White NYC Vote by Burough (Estimated)

[mencken]

I was thinking of calculating the white vote in NYC by borough. I figured 2008 would be better for this purpose than 2012, since Hurricane Sandy arguably had an effect in that year. First, I compared the 2010 census demographics of the state to the demographics of the 2008 exit poll. The breakdown was 58% Non-Hispanic White, 18% Hispanic, 16% Black, 7% Asian in the former, whereas it was 71%, 6%, 17%, 3% in the exit poll. I figured this would be a good estimate for how much each demographic turned out in the NYC buroughs. I gave Obama every black vote (as the exit poll indicated) and roughly 2/3 of Hispanic/Asian/other votes (per the national exit poll).

Staten Island: 64% NHW, 11% B, 17% H, 8% A per 2010 Census data, assuming that translates to turnout of 79% NHW, 12% B, 6% H, 3% A, and assuming the above demographic voting for nonwhites means NHWs here voted 62-38 McCain

Manhattan: 48% NHW, 16% B, 25% H, 11% A per 2010 Census data, assuming that translates to turnout of 66% NHW, 19% B, 9% H, 5% A, and assuming the above demographic voting for nonwhites means NHWs here voted 86-14 Obama*

*The fact that whites voted so overwhelmingly for Obama suggests that the Asian and Hispanic voting assumptions were likely too low for Obama. If we assuming NHW/A/H all voted similarly in Manhattan, this changes to 83-17 Obama

Brooklyn: 36% NHW, 34% B, 20% H, 11% A yields 48% NHW, 39% B, 7% H, 5% A turnout. With original assumptions for nonwhites this gives 65-33 Obama for NHWs here.

Queens: 28% NHW, 19% B, 28% H, 23% A yields 47% NHW, 27% B, 13% H, 13% A turnout. With original assumptions for nonwhites this gives 65-33 Obama for NHWs here.

Bronx: 11% NHW, 36% B, 53% H, 4% A yields 19% NHW, 54% B, 25% H, 2% A turnout. Under original assumptions this would be 90-10 Obama, but the Bronx obviously requires the Manhattan asterisk assumption (if not more lopsided Democratic voting for Hispanics), which yields 70-30 Obama


Do the results of these calculations make intuitive sense, or am I considerably underestimating Asian/Hispanic affinity for Democrats in NYC?

[mencken]

Approaching this problem from another angle...

Excluding the five boroughs of NYC, the rest of New York state voted for Obama 54-44. The demographic model predicts that turnout was 87% NHW, 9% B, 3% H, 1% A*. Under this assumption, the NH-White vote for the rest of New York state was 49-49 between the two candidates. Since the overall White vote in New York state was 52-46 Obama, the overall White vote in NYC would have to be 61-37 Obama. Clearly my suspicions about underestimating Hispanic/Asian Democratic votes were valid.


Now to rectify this problem. Let's look at Staten Island first. With 9% Hispanics and Asians and 79% Non-Hispanic Whites, changing the Hispanic/Asian vote from 67-33 Obama to monolithically Democrat would only change McCain's NHW vote share by ~4%. Given how tiny Staten Island is, this seems to miniscule to have any effect on the overall calculation. Staten Island numbers remain as is.

Manhattan has few non-Black minorities, but with such a large portion of the White vote in NYC may have an effect regardless. Based on the NY/national exit polls, assume Asians gave Obama 10% more and Hispanics 15% more of the vote than Non-Hispanic Whites. This changes Manhattan NHWs to 80-20 Obama.

Knowing the NHW vote in both NYC and Staten Island and Manhattan means that the NHW vote in the three remaining buroughs has to be 55-43 Obama.

Going by CTRattlesnake's assumption that Brooklyn/Queens/Bronx NHWs should vote similarly, this forces Brooklyn H/As and Queens Hs to vote monolithically Democrat and Bronx Hs to vote 87-12 Obama.