Monday, May 11, 2015

Pocketbook Voting, Sociotropic Politics and the Psychological Framing of Voters

“People vote their pocketbook” seems like a rule of physics in political campaigns, as evident as the law of gravity. From breadlines folding out their pockets to make “Hoover flags” in the 1929 to Bill Clinton’s 1992 campaign slogan that “it’s the economy, stupid,” common wisdom suggests that voters cast ballots based on what's going on with their wallets.

Scholars, however, have long wrestled with whether voter psychology can be reduced to “dull science” of economics when the thrills of electoral politics suggest something so more complex, unpredictable, and exciting. Studying election returns, particularly at the presidential level, provides fodder for the debate about how individual, sociotropic, and aggregate attitudes of voters reflect this kind of retrospective voting.

Though pocketbook-voting theory proves effective for predicting the outcomes of elections, it does not tell the whole story about how the country’s economic well being translates to voters’ choices. A successful economy enables candidates to frame successful messages about the economy at large that resonate with voters regardless of their personal economic circumstances. Even as economic considerations largely shape how voters weigh whether to reward or punish incumbents, an analysis of this phenomenon shows that how national economic conditions are framed in media and campaigns can effect how much these factors can influence voters.

How Pocketbook Voting Theory Developed
Michael Lewis-Beck identifies how the debate about pocketbook voting theory has numerous confounding variables since Anthony Downs’ An Economic Theory of Democracy (1957) popularized using economic indicators to predict voting behavior. Lewis-Beck explains that research revealing the determinative effects of partisanship in voting behavior as identified in The American Voter (1960) complicated the purely economic model. Lewis-Beck describes how during this time period, poll questions about voters feeling towards “economic outlook” further confounded the possibility of drawing conclusions differentiating the effects of pocketbook voting and feelings about the economy as a whole.

Morris Fiorina’s Economic Retrospective Voting in National Elections (1978) put the debate in clearer terms, identifying how the sum of pocketbook effects might influence partisanship over time rather than expecting immediate information to create dramatic changes. His model of retrospective voting can be seen throughout the explanations of how pocketbook and sociotropic considerations influence the decisions of voters.[1]

