In a new paper, four researchers look at a dramatic way US elections could be skewed — black Americans tend to die earlier than whites, so a disproportionate amount of them aren't around to vote. If they were around, the researchers argue, several close statewide elections over the past few decades likely would have tipped to the Democrats.
The researchers — Javier Rodriguez, Arline Geronimus, John Bound, and Danny Dorling — came to this conclusion by estimating how many of these missing black voters there are, and how many close elections they would have swung had they been alive. They estimate that if black Americans died at the same time in their lives as whites do, about 1 million more of them would have voted in 2004. And they find that since the 1970s, seven more Senate races and 11 more governor's races would likely have been won by Democrats.
These missing voters alone wouldn't transform US politics, and they wouldn't have been enough to tip the 2004 presidential election to John Kerry. But overall, the researchers' finding fits alongside disenfranchisement of disproportionately black felons as one more way the voting power of the black American population ends up, proportionately, less than that held by whites.
The data: black Americans are much more likely to die earlier
The causes of the different death rates of black and white Americans are complex. But, importantly, the researchers point out that they aren't primarily about young men dying violent deaths. Instead, they write, "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." You can see this in the chart below:
"The mortality gap between blacks and whites is greatest between the ages of about 40 and 65," they write — and that's also when "the probability of turning out to vote is highest." The upshot, they write, is that "blacks are dying off from the electorate at higher rates than whites during the ages of highest voter turnout."
So they pulled state-level data on deaths between 1970 and 2004, in the 32 states with significant black populations plus Washington, DC. Due to data availability, they stopped there, and focused primarily on the 2004 presidential election, which was quite close (George W. Bush beat John Kerry by 2.4 percentage points, and 17 electoral votes could have flipped the outcome). They examined various close Senate and gubernatorial races from that timespan, too.
The findings: 1 million missing black voters in 2004
Overall, they found:
- 2.7 million excess black deaths between 1974 and 2004
- 1.87 million of those would have been alive in 2004
- 1.74 million of those would have been of voting age that year
- About 1 million of those would have voted (assuming similar turnout rates to actual black voters)
- And 86.6 percent of those who voted would have voted for Kerry (assuming a similar partisan breakdown to actual black voters that year)
Those numbers aren't enough to swing most elections
But though those numbers seem impressive, they wouldn't have been enough to swing even the rather close presidential election of 2004.
For instance, in the crucial state of Ohio that would have given Kerry the victory, the researchers project the Democrat would have netted about 24,000 more black votes. This would have made the race closer, but still been quite far from surmounting Bush's actual winning margin of 118,000 votes. Indeed, the researchers don't seem to find any example of a close Electoral College state that would have flipped. Overall, Bush's popular vote margin — 2.4 percentage points — would have shrunk by less than 1 percentage point.
For the missing voters to affect results, the election outcomes would have had to have been both quite close, and in states with rather large black populations. The researchers also found that out of the hundreds of Senate and gubernatorial elections taking place in this 25-year timespan, the missing voters would likely have flipped seven Senate races and 11 gubernatorial races to the Democrats. The authors don't directly discuss the 2000 presidential election, but it would have flipped as well — the results in Florida were close enough that many, many things could have flipped them.
The implications: part of a broader pattern
The researchers do point out that due to limitations in the state-level mortality data, they only looked at 35 years of differential mortality exposure, rather than a full lifetime's worth. So they write that "on balance, our results underestimate the effect that black excess mortality has on the size of the black population and electorate."
Overall, though, it's best to understand mortality differences as one of many contributing factors to why black Americans have less voting power, proportionately, than whites. And of course, as they conclude, "the meaning of lives lost too soon cannot be reduced to aggregate numbers."
(Hat tip to Mother Jones and New Scientist for flagging the study.)