (pocketbook) over the past year, as well as their expectations for the national economy and their personal financial situation over the coming year. These evaluations, if positive, are expected to result in a more favorable view toward the incumbent candidate or party. If these evaluations are negative—for example, if the respondent believes the national economy has gotten worse over the past year—the voter is less likely to vote for the incumbent.

At the very tip of the funnel, the most immediate factor appears to be candidate evaluations: how voters evaluate the candidates or party leaders, whether they like and trust them, whether they have been satisfied with their performance in office, and whether they agree with them on issue positions, ultimately determines how voters decide. Because candidate evaluations are so closely tied to vote choice, this layer is sometimes omitted from models of vote choice. This is because the factors that help us understand candidate evaluations also help us understand how voters decide.

It is important to note that there are other approaches to studying vote choice. One approach that has gained in prominence is the rational choice perspective, which focuses on voters as self-interested actors. Historical and institutional approaches consider the impacts of events and political realignments, registration and voting laws, and governance and constitutional structures across countries on how voters decide. Geospatial, cognitive, and psychological, as well as media, approaches have also presented fruitful avenues in which to look at voter choice (Lewis-Beck et al., 2008).

Ultimately, the funnel of causality is important because of its focus on long-term and short-term influences on the vote decision. Applying the framework in different countries and at different points in time shows both similarities and differences in the magnitudes of the forces in the funnel. In particular, much research has focused on the size of the economic vote and the circumstances that lead to strong or weak economic voting. These differences can often be understood by variations in countries’ political and economic contexts (Stegmaier et al., 2017). In the following country-specific sections, a recent study that uses the funnel of causality framework to analyze vote choice is highlighted, placing it within the broader context in which voters were making their choices. Each section then discusses related research that examines different angles on economic voting in that country. This approach makes possible meaningful comparisons and conclusions on how voters make their decisions in contemporary democracies.

The United States

The United States represents the most studied case in the field of voting behavior. The institutional context—a presidential system with two parties—makes it one of the most straightforward cases in which to analyze the vote decision. It is the country that Campbell et al. (1960) studied and, along with the United Kingdom, was the primary focus for most of the early research on economic voting (Lewis-Beck & Stegmaier, 2007, 2009).

A 2016 study by Lewis-Beck and Stegmaier set out to test the funnel of causality using ANES survey data and the Latino Immigrant National Election Study (LINES) for the 2012 presidential election. While the aim of the study was to assess whether the political choices of Latin American immigrants could be explained in the same way as the decisions of native-born Americans, this article provides a detailed recent example of the funnel of causality in action.

Lewis-Beck and Stegmaier (2016) use a block-recursive model to analyze the contributions of each layer of the funnel. This means that the first model includes the most distant characteristics—sociodemographics. The second model includes sociodemographics and adds partisanship and ideology. The third model builds on the second one, adding issue positions and economic evaluations. The final, full model adds candidate evaluations. This particular modeling strategy not only illuminates the impact of various factors on the vote decision but also explains how much each part (or block) contributes to explaining vote choice.

The dependent variable in these models, the thing to be explained, is presidential approval (how much a person approves/disapproves of President Obama’s performance in office), rather than vote choice itself. The reason for this is that Latin American immigrants overwhelming said they would vote for Obama, meaning there was very little variation in vote choice left to account for (Lewis-Beck & Stegmaier, 2016, p. 167). Thus, to create comparable models for the American public as a whole and Latin American immigrants specifically, the authors use presidential approval. This is similar to the strategy of using “popularity functions” in aggregate-level studies when data limitations prevent “vote functions” that use election results over time (Lewis-Beck & Stegmaier, 2013; Stegmaier et al., 2017).

So, what do Lewis-Beck and Stegmaier (2016) find? Let us look first at results from the general model, as estimated on a representative national sample of American voters in the 2012 presidential election. In the most distant block of variables, sociodemographics, they examine each respondent’s gender, marital status, home ownership, age, education, social class, income, church attendance, health insurance, and race/ethnicity. Of these, race/ethnicity, home ownership, marital status, church attendance, and gender have statistically significant impacts on presidential approval. Women and minorities were more likely to approve of Obama, while those who attend church more often, are married, and own a house were less likely to approve of the president. According to the adjusted R-squared, sociodemographics alone account for nearly 20% of the variation in presidential approval among the American voting public.