Kinder and Kiewiet’s Sociotropic Politics
Donald Kinder and D. Roderick Kiewiet argue in Sociotropic Politics (1981) that these decisions are not as cut and dry as individuals voting strictly according to their current economic condition or even assessing how a candidates’ macroeconomic policy might affect their future. They propose an alternative to the view that pocketbook voting transforms remote policy decisions into “doorstep issues,” heightening the public’s interest and attention through self-interest.[2] They suggest that the rational choice model  “dominated by the Downsian hypothesis—that the citizens whose own economic fortunes have deteriorated will vote for change” misses the way citizens define their economic problems in broader political terms.[3]
            Their alternative—the sociotropic prediction—argues that voters are influenced most of all by the nation’s economic condition, roughly estimating whether the incumbent party’s policies ought to be credited or blamed (132). Kinder and Kiewiet describes this phenomenon “not as one of motivation, but as one of information.” (133) Drawing from the American National Election Surveys from 1972-1974-1976, they find that voters’ estimation of how the country is doing economically is more predictive than personal economic condition in estimating whether they will vote for or against an incumbent.
            Kinder and Kiewitt argue, “Sociotropic judgments do not…serve to rationalize partisanship, nor do they represent mere extrapolations from pocketbook experiences” and rather the evidence “suggests that the electorate is remarkably responsive to changes in general economic conditions.”[4] Kinder and Kiewiet conclude their study by speculating about the psychological process to explain sociotropic voting.  They suggest investigating how the personal and political experiences shape opinions would be a point of departure for their research.
Kramer’s The Ecological Fallacy Revisited
            In 1983, Gerald Kramer argues in The Ecological Fallacy Revisited against Kinder and Kiewiet’s “sociotropic” model for understanding the macroeconomic effects on individual voters. He argues that the main flaw of their study is assuming that the ANES survey generates data on a “single, fixed electorate, whose behavior is governed by a specified behavioral hypothesis” as if the survey were interviewing the same voters in a panel.[5]
            Kramer also argues that the way pocketbook voting matters in predicting elections is the aggregate of the behavioral effects rather than individual-level choices.[6] Kramer argues that ascribing inferences from aggregate data to individual behavior is an “ecological fallacy.”[7]
            Kramer’s individual level study analyzes the effect that economic conditions have on voting in an election by comparing a “normal” election with a “recession” election to demonstrate how pocketbook-voting changes votes for the incumbent. His challenge is to parse what is a “true” economic effect, compared to partisan reaction or “exposure to an intensive media campaign” (ie. a political campaign’s framing of economic questions).[8]
            Kramer identifies variables for pocketbook voting and sociotropic voting to assess the cross-sectional and time-series effects of these variables in predicting elections. He describes voter’s sociotropic understanding of the economy as “indistinguishable” from pocketbook voting.[9] Kramer argues the “real inference problem” in judging pocketbook voting at an aggregate level is understanding individual voters’ net change in incomes affected by government activities.[10]
            One mistake of Kramer’s is that he defines sociotropic thinking as altruistic in nature. Kinder and Kiewett warn in their introduction, “The distinction between pocketbook and sociotropic politics is not equivalent to the distinction between a self-interested and an altruistic politics.”[11]  
            The fact that their effect in Kramer’s regression is “indistinguishable”[12] misses the fact that the “sociotropic voters may construe the incumbent administration's handling of the economy as a public good” and then perceive that as eventually helps their self-interest.[13] They plainly state, “What evidence is currently available cannot convincingly distinguish between these two possibilities.”
            Making the distinction even more difficult for Kramer is his definition of pocketbook voting. As mentioned earlier, his equation is designed around the effect of government policies on income. He argues that this is to account for many personalized and localized problems which affect income but voters do not attribute to government activities. Examples include losses and gains in industry or certain life cycles (ie. young people earning more money in their more productive years) will see increased income in the aggregate, regardless of macroeconomic conditions.[14] However, redefining income as government policy’s effect on income eliminates the clear quality to the pocketbook’s effect on voting, which makes an individual’s bottom line a deciding factor. Kramer’s redefinition writes off a lot of potential issues for individuals to engage in sociotropic judgments.
            Kramer, to his credit, introduces this possibility, suggesting that health problems reducing earning potential, a departing dependent child, an accident, or inheritance might also make income less predictive of a voter’s vote. Kramer’s critique argues that the sociotropic factor is eliminated in the aggregate while an increase in individual income can determines the outcome of election. However, Kramer’s argument loses sight of the psychological question Kinder and Kiewiett consider: do voters vote the way they do because of perception or reality?
            Kramer does admit the distinction, though indistinguishable in his analysis, is not meaningless. He proposes that there could be some intermediate way of assessing the influence of sociotropic voting without reducing individual behavior to a ecological fallacy or rendering it meaningless in the aggregate. While Kramer’s critique of Kinder and Kiewiett shows that economic well being carries a lot of weight in the field of forecasting elections, it misses a key question about how sociotropic voting could potential prove useful to campaigns. Kinder and Kiewiett are more interested in why people are voting the way they do rather predicting how they will vote.
            Kramer rightly notes that sociotropic voting motivations cannot be ascertained in the aggregate but neither can motivations of pocketbook voting. Though aggregate income might effectively predict election results in and out of a recession, it tells us little about whether sociotropic voting mattered in these elections. The now-billion dollar cottage industry of framing political questions suggests that in elections, something beyond the bottom line matters, otherwise why bother to campaign.
The Sears and Lau Hypothesis (1983)
            In an article titled “Pocketbook Voting in U.S. National Election Studies: Fact or Artifact?
” Michael Lewis-Beck takes on the question of how the American National Election Survey might cause respondents to conflate these political and economic questions. He assesses hypothesis posed by Sears and Lau in 1983 that respondents attempt to “rationalize” or “personalize” their political choices when posed close to their economic well being. A respondent who is “worse off financially” might say they will vote against the incumbent party or will voice intent to vote against the incumbent and “politicize” their economic response. Sears and Lau argue this problem for the surveys might produce “artifactually high levels of reported consistency between personal economic situations and political attitudes.”[15] Lewis-Beck suggests that though the theory of rationalizing economic responses invites further study about these results in the aggregate, there is nothing to suggest that the ANES produced a questionnaire where proximity of questions about vote choice and economic conditions influences political preferences.[16]
            In 1991, Lee Sigelman, Carol Sigelman and David Bullock sought to answer more directly why pocketbook effects appear in the aggregate but not in the choices of individuals.[17] They interviewed undergraduate students about their hypothetical considerations in a presidential election, asking them to consider elections with good or bad economic performance nationally, whether or not the president was responsible for such conditions, and where their personal economic condition has either improved or declined.[18]
            The experimental study found that respondents’ personal finances and information about the national economy operated side by side with statistical significance. The study’s authors recognized the limitations of the findings given that sociotropic, pocketbook, and attribution to the president in a real election.[19] In their study,  “evaluations were shaped at least as much by whether outcomes were favorable or unfavorable as by the reasons underlying the outcomes.”[20] Their study was able to keep these forces independent from each other in the subjects mind while real election have a much more complex relationship between these variables.[21]
Economic Voting the County Level
Introducing a study better fit to Kramer’s idea of studying this process intermediate, county level analysis, Daniel Eisenberg and Jonathan Ketcham (2004) explain the importance of drawing such a distinction between the pocketbook and sociotropic responses:
Understanding these issues better would clearly be useful for campaign strategists. From a broader societal perspective, the information could point to some potential welfare-enhancing interventions. For example, if people appear to be evaluating the economy in ways that are inconsistent with the actual ways that an incumbent party influences the economy (which are admittedly hard to determine), then society may benefit from activities to draw attention to more relevant information.
Where voters get their information is essential to Eisenberg and Ketcham’s study of eight different presidential elections (from 1972 to 2000) comparing the effect of that information about economic conditions at the county level and the national level through.         While the dichotomy of local and national economic conditions is not the same as the “pocketbook” versus “sociotropic” argument, Eisenberg and Ketcham argue there are relevant pocketbook effects to glean from studying local economic conditions.[22] They demonstrate the distinction by comparing people’s personal income to national indicators like disposable income, unemployment, and Dow Jones growth.[23] They also assessed how much these indicators predicted an incumbent’s success in an election based on the immediacy of the information by measuring economic indicators by the past year, second year back, third year back and fourth year back through a president’s term.
            Eisenberg and Ketcham’s analysis comes to the conclusion that in presidential elections, local economic conditions evidently do not get attributed to a presidential incumbent, though economic voting has the strongest magnitude at the presidential level. By comparing the influence of different time-frames of economic indicators, Eisenberg and Ketcham discover that long-term economic phenomenon have a greater influence on votes in local elections while short-term economic factors weigh heavier on presidential election. They suggest that this could come from voters getting information from “real-time” experience at the local level while receiving information from the media at the national level.[24]
            The demographic level study suggests that some groups are more responsive to national economic indicators like unemployment and income growth. There are numerous reasons why national indicators might spur increased electoral responsiveness to incumbents: some demographic groups are less susceptible to national economic conditions, some demographics have less affiliation with a particular party, economic influence on voter turnout (which the authors had insufficient data to assess) and political sophistication.[25]
            Eisenberg and Ketcham mention that further research ought to determine whether the overall lack of influence of local economic factors on voters means that individual economic factors also do not have a great effect on voters’ choices.[26] While local pocketbook effects do offer a slight predictive edge in determining national elections, the information that seems to have the greatest effect is national level indicators. The influence of national level funnel of information that voters engage voting with more sophisticated, sociotropic thinking in judging presidential incumbents.

