The Dog That Didn't Bark: A 2016 Presidential Election Mystery Unfolds in Wisconsin’s Juneau County

March 8, 2017

I wrote this geeky analysis of Wisconsin’s traumatic 2016 election landfall in the spring of 2017. Wisconsin is much on the mind of political junkies these days, who rightly view it as the bellwether for the national election in 2020 (listen to the latter 3rd of this recent Pod Save America podcast, in which Dan Pfeiffer interviews Ben Wikler, the chair of the Wisconsin Democratic party). Since Juneau County is a bellwether for Wisconsin, we might view the data explored in this essay as predictive, or at least meaningful, to some degree. I would recommend some special focus on the demographic dynamics in the state since 2016, with special reference to the old-white-guy percentages and to the migration/population growth trends. In the Pod Save America interview, Ben Wikler also provides fascinating (and harrowing) updates on the devastating Scott Walker impacts of union-busting, voter registration rules, campaign finance finagling, and gerrymandering).

In the 2012 presidential election, Barack Obama soundly trounced Mitt Romney in the great state of Wisconsin, harvesting 52.8 percent of ballots cast in the state (compared to Romney’s 45.9 percent), a margin of more than 213,000 votes. While Romney won in 37 of the state’s 72 counties, these were mostly smaller and more rural enclaves, representing only 46 percent of the total votes cast, and since Obama won his counties by an average of 15.6 percent and Romney carried his counties by an average of only 10.3 percent, well, it was game-set-match early in the evening for Barry in the Cheesehead state.


Obama’s solid triumph in Wisconsin in 2012 was nonetheless not a foregone conclusion heading into the election, given the cross-cutting political currents in Wisconsin, a northern and largely white rust belt state that usually tilted Democrat in national elections but which remained pockmarked by provincial (near-antediluvian) resentments that swept Tea Party Republicans Scott Walker, Paul Ryan, and Reince Priebus into positions of power and influence within the state and national party establishments, men of “ideas” who, by virtue of having ideas at all (policies tied to goals, supported by principles and evidence, that one could imagine at least implementing, if perhaps not actually working) within a movement defined by its anti-intellectualism, could present themselves as one-eyed kings in the land of the blind.

Wisconsin delivers only 10 of the 270 electoral votes required to secure the presidency, but remains politically significant because of its reputation for enlightened and progressive governance unbesmirched by either the brass knuckles political mayhem of North Atlantic states such as Rhode Island or New Jersey, or of the primal ooze enveloping politics in Southern States such as Louisiana or Mississippi. Wisconsin, in other words, is a bit of a bellwether for national politics.

Obama’s 2012 victory represented the seventh straight presidential election in which the state had delivered its electoral votes to the Democratic candidate (dating back to 1988), and with Hillary Clinton’s political experience, financial war chest, electoral “ground game”, and semi-progressive, semi-feminist bona fides, and with Obama’s historically high presidential approval ratings, pollsters and pundits alike typically viewed the state as safe for the Democrats heading into the election, far moreso than they had in 2012, a comfortable (if not smug) assumption preserving the general consensus that Clinton could not possibly lose the national election (or, perhaps more accurately, a consensus that Trump could not possibly win the national election).

A Cheesy Election

Clinton’s weakness in Wisconsin on election night as returns began to tally provided one of the clearest and most decisive indications that she was in trouble. Indeed, while Pennsylvania and Michigan were simultaneously tilting toward Trump, it’s probably fair to assume that returns from Wisconsin were what knocked most people sideways In the end, of course, Clinton lost to Trump by only the most slender of margins in Wisconsin, fewer than 23,000 votes out of the nearly 3 million votes cast.

Globalization’s implosion, Russia’s campaign skullduggery, James Comey’s ill-timed Anthony Weiner / laptop / Hillary Clinton / email-investigation / announcement (let’s once again blame our national plunge into the abyss on Anthony Weiner’s ubiquitous weiner because … why not), American white male sexual insecurity and fear of strong women, or Hillary Clinton’s own historically deficient campaign skills – each of these may have tipped the balance toward Trump in Wisconsin (and elsewhere). But focusing on these external forces (viewed as things that happened to us, that we were helpless to defend ourselves against) not only mind-torques the Democratic Party heading into the next round of elections, they distract from things happening to and experienced by real people in Wisconsin that may more accurately explain the election outcome, and surface reproducible lessons for building a sustainable, native political movement in Wisconsin, and elsewhere, that can once again ground our politics in issues of substance.

