NBER Reporter 2012 Number 3: Research Summary
Race, Income, and Political Efficacy
Ebonya Washington *
"[T]he vote is the most powerful instrument ever devised by man for breaking down injustice and destroying the terrible walls which imprison men because they are different from other men," said Lyndon B. Johnson at the signing of the Voting Rights Act in 1965. His statement reflects a long-held, and long-fought-for, belief that political participation can help groups to overcome disadvantage bestowed by history. In my research, I empirically examine the extent to which this is true. I study the ability of minorities and low-income Americans to use the political process to affect policy outcomes and shift the distribution of public resources in their favor. I refer to this as their political efficacy, and ask two broad questions: When are other groups supportive of the policies/candidates that these two minority groups favor? How do American institutions help or hinder these groups' political efficacy?
When Are Other Groups Supportive of the Policies that Low-Income or Black Voters Support?
Because both Blacks and low-income voters are numerical minorities, a central component to their ability to secure passage of their preferred policies is the support of other groups. In my research, I demonstrate circumstances under which that support is and is not forthcoming. For example, for some 60 years before Barack Obama garnered 95 percent of their vote, Blacks have cast their ballots overwhelmingly for the Democratic candidate in two-party elections. But when are non-Blacks more likely to favor the Democratic candidate? Only when the Democratic candidate is not Black, I find by examining Congressional and gubernatorial elections from 1982 to 2000.1 While both Black and White citizens are more likely to turn out to cast a ballot in an electoral contest that includes a Black candidate, the White voters are less likely to vote in favor of the Democratic candidate when s/he is Black. One possible explanation for the White reluctance to vote for Black candidates is that Black candidates (like the Black electorate) tend to be more liberal than their White Democratic counterparts.
Black-White segregation also predicts decreased support among Whites for Black candidates and, in fact, for Democratic candidates more generally, Elizabeth Ananat and I find.2 We uncover two potential explanations for this phenomenon. First, Whites with less liberal attitudes self-select into more segregated communities. Second, contact with Black voters affects White voters' attitudes. In other work, I find additional support for the idea that interactions with others helps to shape one's political attitudes. For example, conditional on the total number of children in his family, a U.S. congressman's propensity to vote liberally, particularly on legislation concerning women's issues, increases with the number of daughters he has.3 Women generally have more liberal attitudes than men; for elite women, this is particularly true. This research suggests that sharing (or at least witnessing) experiences that have led their daughters to grow up to be left leaning also moves Democratic congressmen to cast more liberal votes on the House floor than their counterparts with fewer or no daughters.
Support for policies preferred by the poor also appears to be shaped by experience.4 Eric Brunner, Stephen Ross, and I looked not specifically at whether a person knew someone poor, but rather at how economic circumstances more generally shape views on redistribution. Focusing on California, where voters have the opportunity to weigh in on ballot propositions concerning a variety of issues each year, we show that -- consistent with economic theory -- neighborhood residents are more likely to vote in favor of redistribution and other liberal economic proposals when they are suffering negative economic shocks. We see larger effects in poorer communities, suggesting that those closer to benefitting from economic policies, and/or to observing others benefit from those same policies, have the most malleable opinions. One surprising finding of this study is that negative economic shocks also predict voting for liberal candidates and, to a lesser extent, voting liberally, on non-economic issues.
This co-movement of voting on economic and other issues may come from a desire for party strength, or because individuals strive for consistency across opinions and from opinions to behaviors, as suggested by the psychological theory of cognitive dissonance. In two papers, I find support for the relevance of cognitive dissonance to the voting arena. In the first, Sendhil Mullainathan and I show that the act of voting for a candidate increases one's support for that candidate.5 Of course, the difficulty in trying to tease out this relationship is reverse causality -- those who view the candidate more favorably are more likely to vote for the candidate. We circumvent this difficulty in two ways. First, we exploit the age discontinuity in voting eligibility. That is, we compare those who were just a little too young to vote for president in the focal election year with those just above the age cutoff. Second, we compare feelings about senators most recently elected during a presidential election year (when turnout is greater) with those most recently elected during an off-year. We find that both those who were above the age cutoff and those who most recently voted for a senator in a presidential election show greater polarization in their opinions of the president and the senator, respectively. In other words, the act of voting increases the distance in opinion between those in favor and those against.
In the second paper demonstrating the relevance of cognitive consistency to the voting arena, Alan Gerber, Gregory Huber, and I use a field experiment conducted around the 2008 Connecticut presidential primary.6 The experiment targeted citizens who were registered to vote but unaffiliated with any political party. One group of such individuals was informed via a letter from the Connecticut Secretary of State that they must affiliate with a party in order to vote in the upcoming primary. That experimental "treatment" increased not only party registration, but also support for political figures of the same party. Once again, voters who were prompted to alter a behavior -- this time party registration -- ended up changing not only the behavior but also a related attitude. This co-movement of behaviors and issues suggests that shocks that prompt voters to vote more liberally for candidates, or on any of a variety of issues, may also make these voters more sympathetic to the liberal policies that low-income and minority voters generally favor.
How Do Institutions Impact Political Efficacy?
