Black lives matter: Differential mortality and the racial composition of the U.S. electorate, 1970-2004 lable at ScienceDirect Social Science & Medicine 136-137 (2015) 193e199 Contents lists avai Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed Black lives matter: Differential mortality and the racial composition of the U.S. electorate, 1970e2004 Javier M. Rodriguez a, b, Arline T. Geronimus c, b, d, *, John Bound e, b, d, Danny Dorling f a Mathematica Policy Research, USA b Population Studies Center, Institute for Social Research, University of Michigan, USA c Department of Health Behavior and Health Education, University of Michigan, USA d Center for Advanced Study in the Behavioral Sciences, Stanford University, USA e George E. Johnson Collegiate Professor, Department of Economics, University of Michigan, USA f Halford Mackinder Professor of Geography, School of Geography and the Environment, Oxford University Centre for the Environment, Oxford University, UK a r t i c l e i n f o Article history: Available online 21 April 2015 Keywords: Race Health disparities Premature mortality Electorate Political inequality Voting Health policy * Corresponding author. CASBS, 75 Alta Road, Stanf E-mail address: arline@umich.edu (A.T. Geronimu http://dx.doi.org/10.1016/j.socscimed.2015.04.014 0277-9536/© 2015 Elsevier Ltd. All rights reserved. a b s t r a c t Excess mortality in marginalized populations could be both a cause and an effect of political processes. We estimate the impact of mortality differentials between blacks and whites from 1970 to 2004 on the racial composition of the electorate in the US general election of 2004 and in close statewide elections during the study period. We analyze 73 million US deaths from the Multiple Cause of Death files to calculate: (1) Total excess deaths among blacks between 1970 and 2004, (2) total hypothetical survivors to 2004, (3) the probability that survivors would have turned out to vote in 2004, (4) total black votes lost in 2004, and (5) total black votes lost by each presidential candidate. We estimate 2.7 million excess black deaths between 1970 and 2004. Of those, 1.9 million would have survived until 2004, of which over 1.7 million would have been of voting-age. We estimate that 1 million black votes were lost in 2004; of these, 900,000 votes were lost by the defeated Democratic presidential nominee. We find that many close state-level elections over the study period would likely have had different outcomes if voting age blacks had the mortality profiles of whites. US black voting rights are also eroded through felony disenfranchisement laws and other measures that dampen the voice of the US black electorate. Sys- tematic disenfranchisement by population group yields an electorate that is unrepresentative of the full interests of the citizenry and affects the chance that elected officials have mandates to eliminate health inequality. © 2015 Elsevier Ltd. All rights reserved. 1. Introduction In the United States, after centuries of de jure and de facto disenfranchisement of black Americans, the Voting Rights Act of 1965 resulted in a mass enfranchisement of poor and black Amer- icans. Today, however, erosion of these rights is a great and growing concern. Although the US government acknowledges political participation to be a universal human right, several governmental decisions and practices, often at the state level, appear to be selectively undermining the prohibition against voting rights discrimination on the basis of race, first set forth in the 15th amendment to the US Constitution. Felony disenfranchisement laws in many states have a significant discriminatory impact on ord, CA 94305, USA. s). voting outcomes given race/ethnic variations in prosecution and sentencing of drug-related crimes (Manza and Uggen, 2006; Uggen et al., 2012). Partisan legislative redrawing of electoral boundaries that concentrate racial/ethnic groups into minority districts also has been shown to reduce their political influence (Epstein and O'Halloran, 1999; Trebbi et al., 2008). The trend toward shortened poll hours and more stringent voter ID policies in several states have had or are anticipated to have disproportionately negative effects on voting among the nonwhite and the poor (Barreto et al., 2009). US racial inequalities in excess mortality are another possible threat to the relative voting power of blacks compared to whites, but how important are they? In this analysis we begin to answer that question by estimating the cumulative impact of mortality differentials between US blacks and whites from 1970 to 2004 on the racial composition of the electorate in the general election of 2004 and in close statewide elections during the study Delta:1_given name Delta:1_surname Delta:1_given name Delta:1_surname mailto:arline@umich.edu http://crossmark.crossref.org/dialog/?doi=10.1016/j.socscimed.2015.04.014&domain=pdf www.sciencedirect.com/science/journal/02779536 http://www.elsevier.com/locate/socscimed http://dx.doi.org/10.1016/j.socscimed.2015.04.014 http://dx.doi.org/10.1016/j.socscimed.2015.04.014 http://dx.doi.org/10.1016/j.socscimed.2015.04.014 Fig. 1. Age distributions of the deceased by race (2004). J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199194 period. While voting behavior is influenced by a range of forces, it is certainly true that the longer a person lives, the greater their op- portunity to vote over their lifetime. Throughout the 20th century, the mortality rate of US blacks was, on average, about 60% greater than that of US whites (Kaufman et al., 1998). Although measure- able improvements in black excess mortality were seen mid- century, black-white mortality disparities have changed little over recent decades. For instance, the age-sex standardized mortality rate for blacks was 1.47 in 1960 and 1.41 in 2000 (Satcher et al., 2005). These statistics suggest that significant black-white mor- tality differentials are important social forces shaping the compo- sition of the US electorate. Beyond the compositional impact, a contraction in black voting- age adults might also affect partisan politics and policy, and thereby influence structural inequality. Abundant evidence indicates that race and racial prejudice affect political attitudes (Henry and Reyna, 2007; Sears and Kinder, 1971), candidate preferences (Bobo and Dawson, 2009; Valentino and Sears, 2005), political behavior (Enos, 2011; Sidanius and Pratto, 2001), political ideology (Lane et al., 2011; Pratto et al., 1994), public opinion (Mendelberg, 2008; Valentino et al., 2002), political inclusion (Lavariega Monforti and Sanchez, 2010; Massey and Denton, 1989), and race-based policy preferences (Rabinowitz et al., 2009; Tesler and Sears, 2010). Other evidence ties these racialized political pro- cesses to broad social