Adding party identification and ideology to the model doubles the adjusted R-squared. These two variables alone, both of which are statistically significant, provide a powerful boost in explaining presidential approval.

The next block of variables includes issues and economic evaluations. In the 2012 election, prominent issues included citizenship for illegal immigrants, gay marriage, the death penalty, abortion, and increasing government services. Of these, respondent positions on increasing government services, gay marriage, and the death penalty attained statistical significance, while abortion and citizenship for illegal immigrants appears to have not shaped vote choice in a systematic manner. The standard battery of economic evaluations was included in this model as well: retrospective pocketbook, retrospective national economy, prospective pocketbook, and prospective national economy. All four were statistically significant, but the retrospective national economic evaluation variable dominates the others. The regression coefficient is over five times larger than the next most powerful factor, retrospective pocketbook evaluations. By adding issue positions and economic evaluations to the model, the fit of the model improves to explain 50% of the variation in presidential approval (Lewis-Beck & Stegmaier, 2016, pp. 177–178).

This finding that retrospective sociotropic (national) evaluations matter the most among the economic assessment measures is mirrored in study after study in the U.S. case and in many other democracies. It tells us that voters are evaluating the performance of the incumbent party and rewarding or punishing the party accordingly. We may ask: what goes into this national economic evaluation? As Nadeau and Lewis-Beck (2001) have written elsewhere, “A plausible assumption is that a voter looks at more than one indicator, and places different weights on them” so arriving at a summary evaluation of the economy as “better,” “worse,” or the “same” (p. 160). For example, national measures of earnings, unemployment, inflation, growth, and real disposable income yielded an R-squared of .88 when regressed on a comparable subjective index of national business conditions (Nadeau & Lewis-Beck, 2001, p. 167). Thus, the sociotropic retrospective measure well represents a comprehensive amalgamation of a voter’s view of the economy.

At the tip of the funnel, adding candidate evaluations via feeling thermometer ratings of Obama and Romney, both of which are statistically significant, completes the causal picture. The adjusted R-squared for this full model is .76. Thus, the components of the funnel of causality explain nearly three-quarters of variation in presidential approval, as measured in the general American voting public in 2012. This is remarkable, given how difficult it is to explain or predict how individuals think and behave.

This special 2012 study also examined, in parallel, Hispanic immigrant voters. As immigrants, the process of political socialization (e.g., acquiring party identification and understanding the salient issues in American political life) starts at a later point in their lives—at the time they arrive in the United States. But despite their shorter experience with the U.S. political system, immigrants are assimilating and developing partisanship and views on political issues that shape their support for the president. Comparing Hispanic immigrants to the cross-section of American citizens in the ANES, sociodemographic characteristics and party identification have a weaker impact in structuring presidential support among the immigrant voters. But when it comes to issues, the effects are somewhat stronger for immigrants compared to the American sample. Moreover, Hispanic “immigrants come close to being single-issue voters, in that it is mainly the economy (in its multiple dimensions) that moves them” (Lewis-Beck & Stegmaier, 2016, pp. 178–179).

The economic vote in both the U.S. citizen sample and the Hispanic immigrant sample appears to dominate the issue agenda of these voters. This finding is not really surprising, falling in line with earlier economic voting research. As observed some years ago, “While economic concerns, then, may fall under the rubric of ‘issue voting,’ they certainly represent the premier issue set” (Lewis-Beck, 1988, p. 160).

In the United States, voters hold the president accountable for economic conditions rather than Congress. The mechanism by which the economy plays a role in congressional elections is through accountability via the party of the president. For example, if a Republican is president, then Republican members of Congress running for reelection will feel the same directional effect of credit or blame as the president. Notably, this affect appears for incumbent members of Congress from the president’s party, but it does not exist in open-seat races (Lasley & Stegmaier, 2001).

So, what happens during times of divided government when one party controls the presidency and another controls Congress? Divided government introduces complexity into the voter’s decision, since economic policymaking requires cooperation between Congress and the president. How does the American public sort this out?