To be clear – Donald Trump did not light up the state of Wisconsin. He won with a plurality of only 47.2 percent of the vote, while receiving nearly 3,000 fewer votes than Mitt Romney in 2012, who harvested 45.9 percent of the vote. How could this be? First, despite the fairly obvious significance of this particular election, nearly 100,000 fewer people in the state voted in 2016 than in 2012 – a turnout drop of three percent. Second, candidates not named Trump or Clinton received nearly 200,000 votes in 2016, while third-party candidates in 2012 tallied only about 40,000 votes. Finally, and decisively, Wisconsinites voted for Barack Obama in 2012 as if he were a tangy, ripe Wisconsin cheddar, while in 2016 Hillary Clinton perhaps more closely resembled a somewhat moldy and odoriferous Stilton blue. Clinton received nearly 250,000 fewer votes in 2016 than Obama in 2012, a stunning collapse on the order of 15 percent. Clinton defeated Trump in only 12 of the state’s 72 counties. She outperformed Obama in only two of these 72 counties (on a vote percentage basis). If she had swung about 12,000 of the votes she lost from Obama that went to Trump and the third-party candidates (less than five percent of the vote gap with Obama!), Clinton would have defeated Trump in Wisconsin.

Wisconsin Wobegon – Juneau County

Let’s look at Juneau County, where among all counties in Wisconsin Democratic Party fortunes most steeply plummeted in 2016. Juneau County is a postcard-perfect rural enclave nestled in south-central Wisconsin, about 75 miles northwest of Madison and 140 miles northwest of Milwaukee. In 2012, Barack Obama claimed 52.8 percent of the vote in Juneau County while Mitt Romney received 45.8 percent of the vote, numbers almost identical to the statewide distribution of votes. This “normalcy” matters. As we’ll see, in many other respects, Juneau County truly is “normal” – not quite Lake Wobegon, but not too far off, either. Which makes it all the more surprising that in 2016 Hillary Clinton received only 34.7 percent of the Juneau County vote, while Donald Trump claimed 60.8 percent of the vote.

Much has been made of stark demographic breaches of the national fabric exposed in the course of the 2016 election – rural-urban / white-minority / male-female / old-young / less education-more education. Post-election, Wisconsin journalists explored the Juneau County phenomenon largely along these lines, with an emphasis on support for Donald Trump in rural counties with dispersed, shrinking (and aging) populations, stagnant economies and wages, and encroachment of formerly “urban” problems such as homelessness. However, when one looks closely at demographic, economic, health, and general welfare trends in Juneau County between 2012 and 2016 (looking for, and assuming one will find, causative or correlative movement of some data dial that will explain the electoral hydraulics of the 2016 election), what is striking is the extent to which there is no barking dog. The mystery becomes … not what specific trend caused this tectonic partisan and electoral shift, but how this pronounced shift could have occurred in the absence of any obvious environmental cause.

Demographic Distinctions

Juneau County is an averaged-sized Wisconsin county (about 765 square miles) with a population (26,664) and a density (34.8 people/sq. mi.) almost precisely comparable to averages for Wisconsin’s other 43 counties with populations under 50,000. In 2016, Donald Trump received more than 60 percent of the votes cast in 18 of these counties, compared to only three of the 28 counties with populations exceeding 50,000. Compared to the entirety of the state, the population of Juneau County is substantially more rural, older, and less racially and ethnically diverse, a profile that almost perfectly matches the demographics of the counties that broke hardest for Donald Trump in 2016 (Florence, Green Lake, Oconto, Taylor, and Washington).

The one vector of divergence from these Trump Counties is gender – only 46.9 percent of Juneau County is female, tied with Adams County for the lowest percentage of females among Wisconsin’s 72 counties. Notably, while all five Breaking-Bad-for-Trump counties also broke sharply for Mitt Romney in 2012 (with an average victory margin over Obama of 23.2 percent), Adams County, like Juneau County, broke hard for Obama in 2012, then swung radically against Clinton and for Trump in 2016. A very different dynamic. Is it somehow related to the gender skew? Unclear, but it surely does matter that Adams County and Juneau County are peas in a pod – immediately adjacent to each other and carved in 1858 from the same 60 km-square (approximately 40 mile-square) plot of land (see this interesting history of Adams County for more information).