inequalities (Bonilla-Silva, 2013) including race-based geographic or residential segregation (Dawson, 1995), incarceration rates (Caplow and Simon,1999), and access to and the quality of structural resources such as medical care and welfare (Gilens, 1995) e all factors connected to health outcomes. In the US, where populations with different voting preferences face system- atically unequal life chances, population health inequalities could affect not only the composition of the electorate, but election outcomes and subsequent policy, including policy that influences the health disparities that lead to excess mortality (Blakely et al., 2001; LaVeist, 1992; Purtle, 2013; Rodriguez et al., 2013). Differential mortality by social group has been found to be associated with the composition of the electorate in the United Kingdom (Dorling, 1998, 2010; Smith and Dorling, 1996). For example, individuals living in working class areas in the UK live an average of one general election less than those living in middle or upper class areas (McCartney et al., 2010). Many national UK gen- eral elections have been very closely fought battles and so this difference could have been influential on past electoral outcomes in the UK. However, the possible impact of black excess mortality relative to whites on US election outcomes has not been examined. In this analysis, we estimate the impact of excess deaths among blacks on the racial composition of the electorate in the US presi- dential election of 2004. Because felony disenfranchisement is widely considered significant enough to have changed electoral outcomes, especially in local elections (Manza and Uggen, 2006; Uggen and Manza, 2002), we also explore the impact of black excess mortality on close statewide elections during the study period. 2. Theoretical framework The social, economic and geographical inequalities in mortality found in the US are remarkably large by international standards (Marmot and Bell, 2009) and disproportionately disadvantage blacks relative to whites. And much evidence suggests that US black-white health disparities are persistent at all levels of the so- cioeconomic spectrum (Pearson 2008), and far higher than in other affluent countries with less of a history of racial discrimination. Popularized images portray excess US black deaths as largely occurring to youth e the result of homicide, drug overdoses and other accidents; or reflecting a shorter life expectancy among black relative to white seniors. In fact, the predominant and persistent driving force behind US black/white mortality disparities is the unequal distribution of chronic morbidity among young through middle aged adults (Geronimus et al., 2011, 1999, 1996). Fig. 1 shows the age distributions of all individuals who died in the US in 2004 by race. The area between the curves represents the mortality gap between non-Hispanic blacks and whites. Notably, the distributions do not intersect until approximately age 73, indicating that the mortality gap between blacks and whites per- sists throughout the average life expectancy of blacks. Fig. 1 also shows that the mortality gap between blacks and whites is greatest between the ages of about 40 and 65 e also an age range during which the probability of turning out to vote is the highest, as shown in Fig. 2. Simply put, this creates an especially high political participation disadvantage for the black population because blacks are dying off from the electorate at higher rates than whites during the ages of highest voter turnout. The causes of racial disparities in health are multiple and complex and include social policies and laws that are, at least theoretically, amenable to reinforcement or change depending on political mandates. Among them, residential segregation, cumula- tive disinvestment, and austerity urbanism in predominantly black neighborhoods in the US have contributed significantly to health disparities (Geronimus, 2000; Geronimus et al., 2015; Schulz et al., 2005). Predominantly black neighborhoods are characterized by higher exposure to pollution, fewer recreational facilities, less pedestrian-friendly streets/sidewalks, higher costs for healthy food, and a higher marketing effort per capita by the tobacco and alcohol industries (Diez Roux and Mair, 2010; Diez Roux et al., 2001; Geronimus, 2000; Schulz et al., 2005). In the US, large black- white disparities are also detected in access to and quality of health care resources, including health insurance coverage and health services for preventive screening, diagnostic, diagnosing and treatment, and rehabilitation (Williams and Mohammed, 2009). Moreover, racialization and its subsequent environmental, ma- terial, and health care constraints shape exposure to everyday challenges and coping options. Repeated and high-effort coping with social disadvantage and the contingencies of stereotyped Fig. 2. Age distribution of voting population (2004) (ages 18e84 years). J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199 195 social identity are now thought to contribute to a cumulative physiological toll across the life-course, or weathering (Geronimus, 2013; Geronimus et al., 2006). Weathering reflects stress-mediated physiological damage and dysregulation across body systems. These can result in a relatively steeper age-gradient increase in high allostatic load, adverse health outcomes including early onset of hypertension, diabetes, and disability, and excess death from young through middle adulthood, such as that observed in Fig. 1 (Crimmins et al., 2003; Geronimus et al., 2010; McEwen and Seeman, 1999). To the extent that social stratification processes e which are affected by public policy and political power e sort Americans into different socioeconomic strata and physical environments based on their race, a disproportionate number of blacks are non-randomly exposed to the challenges, physiological stressors, and risks of injury that emanate from social disadvantages, thus contributing to racial inequality in health. It is widely acknowledged that elimi- nating racial disparities in health will require addressing such fundamental social causes and more proximate social determinants of health (Satcher, 2010). Because the social determinants of health are influenced by political forces, the black vote may play a key role in determining both the mechanisms of social stratification and the ultimate exposure of blacks to the psychosocial and environmental threats and challenges that increase racial health inequality (Rodriguez et al., 2013, 2014). Thus, large and persistent US black-white mortality disparities could be both a cause and an effect of political processes. Social and health policies that have population health implications are shaped, in part, by those holding elective office. It is also possible that political mandates or political will in support of policy to diminish disparities in health, are linked to political participation (Keiser et al., 2004; Schneider and Ingram, 1993; Thompson, 2005). Because political representation is a function of the share of the population eligible to vote and participation levels throughout adulthood, racial disparities in age-specific mortality rates may influence political outcomes. In particular, as excess mortality impacts black underrepresentation in the electoral process, it may limit blacks' influence on policy-making and po- litical decision-making processes including those that affect their health. 