Framing the Economy
Explicitly building on the work of Kinder and Kiewitt, Dennis Chong and James Druckman argue in Framing Theory (2007) that the psychology of framing, building on Kinder’s work, has become indispensible in defining political issues and identifying “frames of communication” has become “a virtual cottage industry.”[27]
            Regardless of whether one subscribes to  rational choice theory or the sociotropic theory, Chong and Druckman identify that those seeking to sway public opinion want to estimate both the effect of the personal and the political in order to determine how they can best communicate towards a particular political ends, like promoting a policy or winning an election.
            For example, competing frames about the economy can illuminate how the national economic condition resonates with voters, like Ronald Reagan’s conclusion in the 1980 election in the presidential debate, “are you better off than you were four years ago?” To a less politically sophisticated and less partisan audience, this kind of rhetoric framed national uncertainties about unemployment, inflation, and high interest rates without forcing voters to rationalize a series of policies at a personal level.
            Pocketbook theory implies that personal economic experiences have a great level of salience with the immediate electoral choices of voters. Sociotropic theory argues that those voters’ choices have a stronger connection to their perception of the well being of the country’s economy as a whole. Framing theory posits a mixture of considering particular national values with varying personal enthusiasm or weight.
            The volatile but strong force that economic concerns have in electoral outcomes suggests that the uncertainty of economics fosters beyond partisan affiliations. The winners in a contest of an ideas over the state of the economy depends greatly on economic realities, what information voters receive about those conditions, and how well politicians and activists can frame those facts to draw voters to their conclusion, particularly with voters whose views are less static than partisans or more sophisticated voters. Though economic conditions are not determinative of a persons vote, the commonly held belief that personal feelings about the economy ultimately decide the outcome of elections stems from how campaigns can frame messages about the economy in competing ways.
Recent Theories about Pocketbook vs. Sociotropic

