Talk Elections

Not sure if there's enough data available to analyze this, but has anyone looked into the question of how much the gap between white, Hispanic, and Asian votes in this election (or any recent election) is a consequence of differences between those groups on other demographic parameters, like age and income?

For example, Hispanics are disproportionately younger and poorer, relative to whites.  If you reweighted the Hispanic voting results so that they matched whites in terms of age and income, what would the new Obama/Romney split be?

If California exit polls were right, and 28% of the electorate was 18-29 years old, then at least in California almost all of the trend to Obama could be explained by that. At first I thought that number was completely wrong, but now I am thinking there may have been a huge youth turnout to support Prop 30 and they voted for Obama as well, and the youth in California is much more Asian and Hispanic than the electorate at large.

If you round California right now, it's 60-38 (barely down from 61-37 in 2008). Obama may not cross 60%, but he's very close and I don't know what's still outstanding. The polls definitely understated Obama this time, unlike four years ago. To be honest, the Field Poll is the main reason I predicted Obama under 60% in California, although I was accurate in predicting Romney under 40%. I think California is one major state where age and demographics are seriously conspiring against the Republican Party. Part of me is almost amazed at how the Republican Party in California has so totally collapsed in the state, to the point of almost being practically dead at the moment.

(The CA SoS has the current margin at 59.7-37.9.)

Age and income are part of it, but even excluding those, I think there is still a racial gap.  Lets remember amongst whites 18-29 Romney won this 54-45% while overall it was 60-38%.  Now I wouldn't be surprised if Obama would have won the electoral vote for whites 18-29 as I suspect he would have narrowly carried that group in most Northern states and more solidy in the Pacific Northwest and New England states, while got clobbered amongst whites 18-29 in the South.  I think if you take the South out of the picture where racial divisions are still quite strong, some of the difference could be simply where minorities live.  In the large urban centres, I suspect Obama won the white vote, whereas in rural areas that went GOP the minority population is quite small and generally people who live in smaller communities tend to be more conservative than those living in large urban centres.  Still the gap does exists even when you include other factors.

Yes, it is likely Obama won the white 18-29 vote outside the south. Also Romney won that group 51-44, not 54-45.

You have it backwards. The age gap is because of race. Not vice versa.

Partially true although amongst whites 18-29 it was a lot closer than amongst older whites never mind I suspect Romney's huge majority amongst whites 18-29 in the South probably skewed the results somewhat too.  Lets remember in the New England states and Pacific Northwest or even Upper Midwest which are fairly white, Obama did much better amongst younger voters than older ones.  Off course this is not unique to the US.  In Europe, Canada, Australia, and New Zealand, parties on the left tend to do better amongst younger voters than older voters.  In the 2005 British election, the Conservatives came in third at under 20% amongst the under 30 crowd while won the over 50 crowd.  Likewise here in Canada, the NDP (left wing party) won the under 30 vote while the Conservatives got close to 50% amongst the over 65 crowd.  People tend to be more idealistic when young while get pragmatic and also more resistent to change as they get older.  Also the marriage gap may be a reason too since amongst married men vs. married women, there wasn't much difference, whereas amongst unmarried women; Obama did significantly better than unmarried men although he won both and those groups would be highest amongst the young and the old.

But if you flip the question around, black voters over 65 were actually slightly MORE likely to vote for Obama (91-6) than black voters under 30 (93-8). So being old per se doesn't make you more likely to vote Republican. Being old and white does. To some extent, being old and Hispanic does, in that Romney "only" lost Hispanics over 65 by 30 points, where he lost Hispanics under 30 by 50 points, but clearly even if the average Hispanic was as old as the average white person Obama would still have won Hispanics by a huge margin.

I think the question should be asked the other way around. How much of the age and income voting differences are due to race? After all, poor whites vote not that differently from rich whites, and the same for poor blacks and rich blacks, and for poor Asians and rich Asians. Only Hispanics seem to have much polarization in voting patterns by income. And minorities are much more heavily represented among the youth than the elderly. The 65+ population is something like 85% white. The 18-25 population is maybe 60% white at most (and the infant population, obviously not voting yet, but just wait a couple decades, is minority-majority!).

I would say race is probably the largest factor, although where one lives and age are also factors, not as much as race but more so than income.  After all those 18-29 including those who are making over $100,000 vote more heavily Democrat than the 65+ group making under $30,000.  There is also the location factor as many of the most solid blue states happen to be amongst the wealthier states whereas the poorest states are almost all red states.  There is also the education factor as people with higher incomes are more likely to favour the GOP, but people with higher levels of education are more likely to favour the Democrats and as a general rule people with higher levels of education tend to make more money. 

This isn't entirely true.