Where are Milwaukee's "missing" black voters?

Are 60% of Milwaukee’s black voters from 2008 missing?  That’s what a study from the New Organizing Institute claims, which estimates that as many as 160,000 black voters in the city from the last presidential election can no longer be found.  Slate reports on this claim without any real skepticism or scrutiny, after noting NPR’s speculation that the housing crisis would cause Democrats more trouble in finding their voters:

This spring, the League of Young Voters, which was created to mobilize young minority communities, collaborated with the liberal Wisconsin Voices coalition to dispatch teams of young canvassers. Starting in April, they spent eight weeks knocking on 120,882 doors across 208 of Milwaukee’s 317 wards to raise awareness of the gubernatorial recall election scheduled for June.  The doors had one thing in common: the voter file said they were all home to a registered voter whom a commercial data vendor had flagged as likely to be African-American.

But the voter file represented a fiction, or at least a reality that had rapidly become out of date.  During those eight weeks, canvassers were able to successfully find and interact with only 31 percent of their targets. Twice that number were confirmed to no longer live at the address on file  — either because a structure was abandoned or condemned, or if a current resident reported that the targeted voter no longer lived there.

Based on those results, the New Organizing Institute, a Washington-based best-practices lab for lefty field operations, extrapolated that nearly 160,000 African-American voters in Milwaukee were no longer reachable at their last documented address — representing 41 percent of the city’s 2008 electorate.  It is a staggering figure in a battleground state where Democratic prospects rely on turning out Milwaukee’s urban population, an ever more urgent cause since Paul Ryan’s presence on the ticket could help mobilize core Republican constituencies in the city’s suburbs.  Over half of those identified as displaced were under the age of 35, and thus also less likely to be reachable through traditional landline phones.

Simply put, this is an absurd conclusion to reach.  It starts with an assumption that the “voter file” used in this search was accurate, both in identifying African-American voters and in identifying legitimate registrations.  It ends with the assumption that a voter who no longer lives in the same place in 2012 that he or she did in 2008 is somehow “missing.”  And the result of this absurdity is the claim that 160,000 voters in a city population of only 594,833 — including non-voters — went “missing,” and no one noticed.

If this is actually true, then we have a perfect way to corroborate this claim.  The 2008 election, where Barack Obama became the first African-American President, was generally agreed to boost participation among black voters around the nation.  In Wisconsin, that demo accounted for 5% of the overall vote, according to exit polling. In the 2010 gubernatorial election, it was 4%.  Unlike most other states, Wisconsin held a statewide election as well as a primary — the recall election of Scott Walker, Rebecca Kleefisch, and several state senators.  In exit polling from the June election, black voters once again comprised 5% of the vote, up from 2010’s gubernatorial election.

Obama won Wisconsin handily in 2008 by 13 points in an election where 2,939,604 votes were cast; five percent of that would be 146,980 votes cast by African-American voters in the entire state.  In the 2010 election, 2,133,244 votes were cast, 4% of which would have been 85,330.  The recall election won by Walker had 2,516,065 votes cast — almost perfectly between the two — and five percent of that comes to 125,803.    None of these remotely indicate that Milwaukee had 160,000 extra black voters to go missing in the first place.  That’s more than the number of black voters who turned out statewide in 2008 in Obama’s first presidential election.

By the way, the number of people who voted in Milwaukee County — not the city — in the 2008 election was 475,192.  In 2012, it was 396,183, a difference of only 79,000 in a special election effort.  It also produced almost exactly the same result as in the 2010 gubernatorial election, too.

Finally, let’s go to the best data of all — the Census Bureau.  The state of Wisconsin’s population in the 2011 estimate was 5,711,767 people, of whom 6.5% identify as black.  That’s a statewide total of 371,265 black adults and children, up from 2000’s 304,460 and 5.7%, and slightly higher than 2010’s 359,148 and 6.3% rather than declining.  The city of Milwaukee’s population of 597,867 is 40% African-American, which comes to 239,147 adults and children.  Assuming one child for every two adults, there would only be 157,837 African-American adults of voting age in the city altogether.  In Milwaukee County, the population is 952,532, with 27% being African-American.  That comes to 257,183 African-American adults and children (most of them live in the city itself, obviously), and using the same 2:1 ratio for adults to children, we get 169,741 voting-age-eligible adults.

In order to believe this survey, one would have to believe that every African-American adult had abandoned the city of Milwaukee, and nearly all from the county of Milwaukee, too, which demonstrates the absurdity of its conclusion.   Furthermore, even if the claim were true, those voters would show up elsewhere — either elsewhere in Wisconsin, or elsewhere in the US.  The percentage of black voters in Wisconsin exit polling from 2008 to 2012 should have cratered with that kind of exodus had it occurred, but instead it remained remarkably stable all the way through June of this year.  The only conclusion is that the claim is absolutely and transparently nutty.

There are a few possible reasons for the New Organizing Institute to make this claim, and they’re not mutually exclusive.  One, they may be just that bad at math.  Two, they may just be that bad at door knocking. Three, they may just be that bad at research.  Four, the voter file from 2008 contained massive amounts of fraudulent registrations, and the state’s new voter-ID law has respondents thinking twice about trying it again in 2012 — although 160,000 is as absurd a figure for that as it is for claiming people missing.  Five, the New Organizing Institute wanted to stake a claim that a loss in Wisconsin wasn’t due to their bad GOTV efforts but a result of both the housing crisis and voter suppression, and concocted the data to support the conclusion.

I suspect a bit of all these are at work in this claim, and arguably in the reporting of it too.  It’s very curious that Slate’s Sasha Issenberg didn’t bother to do the math herself to test NOI’s ridiculous figures before writing the lead, “Sixty percent of Milwaukee’s black voters have disappeared.”  This looks like agenda flogging rather than journalism.