While both pocketbook and sociotropic considerations have found their place in the models that political scientists use to predict election outcomes, studying how framing affects how voters weigh particular issues has only begun to unravel its influence on these factors. 

            Theories abound about how these economic considerations influence voters differently. Brad Gomez and J. Matthew Wilson argue that the inconsistency in observing pocketbook effects at an individual may stem from overlooking the heterogenous nature of its effect on voters. They argue that, “less sophisticated voters may not credit or blame government for their own economic circumstances” while “more sophisticated voters focus on people other than the president for the national economy.”[28]

Alternatively, Mitchell Killian, Ryan Schoen and Aaron Dusso argue that how a voter compares their personal outcomes to the rest of the country can influence whether they turn out to vote. Their analysis of ANES data found that “those who perceive that over the last year their own financial situation has improved are less likely to vote than those who think the economy is outperforming their own financial situation.”[29]

            The dichotomy between sociotropic and economic considerations provides for a variety of parse election data to analyze its influence in an innumerable ways, providing the foundation for endless research into the topic. However, given the potential confounding variables and aggregation problems with election surveys, the best way to understand the psychology of these effects on voters might be through a true experiment.

            In 2010, Brian Newman conducted an experiment in priming voters with different information about how responsible President Obama was for the state of the economy, framing the issue as either Obama’s responsibility or not. He found that this framing had an influence on how much economic concerns weighed on respondents choice for the mid-term elections.

            Newman’s study found that the two different frames had suggestive influence on attributing blame to President Obama for the struggling economy. Respondents primed with information that cast President Obama as responsible for the struggling economy had a slightly less favorable view of the president. Those primed with information that cast Obama as not responsible weighed economic concerns less highly with their view of the president. The study also found that those who weighed economic perceptions more heavily also had less favorable perceptions of Obama.[30]

            While pocketbook voting proved predictive in Newman’s experiment and the mid-term elections, his study suggests that a sociotropic framing of the country’s economic condition proved favorable in mitigating the influence of pocketbook considerations.[31]

Questions about Framing’s Influence on Rational Behavior

            A work of James Druckman’s prior to Framing Theory questions how relevant framing theory might prevent drawing conclusions about rational influence on political preferences. He argues that the framing effect occurs when “logically equivalent” frames present the same information in a different way.[32] Talking about the economics seemingly poses a weighing of values, but might violate the premise of framing by presenting measurements on different metrics like one’s personal income or the country’s well being. In order to understand the rational choices voters are making, there needs to be some sort of evidence that places the competing ideas: short-term self interest and long-term overall well-being in an equivalent frame-of-mind for respondents. Druckman says distinctions debates between issue framing (or “emphasis” frames) might not follow the same psychology of framing effects (“valence” frames).[33]

            Michael Lewis-Beck suggests that there are still further ways that economic framing might factor into voter behavior that have not been considered with pocketbook and sociotropic investigations treating economics as a “valence” issue rather than investigating its influence on opinions. Lewis-Beck argues that taking this topic into “new dimensions” of research could better explain how economic considerations effect voter choices. [34]
            Amidst all the possibilities of exploring the “new dimensions” of voters’ economic considerations, pocketbook voting can be a useful tool in predicting outcomes in politics but tell us little about how the individual bottom-line influences votes. Going back to Fiorina’s model for the retrospective voting, we find that the way local and national economic issues get framed by media, campaigns, and common sentiment can change how much any particular economic information sways voters.
            Similarly, the sociotropic explanation cannot fully account for the aggregate result and may be impossible to distinguish from pocketbook effects fully. What economic outcomes we attribute to a president’s policies has as much to do with observable conditions as how well a president or his party can frame the issue. While sociotropic thinking best describe what motivates voters, potentially voting against their empirical self-interest, persuading a majority to support an incumbent requires some measurable economic success to effectively produce an electoral victory.

