The Millennials are the most Democratic cohort in a generation. Some believe attribute their liberal inclinations to the racially diverse demographics within their generation, and while that may be partially true, I prescribe the Millennials' (born after 1980 through the mid-90s) Democratic leanings to the fact they grew up during a strong Clinton administration and a weak Bush one. The Millennials today have mostly replaced the very white, very Democratic voters of the Greatest Generation (who were born 1910-1927) who came of age during the weak years under Hoover and a Roosevelt administration so strong it won FDR a third term.
In both the 2004 and 2012 election, almost all age cohorts voted the same relative to other cohorts, and the Millennials were as Democratic relative to the nation in 2004 and 2012. The Greatest Generation was too small a percentage of the electorate in 2012 to collect poll data, but they voted as Democratic as the Millennnials did in 2004.
Proving the cohort point further, the then +60-year-old Greatest Generation has been the most Democratic cohort all the way back in 1988, when age cohorts didn't differ all that much in how they voted – despite even some members of the Lost Generation still voting. The Greatest Generation crowd was the most Democratic in the 1992 election, as well.
Even the most ardent critics of the cohort theory will admit that a person's views of the presidential administration he or she grew up with will shape their political views going forward. The real question is whether or not the Millennials have a a large enough portion of the electorate to "overwhelm" the rest: would Bush have still won in 2004 with 2012 demographics, and would Obama have still won with 2004 demographics? With the release of the Census Bureau's Current Population Survey (CPS) voter supplement, we can find out.
The CPS is a geographically defined sample that seeks to fully represents each type of voter. Respondents are asked simple questions such as race, registration, and whether or not they voted. It is generally seen as a more comprehensive survey than network exit polls to determine the exact composition of the electorate.
There are, however, reasons to be cautious about making too much of the CPS report. The CPS doesn't have a perfect response rate, and as my friend Sean Trende points out, there are more people who claimed to the CPS to have voted than actually did. The CPS tabulators assume that those who didn't respond didn't vote, which is likely not true, but there's no perfect way to account for the discrepancies. Preliminary examination of board of election data from different states shows that fewer African Americans, for instance, voted than the numbers that the CPS finds.
Either way, the CPS is a very solid starting point, and we can still take a preliminary step in answering whether or not Obama would have won without the Millennials. We can determine this by multiplying the percent of what Obama won among Millennials by the percentage they made up of the electorate. In other words, the percentage of vote the Millennials contributed to Obama's margin of victory. Then we compare this with the percentage of the vote the Greatest Generation contributed in prior years, and we'll have a solid answer.
The Millennials now make up 18% of the electorate, per the CPS. That's less than the exit poll data reports, but exits have been known to count too many young voters. Meanwhile, the CPS data is backed quite well by Pew Research, which most would agree is one of the finest pollsters out there, if not the best.
Those born between 1910 and 1927 were just 2% of the 2012 electorate. In 1996, the election before the Millennials began voting, the Greatest made up 15% of everyone who cast a ballot. By the time 2004 rolled around, Millennials were 8% of the electorate, while the Greatest was down to 7%. Thus, as a percentage of the total electorate, there was a 5pt gain in the Democratic coalition of the Greatest Generation and Millennials from 1996 to 2012, and from 2004 to 2012.
Given Obama's +20pt win among Millennials, what percentage of the vote is that 5pt difference worth? With the 2004 electorate, Obama would have won by 2.6pt instead of 3.85pt last year. Meanwhile, if the 2004 electorate had looked like the 2012 one, George W Bush would have won by 1.2pt instead of 2.45pt. So yes, the Democratic candidate would have done slightly better with the demographic boost, but neither election would have turned out any differently.
Further comparisons to 1996 and 2004 undersell the Greatest Generation's impact. The Greatest were 17% of the electorate in 1992, 21% in 1988, 24% in 1984, and 27% in 1980, but have dropped steeply as more of their members pass away. The Millennials, meanwhile, are still far away from 24%, let alone 27%. If other age cohorts had voted the same relative to the national vote, Obama might have actually won by more in prior years.
To me, the evidence does not suggest the Millennials have "overwhelmed" the Greatest. They are, if anything, a new "Greatest Generation" both in terms of voting patterns and, to a lesser extent, size – though we can still expect the Millennials to grow somewhat as a percentage of the electorate, since people are more likely to vote as they get older. The question going forward is whether Generation Z (born in the mid-90s and later) will follow the voting patterns of the Millennials.
His small re-election margin and his projected historic ratings mean that the Obama administration has largely been seen as mediocre, which in turn suggests that the next age cohort will walk the middle of the road. The polling data agrees that Generation Z will be less Democratic, than the Millenials, but if one believes the racial diversity theory, then the next generation should actually be more liberal, as it will have fewer white voters.
We don't know whether the age cohort or racial diversity theory will end up being more correct in the long run. If it's the diversity one, Republicans are in a lot of trouble. If it's the age cohort theory, then it will be politics as usual. I'd bet on age.