The 2014 Schatz/Hanabusa Senate Primary By Hawaii House District

On August 9th, Hawaii held the crucial Democratic primary for the U.S. Senate seat of the late Daniel Inouye. (A couple of precincts voted a week later, on August 16th, because of a hurricane.)

The Governor of Hawaii, Neil Abercrombie, had appointed his Lieutenant Governor, Brian Schatz, who had previously served as a state legislator and as state party chairman.

Both politicians (who incidentally had won separate primaries for Governor and Lieutenant Governor rather than running as a ticket) are white men, and so Abercrombie appointing Schatz to the seat of one of the most powerful and iconic Asian-American politicians in American history attracted some controversy--especially in a state that is only 23% non-Hispanic white.

In fact, it was Senator Inouye's own wish that Governor Abercrombie appoint Representative Colleen Hanabusa to replace him.

Of course, there was more to the primary, but when Hanabusa declared against Schatz, many presumed that the results would be racially polarized. Ultimately, Schatz defeated Hanabusa by a few percentage points. What did the results look like?

The graphic above compares the Schatz/Hanabusa primary with the estimated non-Hispanic white percentage for each of Hawaii's 51 State House districts. (Estimates are from the 2008-2012 American Community Survey.)

The map on the left shows the primary, with blue districts supporting Schatz and green districts supporting Hanabusa. Darker precincts were stronger for their candidate. The map on the right shows the estimated non-Hispanic white percentage, with darker meaning a larger percentage.

I also included a scatterplot of Hanabusa's percentage against Schatz vs. the estimated non-Hispanic white percentage. Hovering over the numbers will give you a label with the specific breakdown, and it will also highlight the district on the maps. Hovering over a district on the maps will also give the label. (Note: "Japanese" refers to the "one race-Asian-Japanese" estimate, same with "Filipino", while the other racial/ethnic numbers are limited to non-Hispanics, except for "Hispanic" of course.)

(Note: Throughout, the percentage is only Hanabusa vs. Schatz, not Hanabusa's percentage of the total vote--so she won all the districts above the 50% line and lost the rest.)

As you can see, there was a significant correlation between the non-Hispanic white percentage and Hanabusa's percentage. For example, Schatz won all but one House District that was over 28% white. The exception is House District 45, which--unusual in Hawaii--is represented by a Republican, Lauren Kealohilani Matsumoto. It's possible that the Democratic electorate there is significantly less white than the district as a whole, moreso than in other districts, but I'm not sure if that explains it. (Also, see below, re: its large population percentage in group quarters.)

Conversely, most of Hanabusa's districts were under 20% white. In fact, she won every such district except for Districts 20, 21, 23, 24, 25, 26, and 27, which seem to cover much of central Honolulu.

Schatz's single best district was the coastal District 22, the whitest district in downtown Honolulu, followed by the only two majority-white districts in the state: District 50 (Kailua, and also represented by a Republican) and District 11, on Maui.

I had wondered if District 22 was Schatz's home district, but he actually represented the old House District 25, which was a bit up from the coast. District 25 was then represented by Della Au Belatti, who now represents District 24. Schatz's old District 25 was also 19% non-Hispanic white, and 54.7% Asian, according to the 2010 ACS estimates.

Interestingly, while many assumed Hanabusa's best groups would be Japanese-Americans (and Japanese-American women in particular), her two best districts, District 43 and District 44, are not particularly Asian--rather, they have high percentages of "Native American or Pacific Islander" and "Two or More Races".

This might just be familiarity: Hanabusa represented Senate District 21, which covered much or all of the current Districts 43 and 44.

Incidentally: A Twitter conversation with Henry Fung (among others) reminded me I show factor in the population in "group quarters". These include military bases, prisons, and colleges--kind of a broad category, and for our purposes, it lumps together people that are listed in an area's population but who can't vote there (prisoners) with people who might vote there but might vote absentee (college students, the military).

At the moment, I'm not sure how to break the House District group quarters population estimates down into their subcategories, so for now, I'm just including the total percentage.

The total group quarters population percentage only tops 9% in four districts: 23, 30, 45, and 50. I believe District 23 has the University of Hawaii at Manoa, while Districts 30, 45, and 50 have military bases.

Hanabusa did well in Districts 30 and 45, especially relative to their non-Hispanic white population, while Schatz did well in District 23 (and in the surrounding districts). I am not sure if this is because Hanabusa did relatively well with military voters while Schatz did relatively well with students, or if it's a more complicated effect, if any.