Half of the people can be part right all of the time
Some of the people can be all right part of the time
But all of the people can’t be all right all of the time
I think Abraham Lincoln said that.

-Bob Dylan, “Talkin’ World War III Blues”
Works Cited

Carey, Sean. "Election Cycles and the Economic Voter." Political Research Quarterly 59.4 (2006): 543-56. JSTOR. Web. 18 Apr. 2015.

Chong, Dennis, and James N. Druckman. "Framing Theory." Annual Review of Political Science 10.1 (2007): 103-26. JSTOR. Web. 24 Apr. 2015.

Druckman, James N. "Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects." The American Political Science Review 98.4 (2004): 671-86. JSTOR. Web. 27 Apr. 2015.

Gomez, Brad T., and J. Matthew Wilson. "Causal Attribution and Economic Voting in American Congressional Elections." JSTOR. Political Research Quarterly, Sept. 2003. Web. 25 Apr. 2015.

Gomez, B. T., and J. M. Wilson. "Economic Voting and Political Sophistication: Defending Heterogeneous Attribution." Political Research Quarterly 60.3 (2007): 555-58. JSTOR. Web. 25 Apr. 2015.

Killian, Mitchell, Ryan Schoen, and Aaron Dusso. "Keeping Up with the Joneses: The Interplay of Personal and Collective Evaluations in Voter Turnout." Political Behavior 30.3 (2008): 323-40. JSTOR. Web. 25 Apr. 2015.

Kramer, Gerald H. "The Ecological Fallacy Revisited: Aggregate- versus Individual-level Findings on Economics and Elections, and Sociotropic Voting." The American Political Science Review 77.1 (1983): 92. JSTOR. Web. 24 Apr. 2015.

Lewis-Beck, Michael. “Pocketbook Voting in U.S. National Election Studies: Fact or Artifact?” American Journal of Political Science, Vol. 29, No. 2 (May, 1985), pp. 348-356. Midwest Political Science Association. Web. 18 Apr. 2015

Lewis-Beck, Michael S., and Mary Stegmaier. "American Voter to Economic Voter: Evolution of an Idea." Electoral Studies 28.4 (2009): 625-31. JSTOR. Web. 24 Apr. 2015.

Lewis-Beck, Michael, and Richard Nadeau. "Economic Voting Theory: Testing New Dimensions." Special Symposium: Electoral Forecasting Symposium, June 2011. Web. 27 Apr. 2015.

Newman, Brian. "Polls and Elections: Decreasing the Economy’s Impact OnEvaluations of the President: An Experimenton Attribution Framing." Presidential Studies Quarterly 43, no. No 4 (2013): 866-82.

Sigelman, Lee, Carol K. Sigelman, and David Bullock. "Reconsidering Pocketbook Voting: An Experimental Approach." Political Behavior 13.2 (1991): 129-49. JSTOR [JSTOR]. Web. 24 Apr. 2015.

[1] Lewis-Beck , Mary Stegmaier 626
[2] Kinder and Kiewiet 130
[3] Kinder and Kiewitt 133
[4] Kinder and Kiewitt 157
[5] Kramer 95
[6] Kramer 93
[7] Ibid. 94
[8] Ibid. 95
[9] Ibid. 98
[10] Ibid. 99
[11] Kinder and Kiewitt 132
[12] Kramer 98
[13] Kinder and Kiewitt 132
[14] Kramer 98
[15] Lewis-Beck 349
[16] Ibid. 355
[17] Sigelman, Sigelman, and Bullock 139
[18] Ibid. 136
[19] Ibid. 139
[20] Ibid. 141
[21] Ibid. 143
[22] Eisenberg and Ketcham 6
[23] Ibid. 5
[24] Ibid. 14
[25] Ibid. 15
[26] Ibid. 21
[27] Chong and Druckman 106
[28] Gomez and Wilson 555
[29] Killian, Schoen and Dusso 323
[30] Newman 874
[31] Newman 875
[32] Druckman 671
[33] Druckman 672
[34] Lewis-Beck and Richard Nadeau 1