3. Racial mortality gaps and electoral politics The effects of black-white differential mortality on electoral participation are dynamic and influence the demography of politics in at least three quantifiable ways. First, individuals who die before the age of 18 never have the opportunity to vote. Second, in- dividuals who die after the age of 18, but before the age of life expectancy of their cohort have a shorter “electoral life” than those who culminate their normal life span. And third, premature death prevents individuals from voting not only in the election immedi- ately following their death but in all subsequent elections for which they could have expected to have lived, making the effect cumu- lative. This cumulative effect most dilutes the electoral voice of blacks relative to whites. We aim to quantify this cumulative effect of excess mortality on nonparticipation. 4. Research design We evaluate a counterfactual: What would have been the effect on the 2004 general election if blacks had survived at the same rates as whites between the years 1970 and 2004? Considering that differential mortality effects are cumulative, the case study of the 2004 general election allows for the full electoral cumulative effects of excess mortality in the time range of the available data. Because the mortality files we use are only available on the state level until 2007, the 2004 general election represents the latest general elec- tion available for analysis. Ideally we would trace cohort mortality back to birth, but for methodological and data availability reasons, we instead start the clock at 1970 rather than at the birth cohort of the oldest blacks alive in 2004. Although crude versions of the mortality data used in this analysis date back to 1959, it is not until 1970 that the available data allow us to account for consistent state-level mortality sta- tistics in all states and years, and to validly identify the Hispanic or non-Hispanic origin of whites and blacks. In addition, in our statewide calculations we assume no net migration across states. This assumption would be implausible prior to 1970. Between 1910 and 1970, 6.5 million blacks migrated from southern to northern states, 5 million after 1940 during the Great Migration (Lemann, 2011). After 1970 when the Great Migration ended and there were small cross-state net migration rates among blacks, our assumption is reasonable. Using 1970 as the starting point, how- ever, implies that we will underestimate accumulated excess black deaths by 2004 given that all those succumbing to premature death before 1970 e who would have otherwise survived beyond 1970 e are excluded from the analysis. Our calculations use data from four sources. Deaths by state of residence, race, gender, and age were derived from the Multiple Cause of Death files from 1970 to 2004. Population counts by state of residence, race, sex, and age were taken from population esti- mates from the US Census Bureau. Data on the total number of votes by state were taken from the US Elections Project, while state data on the gender, race, and age distribution of voters were taken from the National Election Pool General Election Exit Polls (2004). For more details, see the Data Appendix. All calculations are conducted using stratification by sex, non- Hispanic race origin (i.e., non-Hispanic black and non-Hispanic white), age (0e84 years), state of residence (32 states with signif- icant black populations and the District of Columbia, see Table 1), and year (1970e2004). To exploit the fullness of the available data, we use similar analyses to examine both the presidential election of 2004 and close senate and gubernatorial elections during the study period. Our calculations occur in stages. First, the total number of black excess deaths by sex, age, and state is calculated for each year Table 1 Estimates of excess deceased blacks (1970e2004), hypothetical survivors and votes lost (2004). State Total black excess deceased (1970e2004) Hypothetical black survivors to 2004 Voting-age hypothetical black survivors to 2004 Voting-age hypothetical survivors as percent of black VAP Total black votes lost Black votes lost for Kerry Votes lost for Kerry as percent of total votes lost Black votes lost for Bush Votes lost for Bush as percent of total votes lost Votes lost as percent of total black votes New York 256,004 193,672 183,654 8.7 88,657 77,238 87.1 11,419 12.9 9.5 Illinois 191,485 140,224 130,406 10.0 55,945 50,459 90.2 5486 9.8 10.3 Florida 170,230 107,659 99,705 5.6 54,110 44,364 82.0 9746 18.0 6.2 California 162,482 120,635 112,673 6.8 59,268 48,841 82.4 10,427 17.6 8.2 Texas 158,970 106,493 100,065 5.6 65,143 56,386 86.6 8757 13.4 7.3 Georgia 154,380 102,343 95,052 5.3 47,286 39,728 84.0 7558 16.0 5.8 North Carolina 142,957 92,549 86,960 6.7 59,292 49,271 83.1 10,021 16.9 6.8 Pennsylvania 132,674 90,276 84,579 9.6 71,932 62,099 86.3 9833 13.7 9.8 Michigan 128,294 97,609 90,554 9.1 59,691 54,239 90.9 5452 9.1 9.8 Louisiana 112,055 76,624 71,700 7.2 42,037 37,016 88.1 5021 11.9 8.1 Virginia 107,365 67,084 62,771 6.0 56,231 50,205 89.3 6026 10.7 8.3 New Jersey 103,858 80,155 75,049 9.1 54,188 44,925 82.9 9263 17.1 10.5 South Carolina 101,121 65,970 62,175 7.3 35,223 29,599 84.0 5624 16.0 7.5 Ohio 100,113 68,586 63,685 6.8 37,365 30,659 82.1 6706 17.9 6.9 Maryland 99,806 73,179 68,118 6.0 41,133 36,539 88.8 4594 11.2 7.4 Alabama 90,602 58,313 54,778 6.7 34,753 31,380 90.3 3373 9.7 7.5 Tennessee 75,011 45,494 41,774 6.2 26,181 22,899 87.5 3282 12.5 8.1 D.C. 70,120 48,012 46,425 18.5 21,501 20,735 96.4 766 3.6 17.6 Mississippi 68,700 49,031 45,604 6.4 29,990 27,378 91.3 2612 8.7 7.8 Missouri 57,089 37,961 35,331 7.8 18,933 17,205 90.9 1728 9.1 9.0 Indiana 38,192 26,732 24,663 6.7 13,997 12,696 90.7 1301 9.3 8.2 Arkansas 32,396 20,396 19,254 6.7 11,389 9551 83.9 1838 16.1 7.2 Kentucky 22,080 11,622 10,974 5.0 9565 8370 87.5 1195 12.5 6.4 Connecticut 20,125 16,068 14,902 6.5 6089 5438 89.3 651 10.7 6.7 Massachusetts 16,981 13,999 12,889 4.7 7304 6949 95.1 355 4.9 6.3 Wisconsin 16,511 13,433 11,751 5.6 8915 7499 84.1 1416 15.9 6.1 Oklahoma 14,568 9473 8756 4.6 6853 5003 73.0 1850 27.0 5.5 