Norpoth (2001) looked at how Americans assign credit and blame for economic conditions under a divided government using survey data from the 1992 and 1996 elections. Both elections featured divided governments, but under different parties. In 1992, President George H. W. Bush was the incumbent Republican president running for reelection and the House of Representatives was controlled by the Democrats. In 1996, the tables were turned, with President Bill Clinton running as the incumbent Democratic president while the House of Representatives was controlled by Republicans. Norpoth’s models of individual vote choice for the presidency and House account for party identification, ideology, issue attitudes, and the retrospective evaluation of the national economy. In both elections, voters held the president accountable for the economy and “administered vengeance and rewards to House candidates only by partisan association with the White House occupant” (Norpoth, 2001, p. 426). There was no evidence that a divided government was too complicated for Americans to assign responsibility, nor was there evidence that voters split credit or blame between the president and the House. Ultimately, Americans hold the incumbent president, and not Congress, accountable at the ballot box for economic conditions.

There are times when the assignment of responsibility to the president for economic conditions is diminished. First, when an incumbent president is running for reelection the economic effect appears to be strong, but when the sitting president is not contesting the election, there is no impact (Norpoth, 2002). In these elections, the public cannot punish/reward the incumbent president, since he is not running for reelection. The candidate running from his party seems to be insulated from the public’s reward/punishment, at least retrospectively. However, there is some evidence that “In such a circumstance . . . economic voting is more likely to be prospective” (Nadeau & Lewis-Beck, 2001, p. 174).

A second instance of when presidents are less likely to be held accountable can be seen in presidential approval ratings during the first year of a new presidency. Stegmaier and Norpoth (2018) analyzed monthly unemployment and presidential approval data during the first year of presidencies from Truman to Trump and found that the expected relationship between the economy and approval holds for some presidents and not for others. One reason for this is that the American public still attributes responsibility for the economy during that initial year in office to the previous president and his policies. Thus, the public is savvy and understands that policy effects carry over from one presidency to the next and that it takes time for a new president to impact the direction of the economy.

The economic voting work reviewed here covers a long period of time, from the 1950s to 2018, when this article was written. In terms of explaining American voting behavior, how much has changed? First, the question of whether the Michigan Model still holds merits consideration. The American Voter Revisited, a comprehensive volume of almost 500 pages, tackles the question: “Does the social-psychological explanation of presidential vote choice, developed in The American Voter and symbolized in its famous theory of ‘the funnel of causality,’ still hold? The essential answer drawn from our revisit must be yes” (Lewis-Beck et al., 2008, p. 427). But, what about the more particular question of the impact of the economic vote? Here stands their response: “it need not take much cognitive work for voters to figure out how to vote when the economy is roaring . . . economic issues retain their impact” (Lewis-Beck et al., 2008, pp. 425, 427).


The components of the funnel of causality contribute to vote choice in the United States, but how well do they perform in other democratic contexts? The British institutional context has both similarities and differences with the United States. Britain is a parliamentary system where citizens vote in single-member districts for their member of Parliament, and ultimately the party that wins the majority of seats holds the prime ministership. Thus, Britons are not voting directly for the executive, as in the United States. Yet, the single-member district electoral system, as Duverger (1954) predicted, has meant that Britain has two dominant parties, Labour and Conservative, though regional parties and smaller parties hold some seats in Parliament. The dominance of the Labour and Conservative Parties has typically resulted in one party commanding a parliamentary majority, making it clear to voters which party to hold responsible for current conditions at election time.

In an article examining the relationship between the economy and electoral support, Whiteley, Clarke, Sanders, and Stewart (2014) studied the impact of the economy and economic strategy on voter intentions using pooled data from the Essex Continuous Monitoring Survey, 2004–2013. The study fits well within the funnel of causality framework because it includes most of the common factors including sociodemographic controls, partisanship leanings, perceptions of the national economy and personal finances, and leader evaluations.

The time period of the Whiteley et al. (2014) study includes a Labour majority government (2004–2010) and the first coalition government in Britain since World War II. No party won a majority of parliamentary seats in the May 2010 election, so the Conservative Party, under the leadership of Prime Minister David Cameron, entered a coalition agreement with the Liberal Democrats. These two different types of government—a one-party majority and a coalition government—provide the opportunity to test whether economic evaluations impact vote intention differently in different contexts.

Voter intention models under the period of Labour government control explained essentially two-thirds of voter intention variation. The economic variables measuring voter optimism on the national economy (sociotropic) as well as their personal financial outlook (pocketbook) were significant and of the expected signs. Respondents who were more optimistic were more likely to vote for the incumbent Labour Party and less likely to vote for the opposition Conservatives and Liberal Democrats.

