MacGregor Campbell, New Scientist, May 1, 2015
Dead men cast no votes. A new study has found that the premature death of millions of black voters in the US has affected the outcome of several elections.
Overall, in the US, the mortality rate for blacks, across age and gender, is almost 18 per cent higher than the rate for whites.
When one demographic group dies at a faster rate than another, it skews the makeup of the electorate in favour of the group with the better survival rates. Geronimus and her colleagues wondered what effect this difference might have on US politics.
To find out, the team looked at the 35-year period between 1970 and 2004, and asked how the outcomes of elections in this period, including the 2004 presidential contest between John Kerry and George W. Bush, might have been different if the mortality rate of black and white people had been equal.
Using cause of death data from the US Centers for Disease Control, Geronimus and colleagues calculated that if blacks died at the same rate as whites, 5.8 million African Americans would have died between 1970 and 2004. The actual number of black deaths over that time span was 8.5 million, meaning that African Americans had 2.7 million “excess deaths”, compared with whites.
A million voters
Of those 2.7 million, Geronimus and colleagues calculated that 1.74 million would have been able to vote in the 2004 elections, of whom 1 million would have actually voted.
The researchers then looked at how this extra million might have influenced elections if they had voted in line with actual black voters. African Americans tend to vote overwhelmingly for the Democratic party, so that party’s presidential candidate, John Kerry, missed out on around 900,000 votes. Kerry’s Republican opponent, George W. Bush, lost around 140,000.
The missing black voters alone would not have been enough to change the result–Bush was elected with a majority of more than 3 million votes. But the story is different at state level, especially if another cause of lost black votes is taken into account.
In 2006, Joe Manza of New York University and Christopher Uggen of the University of Minnesota estimated that 1.95 million voting-age African Americans were unable to vote in 2004 due to the fact that they had been convicted of a felony. A 2013 Pew Research Center study found that black men were six times as likely as white men to be incarcerated in 2010–an increase from 1960, when black men were incarcerated five times as often as white men. This disenfranchisement meant that around 15 per cent of voting-age African Americans were excluded from the 2004 national election.
Geronimus and her colleagues estimated that seven senate and 11 gubernatorial election results between 1970 and 2004 would have been reversed had their hypothetical survivors been able to vote. These were close elections in which the margin of victory was less than a third of the number of estimated hypothetical survivors in the state. Accounting for people disenfranchised by felony convictions would have likely reversed three other senate seats. In at least one state, Missouri, accounting for just excess deaths or felony disenfranchisement would not have been sufficient to reverse the senate election–but both sources of lost votes taken together would have.