Delaware 10,460 6984 6437 5.6 5054 4382 86.7 672 13.3 6.8 Washington 8274 6204 5602 3.5 2705 2073 76.6 632 23.4 4.8 Arizona 7159 5024 4440 3.3 2671 2085 78.1 586 21.9 5.5 Colorado 7139 5518 4852 3.7 4759 3770 79.2 989 20.8 6.4 Minnesota 5877 5116 4353 3.1 3694 3176 86.0 518 14.0 4.2 Nevada 3973 3490 3038 2.6 1852 1606 86.7 246 13.3 3.2 Total 2,687,051 1,865,928 1,742,969 7.0 1,043,706 903,763 86.6 139,943 13.4 7.9 J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199196 between 1970 and 2004. Then, applying life tables for whites, we calculate the fraction of black excess deceased who would have survived to 2004 had they faced white mortality rates. Finally, using the election data we then estimate the fraction of these hypothet- ical black survivors who (1) would have voted; and (2) would have voted for each party's nominee e John Kerry (D) or George W. Bush (R) e, assuming that the voting behavior of these hypothetical survivors would have resembled the voting behavior of the existing black population stratified by sex, age, and state of residence. To provide a context for the magnitude of our findings, we compare the number of voting-age hypothetical black survivors in 2004 to the total number of black disenfranchised felons and ex-felons in that year (For detailed descriptions of our estimation procedures, please see the Methodological Appendix.) In addition, we examine the possible partisan-electoral effect of black votes lost to excess mortality at the state level by identifying close senate and gubernatorial elections between 1970 and 2004. In this calculation, we make the conservative assumption that the number of black hypothetical survivors in each year prior to 2004 would be no greater than the number in 2004. Given the stability of the black-white mortality gap, this assumption, while crude, should not bias our conclusions. In addition, as explained above, we un- derestimate the 2004 accumulated excess black deaths by excluding from analysis those excess deaths occurring before 1970. We looked for senate and gubernatorial elections in which the margin of victory for the Republican compared to the Democratic candidate was 35% or less than the number of hypothetical black survivors in that state, positing that these races were sufficiently close that had blacks survived at white rates, it is reasonably likely that the election result would have been reversed. To place our findings in context, we then compared these results to the impli- cations for state elections of total felony disenfranchisement calculated by Uggen and Manza (Manza and Uggen, 2006; Uggen and Manza, 2002). 5. Results As shown in Table 1, implementing our methods, we calculate 2.7 million excess deaths among US blacks (ages 0e84 years) from 1970 to 2004 in the 32 study states plus the District of Columbia (for additional estimations see Appendix Table A2). Considering that the total US black population was 22.6 million in 1970 and 36.1 million in 2004, this number represents 20% of the total national black population growth in this period. The total number of black deaths would have been reduced from 8.5 million to 5.8 million if blacks faced the same mortality schedules as whites. Thus, 1 out of every 3 black deaths occurring within this time period was an excess death. Of the 2.7 million black excess deaths, we project a total of 1.87 million hypothetical survivors to 2004, 1.74 million of voting age, about 1 million of whom would have been voters (Table 1). This number represents 7.9% of the national black vote in the election. Because black Americans vote overwhelmingly Democratic, black excess mortality disproportionally diminishes the Democratic Party voting base. According to our calculations, Democratic pres- idential candidate John Kerry lost 86.6% of the total number of all J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199 197 black votes lost to black excess mortality e or about 900,000 votes e while Republican candidate George W. Bush lost about 13.4%, or 140,000, of these votes. However, despite the closeness of the election, the additional black votes from our estimates of hypo- thetical survivors alone would not have been sufficient to reverse Bush's win. When we combined the effects on the black voting population in 2004 of both excess mortality and felony disenfranchisement, we found that 1 in 7 (15%) of all voting-age blacks did not have the opportunity to vote in that year for one of these two reasons (Table A2, Appendix). In 2004, a total of 166 presidential electoral votes (61.5% of the 270 needed to elect a president) were disputed in states where at least 15% of voting-age blacks did not have the opportunity to vote either due to premature death or felony disenfranchisement. Turning to state level results, we estimate that between 1970 and 2004 the outcomes of 7 close senate elections, and of 11 close gubernatorial elections would have been reversed from Republican to Democratic victors with the addition of black hypothetical sur- vivors alone (see Tables 2 and 3). Uggen and Manza's estimations of the impact of total felony disenfranchisement on senate elections (Manza and Uggen, 2006) suggest it could have reversed 7 senate races between 1978 and 2004, of which 4 overlap with the 7 we have identified. In one additional election, the 2002 senate election in Missouri, adding excess mortality and felony disenfranchisement effects together, we infer that the Democrat would have won, even though neither effect alone would have been sufficient to change the election. 6. Discussion In this study, we provide the first estimates of the impact of racial mortality differentials on political participation in the US. We find that premature deaths among blacks have had a significant impact on the racial composition of America's electorate and, dur- ing the study period, may have been a key influence on several state election outcomes. State level findings suggest that our estimated effects could have had political potency at the national level, as well, given that the predicted reversal of specific senate elections would have sustained Democratic control of the Senate from 1986 to 2002 (Manza and Uggen, 2006; Purtle, 2013). In our calculations we were able to account for only 35 years of mortality exposure rather than the ideal of 84 years. Thus, while somewhat crude, on balance, our results underestimate the effect that black excess mortality has on the size of the black population and electorate. Even with our truncated years of data, we estimated 1.74 million total black voting-age hypothetical survivors in 2004 Table 2 Crude test of hypothetical effect of mortality gaps on US senate elections. Territory e year Republican candidate votes (two-party vote share %) Democratic can (two-party vot Georgia e 1992 635,118 618,7 50.7 49. Florida