Other variables like partisanship and leadership evaluations were also significant and of the expected signs meaning that Labour and Conservative partisans were more likely to vote for their party and for their party’s leaders if they held favorable opinions of them and were less likely to vote for the opposition. Some demographic variables were significant, such as age, education, ethnic minority, and Scottish residency, while others, like employment status, occupation, and gender, were not.

Under the Conservative and Liberal Democrat coalition government that took power after the 2010 general election, economic evaluations are more muddled in predicting vote choice. In this period the economic variables for national and personal optimism are insignificant. However, a voter’s expressed view that no party could best manage the economy is a significant predictor of voting against Conservatives while parties outside the coalition show gains. Similar to the other models with Labour in control of government, evaluations of party and party leadership, and the sociodemographic variables are still significant and follow the expected patterns for Conservative and Labour support.

While national economic optimism was increasing during the period after the election of the coalition government, this did not increase support for the coalition parties. Whiteley et al. (2014) propose two explanations. First, during this period there was a perception that no party could effectively deal with Britain’s economic woes, and second, the improving national economy had not translated into positive evaluations of the personal (pocketbook) economy. At the same time, Labour was experiencing only marginal gains in voter intention, which Whiteley et al. (2014) attributed to the fact that the coalition parties were still using the recession as an issue with which to attack Labour.

This highlights some of the complexities in how economic evaluations get translated into party support depending on the different institutional and situational environments. Indeed, under Britain’s coalition government, it was likely difficult for citizens to figure out which of the two coalition parties to hold accountable. Coalition governments, which we discuss in more detail in the section “Germany,” make the assignment of responsibility for outcomes more difficult than when just one party governs alone (Lewis-Beck, 1988; Powell & Whitten, 1993).

While most studies examine the vote choice of all voters in one model, Palmer, Whitten, and Williams (2013) consider how voters in different income groups have different assessments of economy management by party leadership. Using the Continuous Monitoring Survey from 2004 to 2010, they look at whether voters who are high, middle, and low income respond to different economic signals. The authors find that low-income voters are more responsive in their assessment of the incumbent Labour government based on evaluations of unemployment, while high-income voters are more responsive to inflation. This demonstrates that in the same economic setting, different income groups make different evaluations of economic leadership.

Another British study, by Tilley, Neundorf, and Hobolt (2018), takes a closer look at pocketbook voting, using data from the British Household Panel Survey for the years 1991 through 2008 covering both the Conservative government from 1991 to 1997 and the Labour government from 1997 to 2008. The study’s authors found that “[n]o matter the color of the government, when individuals’ household finances improve, they reward the governing party and punish the opposition; and when their household finances deteriorate, they punish the governing party and reward the opposition” (p. 560). The authors also found support linking perceptions of government transfers with perceptions of personal finances. If voters believed their financial situation improved and attributed that improvement to welfare transfers by the party in power, they were more likely to support that party. Conversely, if voters believed their finances were hurt by the policies of the ruling party, they were more likely to support the opposition.

In Britain, then, voters make their decisions much the same way as American voters. Per the “funnel of causality,” their demographic characteristics, socioeconomic status, and partisanship matter, but economic issues weigh heavily on their choice, especially when parliament is led by one party.


Germany presents a more complicated context in which to study vote choice compared to the United States or Britain. Like Britain, it is a parliamentary system with the chancellor, the equivalent of the British prime minister, heading the government. The mixed electoral system in Germany—a combination of single-member districts and proportional representation—affects the strategy voters use in casting their votes (Gschwend, 2007) and has produced a multiparty system, without parties strong enough to command a majority of seats alone. Thus, parties must join together to form a majority governing coalition. This makes the vote decision more complicated for voters. If two or more parties govern together, do voters credit/punish the coalition as whole, or is one party held more accountable than others?

The question of vote choice in multiparty systems had, for many years, focused on the choice between the government and the opposition. In other words, a voter voted for a party within the coalition or a party in the opposition (Goergen & Norpoth, 1991; Lewis-Beck, 1988; Lewis-Beck & Nadeau, 2012). With this setup, the statistical modeling could proceed just like an analysis of vote choice in a two-party system. What political scientists observed, however, was that the effects of the economy on vote choice were sometimes muted, or weaker, than in the United States or United Kingdom.

Coalition governments represent a clear case where “clarity of responsibility” is lacking (Powell & Whitten, 1993). Lewis-Beck (1988, pp. 108–109) recognized this phenomenon in his study of vote choice in the major Western democracies. The economy had the strongest impact on vote choice in the United Kingdom, where one party controlled government; it had the weakest impact in Italy with five parties in the 1983 governing coalition. Germany, with its two-party government, fell between the two extremes.