Social Welfare Indicators – Positive Trend Lines

Social indicators include measures of well-being determined by health, education, and child welfare factors. Given Juneau County’s radical political tilt, between 2012 and 2016, away from Obama and the Democratic Party and toward Trump and the Republican Party, one might except some disjunctive negative movements in some (or many) of these indicators. In fact, what is striking is how little these measures changed in four years, and to the degree change did occur, how much of it was in a positive direction. Here are some examples.

Between 2012 and 2016, in Juneau County,

  • Median household Income rose (from $40,228 to $45,158).

  • Unemployment fell (from 9.9 percent to 6.7 percent).

  • The teen birth rate fell (from 39.2 per 1,000 births to 35.64 per 1,000 births).

  • The percentage of low birth weight babies fell (from 6.0 percent to 5.7 percent).

  • The percentage of children living in poverty fell (from 25.7 percent to 20.8 percent).

  • The violent crime rate remained stable.

  • The percentage of the population receiving at least some post-secondary increased slightly (from 49.4 percent to 50.7 percent).

  • Per capita health care costs declined slightly, even with an aging population (from $7,644 to $7,541).

  • The number of uninsured adults fell (from 2,455 to 2,351).

  • The percentage of people in “poor” or “fair” health dropped (from 19.5 percent to 14.5 percent – even while it rose across the state, generally, from 12 percent to 14.9 percent).

  • The percentage of smoking adults decreased sharply (from 28.4 percent to 19.2 percent).

  • Reported cases of Chlamydia fell (from 199 per 100,000 people to 184 per 100,000).

  • The percentage of adults who drink excessively fell (from 26 percent to 22.7 percent).

  • Diabetic screenings increased.Mammography screenings increased.

On the downside,

  • Between 2012 and 2016, in Juneau County, the percentage of children qualifying for free lunch programs increased from 31.2 percent to 46.6 percent, a figure far higher than the state average of 35.3 percent.

  • The percentage of the population with some form of diabetes rose (from 9.2 percent to 10.7 percent).

  • The percentage of obese adults increased (from 29.9 percent to 33.9 percent).

  • Preventable hospital stays for Medicare enrollees increased (from 3,350 to 3,849).

  • The percentage of county residents with limited access to healthy food rose (from 4.9 percent to 5.4 percent).

Migration Patterns – Stagnation and Turbulence

Population dynamics and migration patterns matter enormously for any effort to assess underlying political stress points. Geographic pockets characterized by declining, aging populations are typically an important source of support for reactionary political movements. Locations defined by measurable population flows and dynamic migration patterns are typically more turbulent, socially and politically, but also more forward-looking and less trapped by nostalgia or by fever dreams.

Between 2012 and 2016, in Wisconsin, the state’s population remained steady, while international migration into the state averaged 0.32 percent of the population across counties. By contrast, in Juneau County, the population fell by 1.65 percent and international migration increased by 0.18 percent. These population and migration numbers are consistent with population trends in the five counties listed above that delivered the highest percentages of votes to Donald Trump in the 2016 election, which registered, on average, virtually no change to their population totals in the four years between presidential elections, while registering only a minuscule amount of international migration as a percentage of county population (0.1 percent).

By contrast, the five counties where Hillary Clinton received the largest vote percentages (Ashland, Dane, Iowa, Menominee, and Milwaukee) averaged 3 percent population growth and international migration flows exceeding 1.1 percent of the counties’ populations (with many of these arrivals landing in professional, government, and academic centers such as Dane County and Milwaukee County). This turbulent migration profile also closely tracked counties that favored Barack Obama in 2012. Ironically, perhaps, locations less directly confronted by the reality of dynamic population flows often react more intensely to fears about its impact on their communities. Given Juneau County’s dramatic pivot toward Donald Trump in 2016, we can plausibly wonder about the impact of events outside the actual boundaries of the county, particularly given the absence of obvious triggers within the county.