e 1988 2,051,071 2,016 50.4 49. Georgia e 1980 803,686 776,1 50.9 49. North Carolina e 1980 898,064 887,6 50.3 49. Texas e 1978 1,151,376 1,139 50.3 49. Virginia e 1978 613,232 608,5 50.2 49. Nevada e 1974 79,605 78,9 50.2 49. (see Table 1). This number is close to the 1.95 million black voting- age disenfranchised felons and ex-felons in the year 2004 esti- mated by Manza and Uggen (2006). Manza and Uggen's estimated figure is widely considered significant enough to have changed electoral outcomes, especially in local elections (Manza and Uggen, 2006; Uggen and Manza, 2002). Blacks having the same mortality schedules as whites during the study period could have yielded different results in other political arenas as well e state legislatures, cities, counties, and congres- sional districts, although data limitations precluded us from these calculations. If so, they also may have acted to reduce the gerrymandering and redistricting by the majority that dilutes the political power of racial minorities, potentially altering national congressional majorities. The impact of these hypothetical black survivors might have been felt in additional aspects of the demo- cratic process, such as Democratic primaries, in which blacks manifest a high vote share in key Southern states (some over 40%), or in the electoral college presidential vote, especially when a small number of swing states decide the election. The current study findings suggest that excess black mortality has contributed to imbalances in political power and representa- tion between blacks and whites. Politics helps determine policy, which subsequently affects the distribution of public goods and services, including those that shape the social determinants of health, which influence disenfranchisement via excess mortality. In the United States, especially after the political realignment of the 1960s, policy prescriptions emanating from government structures and representing ideologically divergent constituencies have influenced the social determinants of health, including those that affect racial disparities. And given the critical role of elected poli- ticians in the policy-making apparatus, the available voter pool is an essential mechanism for the distribution of interests that will ultimately be represented in the policies and programs that affect us all. Thus, our examination suggests that large and persistent black- white mortality disparities have been both a cause and a conse- quence of partisan US politics over the past 40 years. In our polarized electoral environment, partisan electoral implications can translate into important policy differences. On a speculative level, there are a huge number of ‘what might have been?’ hy- potheses. For example, our estimates suggest that some recent Republican governors may have been defeated by their Democratic opponent, if hypothetical survivors were included in the electorate. The state-by-state question of whether or not to incorporate the Medicaid expansions provided by the Affordable Care Act has proven to be a highly partisan issue, suggesting that different electoral outcomes in states with Republican governors might have didate votes e share %) Two-party vote difference (%) 35-year votingage hypothetical black survivors (% black VAP) 74 16,344 95,052 3 1.3 5.3 ,553 34,518 99,705 6 0.8 5.6 43 27,543 95,052 1 1.7 5.3 53 10,411 86,960 7 0.6 6.7 ,149 12,227 100,065 7 0.5 5.6 11 4721 62,771 8 0.4 6.0 81 624 3038 8 0.4 2.6 Table 3 Crude test of hypothetical effect of mortality gaps on US gubernatorial elections. State e year Republican candidate votes (two-party vote share %) Democratic candidate votes (two-party vote share %) Two-party vote difference (%) 35-year voting-age hypothetical black survivors (% black VAP) Alabama e 2002 672,225 669,105 3120 54,778 50.1 49.9 0.2 6.7 New Jersey e 1997 1,133,394 1,107,968 25,426 75,049 50.6 49.4 1.1 9.1 Alabama e 1994 604,926 594,169 10,757 54,778 50.4 49.6 0.9 6.7 New Jersey e 1993 1,236,192 1,210,130 26,062 75,049 50.5 49.5 1.1 9.1 Mississippi e 1991 361,296 349,775 11,521 45,604 50.8 49.2 1.6 6.4 Michigan e 1990 1,276,134 1,258,539 17,595 90,554 50.3 49.7 0.7 9.1 New Jersey e 1981 1,145,999 1,144,202 1797 75,049 50.0 50.0 0.1 9.1 Louisiana e 1979 690,691 681,134 9557 71,700 50.3 49.7 0.7 7.2 Texas e 1978 1,183,828 1,166,919 16,909 100,065 50.4 49.6 0.7 5.6 Ohio e 1974 1,532,214 1,520,554 11,660 63,685 50.2 49.8 0.4 6.8 South Carolina e 1974 266,338 248,861 17,477 62,175 51.7 48.3 3.4 7.3 J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199198 affected the health insurance prospects of millions of poor resi- dents of those states. In general terms, what difference might it make for American democracy, if we were able to diminish or eliminate the several sources of disproportionate black disenfranchisement? Would minority interests be better reflected in government and policy? Suggestive evidence comes from a recent study where Fowler (2013) investigated the implementation of compulsory voting in Australia � a democracy with past turnout inequalities between different social groups similar to present-day inequalities in the US. The investigation was undertaken to respond to the question of what would happen to public policy and the partisan composition of government if the electorate were substantially expanded. Findings from this research suggest that both policy and govern- ment became more representative of the aggregate interests of the citizenry, including the most disadvantaged who did not vote before the adoption of compulsory voting. Our findings highlight that black excess deaths are a challenge to democracy. While we have presented the results of a statistical exercise, the meaning of lives lost too soon cannot be reduced to aggregate numbers. As with all human beings, it matters whether a black person is alive or dead. Each of the hypothetical survivors represented in our results had a name, a personal history, a family, a community, human rights, and the potential to continue to contribute. They matter. After reconstruction, along with literacy tests and poll taxes, the lynch mob was used explicitly to rob blacks of their votes and to intimidate surviving blacks from fully exer- cising their rights, including to vote (Markovitz, 2004). Although less spectacular or overtly intentional than the noose, the culture of impunity that allows us to escape accountability for the structural violence that disproportionately cuts black lives short e whether through acute injury, a discriminatory and militarized criminal justice system, or the accumulated physiological insults inherent to everyday life at the margins of a race-conscious society e remains a moral failure and threatens democracy. Funding This research was supported in part by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (Grant #T32 HD007339) and by the Center for Advanced Study in the Behavioral Sciences, Stanford University. Acknowledgments The authors would like to acknowledge helpful conversations with James DeNardo, Teresa E. Seeman, Mark Q. Sawyer, David O. Sears, and Peter M. Bentler, the support of Libbie Stephenson and Jamie Jamison at the UCLA Social Science Data Archive, and the editorial assistance of N.E. Barr. The findings and conclusions expressed are solely those of the authors and do not represent the views of Mathematica Policy Research. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.socscimed.2015.04.014. References Barreto, M.A., Nuno, S.A., Sanchez, G.R., 2009. The disproportionate impact of voter- ID requirements on the electoratednew evidence from Indiana. PS Polit. Sci. Polit. 42, 111e116. Blakely, T.A., Kennedy, B.P., Kawachi, I., 2001. Socioeconomic inequality in voting participation and self-rated health. Am. J. Public Health 91, 99. Bobo, L.D., Dawson, M.C., 2009. A change has come. Du Bois Rev. Soc. Sci. Res. Race 6, 1e14. Bonilla-Silva, E., 2013. Racism without Racists: Color-Blind Racism and the Persis- tence of Racial Inequality in America. Rowman & Littlefield Publishers. Caplow, T., Simon, J., 1999. Understanding prison policy and population trends. Crime Justice 63e120. Crimmins, E.M., Johnston, M., Hayward, M., Seeman, T., 2003. Age differences in allostatic load: an index of physiological dysregulation. Exp. Gerontol. 38, 731e734. Dawson, M.C., 1995. Behind the Mule: Race and Class in African-American Politics. Princeton University Press. Diez Roux, A., Mair, C., 2010. Neighborhoods and health. Ann. N. Y. Acad. Sci. 1186, 125e145. Diez Roux, A., Merkin, S.S., Arnett, D., Chambless, L., Massing, M., Nieto, F.J., et al., 2001. Neighborhood of residence and incidence of coronary heart disease. N. Engl. J. Med. 345, 99e106. Dorling, D., 1998. Whose voters suffer if inequalities in health remain? J. Contemp. Health 7, 50e54. Dorling, D., 2010. All connected? Geographies of race, death, wealth, votes and births. Geogr. J. 176, 186e198. Enos, R.D., 2011. Racial Threat: Field-Experimental Evidence that a Proximate Racial http://dx.doi.org/10.1016/j.socscimed.2015.04.014 http://dx.doi.org/10.1016/j.socscimed.2015.04.014 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref57 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref57 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref57 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref57 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref57 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref1 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref1 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref2 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref2 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref2 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref3 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref3 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref3 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref4 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref4 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref4 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref5 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref5 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref5 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref5 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref6 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref6 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref7 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref7 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref7 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref8 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref8 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref8 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref8 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref9 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref9 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref9 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref10 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref10 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref10 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref11 J.M. Rodriguez et al. / Social Science & Medicine 136-137 (2015) 193e199 199 Outgroup Activates Political Participation. Harvard University, Unpublished Manuscript, pp. 1e50. Epstein, D., O'Halloran, S., 1999. Measuring the electoral and policy impact of majority-minority voting districts. Am. J. Polit. Sci. 367e395. Fowler, A., 2013. Electoral and policy consequences of voter turnout: evidence from compulsory voting in Australia. Q. J. Polit. Sci. 8, 159e182. Geronimus, A.T., 2000. To mitigate, resist, or undo: addressing structural influences on the health of urban populations. Am. J. Public Health 90, 867. Geronimus, A.T., 2013. Jedi public health: leveraging contingencies of social identity to grasp and eliminate racial health inequality. In: Gomez, L., N Lopez (Eds.), Mapping Race and Inequality: A Critical Reader on Health Disparities Research. Rutgers University Press. Critical Issues in Health and Medicine series: chapter 11. Geronimus, A.T., Bound, J., Colen, C.G., 2011. Excess black mortality in the United States and in selected black and white high-poverty areas, 1980e2000. Am. J. Public Health 101, 720. Geronimus, A.T., Bound, J., Waidmann, T.A., 1999. Poverty, time, and place: variation in excess mortality across selected US populations, 1980e1990. J. Epidemiol. Community Health 53, 325e334. Geronimus, A.T., Bound, J., Waidmann, T.A., Hillemeier, M.M., Burns, P.B., 1996. Excess mortality among blacks and whites in the United States. N. Engl. J. Med. 335, 1552e1558. Geronimus, A.T., Hicken, M., Keene, D., Bound, J., 2006. ‘Weathering’ and age- patterns of allostatic load scores among blacks and whites in the United States. Am. J. Public Health 96, 826e833. Geronimus, A.T., Hicken, M.T., Pearson, J.A., Seashols, S.J., Brown, K.L., Cruz, T.D., 2010. Do US black women experience stress-related accelerated biological ag- ing? Hum. Nat. 21, 19e38. Geronimus, A.T., Pearson, J.A., Linnenbringer, E.P., Schulz, A.J., Reyes, A., Epel, E., Lin, J., Blackburn, E., 2015. Race/Ethnicity, poverty, urban stressors and telomere length in a Detroit community-based sample. J. Health Soc. Behav. (in press). Gilens, M., 1995. Racial attitudes and opposition to welfare. J. Polit. 57, 994e1014. Henry, P.J., Reyna, C., 2007. Value judgments: the impact of perceived value viola- tions on American political attitudes. Polit. Psychol. 