Mortality, Incarceration, and African-American
Disenfranchisement in the Contemporary United States1
David Cottrell2
Michael C. Herron3
Daniel A. Smith5
Javier M. Rodriguez4
January 24, 2017
1A
previous draft of this paper was presented at the 2016 Annual Meeting of the Midwest Political
Science Association. The authors thank Josiah Alexakos and Brendan Krimsky for research assistance
and Matthew Hayes, Melissa Herman, Sean Westwood, and seminar participants at Dartmouth College,
Freie Universität Berlin, The Ohio State University, the University of California at Los Angeles, and Yale
University for helpful comments.
2 Postdoctoral Research Fellow, Program in Quantitative Social Science, Dartmouth College. 6108
Silsby Hall, Hanover, NH 03755 ([email protected]).
3 Visiting Scholar, Hertie School of Governance, Berlin, Germany, and Professor of Government, Dartmouth College. 6108 Silsby Hall, Hanover, NH 03755-3547 ([email protected]).
4 Researcher, Mathematica Policy Research, 505 14th Street, Suite 800, Oakland, CA 94612-1475, and
Research Affiliate, Population Studies Center, Institute for Social Research, University of Michigan. 426
Thompson St., Ann Arbor, MI 48103-2321 ([email protected]).
5 Professor of Political Science, University of Florida, 234 Anderson Hall, Gainesville, FL 32611-7325
([email protected]).
Abstract
African-Americans in the United States experience disproportionately high rates of mortality and
incarceration compared to Whites. This has profoundly diminished the number of voting-eligible
African-Americans in the country, and as of mid-2010 deleterious living conditions have cost
approximately 3.9 million African-American men and women the right to vote. This amounts
to a national African-American disenfranchisement rate of 13.2%. While many disenfranchised
African-Americans have been stripped of their voting rights by felony disenfranchisement laws, the
majority are literally “missing” from their communities due to premature mortality and incarceration. Leveraging variation in gender ratios across the United States, we show that disenfranchised
African-Americans are concentrated in the Southeast and that African-American disenfranchisement rates in some legislative districts lie between 20% and 40%. Despite the many successes of
the Voting Rights Act and the civil rights movement, high levels of African-American disenfranchisement remain a continuing feature of the American polity.
Introduction
There is a dramatic gap in living conditions in the United States between Whites and AfricanAmericans. The latter are disproportionately poor (Meyer and Wallace 2009), less healthy (Orsi,
Margellos-Anast and Whitman 2010), receive less education (Jencks and Phillips 1998), face
greater barriers to success in the labor market (Fairlie and Kletzer 1998; Fryer, Jr., Pager and
Spenkuch 2013), are imprisoned at greater rates (Hero 2007; Alexander 2012), and ultimately die
younger (Geronimus et al. 1996; Geronimus, Bound and Colen 2011). As of mid-2010, such deleterious living conditions have caused an excess 2.74 million voting-age African-American men
and women to be “missing” from their communities due to premature death and incarceration. In
conjunction with an excess 1.17 million African-Americans who have been released from prison
yet are not allowed to vote on account of laws restricting the rights of ex-felons, we estimate that,
were African-American and White living conditions equivalent across the United States, approximately 3.9 million more African-Americans would have been enfranchised as of 2010. All told,
this amounts to an effective African-American disenfranchisement rate of 13.2%.
We take as a baseline White living conditions in the country and assume that the contemporary
White disenfranchisement rate in the United States constitutes the country’s normal rate of disenfranchisement. Thus, if 5% of Whites and 7% of African-Americans in the United States are disenfranchised, then African-Americans are disproportionately—or excessively—disenfranchised by
two percentage points relative to Whites. To keep our language as straightforward as possible,
we refer to disproportionatley disenfranchised African-Americans simply as disenfranchised, and
thus what we call the African-American disenfranchisement rate—namely, the 13.2% figure noted
above—should be understood as the rate at which African-Americans lack the right to vote over
and beyond that of Whites.
The sheer number—around 3.9 million—of disenfranchised African-Americans that we identify is striking and indicative of ongoing structural barriers to electoral participation. Political
1
science literature has, for the most part, not engaged the question of whether the absence of
several million members of a racial group from the American polity has implications for voting rights. It does—not just theoretically but, as we show here, practically. To be clear, our
results on African-American disenfranchisement are not characterizations of the pre-Voting Rights
Act South, where legal barriers to African-American electoral participation were both overt and
comprehensive (Kousser 1974). Rather, as of mid-2010, African-American disenfranchisement
is woven into states, Congressional Districts, and state legislative districts, both upper and lower,
throughout the United States. Of the 50 states, we find that 22 have African-American disenfranchisement rates of at least ten percent. While many federal and state legislative districts have
very few African-Americans and are not substantively affected by African-American disenfranchisement, others are dramatically affected. For example, in Virginia’s 3rd Congressional District,
where African-Americans comprise nearly 53% of the district’s voting age population, we estimate an African-American disenfranchisement rate of just under 39%; this figure dwarfs the rate
at which strict voter identification laws can be thought of as disenfranchising the poor (Barreto
et al. 2009). We find as well that city council districts in some major cities reflect the realities of
African-American disenfranchisement in a profound way. Among Chicago’s 50 wards, for example, the three with the largest African-American populations have an aggregate African-American
disenfranchisement rate of 20%. Like the Virginia example above, such a disenfranchisement rate
is large by any modern standard.
African-Americans who die prematurely or are disproportionately incarcerated—two key
mechanisms of what can be thought of as political exclusion—cannot vote to alleviate the deleterious living conditions that led to their exclusion. We interpret this as a violation of the principle of
affected interests, which states that individuals affected by a government decision should have the
right to participate in the making of said decision (Dahl 1990; Goodin 2007; Fung 2013). With this
principle as background, the extent to which all citizens in a democracy can avail themselves of
the right to vote is thus an important indicator of democratic health. For this reason alone, under2
standing the nature of contemporary African-American disenfranchisement in the United States is
important. Moreover, our results show that the form of African-American disenfranchisement we
identify has the potential to change election outcomes.
Our contributions are twofold. First, we count disproportionately disenfranchised AfricanAmericans in the United States as of mid-2010. Second, we build on Rodriguez et al. (2015) by
locating disenfranchised African-Americans in Congressional Districts, state legislative districts,
and some city council districts. Voting is inherently a local activity, and any assessment of the
consequences of African-American disenfranchisement must contend not only with how many
African-Americans are disenfranchised overall in the United States but also where these disenfranchised African-Americans live—or, in the case of missing African-Americans, where they would
live if they were neither deceased nor incarcerated. To this end, we offer a model that uses gender
ratios to distribute disenfranchised African-Americans across the United States. The model assumes that the locations in the United States with the most skewed gender ratios—meaning, places
where the ratios of men to women differ drastically between Whites and African-Americans—are
those where African-American “missingness” is most severe.
Our results constitute an imperative to re-conceptualize theoretically the understanding of disenfranchisement and representation. Many of the African-Americans we label as disenfranchised
are deceased, and one might argue that our characterizing these individuals as disenfranchised
stretches this term’s meaning. Nonetheless, we believe that a complete reckoning of disenfranchisement in any country cannot disregard members of a group who, on account of marginalization,
are physically deprived of the right to vote.
In what follows we provide a theoretical motivation for our analysis and review current debates
on voting rights and contemporary African-American living conditions. We then provide intuition
for how we count the number of disenfranchised African-Americans in the United States; detailed
calculations are in an appendix. Next, we present results on African-American disenfranchisement across states, Congressional Districts, state legislative districts, and selected cities. Our final
3
section concludes with suggestions for future research on African-American disenfranchisement
and political participation and with an implicit call for re-examining how we should think about
disenfranchisement.
Voting rights and disenfranchisement
Who has the right to vote? This question has bedeviled—both theoretically and practically—
the United States since its founding and continues to do so today (Keyssar 2000; Hasen 2012;
Wang 2012). All non-incarcerated adult citizens are now legally enfranchised in the country, although there are exceptions in some states of ex-felons and those adjudged mentally incompetent (Schriner, Ochs and Shields 2000; Manza and Uggen 2006). Beyond formal mechanisms of
incarceration, the country has witnessed debates as to whether requirements for voter identification (Barreto et al. 2009; Bentele and O’Brien 2013; Rocha and Matsubayashi 2014), congested
precincts (Mukherjee 2009; Herron and Smith 2015), and difficult registration procedures (Wolfinger and Rosenstone 1980; Herron and Smith 2013) are so onerous that they can be thought of
disenfranchising.
These matters notwithstanding, the right to vote in the United States has been enshrined by the
notion of, “one person, one vote.” While this aphorism guarantees that each cast vote is treated
equally, it does not specify who precisely may participate in electoral politics and who should be
excluded. In what follows, our perspective on the matter of inclusion and exclusion draws on the
principle of affected interests. According to this principle, which we mentioned in the introduction,
voting rights should be accorded to those individuals in a polity who are affected by government
decisions. This principle motivates our consideration of the mechanisms which exclude AfricanAmericans from American political life and leads us to describe African-Americans not able to
vote as disenfranchised—even if some of these individuals are, for example, deceased.
As will be clear shortly, we focus here on sweeping mechanisms of disenfranchisement,
4
namely, premature mortality, incarceration, and the aftermath of incarceration. However, contemporary voting rights debates in the United States, both academic and otherwise, turn for the
most part on technical matters like voter identification, voting technology, and allowable days of
early voting (e.g., Tomz and van Houweling 2003; Highton 2004; Herron and Smith 2012). The
literature’s focus on technical concerns and progress in the area of minority voting rights (e.g.,
Issacharoff 2004; Ansolabehere 2006; Leighley and Nagler 2014) suggest that broader concerns
about structural disenfranchisement, the outright exclusion of a group of individuals from voting,
are, in the modern United States, anachronistic. The most egregious example of structural disenfranchisement in the United States has historically involved African-Americans. At the nation’s
founding, the franchise was limited to White men, and African-Americans lacked the right to vote
until the 15th Amendment, ratified in 1870 after the Civil War, enfranchised African-American
men de jure. Beginning in Reconstruction, strong barriers to voting prevented African-American
participation in American elections and the polity more broadly (Kousser 1974; Keyssar 2000),
and not until the Voting Rights Act (1965) and notable Supreme Court decisions like Harper v.