The Dog That Didn’t Bark

Nate Silver’s election post-mortem at FiveThirtyEight has included analysis of voting patterns that indicates “swing counties” pivoting hard from Obama to Trump tended to reflect lower levels of education rather than lower levels of income. Juneau County, with post-secondary education levels well below the mean for the state, largely conforms to this model, but the education variable cannot itself serve as an explanation for this electoral pivot. What we need are plausible explanations for the pivot that factor in education, but position it as a source of susceptibility, not in itself as a cause. In other words, lower levels of education may allow for more voter volatility or unpredictability based on Factor X or Factor Y.

And so, given the absence of any contributing “shock” to Juneau County – indeed, with measures of well-being generally high and/or on the rise – we are left with an unsolved mystery. What might account for the dramatic vote shifting in Juneau county in the 2012 and 2016 presidential elections, from Obama to Trump, from Hope to Dope? With respect to endogenous factors that give Juneau County itself some agency in the voting outcomes, I am inclined to propose two contributing factors: a) the breaching of some critical gender/age boundary (which we might call the Get Off My Lawn effect); and b) the counterintuitive idea that increased stability and well-being in Juneau County may have liberated many of its residents from attention to the empirical reality of their lives (the realm of necessity), and presented them with license to travel more deeply into the darker and more speculative and imaginative recesses of their minds (which we might call the Richard Hofstadter effect).

  • Get Off My Lawn – There are a lot of older white dudes in Juneau County (and even moreso in Adams County). More than 30 percent of Juneau County are white males 35 years of age and older. By contrast, the same cohort in Madison’s Dane County comprises only 22 percent of the population. And the pace of this aging has accelerated in recent years, even while it has skewed increasingly male. This older-white-male skew is not uncommon among generically Republican-leaning counties (the skew is is even more pronounced in Oconto County, Wisconsin, for example, where Trump defeated Clinton by more than 36 percentage points). But what I want to suggest is that the impact of this shift may not be linear. It is possible that when the less-than-well-educated-older-white-guy percentage crosses a specific threshold (I am going to arbitrarily say 30 percent of the population!), a susceptibility to a cranky, reactionary, centrifugal politics may rapidly establish itself, independently of the objective circumstances of the life surrounding one (imagine a lot of bitched dudes with too much time on their hands enlisting in the online Breitbart jihad – these message boards don’t write themselves).

  • Richard Hofstadter – This is where we can recall our debt to the redoubtable American historian Richard Hofstadter, whose books on Anti-Intellectualism in American Life and on The Paranoid Style in American Politics can provide us with a kind of solace, in that we learn to what extent the politics we are now experiencing, no matter how bizarre they seem to us in the moment, are definitely in the American grain. Hofstadter’s specific contribution to understanding this pickle in which we find ourselves is to consider how politics can target and address emotional and psychological states with reference to matters of status, which is relative, versus interests. This summary is hopelessly oversimplified, but for our purposes, the tantalizing possibility is that the Obama-Trump pivot in Juneau County takes place in the mind’s eye, not with reference to what is actually outside one’s window, but in association with what is just beyond one’s view, just down the road – the police shootings in Madison, the riots in Milwaukee, a crawling anxiety that things are not quite right at home, even if they seem just fine. And so much of these imaginings may revolve around relative perceptions of status and well-being, which can of course leave one vulnerable to a sense that all is in flux around you, partly because there is so little flux in your stagnant little frog pond – that folks different from you and your family and friends, folks you don’t know but who (because you don’t know them) are more easily imagined as alien, invading hordes, marshaling in the fetid urban zones surrounding Madison and Milwaukee; an unruly, unworthy, unsanctified mob receiving undeserved benefits and favors and handouts that never accrued to you or to your little peace of heaven in Juneau County, Wisconsin.

Speculative, yes, but the recipe makes sense: A critical mass of less educated, aging white males + Local stasis or stagnation + Multi-ethnic, multi-racial, and multi-cultural flux just beyond one’s borders = the Donald Trump Fever Dream.

Note on sources – The analysis in this essay uses data from the U.S. Census Bureau and from the fantastic County Health Rankings & Roadmaps program, a collaboration between the Robert Wood Johnson Foundation and (appropriately) the University of Wisconsin Population Health Institute.