28, 273e298. Kaufman, J.S., Long, A.E., Liao, Y., Cooper, R.S., McGee, D.L., 1998. The relation be- tween income and mortality in US blacks and whites. Epidemiology 147e155. Keiser, L.R., Mueser, P.R., Choi, S.W., 2004. Race, bureaucratic discretion, and the implementation of welfare reform. Am. J. Polit. Sci. 48, 314e327. Lane, K.A., Jost, J.T., Parks, G.S., Hughey, M.W., 2011. Black man in the White House: Ideology and implicit racial bias in the age of Obama. Obamas a (Post) Racial Am. 48e69. Lavariega Monforti, J., Sanchez, G.R., 2010. The politics of perception: an investi- gation of the presence and sources of perceptions of internal discrimination among Latinos. Soc. Sci. Q. 91, 245e265. LaVeist, T.A., 1992. The political empowerment and health status of African- Americans: mapping a new territory. Am. J. Sociol. 1080e1095. Lemann, Nicholas, 2011. The Promised Land: The Great Black Migration and How it Changed America. Vintage. Manza, J., Uggen, C., 2006. Locked Out: Felon Disenfranchisement and American Democracy. Oxford University Press, USA. Markovitz, Jonathan, 2004. Legacies of Lynching: Racial Violence and Memory. University of Minnesota Press. Marmot, M., Bell, R., 2009. Action on health disparities in the United States: com- mission on social determinants of health. JAMA: J. Am. Med. Assoc. 301, 1169e1171. Massey, D.S., Denton, N.A., 1989. Hypersegregation in US metropolitan areas: black and Hispanic segregation along five dimensions. Demography 26, 373e391. McCartney, G., Collins, C., Dorling, D., 2010. Would action on health inequalities have saved New Labour? Br. Med. J. 340. McEwen, B.S., Seeman, T., 1999. Protective and damaging effects of mediators of stress: elaborating and testing the concepts of allostasis and allostatic load. Ann. N. Y. Acad. Sci. 896, 30e47. Mendelberg, T., 2008. Racial priming revived. Perspect. Polit. 6, 109e123. Pearson, J.A., 2008. Can't buy me Whiteness. Du Bois Rev. Soc. Sci. Res. Race 5, 27e47. Pratto, F., Sidanius, J., Stallworth, L.M., Malle, B.F., 1994. Social dominance orienta- tion: a personality variable predicting social and political attitudes. J. Personal. Soc. Psychol. 67 (4), 741e763. Purtle, J., 2013. Felon disenfranchisement in the United States: a health equity perspective. Am. J. Public Health 103, 632e637. Rabinowitz, J.L., Sears, D.O., Sidanius, J., Krosnick, J.A., 2009. Why do White Amer- icans oppose race e targeted policies? Clarifying the impact of symbolic racism. Polit. Psychol. 30, 805e828. Rodriguez, J.M., Bound, J., Geronimus, A.T., 2013. US infant mortality and the President's party. Int. J. Epidemiol. 43, 818e826 dyt252. Rodriguez, J.M., Bound, J., Geronimus, A.T., 2014. Rejoinder: time series analysis and US infant mortality: de-trending the empirical from the polemical in political epidemiology. Int. J. Epidemiol. 43, 831e834 dyt284. Satcher, D., 2010. Include a social determinants of health approach to reduce health inequities. Public Health Reports 125, 6. Satcher, D., Fryer, G.E., McCann, J., Troutman, A., Woolf, S.H., Rust, G., 2005. What if we were equal? A comparison of the black-white mortality gap in 1960 and 2000. Health Aff. 24, 459e464. Schneider, A., Ingram, H., 1993. Social construction of target populations: implica- tions for politics and policy. Am. Polit. Sci. Rev. 87, 334e347. Schulz, A.J., Kannan, S., Dvonch, J.T., Israel, B.A., Allen, A., James, S.A., House, Lepkowski, J., 2005. Social and physical environments and disparities in risk for cardiovascular disease: the healthy environments partnership conceptual model. Environ. Health Perspect. 113, 1817e1825. http://dx.doi.org/10.1289/ Ehp.7913. Sears, D.O., Kinder, D.R., 1971. Racial Tension and Voting in Los Angeles: Institute of Government and Public Affairs. University of California. Sidanius, J., Pratto, F., 2001. Social Dominance. Cambridge University Press. Smith, G.D., Dorling, D., 1996. “I'm all right, John”: voting patterns and mortality in England and Wales, 1981e92. Br. Med. J. 313, 1573. Tesler, M., Sears, D.O., 2010. Obama's Race: The 2008 Election and the Dream of a Post-racial America. University of Chicago Press. Thompson, J.P., 2005. Double Trouble: Black Mayors, Black Communities, and the Call for a Deep Democracy. Oxford University Press, USA. Trebbi, F., Aghion, P., Alesina, A., 2008. Electoral rules and minority representation in US cities. Q. J. Econ. 325e357. Uggen, C., Manza, J., 2002. Democratic contraction? Political consequences of felon disenfranchisement in the United States. Am. Sociol. Rev. 777e803. Uggen, C., Shannon, S., Manza, J., 2012. State-level Estimates of Felon Disenfran- chisement in the United States, 2010: Sentencing Project. Valentino, N.A., Hutchings, V.L., White, I.K., 2002. Cues that matter: how political ads prime racial attitudes during campaigns. Am. Polit. Sci. Rev. 96, 75e90. Valentino, N.A., Sears, D.O., 2005. Old times there are not forgotten: race and partisan realignment in the contemporary South. Am. J. Polit. Sci. 49, 672e688. Williams, D.R., Mohammed, S.A., 2009. Discrimination and racial disparities in health: evidence and needed research. J. Behav. Med. 32, 20e47. http://refhub.elsevier.com/S0277-9536(15)00243-9/sref11 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref11 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref11 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref12 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref12 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref12 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref13 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref13 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref13 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref14 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref14 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref15 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref15 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref15 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref15 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref15 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref16 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref16 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref16 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref16 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref17 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref17 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref17 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref17 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref17 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref18 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref18 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref18 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref18 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref19 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref19 