Virginia Board of Elections (1966), ruling that poll taxes in state elections were unconstitutional,
were African-Americans given the de facto right to vote (May 2013). Here we are referring to all
adult African-Americans, as women were enfranchised by the 19th Amendment.
African-American living conditions and disproportionate disenfranchisement
Notwithstanding trends in scholarship and the majority opinion in Shelby County v. Holder (2013),
both of which draw attention to the progress that African-Americans have made since the 1960s,
due to disparities in White and African-American living conditions a disproportionate number of
African-Americans cannot participate in electoral processes because they have died prematurely or
are imprisoned. “Disproportionate,” here, refers specifically to deceased or incarcerated AfricanAmericans who would be alive and free to vote, respectively, were African-American and White
living conditions equivalent. African-Americans who are not in their communities due to dispro5
portionate death or incarceration—but who would be alive and in their communities if AfricanAmerican living conditions were equivalent to White conditions—are said to be missing.1 Missing
African-Americans cannot in general vote.2 Beyond those missing is a second group of AfricanAmericans who have been disproportionately disenfranchised by state laws directed at formerly
incarcerated individuals. These individuals cannot vote because they reside in states where voting
rights are restricted for ex-felons.
All African-Americans who would have been at least 18 years old as of 2010, but who had
nonetheless died prior to this year, could not vote in 2010. However, it does not follow that all of
these deceased African-Americans are disenfranchised anymore than it implies that all deceased
Whites (or deceased Hispanics, or deceased Native Americans, and so forth) are disenfranchised.
Rather, only the deceased African-Americans who in 2010 would have been alive in that year had
African-American living conditions been equivalent to White living conditions can be said to be
disproportionately disenfranchised. Similarly, of ex-felon African-Americans whose voting rights
are abridged, there is a set therein whose rights would not be so abridged were African-American
1
We believe that the first use of the term “missing” in the context of African-Americans
can be found in an article on African-American men in Ferguson, MO, by Stephen Bronars.
See “Half Of Ferguson’s Young African-American Men Are Missing,” Forbes, March 18,
2015, available at http://www.forbes.com/sites/modeledbehavior/2015/03/
18/half-of-fergusons-young-african-american-men-are-missing (last accessed March 29, 2016).
2
While almost all incarcerated felons across the United States are not permitted to vote, there
are some exceptions. In Maine and Vermont, convicted felons who are incarcerated are allowed
to vote. However, there is a relatively small number of African-American men and women who
are incarcerated in these two states. As of the 2010 Census, there were in Maine ten incarcerated
African-American women and 234 incarcerated African-American men; in Vermont there were 14
incarcerated African-American women and 125 incarcerated African-American men.
6
living conditions equivalent to White conditions. It is only these non-missing African-Americans
who we say are disproportionately disenfranchised by ex-felon voting laws. With respect to these
laws in particular, it should be evident that our interest is not in the disenfranchising effect of
these laws per se; rather, we care about the disenfranchising effect of African-American living
conditions, holding fixed laws restricting ex-felon voting rights.
Traditional studies of African-American disenfranchisement, like Manza and Uggen (2006)
and Alexander (2012), tend to focus on living African-Americans; Rodriguez et al. (2015) is an
exception on which we comment below. This focus is understandable: the living African-American
population represents individuals with the capacity to vote. However, we show that ignoring disproportionate mortality within the African-American community leads to a vast underestimation of
the effect of African-American missingness on politics: the number of African-Americans missing
from the American electorate due to mortality is much greater than the number disenfranchised
by ex-felon laws. Per our invocation of the principle of affected interests, since missing AfricanAmericans are unable to participate in debates on African-American living conditions, they should
be counted as disenfranchised.
Our consideration of missing individuals is similar to that of Pettit (2012), who shows that
evidence on the educational progress of African-American men is confounded by the fact that
incarcerated African-American men, whose education levels tend to be relatively low, are often
not sampled by researchers studying educational attainment. Consequently, advances in AfricanAmerican education levels are not as high as they appear. Pettit refers to non-sampled AfricanAmerican men as “invisible,” and our notion of missing is a stronger version of this term. Missing
African-Americans are indeed invisible in that they are actually deceased and therefore excluded
from estimates in the traditional sense. In line with Pettit, a natural implication of our findings is
that measures like African-American voter registration rates may overstate the extent of AfricanAmerican electoral participation.
7
Consequences of African-American living conditions for political activity
The question of whether African-American living conditions affect political participation and representation directly through missingness is relatively unexplored. However, there is an extensive
literature on race and politics which considers the effect of living conditions on present and living
African-Americans. Many studies analyze how living conditions affect African-American turnout
rates (e.g., Rusk and Stucker 1978; Wolfinger and Rosenstone 1980; Piven and Cloward 2000) and
levels of political engagement (Squire, Wolfinger and Glass 1987; Teixeira 1987). A related line
of research examines African-American turnout as a function of the racial composition of electoral
districts (e.g., Gay 2001; Fraga 2016). Beyond turnout, African-Americans are less likely than
Whites to possess underlying documentation needed to provide proof of citizenship in order to
register to vote (Barreto, Nuño and Sanchez 2007) and are more likely to be required to cast provisional ballots (Kimball and Foley 2009). Not only are African-American electoral experiences
different than those of Whites (e.g., Herron and Smith 2012), numerous scholars have found significant differences between African-Americans and Whites in terms of policy preferences (e.g.,
Tate 1993; Kinder and Sanders 1996; Gilens 2005; Tesler and Sears 2010), even beyond social
welfare and race-related policies (e.g., Mendelberg 2008; Abrajano and Poole 2011).3
In terms of the direct effect of missingness on African-American political life, there is a growing literature on incarceration and African-American electoral participation, and a prominent example is Alexander (2012), who argues that incarceration rates in the United States of AfricanAmerican men in particular are effectively removing a large segment of the African-American
population from society in a way redolent of Jim Crow. Burch (2012) argues that Florida’s ex3
Given that African-Americans are on average poorer than Whites, analyses of socioeconomic
influences on turnout and voting have racial connotations. For example, Bartels (2008) finds
that income-related voting disparities are consequential for political representation and policy outcomes, although he rarely discusses race explicitly.
8
felon disenfranchisement laws, which disproportionately affect African-Americans, helped swing
the disputed 2000 presidential election in favor of George W. Bush. The interaction between these
laws and racial segregation, which exacerbates incarceration rates (Burch 2014), certainly has the
potential to diminish African-American political representation.
We build upon Rodriguez et al. (2015), who identify approximately 2.7 million prematurely
deceased African-American between 1970 and 2004, of whom 1.7 million would have been of
voting-age in 2004, contributing 900,000 and 100,000 votes to the Democratic and Republican
candidates, respectively, in the 2004 General Election. Rodriguez et al. argue that at least seven
senate and 11 gubernatorial elections around 2004 would have shifted from a Republican victor
to a competing Democrat. Our contribution here extends Rodriguez et al.’s line of reasoning into
2010 and in addition introduces a model that allows us examine sub-state jurisdictions, thus providing insight into how missing African-Americans are connected to political environments—like
Congressional Districts and city council districts—that are key to American democracy.
In summary, the fundamental question underneath Alexander (2012), Uggen, Shannon and
Manza (2012), Rodriguez et al. (2015), and our research is, what is the aggregate effect of AfricanAmerican living conditions on the political landscape of the United States? This question has many
facets, and our objective here is relatively simple: we want to know how many African-American
individuals would be enfranchised if contemporary African-American and White living conditions
were equivalent, and we want to know which political geographies are most affected by AfricanAmerican disenfranchisement. For the most part we limit ourselves to counting and apportioning,
and in the conclusion we consider research questions that build on findings presented here.
Counting and allocating disenfranchised African-Americans
We now describe how we count the number of disproportionately disenfranchised AfricanAmericans in the United States as of mid-2010 and allocate these African-Americans across the
9
country, to states, Congressional Districts, state legislative districts, and municipal districts. Counting and allocating are distinct: we want to know the total number of African-Americans who could
vote in the absence of a gap in African-American and White living conditions and where they could
in principle cast ballots. Throughout our analysis we assume that African-American living conditions are prior to disenfranchisement.
Determining where disenfranchised African-Americans are located requires that we identify
where missing African-Americans would have lived if they were not missing. With particular
attention to African-Americans who are missing on account of being incarcerated, if we were
able to identify the pre-imprisonment communities of these individuals, then we would be able
to locate them where they would have lived, were they not incarcerated. However, locating incarcerated African-Americans in pre-incarceration locations is not in general possible because
neither states nor the federal government publishes prior addresses of the individuals whom they
have incarcerated.4 The implication of this is that we will need a model to allocate all missing
African-Americans, both those who are prematurely deceased and those who are disproportionately incarcerated, to local areas.5
4
As of 2010, a New York state law known as Part XX prohibits so-called prison gerryman-
dering, the practice in which, for redistricting purposes, incarcerated inmates are counted as living where they are incarcerated. Details on this law can be found at http://www.latfor.
state.ny.us/data/?sec=2010amendpop (last accessed March 25, 2016).
5
In some states, the counties of incarcerated individuals are publicly available; this information
is not particularly useful, though, given the sizes of many counties. Moreover, federal prisons do
not necessarily house inmates in the states where they committed crimes and/or previously lived.
We contacted the Federal Bureau of Prisons in an attempt to gather data on the distribution of
federal prisoners’ original states or regions of sentencing, but the Bureau declined to provide us
this information. Our Bureau of Prisons Freedom of Information Act request number was 201507778, denied on October 15, 2015.
10
What we call “allocation” here is a form of counterfactual analysis in the sense of Ziblatt and
Capoccia (2010). We cannot conduct an experimental study of missing African-Americans in
the United States as there is only one United States; a “control” United States simply does not
exist. Thus, in order for us to estimate the consequences of African-American missingness on the
American polity, we have to establish a counterfactual United States and then compare this version
of the country with the realized United States (e.g., Bernhard 2009). This, broadly construed, is
the methodological background of this project.