So, what happens when the public’s assignment of responsibility is examined separately for the parties within the governing coalition? Debus, Stegmaier, and Tosun (2014) use the German Elections Studies survey data to look at vote choice in six German parliamentary (Bundestag) elections between 1987 and 2009. Each of these elections featured a two-party coalition government but with different party configurations.

Debus et al. (2014, p. 56) acknowledge the work of Campbell et al. (1960) in shaping their approach to studying vote choice in Germany, which guides their selection and construction of variables. Their vote choice models include the following set of variables: church attendance, labor union membership, ideological distance between the voter and the parties, national economic evaluation, assessment of which party is most capable of solving the country’s most important problem, and preferred chancellor. In this model, the funnel of causality is at work: sociodemographics (church, unions), ideology as a substitute for party identification, issues (economic evaluation and which party is best on the most important issue), and candidate evaluations.

The results are clear across each of these elections—elections that featured different coalition configurations (Christian Democrats–Free Democrats; Social Democrats–Greens; and a grand coalition of the Christian Democrats–Social Democrats). The church, union, ideology, problem-solving, and chancellor preference measures were nearly always statistically significant and in the expected direction, meaning that they contribute to the explanation of vote choice for both the chancellor’s party and the junior coalition member. For example, voters who believed that the Green Party, which served as a junior coalition member in the early 2000s, was best able to solve the most important problems were likely to vote for the Green Party. It didn’t matter that the Greens were the junior partner in government. But, the important distinction in the results between the chancellor’s party and the smaller coalition member is with the economic evaluation variable. This variable is always statistically significant for the chancellor’s party in the expected direction (people who evaluated the economy positively were more likely to vote for that party), but for the smaller coalition member, the results in all six elections were either insignificant or in the wrong direction. For example, in 1987, 1990, and 1994 elections, when the Free Democrats (FDP) were the junior coalition member, they were punished by voters who believed the economy had improved!

One of these German elections, in 2009, featured a grand coalition—the term used for a coalition between the country’s two largest parties: the Christian Democrats (CDU/CSU) which held the chancellorship and the Social Democrats (SPD). It is in this election that the effect of economic evaluations on vote choice was weak. The chancellor’s party reaped some reward from those who evaluated the economy positively, but compared to the other elections, the coefficient is the smallest. Additionally, unlike in the earlier elections, the impact of the economy on voting for the coalition partner is statistically insignificant. The dynamics of the two large rival parties in coalition together seems to alter the assignment of economic responsibility, reducing the magnitude of the economic vote.

The Debus et al. (2014) study demonstrates quite conclusively that the chancellor’s party is held accountable by voters. But what about the other parties not in government? If voters are shifting their support, are they shifting within the coalition or are they opting for opposition parties? Williams, Stegmaier, and Debus (2017) examine this question at the aggregate level using monthly party support for the CDU/CSU, SPD, FDP, and the Greens from 1993 to 2011. Their modeling approach allows them to assess how conditions affect the different parties simultaneously, since increasing support levels for one party mean that at least one other party is losing support. What they find is that the two large parties, the CDU/CSU and SPD, are more affected by economic evaluations than the smaller parties, the FDP and the Greens. For the large parties, when one held the chancellorship, it reaped rewards for improved economic evaluations and was punished when the evaluations worsened. When one of the large parties was in opposition, the impact of the economy was the opposite. As would be expected, the large opposition party benefited in support when evaluations worsened and lost support when the economy improved.

The shifts in support from the party of the chancellor and the large opposition party exist when the CDU/CSU and the SPD are split between the government and the opposition. Matters get more complicated when the two parties are joined in a grand coalition. The grand coalition included in this study is the one that existed from 2005 to 2009, led by Chancellor Angela Merkel of the CDU/CSU, with the SPD as the coalition partner. Under this configuration “the chancellor’s party is the sole beneficiary of the rewards from positive economic evaluations. The SPD, as the coalition partner, gets an initial small boost which rapidly deflates” (Williams et al., 2017, pp. 293–294). Thus, governing with the chancellor’s party does very little to enhance a party’s support.