Money is thought to be a potent force in the American political process. For poor voters, the common wisdom is that money is an obstacle to having their viewpoints heard. Brunner, Ross, and I examine the relevance of this view to legislative voting in California.7 Because of the aforementioned numerous ballot propositions there, we have good data on how voters in both low- and high-income areas of a district feel about various issues that will be considered by the public and by the legislator. Using these data, we calculate the extent to which legislative voting coincides with the majority view of low- and high-income constituents. Contrary to popular wisdom, we find that less income does not mean less representation. In fact, the opinions of high- and low-income voters are highly correlated, and the legislato's vote most often represents the views of both groups of voters in his/her district. Any differences in representation by income that do exist vary by the legislator's party. Republicans vote the will of their higher income over their lower income constituents more often; Democratic legislators do the reverse. We find that these patterns of representation by income are largely explained away by partisanship. Republicans vote like high-income voters in their district not because those voters are high income, per se, but because they are highly likely to vote Republican. Thus, rather than finding evidence for underrepresentation of the financially disadvantaged, we confirm underrepresentation of the politically disadvantaged — those who are represented by a politician of a differing party.
Of course, legislative voting is just one type of one representation and California is but one state. An important topic for future work is to examine whether these findings generalize to other legislative behaviors (constituent service, agenda setting, "pork" distribution) and to other geographic settings.
For Blacks, one alleged impediment to representation is race-based legislative redistricting. A majority Black legislative district is a congressional district in which a majority of residents are Black. When a state creates such a district, there are, by definition, fewer Blacks in the remaining districts. The conventional view (espoused by political scientists and both major political parties) is that the creation of these districts in a state leads that state's House delegation to vote more conservatively. The idea is that the majority Black district will elect a representative who is more liberal than average, but the remaining districts (with a lower percentage of Black voters) will elect correspondingly more conservative representatives, on balance moving the delegation's average vote in a more conservative direction. I investigate this common wisdom in regard to the 1990 congressional redistricting, the redistricting period that saw the largest increase in majority-minority districts.8 This increase was effectively mandated in some states by a 1982 amendment to the Voting Rights Act (VRA). Comparing southern states that were forced to increase the number of Black districts with those that were not, I find no evidence that majority Black districts move the state's congressional delegation in a more conservative direction. In fact the results, although largely insignificant, point in a more liberal direction. Thus, the creation of majority-minority districts seems a net positive for Black representation. These districts serve to increase both Black descriptive representation--the number of Blacks in Congress--and Black substantive representation--the number of congresspersons who vote as Blacks hope they will.
The majority-minority district mandate is only one part of one reauthorization of the VRA. In other work examining the impact of American institutions on minority representation, Elizabeth Cascio and I look at the impact of the Act’s original passage in 1965 on the distribution of public resources.9 The Act dismantled barriers to Black voter registration, chief among them literacy tests. Those tests, despite their name, might be more aptly characterized as tests of race than of reading ability. Thus there were greater numbers of disenfranchised voters in literacy-test states in counties with larger shares of Black residents. We find that, post-VRA, not only did these counties see large increases in enfranchisement, but they also saw increases in their share of state transfers, which were largely earmarked for public education. Of course, the period around the passage of the VRA was notably turbulent in the American south, but we are able to rule out competing explanations for the finding including desegregation, black political activism, and basic changes in need. Shortly before the passage of the VRA, Reverend Martin Luther King, Jr., wrote, "Voting is the foundation stone for political action. With it the Negro can eventually vote out of office public officials who bar the doorway to decent housing, public safety, jobs and decent integrated education."10 Our empirical evidence seems to back his early assertion.
In conclusion, my work has established a few predictors of political efficacy for low-income and Black Americans. The ongoing goal is to examine when and how marginalized populations can use the political system to fulfill economic needs. President Johnson argued in the quote with which I began that voting is a powerful instrument. My work seeks to understand the circumstances and methods in which the instrument is most effectively wielded.
* Washington is a Faculty Research Fellow in the NBER's Program on Political Economy and the Henry Kohn Associate Professor of Economics at Yale University.
3. E. Washington, "Female Socialization: How Daughters Affect Their Legislator Fathers' Voting on Women's Issues," NBER Working Paper No. 11924, January 2006, and American Economic Review, Vol. 98, No. 1, March 2008, pp. 311-32.
4. E. Brunner, S. Ross, and E. Washington, "Economics and Ideology: Causal Evidence of the Impact of Economic Conditions on Support for Redistribution and Other Ballot Proposals," NBER Working Paper No. 14091, June 2008, and as E. Brunner, S. Ross, and E. Washington, "Economics and Policy Preferences: Causal Evidence of the Impact of Economic Conditions on Support for Redistribution and Other Ballot Proposals," Review of Economics and Statistics, Vol. 93, No. 3, August 2011, pp. 888-906.
5. S. Mullainathan and E. Washington, "Sticking with Your Vote: Cognitive Dissonance and Voting," NBER Working Paper No. 11910, January 2006, and as S. Mullainathan and E. Washington, "Sticking with Your Vote: Cognitive Dissonance and Political Attitudes," American Economic Journal: Applied Economics, Vol. 1, No. 1, January 2009, pp. 86-111.
6. A. Gerber, G. Huber, and E. Washington, "Party Affiliation, Partisanship, and Political Beliefs: A Field Experiment," NBER Working Paper No. 15365, September 2009, and American Political Science Review, Vol. 104, No. 4, November 2010, pp. 720-44.
8. E. Washington, "Do Majority Black Districts Limit Blacks' Representation? The Case of the 1990 Redistricting," NBER Working Paper No. 17099, May 2011, and Journal of Law and Economics, forthcoming.
10. M.L. King, Jr., "Civil Right No. 1–The Right to Vote," New York Times, March 14, 1965, p. SM26.