Missing African-Americans and gender ratios
How might we know when a location in the United States has missing African-Americans? Typical
measures of populations count people who are present in localities—not individuals absent on
account of premature death or incarceration. Our answer to this conundrum lies in gender ratios.
Other scholars have exploited gender ratios in studies of missing populations (e.g., Sen 1992; Oster
2005), and gender ratio-based research is anchored in the premise that skewed gender ratios are
indicative of missing individuals.
In the United States the male-female gender ratio at birth is approximately 1.05 and, among
African-Americans, 1.02 (Branum, Parker and Schoendorf 2009). However, boys and men tend
to die and become incarcerated at greater rates than girls and women, respectively, and thus malefemale ratios for non-incarcerated individuals beyond adolescence tend to favor females.6
Table 1 reports for individuals age 18 and older the non-incarcerated male-female ratio in the
20 most heavily African-American counties in the United States (data are from the 2010 Census).
Among Whites in these counties, the gender ratio lies typically somewhat under one, and thus
women somewhat outnumber men among non-incarcerated Whites who are at least 18 years of
age. With respect to African-Americans, though, women heavily outnumber men. The African6
The literature on gender differences in mortality and incarceration is extensive (e.g., Case and
Paxson 2005; Rogers et al. 2010), exploring biological factors and behavioral factors like smoking.
11
American gender ratios in Table 1 are all lower than corresponding White ratios, and gaps between
African-American and White gender ratios are sizable. In Cook County, Illinois, which includes
the city of Chicago, there are only 76 non-incarcerated men of age at least 18 for every 100 comparable women; the relevant number of White men is 93.
Table 1: Non-incarcerated gender ratios among counties with the largest African-American populations, individuals age 18 and up
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
State
Illinois
New York
California
Texas
Michigan
Pennsylvania
Maryland
Texas
New York
Florida
Tennessee
Florida
New York
Georgia
Maryland
Ohio
Georgia
District Of Columbia
New Jersey
North Carolina
County
Cook
Kings
Los Angeles
Harris
Wayne
Philadelphia
Prince George’s
Dallas
Bronx
Miami-Dade
Shelby
Broward
Queens
Fulton
Baltimore City
Cuyahoga
DeKalb
District Of Columbia
Essex
Mecklenburg
White
0.92
0.92
0.99
0.96
0.94
0.92
1.00
0.93
0.92
1.03
0.91
0.97
0.93
0.99
0.94
0.91
0.92
0.97
0.92
0.93
African-American
0.76
0.72
0.82
0.81
0.80
0.74
0.80
0.78
0.76
0.84
0.77
0.80
0.76
0.79
0.76
0.76
0.77
0.77
0.76
0.76
Note: gender ratios are male counts divided by female counts.
Race-based differences in gender ratios are apparent at levels well below that of the county. Figure 1 describes gender ratios across 13,567 census blocks in the city of Baltimore, Maryland. We
will return to Baltimore later insofar as it was the site of Freddie Gray’s 2015 arrest and subsequent
death, an incident that has focused public attention on racial disparities and African-American living conditions. The horizontal black line in Figure 1 is at 50%, which is equivalent to a balanced
12
gender ratio. The figure’s blue line smooths gender ratios as block population becomes increasingly African-American, showing that the more African-American the population of a block, the
more it is female-dominated. While there is a potential ecological fallacy here, the racially homogeneous blocks at the ends of the spectrum in Figure 1 imply otherwise. Baltimore blocks that are
entirely White have populations that are almost 50% male, and whereas blocks that are entirely
African-American have populations around 43% male.
Figure 1: Gender ratios in Baltimore, Maryland, across census blocks
100
●
●
●
95
●
●
90
●
85
80
●
●
75
●
70
Percent male
55
50
45
40
35
30
25
20
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
● ● ●
●
●
●
●
●●
●
● ●
●
●
●
●
●
●●
●●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
● ●
●● ●
●
●
●
●
● ●
●
●●
●
● ●
● ● ● ●●●
●
● ●
●
●●●
●●
● ●
● ●
●●
● ●
●
●●● ● ● ●
● ●
●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
● ●
●
●
●
● ●●●●●
●●
●
●●
●
●●
● ● ●●
●
●
●
●
●
● ●
●
● ● ●
● ●
●
●
●
●
●
●
●
●
● ●●●
●
●
●
●
●●●●
●
●
● ●● ● ●
● ●● ●
●
●
●
● ●
● ●●
●
●
● ●●
● ●
●
●
● ●
●
● ●
●●
●
● ●
●
●●
●● ●
●●
●
● ● ●
●
●
●●
●●
●
●●
●
●
●
● ●●
●●●
●●
●
●
●●●●●●● ●●● ●● ●●● ●
●
●
●● ●
●
●
●
●● ●
●●
●
● ●
●● ●
●●● ● ●
●
●●
● ● ● ● ●● ●● ●
●
●
●
●
●
●
●
● ● ●● ●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
● ●●
● ● ●● ●
●●
● ●
●
●
● ●
●
●●●
● ● ● ●
●●
● ● ●● ● ● ● ●
●
●●
●●
●●
●
● ● ●
●
●
●
●●
● ●●
●
●
●
● ● ● ●● ● ●
●● ●
●
●
●
●
●
● ● ●
●● ● ● ●●
●●
●
●
●●
●
●
●
●
●● ●
● ●
●
●
●
●
● ●
●
● ● ● ● ●● ●●
●
● ●●●●
●
●
●
●●
●
● ●●
●
●● ●●
●
●
●
●
●
●
●
●
●
● ●
●
●
● ●
● ● ●● ●
●●
●●
●●● ●
●●●● ●
● ● ● ● ● ●● ● ● ●
● ●●
●
●
●
●
●
●
●
● ●
●
●
● ● ●
●
●
●●● ●
●
●●●●●●● ●
●●
●
●● ●
●
●
●●
●
●
● ●
●●
●
●●
●
●
●
●
●●●
●
● ●
●
●
●
● ●● ● ●
● ●
●● ● ● ●
● ●● ●●
● ●●
●
●●
●
●●●●
●●●
●● ●
● ●● ● ● ●●●
●
●
●●
● ●●
●
●
●
●● ●
●
●
●
● ●●
●●
●
●
● ●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●
●● ●
●
●
● ●
● ●●
●●
●●
● ●●
●●
●● ●
●●●
● ●
●
●●●
●
●●
●
●
●●●
●
●● ● ●
● ●●
●●
●●●
●
●●
●●●● ● ● ●● ●
●●● ● ●
●● ● ●
●
●
●● ●
●●●●● ●
● ●● ●
●●●
●●●●
●
● ● ● ●
● ●
●
●
● ● ●
● ●●
●
● ●
●
● ●● ● ●
● ● ● ●●●●
●
●
●● ●●
●
●
●
●●
●●
●
● ●
●●
●●
●
● ●●
●●
●
●
●
●●●
●● ●● ● ●● ●
●●
●●
●
●
●
●
● ● ●●
● ●
●
●
●
●
●
●●
●●●●
●
● ● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
● ● ●
●
●
● ● ●
●
●●
●● ●
●● ●●
● ● ● ●●●● ●● ●●
● ●
● ●
●
● ●●
●
●
●
●
●
●● ●
●
●
●●
●●●●●
●●●
●
● ●
●
●
●
● ●●
●
●
●
●●
●●
●
●●
●
●●
●
●
●
●
●
●
● ●● ●●●●●
●●●
●
●●
●
●
●●
●●● ●
●
●
●
●
●
●
●
●
●● ●
●
●● ●
●●●
●
●●
●● ● ●●
●
●
●
●
●●
●
●
●
●●●
●●
●
●
●
● ● ●
● ● ● ●
●
●●●●● ●
●
●
●
●
●
●●
● ● ●●
●●●
●●● ● ●
●
●
● ●●
●
●
●
●
● ●
● ● ●
● ●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●●
●
●
●●
● ●
●
● ●●
● ●●●● ●●●●
●
●
●●
●
●
●●
●
●
●
●
● ● ●
●●●
● ● ●●
●●●
●
●
●●
●●
●●
● ●● ●● ● ●● ●
●
●●
●●
●●
●
●
●
●●
●●
●
●
● ●●● ●
●
●
●
●
●●
●●● ●
● ● ● ●●
●
●● ●
●
●●
●● ●●●
●
●●
●
●
●
●●
●
●
●
●
●
●●
●●
●
●
●
●● ●●●
●
●
●
●
●
●●● ●●●
●
●
●
●●
●●
●
●
● ● ● ●●
●
●
●
●
●
●
● ● ●● ●
● ●●
●
● ●●●●
●
●
●
●
●●●
● ●● ● ●
● ● ●
●
●
●
●
●
●
● ● ●●
●
●
●
●
●
●
●
●
●
●
●●●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●● ● ●
●● ●●
●
●●
●●●
●
●
● ●
● ●
●
●
●
●●
●
●● ●
● ●
●
●●●
●●
●●●
●
●
● ●
● ●● ● ●● ●●
●
●●
●
●● ●
●
●●
●●
●
●● ●●
●
● ●
● ● ●●
●
●
●●
●●●
●●● ●
●
●
●●
●
●● ●
●●
●●
●
●
●●●
●
●●
●●● ●●●●
●●
●
● ●● ●
●
●●
● ●
● ●●●
●●
● ● ●●●●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●●
● ●●
●●
●●
●
●
●
●
● ● ●
●
●
●
●
●● ● ●●●●
●
●
●
● ●●
●
● ●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●●●
●
●●●●
●●
●
●
● ● ●●
● ●● ● ●●
●
●
●●●
● ●●
●
●●
●
●
●
●
●●
●●
●●
●
●
●●
● ● ● ●●
● ●● ●
●
●
●
●●
●
●●
●
●
● ●●
●
●● ●
●
● ●●
●
●●
●
●
●
●●
●
● ●
●
●
●●
●●● ●●
●
●
●
●
●●●
●
●
●
●
●
●
● ●
●
●
●
●● ●
●
●
●
● ●● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
● ●
●●
●
●
●
●
●●
●
●●
●●
●● ●
●
●
●
●●●
●
●
●
●
● ●● ● ●
● ● ●●
●●
●
●
● ●● ●
●
●
●
●
●
●
●
● ●●
●
●
●
●
● ●
●
●
●●
●
●
●●
●
●
●
● ●● ●
●
●
●●●●
●
●●●
●
●
●●
●●
●
●●●●
●●
●●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●● ●●●
● ●
●
● ●
●●●●●
●
●
●
●●
●●
● ●
●
● ●
●
●●
●●●
●
●
● ●
● ●
●
● ● ●
●●
●
●
●
●
●
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●
●
●
●
● ● ●●● ● ● ●● ●
●
●
● ● ●
●
●
●
●
●
●●● ●● ●
●
● ●
●
●
● ● ●●
●
●
●
●
●
●
● ●
●●
●●
●● ●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●●
●
●●
●
●
● ●
●
●
●
●●●
●● ●●●●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●●
●●●
●
●
● ●
●
●
●
●●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ● ●●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
60
●
●
●
●
●
●
65
●
●
●
●
●
●●
15
●
10
5
0
●
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95 100
Percent African−American
Note: includes blocks with at least 50 African-American and White residents, age 18 and up.