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref19 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref19 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref20 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref20 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref20 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref20 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref21 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref21 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref21 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref22 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref22 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref23 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref23 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref23 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref24 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref24 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref24 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref25 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref25 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref25 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref26 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref26 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref26 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref26 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref27 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref27 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref27 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref27 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref28 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref28 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref28 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref29 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref29 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref30 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref30 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref31 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref31 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref32 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref32 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref32 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref32 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref33 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref33 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref33 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref34 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref34 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref35 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref35 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref35 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref35 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref36 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref36 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref58 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref58 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref58 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref37 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref37 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref37 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref37 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref38 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref38 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref38 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref39 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref39 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref39 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref39 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref39 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref40 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref40 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref40 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref41 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref41 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref41 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref41 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref42 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref42 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref43 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref43 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref43 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref43 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref44 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref44 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref44 http://dx.doi.org/10.1289/Ehp.7913 http://dx.doi.org/10.1289/Ehp.7913 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref46 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref46 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref47 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref48 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref48 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref48 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref49 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref49 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref50 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref50 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref51 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref51 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref51 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref52 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref52 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref52 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref53 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref53 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref54 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref54 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref54 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref55 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref55 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref55 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref56 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref56 http://refhub.elsevier.com/S0277-9536(15)00243-9/sref56 Black lives matter: Differential mortality and the racial composition of the U.S. electorate, 1970–2004 1. Introduction 2. Theoretical framework 3. Racial mortality gaps and electoral politics 4. Research design 5. Results 6. Discussion Funding Acknowledgments Appendix A. Supplementary data References