Plescia and Kritzinger (2017) focus specifically on the 2013 German election, where the CDU/CSU was in coalition with the FDP. Like Williams et al. (2017), they are interested in how voters make their vote decision—a vote for one party—when they simultaneously have more than one party in government and more than one party in opposition. Which coalition party do voters credit for improved economic conditions? And for voters who are dissatisfied with the government’s performance and turn away from the coalition parties, how do they choose from among the opposition parties?

Ultimately, the Plescia and Kritzinger (2017) findings buttress the findings of Debus et al. (2014) and Williams et al. (2017). They find that “electoral accountability occurs at the party level with party size playing an important role in responsibility attribution to single parties, regardless of whether they were in government or opposition. It is not only the prime minister or chancellor’s party which is affected the most by retrospective voting . . . but also the largest opposition party” (Plescia & Kritzinger, 2017, pp. 157–158).

From these German voting studies, we see the importance of looking beyond the government/opposition vote dichotomy. Not all parties within the coalition are treated the same by voters, nor are all opposition parties. In Germany, voters appear to target the two largest parties, especially when one leads the government and the other is in opposition. While many other democracies have periods of coalition governments, the number and ideological positions of parties within the coalition, and the electoral viability of opposition parties (Rowe, 2013), might produce different results. The party of the prime minister would be expected to experience credit/blame for economic conditions, but the effects on junior coalition parties and the different opposition parties could differ across political contexts.

When studies focus on voting for specific parties, in particular support for the chancellor’s party, the results show that theoretical framework of The American Voter holds. German vote decisions are shaped by their personal characteristics, ideological position, and economic issues. Voters use their evaluation of the economy to punish or reward the party of the chancellor.


In this review of economic voting in three leading Western democracies, the economy, much like water running downhill, finds its way to the stream, in this case the stream of votes shaping governments. In these nations, when voters perceive an economic downturn, they tend to vote against the ruling party and, in opposite fashion, when they perceive an economic upturn, they tend to vote for the ruling party.

Does this classic reward–punishment pattern of economic voting hold in other democracies? Yes is the short answer. It holds, for example, in the complex democracy that is France, with its many, and frequently changing, political parties (Lewis-Beck, Nadeau, & Bélanger, 2012). It holds, as well, in the small, globally open political economy of Denmark, with its frequent minority governments (Stubager, Botterill, Lewis-Beck, & Nadeau, 2014). Further, the economic vote presses hard in times of crisis, appearances sometimes to the contrary, as the Spanish case illustrates (Fraile & Lewis-Beck, 2014). Of course, these are all established democracies. However, the economic vote does not stop at that water’s edge, so to speak. In the transitional democracies it shows steady presence (Lewis-Beck & Stegmaier, 2008). For Latin American democracies, an object of special study, the economic vote also makes itself strongly felt (Lewis-Beck & Ratto, 2013).

In fact, the effects of the economic vote appear sharp and persistent around the democratic world. That proposition sustains itself, overcoming supposed difficulties of accurate economic perceptions, or the blurring of the clarity of responsibility (see, respectively, Nadeau et al., 2013; Dassonneville & Lewis-Beck, 2017). It has also overcome the concern that economic evaluations are biased by partisanship (Lewis-Beck, Martini, & Kiewiet, 2013; Lewis-Beck, Stubager, & Nadeau, 2013). Further, it actually functions in two neglected dimensions of economic voting, patrimony and policy. (For a relevant example, which combines all three dimensions of the economic vote, see a recent investigation of the 2010 general election in Britain (Lewis-Beck, Nadeau, & Foucault, 2013).) Obviously, the several electoral rivulets flowing from the economy remain to be tapped by enthusiastic students and future researchers.