Table 1 and Figure 1 are evidence of African-American male missingness.7 As a thought
7
The underlying idea behind the table can be found in “1.5 Million Missing Black Men,” The
New York Times, September 23, 2016, available at http://www.nytimes.com/2015/04/
21/upshot/the-methodology-1-5-million-missing-black-men.html.
The
authors of this article, Justin Wolfers, David Leonhardt, and Kevin Quealy, estimate using gender ratios that there are 1.5 million missing African-American men between the ages of 25 and 54
13
exercise, one might infer the number of missing African-American males in Cook County, Illinois, based on equating White and African-American gender ratios. According to the 2010
Census, Cook County had 527,867 non-incarcerated African-American women and 399,833
non-incarcerated African-American men. African-American men are therefore 77% of AfricanAmerican women, far fewer than the 93% ratio of White men to White women. If one were to add
87,283.5 African-American men to Cook County, the African-American gender ratio there would
equal the White gender ratio. Based on this logic, there are 87,283.5 missing African-American
men in Cook County.
The exercise is a useful thought experiment, yet it assumes that no African-American females
are missing (this is wrong because African-American women die prematurely compared to White
women), that all non-missing African-American men are enfranchised (also wrong, because nonmissing African-Americans can be disenfranchised by ex-felon laws), and that there is no undercount in census values (which is wrong as well). The appendix explains how we calculate the total
number of missing African-Americans in the United States without imposing these assumptions.
Our method relies on 2010 Census data and life tables published by the National Center for Health
Statistics, part of the Centers for Disease Control. Life tables provide mortality rates, broken down
by age, gender, and race, and we use life tables from 1970 through 2010 to estimate the number
of African-Americans who would have been alive in 2010 had they been exposed to White mortality rates. The appendix also explains how we allocate missing African-Americans to sub-state
jurisdictions like Congressional Districts, state legislative districts, and municipal districts.
Intuitively, our method calculates the number of missing African-Americans by considering
what the African-American population in the United States in 2010 would look like if, since 1970,
African-American and White mortality rates were the same. African-Americans who would have
been alive in 2010 if African-American and White mortality rates were identical are missing due
to premature mortality. The same logic applies to incarceration: disproportionately incarcerated
in the United States. On the origins of the term “missing,” see fn. 1.
14
African-Americans are those who are incarcerated beyond the level of Whites, holding fixed gender
and age. Counts of missing African-Americans are apportioned to legislative districts based on
gender ratios where, as we noted already, we assume poorly gender-balanced districts (too many
African-American women compared to African-American men given the ratio of White women to
White men) are those with the most missing African-Americans.
Figure 2: Age-gender distribution of the missing African-American population as of 2010
Incarceration
Men
0−
5
5−
10 − 9
1
15 − 4
1
20 − 9
2
25 − 4
2
30 − 9
3
35 − 4
3
40 − 9
4
45 − 4
4
50 − 9
5
55 − 4
5
60 − 9
6
65 − 4
6
70 − 9
7
75 − 4
7
80 − 9
84
85+
Women
0−
5
5−
10 − 9
1
15 − 4
1
20 − 9
2
25 − 4
2
30 − 9
3
35 − 4
3
40 − 9
4
45 − 4
4
50 − 9
5
55 − 4
5
60 − 9
6
65 − 4
6
70 − 9
7
75 − 4
7
80 − 9
84
85+
Total missing, in thousands
Mortality
250
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Age
Based on calculations detailed in the appendix, there were approximately 2.8 million missing
African-Americans in the United States as of mid-2010, and Figure 2 breaks down this number
by age-gender and by disproportionate mortality (left panel) and incarceration (right panel). Both
male and female mortality distributions are skewed to the left, connoting higher excessive mortality
among older adults; for example, as of 2010, there were about 340,000 excess African-American
deaths (250,000 men and 90,000 women) for individuals age 60-64 years.8 The right panel of
8
There is a slight drop in Figure 2 around the 40-44 age group. This drop occurs because our
data capture only disproportionate African-American deaths that occurred after 1970. Although
15
Figure 2 describes by age and gender the excess number of incarcerated African-Americans.
As of mid-2010, African-American missingness due to incarceration is a smaller problem than
missingess due to mortality, and this is evident in a comparison of bar heights across the left and
right panels of the figure. Figure 2 shows as well that disproportionate African-American incarceration is almost entirely a male phenomenon.
Non-missing yet disenfranchised African-Americans
To analyze the African-American population that is non-missing yet disproportionately disenfranchised due to ex-felon voting laws, we consult Uggen, Shannon and Manza (2012)’s state-level
estimates of felon disenfranchisement. We use these estimates to determine the number of exfelon African-Americans who would have voting rights if African-Americans and non-AfricanAmericans were similarly affected by felon disenfranchisement laws. In other words, we estimate
the difference between the number of African-Americans who are actually disenfranchised by
state laws and the number of African-Americans who would be disenfranchised by said laws if
African-American and non-African-Americans had identical disenfranchisement rates.
Uggen, Shannon and Manza provide estimates at the state level of African-Americans disenfranchised by ex-felon voting laws, and we are interested in estimating disproportionate disenfranchisement at geographic levels below the 50 states.9 To allocate ex-felons within states,
not evident in Figure 2, there is a large drop in missing blacks who are exactly 41 years old in 2010,
and this reflects African-Americans who were born in 1969 and died in excess that year. These
individuals, as well as those survivors who died in excess in years previous to 1969, ideally would
be captured in the surviving population depicted in Figure 2. Because they are not, our estimates
of African-American missingness as of mid-2010 are conservative.
9
We treat the category of “non-African-American” in Uggen, Shannon and Manza (2012) as
equivalent to White. Therefore, because Hispanics make up a relatively large percent of the
disenfranchised, non-African-American population, we overstate the true rate at which Whites
16
we assume that African-Americans disproportionately disenfranchised by ex-felon laws are distributed across each state in the same way that missing African-Americans are distributed. With
this assumption, we apportion the disproportionately disenfranchised African-American ex-felon
population to sub-state geographies according to the percent of the missing population in the state.
For example, if our missingness calculations show that approximately 30% of Maryland’s missing
African-American population belongs in Baltimore City, then we assume that 30% of Maryland’s
disproportionately disenfranchised ex-felons also live in Baltimore City.
Figure 3 displays by state the total number of disenfranchised African-American ex-felons and
the number disproportionately disenfranchised. As the black bars in the figure make clear, states
with the greatest numbers of disenfranchised ex-felons can be found in the South, in states like
Florida and Virginia with large African-American populations and strict ex-felon disenfranchisement laws. The shorter red bars in Figure 3 show that, across every state pictured, the majority
of African-Americans disenfranchised by ex-felon voting laws are disproportionately disenfranchised. This is evidence that ex-felon voting laws have an excessive effect on African-Americans.
If ex-felon African-Americans in Florida and Virginia were disenfranchised at the same rate as
ex-felon non-African-Americans, then there would be an increase in the voting-eligible AfricanAmerican population in these two states by 308,313 and 177,068, respectively. There are five other
states—Tennessee, Georgia, Alabama, Texas, and Mississippi—where, in the absence of deleterious African-American living conditions, the number of African-Americans who would otherwise
be able to vote is over 50,000.10
have been disenfranchised by ex-felon laws. This makes our estimate of the number of AfricanAmericans who have been disproportionately disenfranchised conservative.
10
In Spring, 2016, Virginia Governor Terry McAuliffe restored voting rights to exfelons in his state.
our analysis.
Because our results date to mid-2010, we do not incorporate this in
See “Virginia Governor Restores Voting Rights to Felons,” The New York
Times, April 22, 2016, available at http://www.nytimes.com/2016/04/23/us/
17
Figure 3: African-Americans disenfranchised by ex-felon voting laws, by state
Florida
Virginia
Tennessee
Georgia
Alabama
Texas
Mississippi
Kentucky
Louisiana
New Jersey
California
Maryland
Missouri
New York
South Carolina
North Carolina
Nevada
Arkansas
Arizona
Wisconsin
Minnesota
Delaware
Oklahoma
Washington
Kansas
Iowa
Nebraska
Colorado
Connecticut
New Mexico
Disenfranchised
Disproportionately disenfranchised
0
50
100
150
200
250
300
350
400
450
500
Individuals, in thousands
Note: lists only states with at least 1,000 African-American ex-felons.
African-American disenfranchisement across the United States
We now turn to an assessment of the political consequences of African-American missingness and
disproportionate exposure to ex-felon voting laws. We first consider the United States as a whole,
and our results are thereafter ordered by jurisdiction, starting with states, moving to legislative
districts, and finally to city council districts in several cities.
governor-terry-mcauliffe-virginia-voting-rights-convicted-felons.
html (last accessed April 30, 2016).
