Sunday, August 10, 2014

Two Missing Hawaii Precincts Probably Can't Save Hanabusa

On Saturday night (or Sunday morning for those of us on the East Coast), Senator Brian Schatz and Congresswoman Colleen Hanabusa fought to a draw in the Democratic primary for Hawaii senator. As of 4pm ET on Sunday, the Associated Press had the results at 113,789 for Schatz and 112,154 for Hanabusa, with 245 of 247 precincts reporting.

So what's the matter with those last two precincts? Well, they're not going to report for a very long time, since they didn't actually vote on Saturday. Hawaii Paradise Community Center and Keoneopoko Elementary School polling places had to be closed after the roads to them were blocked, rendering them inaccessible on Saturday. (The two precincts are in the Puna district of the Big Island, which sustained heavy damage from Tropical Storm Iselle on Friday.) Their voters who didn't get the chance to vote early will now vote by mail over the next couple weeks instead of going to the polling places—setting up a Swing Vote–type scenario where this small group of voters could determine who goes to the US Senate.

The question everyone is asking today is simple: who has the advantage with this last group of voters? If you go by demographics, it's Hanabusa. Hawaii politics has long been defined by its racial tensions; in a state where Republicans are a mere nuisance to be brushed aside, the real dividing line is between white politicians and Asian and Native Hawaiian politicians. If you put stock in the theory of identity politics, that means we should look at the racial demographics of the outstanding areas. According to the US Census, the communities in the two precincts are almost perfectly evenly divided:

Census-Designated Place White Asian Native Hawaiian/Pacific Islander Two or More Races Other
Hawaiian Paradise Park 3,958 2,148 1,341 3,679 278
Hawaiian Beaches 1,360 637 634 1,534 115
Total 5,318 (33.9%) 2,785 (17.8%) 1,975 (12.6%) 5,213 (33.2%) 393 (2.5%)

These data are inconclusive, since there are a lot of multiracial residents. (Portions of the precincts also lie in unincorporated areas, so these population statistics are not totally complete.) However, election data might break the tie by revealing which race of candidate they are inclined to vote for. In 2012, Hawaii saw a situation similar to the current election: a close Democratic primary for a US Senate seat that featured one white, male candidate and one Asian, female candidate. Here's how the voting broke down in the two precincts:

Precinct Ed Case Mazie Hirono Other
04-01 Hawaii Paradise Community Center 696 1,203 24
04-02 Keonepoko Elementary School 554 1,145 47
Total 1,250 (34.1%) 2,348 (64.0%) 71 (1.9%)

Hirono, the Asian candidate, crushed Case 64.0% to 34.1%—indicating an electorate 6–7 points more "pro-Asian" (to put it bluntly) than the state as a whole, where Hirono won the primary 56.8% to 40.3%. If we assume these precincts will behave the same way in 2014, we can estimate that Hanabusa will win them approximately 56% to 43%.

However, that is not a big enough margin for Hanabusa to overcome the deficit she is already running: 1,635 votes, remember, statewide. If the 56%–43% projection is accurate, Hanabusa would need 12,577 voters to turn out in these two precincts in order to gain a net total of 1,635 and pull her into a tie statewide. However, the precincts are only home to about 8,255 registered voters. In the 2012 primary, only 4,429 turned out to vote—and just 3,741 of those participated in the Democratic primary. Even if you assume that the Swing Vote circumstances, the ease of voting by mail, and the campaigns' intense focus on the precincts in the next couple weeks will all drive turnout through the roof, it seems unlikely to exceed turnout in the 2012 general election (a presidential high-water mark), which was 6,556. (Hawaii has a notorious turnout problem.)

Then you have to take into account that early voting means a lot of the precincts' votes have already been banked. The universe of voters who haven't yet voted and thus can still be persuaded could be as small as 3,880 (those who waited to vote in person on Election Day in the 2012 general), 2,531 (those who waited to vote in person on Election Day in the 2012 primary), or 2,149 (those who waited to vote in person on Election Day in the 2012 Democratic primary).

All in all, our calculations look grim for Hanabusa. If we go by 2012 Democratic primary turnout (the most likely model in my opinion), Hanabusa would need to win the two precincts 72% to 28%. Despite the precincts' favorable demographics, that's a daunting task.

UPDATE: A local viewpoint takes issue with my conclusions:

Lots of localities exhibit such idiosyncracies, and if Puna's progressivism overcomes the racial demographics my projections could easily be wrong. We'll find out the final totals in a couple weeks, but regardless the situation doesn't look any less dire for Hanabusa.

No comments:

Post a Comment