Further Reading

  • Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York, NY: John Wiley & Sons.
  • Lewis-Beck, M. S. (1988). Economics and elections: The major Western democracies. Ann Arbor: Michigan University Press.
  • Lewis-Beck, M. S., Jacoby, W. G., Norpoth, H., & Weisberg, H. F. (2008). The American voter revisited. Ann Arbor: University of Michigan Press.
  • Lewis-Beck, M. S., & Lobo, M. C. (2017). The economic vote: Ordinary vs. extraordinary times. In K. Arzheimer, J. Evans, & M. S. Lewis-Beck (Eds.), The Sage handbook of electoral behavior: Volume 2 (pp. 606–630). London, UK: SAGE.
  • Lewis-Beck, M. S., & Stegmaier, M. (2000). Economic determinants of electoral outcomes. Annual Review of Political Science, 3(1), 183–219.
  • Lewis‐Beck, M., & Stegmaier, M. (2007). Economic models of voting. In R. J. Dalton & H-D. Klingemann (Eds.), The Oxford handbook of political behavior (pp. 518–537). Oxford, UK: Oxford University Press.
  • Lewis-Beck, M. S., & Stegmaier, M. (2008). Economic voting in transitional democracies. Journal of Elections, Public Opinion and Parties, 18(3), 303–323.
  • Lewis-Beck, M. S., & Stegmaier, M. (2013). The VP-function revisited: A survey of the literature on vote and popularity functions after over 40 years. Public Choice, 157(3–4), 367–385.
  • Stegmaier, M., Lewis-Beck, M. S., & Park, B. (2017). The VP-function: A review. In K. Arzheimer, J. Evans, & M. S. Lewis-Beck (Eds.), The Sage handbook of electoral behavior: Volume 2 (pp. 584–605). London, UK: SAGE.