18
The United States
In the country overall, we identify 3,902,862 African-Americans disenfranchised as of mid-2012
due to disproportionate mortality, incarceration, and prior-felony conviction. This amounts to
13.24% of the African-American population that would be enfranchised if African-Americans and
Whites had comparable living conditions and 1.684% of the total United States population that
would be so enfranchised.11 Of the approximately 3.9 million disenfranchised African-Americans,
2,017,344 are missing due to excessive mortality and 718,380 due to excessive incarceration.12
These national totals are striking, but they understate the degree to which African-American
living conditions have altered African-American enfranchisement at sub-national levels. Due to the
concentrated nature of the African-American population in the United States, the rate of AfricanAmerican disenfranchisement in some areas dwarfs the national rate of approximately 13%. As is
well known, African-Americans are not uniformly distributed across the United States but rather
are highly concentrated in what is historically called the “Black Belt” (Key 1949, pp. 5-7) as well
11
To calculate the national rate of African-American disenfranchisement, we divided the num-
ber of disenfranchised African-Americans (missing plus non-missing) in the country by the number of enfranchised African-Americans who would be present were African-American and White
living conditions identical. The enfranchised population is calculated by subtracting the number
of African-Americans restricted from voting due to a felony conviction from the total number of
non-incarcerated African-Americans. Then we add the excess disenfranchised African-American
population back—those disenfranchised due to excessive mortality rates, incarceration, and felon
conviction—to the population to obtain the population that would be enfranchised if AfricanAmericans and Whites shared similar disenfranchisement rates.
12
What we characterize as an overall African-American disenfranchisement rate is not necessarily the full rate of African-American disenfranchisement in the United States as of mid-2010.
For example, we ignore adult African-Americans who cannot vote on account of not possessing
required forms of voter identification.
19
as in major metropolitan areas, e.g., Baltimore, Chicago, Cleveland, and Detroit. And within many
American cities, there are areas of high African-American concentration and areas that have almost
no African-Americans (e.g., Boustan 2011). The effects of African-American disenfranchisement
thus vary tremendously across the United States, and this is readily apparent in the results below.
States
We identify 28 states with positive African-American disenfranchisement rates.13 These states
are listed in Figure 4, and the bar lengths in the figure describe the fraction of a state’s AfricanAmerican population that would have had been enfranchised had African-Americans and Whites
shared comparable living conditions.
Consistent with the concentration of African-Americans in the United States, the southeast
is heavily represented in Figure 4. Moreover, disproportionate disenfranchisement rates in five
states—Tennessee, Virginia, Florida, Alabama, and Kentucky—are at least one and a half times
that of the United States. Tennessee has the highest African-American disenfranchisement rate,
and we find that 193,401 African-Americans—just under a quarter of the African-American population who would have otherwise been able to vote in this state—are disenfranchised. Although
this is not apparent in the figure, nearly half (93,868) of those disenfranchised in Tennessee are
no longer in their communities because they are missing; the other half consists of individuals
disproportionately disenfranchised by Tennessee’s felony laws.
Figure 4’s disenfranchisement rates imply that structural disenfranchisement of African13
Negative missingness is explained in the appendix. Briefly, some states have a surfeit of
African-American men due in part to features of the national labor market. Since these states
tend to have “too many” men compared to the number of women, we conclude that the number
of missing individuals in such states is negative. The sum of the number of negative missing
African-Americans is very small compared to the sume of the positive number of missing AfricanAmericans.
20
Figure 4: African-American disenfranchisement rates, by state, as of mid-2010
24.3%
Tennessee
23.6%
Virginia
22.8%
Florida
21.5%
Alabama
20.7%
Kentucky
18.7%
Mississippi
17.2%
Georgia
15.3%
New Jersey
15.3%
New York
14.5%
Missouri
14.3%
Delaware
13.8%
Arkansas
13.8%
Maryland
13.5%
Louisiana
13.4%
South Carolina
13.0%
Illinois
12.7%
North Carolina
12.3%
Wisconsin
11.8%
Pennsylvania
10.4%
Texas
Michigan
10.2%
Connecticut
10.1%
9.6%
Indiana
9.2%
Ohio
4.9%
Massachusetts
4.1%
Oklahoma
1.8%
California
1.4%
Nevada
0%
5%
10%
15%
20%
25%
30%
Americans has a far greater effect on voting rights than many other features of African-American
political life that are considered to have significant effects. For example, Barreto et al. (2009) find
in Indiana an 11.5 percentage point gap between African-Americans and Whites in the percent of
the voting age population that has valid voter identification. This is a sizable difference, but it is
around half of what we estimate to be the African-American disenfranchisement rate in Tennessee
due to African-American living conditions. In fact, we identify 19 states (Pennsylvania and up)
whose African-American disenfranchisement rates exceed 11.5 percentage points.
A more nuanced view of African-American disenfranchisement can be see in Figure 5, which
21
Figure 5: African-American disenfranchisement across counties as of mid-2010
Percent of African−Americans
disenfranchised
50
40
30
20
10
0
describes county-level disenfranchisement rates. The importance of county government varies
across the United States, but many counties possess law enforcement, election administration, and
financial powers. The Southeast is prominent in Figure 5 as are some urban areas that have many
African-Americans, e.g., Chicago and Detroit. Moreover, Figure 5 illustrates very nicely how
much variance there is across the United States in African-American disenfranchisement rates. It
is fair to say that most counties in the country have African-American disenfranchisement rates
that are effectively zero. In others, we observe disenfranchisement rates of over 30 percent. The
extent to which political institutions are attempting to ameliorate the systemic problems exposed in
Figure 5 is beyond the scope of our analysis, but the figure makes it clear that the political salience
of African-American disenfranchisement is a matter that will more local than national.
How can we interpret our result that, in Tennessee and Virginia, over one-fifth of eligible
African-Americans are not able to vote? In our judgment, the sheer qmagnitudes of the AfricanAmerican disenfranchisement rates in some states and counties make it very difficult to speculate
on potential consequences for turnout among non-missing African-Americans, for example. We
22
return to this issue in the conclusion where we seek to put our results in perspective.
Congressional Districts
The most important lawmaking institution in the United States is Congress, and we thus turn to
estimates of African-American disenfranchisement rates across Congressional Districts. Figure
6’s histogram of African-American disenfranchisement rates across Congressional Districts in the
112th Congress shows that, while most Districts are estimated to have a tiny (near zero) number
of disenfranchised African-Americans, there are a number of Districts estimated to have tens of
thousands of disenfranchised African-American men and women. The average Congressional District in 2010 had just over 500,000 voting-age residents, and there are 73 Districts where at least
20,000 African-American men and women have been disproportionately disenfranchised due to
premature mortality, incarceration, or prior felony convictions. There are also 29 Districts with
more than 40,000 such individuals and three with over 80,000. The most dramatic example of
African-American disenfranchisement at the Congressional District level is Virginia’s 3rd District,
where the number of disenfranchised African-Americans reaches over 112,000.
Figure 6: African-American disenfranchisement across Congressional Districts as of mid-2010
160
140
120
Frequency
100
80
60
40
20
0
−10
0
10
20
30
40
50
60
70
Disenfranchised, in thousands
23
80
90
100
110
120
As shown in Figure 6, our estimates of the African-American disenfranchised population are
not always positive (see the previous fn. 13). In fact, there are 179 Districts where our estimates of
the disenfranchised population are less than zero. This occurs because gender ratios tend to vary
across geographies in ways that are not always a result of incarceration or premature mortality. For
example, the at-large Congressional District in Montana has twice as many African-American men
as it does African-American women. This is not only because the African-American population in
Montana is very small (totaling 2,641 individuals) but also because African-Americans who have
migrated to Montana are mostly men seeking employment in extractive industries. In general,
Districts with the greatest ratios of African-American men to African-American women are in
West and Midwest states, places with small African-American populations mostly composed of a
disproportionate number of men who have moved there for labor opportunities.
The negative missing values in Figure 6 are mostly very small. Among Congressional Districts with negative numbers of missing African-Americans, as many as 90% have fewer than
4,445 such individuals. In contrast, among Districts with a positive number of missing AfricanAmericans, only 30% are below that magnitude. This suggests that, although our estimates of
African-American missingness can be biased by the geography of labor markets, this bias translates into small absolute numbers of individuals.
Figure 7 is a map in which Congressional Districts are shaded according to the size of their
disenfranchised populations. The average Congressional District has nearly 9,000 disenfranchised
African-Americans, but this number is somewhat misleading on account of the concentration of
the African-American population in the South and in large, metropolitan areas. In Southern and
urban Congressional Districts, African-American disenfranchisement can be excessive. As Table
2 shows, there are five Congressional Districts—all in Virginia and Florida, where many ex-felons
are disproportionately disenfranchised—where the African-American disenfranchisement rate is
above 30%.
We alluded earlier to the issue of whether African-American disenfranchisement could change
24
Figure 7: African-American disenfranchisement across Congressional Districts as of mid-2010
Excess disenfranchised
50000+
40000
30000
20000
10000
election outcomes, and to this end Figure 8 describes African-American disenfranchisement rates
across Congressional Districts in four states. Districts are ordered by the African-American fraction of the voting age population that would be enfranchised had African-Americans and Whites
shared similar living conditions. Each solid black bar in the figure represents the actual number
of enfranchised African-Americans in a District, and the empty/not-shaded extension of each bar
reflects the size of the disenfranchised population.
