The Missing Voters, The Missing White Voters

In the 2008 Presidential election, the vote outcome was (ignoring minor parties):

Obama, 69,456,897
McCain, 59,934,814
Total, 129,391,711

In the 2012 Presidential election, the vote outcome was (ignoring minor parties):

Obama, 61,164,405
Romney, 58,159,408
Total, 119,323,813
Difference between the candidates, 3,004,997

There were 10,067,898 fewer votes cast for the two major party candidates in the 2012 election despite a growing population.  If the same proportion of the population had voted in 2012 as did in 2008, there should have been about 15.2 million more votes cast in 2012 than in 2008.

The difference in votes for the Democrat and Republican presidential candidates in 2012 compared to 2008 was:

Democrat, Obama, lost 8.29 million votes
Republican, lost 1.76 million votes
Total Lost Votes, 10.05 million votes

According to the CNN exit polls, the vote by race in 2012 was:

White, 72%, down from 74% in 2008
Black, 13%, same as in 2008
Latino, 10%, up from 9% in 2008
Asian, 3%, up from 2% in 2008

The number of white votes can be calculated for each year and we find that there were 9.84 million fewer white voters in 2012 than in 2008.

2008 White Vote:
Obama, 43%
McCain, 55%

2012 White Vote:
Obama, 39%
Romney, 59%

There was a 20% differential in this election of white voters for Romney.  If the 9.84 million white voters who stayed home had voted in the same percentages for Romney and Obama as those who voted did, then Romney would have had (0.20)(9.84 million) = 1.97 million more votes.  This would have only been two-thirds of the difference in the vote for Obama and Romney, but it would be very interesting to understand why nearly 10 million white voters stayed home.  In fact, while white voter numbers are not increasing as fast as are black and Latino numbers, they did increase somewhat, so this analysis understates the Republican problem with potential white voters.

Demographically, Republicans do have a serious need to win Latino votes.  But it is just as bad a problem that they could not motivate many white voters to come to the polls and vote and did even worse this time than they did with the very lackluster Senator John McCain.  Those who say that the Republican problem is simply that they only got 27% of the Latino vote are missing another critical part of the puzzle.  The 1% increase in the Latino voting percentage among about 10 million fewer voters, means there were only about 0.3 million more Latino votes.  Obama's 71% portion of that added vote gave him about 0.2 million more votes.  As a result, Obama's advantage due to more Latino voters was only about 0.1 times Romney's loss of votes due to not turning out white voters.

To have actually won the election, Romney would still have needed to get a significantly larger percentage of white and/or Latino votes.  The percentage of black voters stayed the same and given the lower number of total voters, the actual number of black voters decreased compared to their number in 2008.  The actual number was about 1.3 million fewer black voters.  Those black voters who voted, voted 91% for Obama.  Thus, a lower turnout of black voters may have cost him about 1.2 million votes.  So, Latino voters increased very slightly and white voters decreased hugely compared to 2008.

Summarizing the effects of voter turnout by race alone, and recognizing that using percentages only to the nearest 1% causes large errors, we have:

Obama's lost votes, 1.2 million black votes minus 0.2 million Latino votes = 1.0 million lost votes.

Romney's lost votes, 2.0 million white votes

So Romney nominally lost twice as many votes as Obama did due to the changes in numbers of voters choosing to vote, assuming each race voted in the same percentage as those of that race who turned up to vote.  Admittedly, that may very well not be a good assumption.  Nonetheless, it makes it clear that much more than just changing demographics is responsible for the sorry showing of Romney.