  • Butler, D., & Stokes, D. (1969). Political change in Britain. London, U.K.: Macmillan.
  • Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter. New York, NY: John Wiley & Sons.
  • Dassonneville, R., & Lewis-Beck, M. S. (2017). Rules, institutions and the economic vote: Clarifying clarity of responsibility. West European Politics, 40(3), 503–515.
  • Debus, M., Stegmaier, M., & Tosun, J. (2014). Economic voting under coalition governments: Evidence from Germany. Political Science Research and Methods, 2(1), 49–67.
  • Downs, A. (1957). An economic theory of democracy. New York, NY: Harper & Row.
  • Duverger, M. (1954). Political parties: Their organization and activity in the modern state. London, U.K.: Metheun.
  • Fraile, M., & Lewis-Beck, M. S. (2014). Economic vote instability: Endogeneity or restricted variance? Spanish panel evidence (2008 and 2011). European Journal of Political Research, 53(1), 160–179.
  • Goergen, C., & Norpoth, H. (1991). Government turnover and economic accountability. Electoral Studies, 10(3), 191–207.
  • Goodhart, C. A. E., & Bhansali R. J. (1970). Political economy. Political Studies, 18, 43–106.
  • Gschwend, T. (2007). Ticket-splitting and strategic voting under mixed electoral rules: Evidence from Germany. European Journal of Political Research, 46(1), 1–23.
  • Key, V. O. (1966). The responsible electorate: Rationality in presidential voting, 1936–1960. Cambridge, MA: Harvard University Press.
  • Kramer, G. H. (1971). Short-term fluctuations in US voting behavior, 1896–1964. The American Political Science Review, 65(1971), 131–143.
  • Lasley, S., & Stegmaier, M. (2001). A comparison of national effects on individual U.S. Senate elections. Politics & Policy, 29(3), 545–567.
  • Lewis-Beck, M. S. (1988). Economics and elections: The major Western democracies. Ann Arbor: Michigan University Press.
  • Lewis-Beck, M. S., Jacoby, W. G., Norpoth, H., & Weisberg, H. F. (2008). The American voter revisited. Ann Arbor: University of Michigan Press.
  • Lewis-Beck, M. S., & Lobo, M. C. (2017). The economic vote: Ordinary vs. extraordinary times. In K. Arzheimer, J. Evans, & M. S. Lewis-Beck (Eds.), The Sage handbook of electoral behavior: Volume 2 (pp. 606–630). London, U.K.: SAGE.
  • Lewis-Beck, M. S., Martini, N. F., & Kiewiet, D. R. (2013). The nature of economic perceptions in mass publics. Electoral Studies, 32(2), 524–528.
  • Lewis-Beck, M. S., & Nadeau, R. (2012). PIGS or not? Economic voting in Southern Europe. Electoral Studies, 31(3), 472–477.
  • Lewis-Beck, M. S., Nadeau, R., & Bélanger, E. (2012). French presidential elections. London, UK: Palgrave Macmillan.
  • Lewis-Beck, M. S., Nadeau, R., & Foucault, M. (2013). The compleat economic voter: New theory and British evidence. British Journal of Political Science, 43(2), 241–261.
  • Lewis-Beck, M. S., & Ratto, M. C. (2013). Economic voting in Latin America: A general model. Electoral Studies, 32(3), 489–493.
  • Lewis-Beck, M. S., & Stegmaier, M. (2000). Economic determinants of electoral outcomes. Annual Review of Political Science, 3(1), 183–219.
  • Lewis-Beck, M., & Stegmaier, M. (2007). Economic models of voting. In R. J. Dalton & H-D. Klingemann (Eds.), The Oxford handbook of political behavior (pp. 518–537). Oxford, U.K.: Oxford University Press.
  • Lewis-Beck, M. S., & Stegmaier, M. (2008). Economic voting in transitional democracies. Journal of Elections, Public Opinion and Parties, 18(3), 303–323.
  • Lewis-Beck, M. S., & Stegmaier, M. (2009). American voter to economic voter: Evolution of an idea. Electoral Studies, 28(4), 625–631.
  • Lewis-Beck, M. S., & Stegmaier, M. (2013). The VP-function revisited: A survey of the literature on vote and popularity functions after over 40 years. Public Choice, 157(3–4), 367–385.
  • Lewis-Beck, M. S., & Stegmaier, M. (2016). The Hispanic immigrant voter and the classic American voter: Presidential support in the 2012 election. The Russell Sage Foundation Journal of the Social Sciences, 2(3), 165–181.
  • Lewis-Beck, M. S., & Stegmaier, M. (2019). Economic voting. In R. Congleton, B. Grofman, & S. Voigt (Eds.), Oxford handbook of public choice, Volume I (pp. 247–265). Oxford, U.K.: Oxford University Press.
  • Lewis-Beck, M. S., Stubager, R., & Nadeau, R. (2013). The Kramer problem: Micro–macro resolution with a Danish pool. Electoral Studies, 32(3), 500–505.
  • Lewis-Beck, M. S., & Whitten, G. (Eds.). (2013). Economics and elections: Effects deep and wide [Special issue]. Electoral Studies, 32(3).
  • Mueller, J. E. (1973). War, presidents and public opinion. New York, NY: John Wiley & Sons.
  • Nadeau, R., & Lewis-Beck, M. S. (2001). National economic voting in U.S. presidential elections. Journal of Politics, 63(1), 160–181.
  • Nadeau, R., Lewis-Beck, M. S., & Bélanger, E. (2013). Economics and elections revisited. Comparative Political Studies, 46(5), 551–573.
  • Norpoth, H. (2001). Divided government and economic voting, Journal of Politics, 63(2), 414–435.
  • Norpoth, H. (2002). On a short-leash: Term limits and the economic voter. In H. Dorussen & M. Taylor (Eds.), Economic voting (pp. 121–136). Oxford, U.K.: Routledge.
  • Palmer, H. D., Whitten, G. D., & Williams, L. K. (2013). Who should be chef? The dynamics of valence evaluations across income groups during economic crises. Electoral Studies, 32(3), 425–431.
  • Plescia, C., & Kritzinger, S. (2017). Retrospective voting and party support at elections: Credit and blame for government and opposition. Journal of Elections, Public Opinion and Parties, 27(2), 156–171.
  • Powell, G., & Whitten, G. (1993). A cross-national analysis of economic voting: Taking account of the political context. American Journal of Political Science, 37(2), 391–414.
  • Rowe, K. T. (2013). Making voice count: Economic voting and the number of parties. Party Politics, 21(5), 803–812.
  • Stegmaier, M., Lewis-Beck, M. S., & Park, B. (2017). The VP-function: A review. In K. Arzheimer, J. Evans, & M. S. Lewis-Beck (Eds.), The Sage handbook of electoral behavior: Volume 2 (pp. 584–605). London, U.K.: SAGE.
  • Stegmaier, M., & Norpoth, H. (2018, January 31), Trump isn’t benefiting from economic growth. But he’s not unusual. The Washington Post.
  • Stubager, R., Botterill, N. W., Lewis-Beck, M. S., & Nadeau, R. (2014). Scope conditions of economic voting: The Danish exception? Electoral Studies, 34, 16–26.
  • Tilley, J., Neundorf, A., & Hobolt, S. B. (2018). When the pound in people’s pocket matters: How changes to personal financial circumstances affect party choice. The Journal of Politics, 80(2), 555–569.
  • Whiteley, P., Clarke, H. D., Sanders, D., & Stewart, M. C. (2014). The economic and electoral consequences of austerity policies in Britain. Parliamentary Affairs, 68, 4–24.
  • Williams, L. K., Stegmaier, M., & Debus, M. (2017). Relaxing the constant economic vote restriction: Economic evaluations and party support in Germany. Party Politics, 23(3), 286–296.