Each bar in Figure 8 has two associated dots, and these dots reflect the number of enfranchised
African-Americans required for the District to be majority African-American. Empty/not-shaded
dots reflect counterfactual calculations—i.e., the number of African-Americans who would live
in a District were White and African-American living conditions equivalent—and solid dots reflect actual numbers of African-Americans. When a solid dot is above a District’s solid bar, then
25
Percent disenfranchised
District
among African-Americans
District overall
Virginia 3
38.6
21.5
Florida 2
32.6
7.0
Florida 3
32.1
16.6
Virginia 4
31.4
10.8
Florida 11
30.1
8.4
Tennessee 8
29.5
7.0
Tennessee 9
27.8
17.8
Tennessee 5
26.6
6.8
Florida 17
25.4
14.9
Alabama 7
25.0
15.9
Alabama 2
22.6
7.2
New York 11
22.4
13.8
Mississippi 2
21.6
14.4
Virginia 5
21.5
4.7
New York 10
21.4
14.6
Missouri 1
21.2
12.0
Alabama 1
20.5
5.8
Alabama 3
20.4
6.8
Georgia 13
20.0
11.7
New Jersey 10
20.0
12.0
Note: considers only Congressional Districts that are at least 20% African-American
Table 2: Congressional Districts with the greatest African-American disenfranchisement rates
African-Americans did not have an enfranchised majority in the District in 2010; for example, the
7th Congressional District in Alabama was majority African-American in 2010, but the 6th was
not. If an empty/non-shaded dot lies below the top of a District’s empty extension, then the disenfranchised population is large enough to give African-Americans the majority in the District. Key
here are Congressional Districts with solid dots that lie above solid bars (these Districts are not
majority African-American based on actual conditions) yet have empty dots below corresponding
empty bar extensions (the Districts are majority African-American when disenfranchised AfricanAmericans are counted). In Florida, for example, District 3 becomes majority African-American
when the disenfranchised population is counted, and in Georgia, Districts 2 and 5 become majority
African-American when disenfranchised African-Americans are reinserted in these jurisdictions.
26
Figure 8: African-American disenfranchisement across Congressional Districts in four states
350
●
●
300
●
●
275
250
300
●
●
●
●
●
●
●
●
●
275
●
●
●
225
250
●
●
● ●
●
● ●
●
●
●
●
●
●
●
● ●
●
●
● ●
●
●
●
●
●
● ●
●
● ●
● ●
●
● ●
●
225
Total (In Thousands)
200
Total (In Thousands)
●
●
325
175
150
125
200
● ● ●
●
●
●
175
150
125
100
100
75
75
50
50
25
25
0
0
04
06
05
01
02
03
07
10 13 09 22 05 18 14 21 24 16 15 25 07 08 20 01 19 04 06 12 02 11 03 23 17
District
District
(a) Alabama
(b) Florida
350
●
●
●
275
●
●
●
●
Total (In Thousands)
250
●
275
●
●
●
250
●
●
●
●
●
●
225
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
225
●
●
200
●
Total (In Thousands)
300
●
300
●
●
325
200
175
150
175
150
125
125
100
100
75
75
50
50
25
25
0
0
09
06
11
10
01
07
03
08
12
02
05
04
13
09
District
10
06
11
08
07
02
01
05
04
03
District
(c) Georgia
(d) Virginia
The implication of Figure 8 is that African-American disenfranchisement has changed the
majority-minority status of existing Congressional Districts. As majority-minority Districts are
disproportionately likely to yield African-American members of Congress (Lublin 1997; Valelly
2004), it follows that the racial makeup of the United States House as of 2010 reflects the extent of African-American disenfranchisement documented here. One might posit that this point is
27
somewhat superfluous insofar as abrogations of a group’s voting rights are not generally judged
based on the possibility that they have affected election results. Nonetheless, if one were to argue
that African-American disenfranchisement matters only if it has the potential to change election
outcomes, then Figure 8 is highly germane.
State legislative districts
We now turn to state legislative districts, which are akin to federal Congressional Districts in that
they are used to staff state legislatures. The nature of federalism in the United States is such
that many laws affecting African-American welfare are heavily influenced by state laws, i.e., laws
regulating tax policy, education, healthcare, election administration, and so forth. State legislative
districts contain smaller and hence more concentrated populations than Congressional Districts,
and as a result they are more vulnerable to dramatic rates of African-American disenfranchisement.
Tables 3 (upper chamber) and 4 (lower chamber) display the 20 state legislative districts
with the highest rates of African-American disenfranchisement (excluding districts less than 20%
African-American). For both chamber types, there are districts where the disenfranchised AfricanAmerican population exceeds 40% of the district’s total African-American population. Moreover,
in every district in Tables 3 and 4, over a quarter of potential African-American voters have been
disenfranchised. Finally, among the lower chamber districts in Table 4, where constituencies are
relatively small, almost a third of the African-American population is disenfranchised.
The legislative districts in Tables 3 and 4 are in Southern states with large African-American
populations, large incarcerated populations, and strict laws that disenfranchise ex-felons; this explains the recurrence of Alabama, Florida, Mississippi, Tennessee, and Virginia. Overall, our state
legislative disenfranchisement rates illustrate that inequalities in health and incarceration have major representational consequences for African-American voting blocs in many Southern state legislative districts. This would be missed were we to consider only national or state-level estimates
of African-American disenfranchisement.
28
Table 3: African-American disenfranchisement in upper chamber state legislative districts
State
Virginia
Florida
Virginia
Florida
Tennessee
Florida
Tennessee
Tennessee
Virginia
Florida
Tennessee
Tennessee
Mississippi
Virginia
Virginia
Alabama
Mississippi
Virginia
Alabama
Tennessee
District
9
32
11
1
28
6
10
31
18
18
30
33
25
2
16
19
12
19
35
19
Percent disenfranchised
African-Americans Overall
44.3
26.0
42.3
10.8
36.9
10.2
35.9
17.7
35.2
20.8
34.0
10.5
33.9
11.2
31.8
11.0
30.9
18.5
30.7
12.6
30.1
14.5
29.8
22.4
29.7
9.7
29.5
18.1
29.4
16.7
29.2
21.4
28.2
22.5
27.9
6.8
27.6
10.1
27.4
15.8
As a final point about legislative districts in the United States, Table 5 describes changes to
African-American majority status induced by enfranchising African-Americans in the three legislative district types we have considered. There are 22 Congressional Districts where AfricanAmericans make up the majority of the enfranchised population. However, if the disenfranchised
population were to be considered, there would be 28 Districts where African-Americans held the
majority of the disenfranchised population. This represents a 27% increase in majority AfricanAmerican districts. Percentage changes for state upper and state lower house districts are 17% and
20%, respectively.
With these percentages in mind, the number of federal and state legislative districts across the
United States that become majority-minority when disenfranchised African-Americans are considered is dramatic. Approximately 14% of the population of the United States identified as African29
Table 4: African-American disenfranchisement in lower chamber state legislative districts
State
Virginia
Tennessee
Virginia
Virginia
Tennessee
Virginia
Virginia
Tennessee
Virginia
Virginia
Tennessee
Virginia
Mississippi
Virginia
Tennessee
Mississippi
Tennessee
Mississippi
Mississippi
Tennessee
District
69
83
74
90
50
14
77
28
79
70
93
27
41
23
84
26
97
64
61
54
Percent disenfranchised
African-Americans Overall
42.9
26.1
39.8
6.7
39.5
26.3
39.0
23.8
38.6
9.5
35.9
14.8
35.3
21.7
35.1
19.2
35.0
15.0
34.4
22.5
33.4
15.7
33.2
9.6
32.8
23.9
32.2
9.4
32.0
27.7
32.0
25.9
31.9
9.5
31.9
12.7
31.5
7.2
31.4
20.0
American, either in part or exclusively, and around 5% of Congressional Districts are majority
African-American. Without African-American disenfranchisement, this percentage would rise to
6%. Both of these numbers are small, and this explains why the change of 5% to 6% represents
a 27% increase in the raw number of majority African-American Congressional Districts. Overall, Table 5 shows that, were disenfranchised African-Americans part of the American electorate,
there would be a significant increase in the number of legislative districts expected to elect AfricanAmerican representatives.
30
Table 5: Legislative district changes induced by African-American enfranchisement
Legislature
Total
districts
Congress
435
State upper
1934
State lower
4720
African-American
current
22
108
305
African-American
after
28
126
365
Difference
6
18
60
Note: describes counts of African-American majority districts based on existing populations
(“current”) and populations augmented with disenfranchised African-Americans (“after”).
Cities
Our final set of results focuses on cities, and here we consider Baltimore, Chicago, Cleveland,
and Detroit.14 Our interest in these cites reflects the importance of race relations in their political
landscapes, and a complete analysis of African-American disenfranchisement in the urban United
States would consider the nearly 40 American cities that have at least 100,000 African-American
residents. For the sake of brevity, we focus here on four metropolitan areas.
Table 6 lists African-American disenfranchisement rates in Baltimore, Chicago, Cleveland,
and Detroit, and these rates range from approximately 10% of potential African-American voters
to over 16%. Given the large African-American populations in these four cities, these disenfranchisement rates correspond to thousands of individuals who have been deprived of voting rights.
Particularly notable about the counts in Table 6 is that most of the disenfranchised individuals in
our four cities are disenfranchised because they are missing. Only in Baltimore did Maryland state
laws disenfranchise non-incarcerated individuals for past felony convictions.15
14
The Census does not provide the incarcerated populations broken down by sex, race and age
at the city council level. However, it does indicate the census blocks where individuals are incarcerated. We calculate the ratio of non-incarcerated African-American men to non-incarcerated
African-American women by first removing these blocks and then aggregating the remaining population by sex, race and age to the city council level.
15
Maryland
Senate
Bill
340/House
Bill
980
ended
this
practice.
See
https://www.brennancenter.org/analysis/
31
Table 6: African-American disenfranchisement in four cities
Total African-American
City
disenfranchised
Baltimore
56,316
Chicago
118,031
Cleveland
17,178
Detroit
47,031
Percent disenfranchised
among African-Americans overall
16.6
10.6
15.5
5.4
10.3
5.5
9.9
8.3
Earlier we alluded to Freddie Gray’s 2015 arrest and death in Baltimore. This incident led to
riots in the city, and Freddie Gay’s death and the public reaction to it are now part of the national
discussion on policing, the use of deadly force, and African-Americans, in particular AfricanAmerican men.16 We cannot draw causal connections between African-American disenfranchisement in Baltimore and the likelihood of race-based violence in the city, but it is hard to think that
African-American missingness in the city is wholly unrelated to what transpired there regarding
Freddie Gay. Literally thousands of African-American men are missing from Baltimore, and, more
importantly, African-American disenfranchisement effects almost all of Baltimore’s city council
districts. The question implicit here is, would Baltimore’s city council members have enacted
different policies regarding local policing if the city’s disenfranchised policy were instead enfranchised? Given our objective simply of characterizing and allocating disenfranchised AfricanAmericans, this question is beyond our scope. We expect that future research will nonetheless seek
to tackle it.
As in our analysis of Congressional districts, we now consider whether the disenfranchised
populations in our four cities of interest are large enough to have electoral consequences in city
council districts. Figure 9 shows that, in many such districts, there is a large fraction of potential
African-Americans who are unable to vote. However, in our four cities disenfranchisement is
voting-rights-restoration-efforts-maryland (last accessed on May 17,
2016).
16
For coverage of the Freddie Gray situation in Baltimore, see http://data.
baltimoresun.com/news/freddie-gray/ (last accessed March 21, 2016.)
32
Figure 9: African-American disenfranchisement across city council districts
45
35
40
●
30
35
●
30
20
●
●
●
●
●
●
●
●
●
●
●
●
●
15
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Total (In Thousands)
Total (In Thousands)
25
●
●
●
●●
●
25
20
●
●●
●
●
●
●
●
●
15
●
●●
●
●
●
●
●●●
●
●●
●●
●
● ●
●
● ●
● ●
●●
●
●
●●
●
● ● ● ● ● ●●
● ●●
●
● ● ●● ●●●
● ●
●
●●●
●
● ●
●●
●
●
●●
● ●
● ●
●
●
●
10
10
5
5
0
0
1
11
14
10
5
3
12
2
13
4
7
9
6
8
41 45 38 14 11 12 47 39 44 32 33 31 30 35 1 23 22 13 25 43 40 36 50 42 26 48 46 10 49 19 16 27 15 2 37 3 29 20 28 18 5 4 24 9 17 7 34 8 21 6
District
Ward
(a) Baltimore
(b) Chicago
90
20
85
80
75
70
15
65
●
●
●
●
●
10
Total (In Thousands)
Total (In Thousands)
60
●
● ● ● ● ●
●
● ● ●
● ●
● ● ● ● ●
●
● ● ● ●
●
●
●
●
●
●
●
55
50
45
40
35
●
●
●
●
●
●
●
●
●
●
●
●
7
5
1
●
●
30
25
5
20
15
10
5
0
0
19
13
14
15
18
3
16
17
12
9
8
11
7
5
2
4
6
10
1
6
Ward
3
4
2
District
(c) Cleveland
(d) Detroit
rarely pivotal to a district’s majority-minority status. Among the cities there is only one district—
the 10th City Council District in Baltimore—whose majority-minority status would change were
disenfranchised African-Americans to be enfranchised.
This result reinforces what can be interpreted either as a limitation of the present study or a
33
Table 7: Baltimore’s City Council Districts
District
5
10
4
14
3
2
13
6
11
12
7
8
9
1
Percent African-American
Existing
51.36
45.29
71.24
39.26
59.45
66.13
78.90
78.03
31.64
71.30
78.93
88.18
86.65
10.74
Percent African-American
After
57.39
51.07
76.24
44.07
63.60
70.14
82.19
80.84
34.19
73.71
81.33
90.37
87.64
11.42
Difference
6.03
5.78
5.00
4.81
4.15
4.01
3.29
2.82
2.55
2.41
2.40
2.19
0.99
0.67
Note: districts ordered by difference in percentage African-American where differences are based
on current districts versus districts after disenfranchised African-Americans are enfranchised.
notable feature of the contemporary United States. To the point, we have examined the effects
of African-American disenfranchisement in four prominent cities while holding fixed the (often)
racially segregated locations of African-Americans, Whites, and other groups. We have also held
fixed city council district boundaries, some of which are racially gerrymandered. This means that
our analysis will almost certainly understate the extent of the effect of African-American disenfranchisement on city political environments (and, similarly, on other legislative environments).
Many urban city council districts are heavily African-American to start and become only more
so when we insert in them enfranchised African-Americans. In the case of Baltimore, this is illustrated in Table 7, which lists the city’s council districts as well as their percent African-American.
These districts do not “flip” to become majority African-American in the absence of disenfranchisement, but this reflects high levels of racial segregation, which itself is presumably a feature of
African-American living conditions. Similarly, many non-African-American city council districts
34
do not “flip” because they are so heavily non-African-American to begin with. With this in mind,
Table 6 and Figure 9 are illustrative of the types of results that our study can bring to bear on urban environments in the United States, but a full understanding of the effect of African-American
living conditions on cities in the United States requires a counter-factual understanding of these
environments that transcends the analysis we are offering here.
Conclusion
Disenfranchisement has traditionally been understood as the revocation of political rights, either
the right to vote or the right to equal representation. Formally, large-scale disenfranchisement in the
United States is an artifact of the past: with the exception in some jurisdictions of felons, ex-felons,
and those judged to be mentally incompetent, all adult citizens in the United States have the right
to vote. We nonetheless argue that disenfranchisement can transcend formal legal institutions, and
to this end we have calculated the number—approximately 3.9 million—of African-Americans
who, as of mid-2010, cannot vote due to premature mortality, incarceration, or ex-felon status.
The principle of affected interests states that all African-Americans affected by deleterious living
conditions should be afforded the opportunity to participate in debates about these conditions, and
from this it follows that the 3.9 million African-Americans we have identified should be recognized
as disenfranchised. This number is evidence of structural racism or what Ellis (2015) might call
“Lockout.” Ellis’s characterization of lockout refers to disenfranchisement channeled through the
legal system, and here we build on this notion by considering the disenfranchising effects of excess
mortality as well.
Not only have we argued that the national African-American disenfranchisement rate is 13.2%,
but we have also shown how African-American disenfranchisement is woven into the political
fabric of the nation. Congressional Districts, state legislative districts, and selected city council
districts are confounded by the several million African-American men and women who cannot
35
vote on account of their living conditions. These findings reveal a troubling aspect of American
democracy.
We motivated this paper in part by noting that contemporary literature on American elections
tends to examine narrow barriers to voting like voter identification laws. We do not disagree that
these barriers are important and worthy of study. Nonetheless, our results imply that the most
dramatic barriers to African-American electoral participation in the United States are neither voter
identification laws nor laws restricting early voting, for example. Rather, disenfranchisement via
premature death and excessive incarceration is a feature of contemporary African-American life in
the United States that dominates all other barriers to political activity that we can think of.
Our twin objectives have been enumerating and apportioning disenfranchised AfricanAmericans across key political geographies like states and various types of legislative districts.
Thinking just about our enumeration exercise, our findings on African-American disenfranchisement almost certainly understate the effects of African-American living conditions on the United
States franchise. We take as given where Whites and non-incarcerated African-Americans live,
and this incorporates in our analysis historical legacies of racial segregation, African-American
impoverishment, enduring patterns of violence, and White flight from urban centers. One would
imagine that, in a world with equal living conditions, African-Americans would not be as concentrated as they are in the contemporary United States, and this means that our estimates of the
effects of African-American disenfranchisement are likely too narrow.
We hope that our results on national and district-level disenfranchisement rates to spur research
on the political consequences of the local disenfranchisement. How does extensive—yet legal—
African-American disenfranchisement affect turnout among enfranchised African-Americans?
Does the dramatic disenfranchisement we have documented lead enfranchised African-Americans
to vote at relatively high rates in attempts to change their living conditions; or, does disenfranchisement damage communities and lead to low levels of turnout? How different would legislative
district lines be if missing African-Americans were no longer missing? Missingess by construc36
tion diminishes the pool of available African-American candidates for office; how many potentially
valuable African-American legislators are not in office because they died prematurely? Finally, on
account of African-American missingness, legislative district median voters are almost certainly
too conservative; in future research we intend to try to characterize the extent to which federal and
state legislative medians would change were disenfranchised African-Americans enfranchised.
We conclude by emphasizing that we have restricted our attention to the gap in living conditions between two racial groups, Whites and African-Americans. Nonetheless, one perspective on
our results is that we have estimated the consequences for the United States franchise of a particular form of socioeconomic inequality, namely, explicitly race-based poverty and marginalization.
Inequality in the United States transcends racial lines, and in the future we anticipate extending our
analysis to characterizing the extent of missingness—and consequent disenfranchisement—based
solely on poverty. That is to say, we want to estimate the number of poor individuals in the United
States who would be alive, and thus could vote, if not for premature mortality. A characterization
of the relationship between poverty and disenfranchisement will help us understand the nature of
democracy in the United States and the extent to which the right to vote is truly available to all
citizens in the country.
References
Abrajano, Marisa and Keith T. Poole. 2011. Assessing the Ethnic and Racial Diversity of American
Public Opinion. In Who Gets Represented?, ed. Peter K. Enns and Christopher Wlezien. New
York: Russell Sage Foundation pp. 32–60.
Alexander, Michelle. 2012. The New Jim Crow. New York: The New Press.
Ansolabehere, Stephen. 2006. Election Administration and Voting Rights. In The Future of the
37
Voting Rights Act, ed. David L. Epstein, Richard H. Pildes, Rodolfo O. de la Garza and Sharyn
O’Halloran. New York: Russell Sage Foundation pp. 203–222.
Barreto, Matt A., Stephen A. Nuño, and Gabriel R. Sanchez. 2009. “The Disproportionate Impact of Voter-ID Requirements on the Electorate—New Evidence from Indiana.” PS: Political
Science & Politics 42(1):111–116.
Barreto, Matt A., Stephen A. Nuño and Gabriel A. Sanchez. 2007. “Voter ID Requirements and
the Disenfranchisement of Latino, Black and Asian Voters.” Paper prepared for presentation at
the annual meeting of the American Political Science Association, Chicago, IL.
URL:
https://www.brennancenter.org/sites/default/files/legal-
work/63836ceea55aa81e4f hlm6bhkse(1).pdf
Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age.
Princeton, NJ: Princeton University Press.
Bentele, Keith G. and Erin E. O’Brien. 2013. “Jim Crow 2.0? Why States Consider and Adopt
Restrictive Voter Access Policies.” Perspectives on Politics 11(4):1088–1116.
Bernhard, Michael. 2009. “Methodological Disputes in Comparative Politics.” Comparative Politics 41(4):495–515.
Boustan, Leah Platt. 2011. Racial Residential Segregation in American Cities. In The Oxford
Handbook of Urban Economics and Planning, ed. Nancy Brooks, Kieran Donaghy and GerritJan Knaap. New York: Oxford University Press pp. 318–339.
Branum, Amy M., Jennifer D. Parker and Kenneth C. Schoendorf. 2009. “Trends in US sex ratio
by plurality, gestational age and race/ethnicity.” Human Reproduction 24(11):2936–2944.
Burch, Traci. 2012. “Did Disfranchisement Laws Help Elect President Bush? New Evidence on
the Turnout and Party Registration of Floridas Ex-Felons.” Political Behavior 34(1):1–26.
38
Burch, Traci. 2014. “The Old Jim Crow: Racial Residential Segregation and Neighborhood Imprisonment.” Law & Policy 36(3):223–255.
Case, Anne and Christina Paxson. 2005. “Sex differences in morbidity and mortality.” Demography
42(2):189–214.
Dahl, Robert. 1990. After the Revolution? Authority in a Good Society. Rev. ed. New Haven, CT:
Yale University Press.
Ellis, Atiba R. 2015. “Race, Class, and Structural Discrimination: On Vulnerability within the
Political Process.” Journal of Civil Rights and Economic Development 28(1):33–59.
Fairlie, Robert W. and Lori G. Kletzer. 1998.
“Jobs Lost, Jobs Regained: An Analysis of
Black/White Differences in Job Displacement in the 1980s.” Industrial Relations 37(4):460–
477.
Fraga, Bernard L. 2016. “Candidates or Districts? Reevaluating the Role of Race in Voter Turnout.”
American Journal of Political Science 60(1):97–122.
Fryer, Jr., Roland G., Devah Pager and Jorg L. Spenkuch. 2013. “Racial Disparities in Job Finding
and Offered Wages.” Journal of Law and Economics 56(3):633–689.
Fung, Archon. 2013. The Principle of Affected Interests and Inclusion in Democratic Governance.
In Representation: Elections and Beyond, ed. Jack Nagel and Rogers Smith. Philadelpha, PA:
University of Pennsylvania Press pp. 236–268.
Gay, Claudine. 2001. “The Effect of Black Congressional Representation on Political Participation.” American Political Science Review 95(3):589–602.
Geronimus, Arline T., John Bound and Cynthia G. Colen. 2011. “Excess black mortality in the
United States and in selected black and white high-poverty areas.” American Journal of Public
Health 101(4):720729.
39
Geronimus, Arline T., John Bound, Timothy A. Waidmann and Marianne M. Hillemeier. 1996.
“Excess Mortality among Blacks and Whites in the United States.” New England Journal of
Medicine 335(21):1552–1558.
Gilens, Martin. 2005. “Inequality and Democratic Responsiveness.” Public Opinion Quarterly
69(5):778–796.
Goodin, Robert. 2007. “Enfranchising all Affected Interests, and its Alternatives.” Philosophy and
Public Affairs 35(1):40–68.
Hasen, Richard L. 2012. The Voting Wars: From Florida 2000 to the Next Election Meltdown.
New Haven, CT: Yale University Press.
Hero, Rodney E. 2007. Racial diversity and social capital: Equality and community in America.
Cambridge: Cambridge University Press.
Herron, Michael C. and Daniel A. Smith. 2012. “Souls to the Polls: Early Voting in Florida in the
Shadow of House Bill 1355.” Election Law Journal 11(3):331–347.
Herron, Michael C. and Daniel A. Smith. 2013. “The Effects of House Bill 1355 on Voter Registration in Florida.” State Politics & Policy Quarterly 13(3):279–305.
Herron, Michael C. and Daniel A. Smith. 2015. “Precinct Closing and Wait Times in Florida
during the 2012 General Election.” Election Law Journal 14(3):220–238.
Highton, Benjamin. 2004. “Voter Registration and Turnout in the United States.” Perspectives on
Politics 2(3):507–515.
Issacharoff, Samuel. 2004. “Is Section 5 of the Voting Rights Act a Victim of Its Own Success?”
Columbia Law Review 104(6):1710–1731.
40
Jencks, Christopher and Meredith Phillips. 1998. The Black-White Test Score Gap: An Introduction. In The Black-White Test Score Gap, ed. Christopher Jencks and Meredith Phillips. Washingt, DC: Brookings Institution Press.
Key, V.O. 1949. Southern Politics. New York: Knopf.
Keyssar, Alexander. 2000. The Right to Vote: The Contested History of Democracy in the United
States. New York: Basic Books.
Kimball, David and Edward B. Foley. 2009. Unsuccessful Provisional Voting in the 2008 General
Election. In Provisional Ballots: An Imperfect Solution. Pew Center on the States.
Kinder, Donald and Lynn Sanders. 1996. Divided By Color: Racial Politics and Democratic Ideals.
Chicago: University of Chicago Press.
Kousser, Morgan J. 1974. The Shaping of Southern Politics: Suffrage Restriction and the Establishment of the One-Party South, 1880-1910. New Haven, CT: Yale University Press.
Leighley, Jan E. and Jonathan Nagler. 2014. Who Votes Now? Demographics, Issues, Inequality,
and Turnout in the United States. Princeton, NJ: Princeton University Press.
Lublin, David M. 1997. The Paradox of Representation: Racial Gerrymandering and Minority
Interests in Congress. Princeton: Princeton University Press.
Manza, Jeff and Christopher Uggen. 2006. Locked Out: Felon Disenfranchisement and American
Democracy. New York: Oxford University Press.
May, Gary. 2013. Bending Toward Justice: The Voting Rights Act and the Transformation of
American Democracy. New York: Basic Books.
Mendelberg, Tali. 2008. “Racial Priming Revived.” Perspectives on Politics 6(1):109–123.
41
Meyer, Daniel R. and Geoffrey L. Wallace. 2009. Poverty Levels and Trends in Comparative
Perspective. In Changing Poverty and Changing Antipoverty Policies, ed. Maria Cancian and
Sheldon Danziger. New York: Russell Sage Foundation pp. 35–62.
Mukherjee, Elora. 2009.
“Abolishing the Time Tax on Voting.” Notre Dame Law Review
85(1):177–246.
Orsi, Jennifer M., Helen Margellos-Anast and Steven Whitman. 2010. “Black-White Health Disparities in the United States and Chicago: A 15-Year Progress Analysis.” American Journal of
Public Health 100(2):349–356.
Oster, Emily. 2005. “Hepatitis B and the Case of the Missing Women.” Journal of Political Economy 113(6):1163–1216.
Pettit, Becky. 2012. Invisible Men: Mass Incarceration and the Myth of Black Progress. New
York: Russell Sage Foundation.
Piven, Francis F. and Richard A. Cloward. 2000. Why Americans Still Don’t Vote and Why Politicians Want It That Way. Boston: Beacon Press.
Rocha, Rene R. and Tetsuya Matsubayashi. 2014. “The Politics of Race and Voter ID Laws in the
States: The Return of Jim Crow?” Political Research Quarterly 67(3):666–679.
Rodriguez, Javier M., Arline T. Geronimus, John Bound and Danny Dorling. 2015. “Black lives
matter: Differential mortality and the racial composition of the U.S. electorate.” Social Science
& Medicine 136-137:193–199.
Rogers, Richard G., Bethany G. Everett, Jarron M. Saint Onge and Patrick M. Krueger. 2010.
“Social, Behavioral, and Biological Factors, and Sex Differences in Mortality.” Demography
47(3):555–578.
42
Rusk, Jerrold G. and John J. Stucker. 1978. The effect of the southern system of election laws on
voting participation. In The history of American electoral behavior, ed. Joel H. Silbey, Allan G.
Bogue and William H. Flanigan. Princeton, N.J.: Princeton University Press pp. 198–250.
Schriner, Kay, Lisa Ochs and Todd Shields. 2000. “Democratic Dilemmas: Notes on the ADA
and Voting Rights of People with Cognitive and Emotional Impairments.” Berkeley Journal of
Employment & Labor Law 21(1):437–472.
Sen, Amartya. 1992. “Missing Women: Social Inequality Outweighs Womens Survival in Asia
and North Africa.” British Medical Journal 304(6827):587–588.
Squire, Peverill, Raymond E. Wolfinger and David P. Glass. 1987. “Residential mobility and voter
turnout.” American Political Science Review 81(1):4565.
Tate, Katherine. 1993. From Protest to Politics: The New Black Voters in American Elections.
Cambridge, MA: Harvard University Press.
Teixeira, Ruy A. 1987. Why Americans Don’t Vote: Turnout Decline in the United States, 19601984. Westport, CT: Greenwood Press.
Tesler, Michael and David O. Sears. 2010. Obamas Race: The 2008 Election and the Dream of a
Post-racial America. Chicago: University of Chicago Press.
Tomz, Michael and Robert P. van Houweling. 2003. “How Does Voting Equipment Affect the
Racial Gap in Voided Ballots?” American Journal of Political Science 47(1):46–60.
Uggen, Christopher, Sarah Shannon and Jeff Manza. 2012. “State-Level Estimates of Felon
Disenfranchisement in the United States, 2010.” Report published by The Sentencing Project.
URL: http://sentencingproject.org/doc/publications/fd State Level Estimates of Felon Disen 2010.pdf
Valelly, Richard M. 2004. The Two Reconstructions: The Struggle for Black Enfranchisement.
Chicago: University of Chicago Press.
43
Wang, Tova. 2012. The Politics of Voter Suppression: Defending and Expanding Americans’ Right
to Vote. Ithaca, NY: Cornell University Press.
Wolfinger, Raymond E. and Steven J. Rosenstone. 1980. Who Votes? New Haven: Yale University
Press.
Ziblatt, Daniel and Giovanni Capoccia. 2010. “The Historical Turn in Democratization Studies.”
Comparative Political Studies 43(8/9):931–968.
44