The Flip Counties Felt It First

The roughly two hundred counties that swung from Obama to Trump have carried more interpretive weight per capita than any other patch of American ground. Were they economically wrecked? Status-anxious? Simply contrarian? Most of the evidence marshaled on every side is aggregate — trade exposure, mortality, median wage — or collected after 2016, when memory had already been reorganized by the result. Almost none of it asks what people in those places said about their own lives, year by year, before anything flipped.

Gallup's daily tracking poll did ask. From 2008 through 2016 it interviewed Americans every night — 3,530,270 interviews in this extract, geocoded to county, including 213,784 in the 223 counties that voted Democratic in 2012 and Republican in 2016. Interviewers asked people to place their current life on a 0-to-10 ladder, to guess where their life would stand in five years, and to say whether their standard of living was getting better or worse. Because the questions were asked in real time, they cannot be contaminated by hindsight.

Comparing flip-county residents with residents of demographically similar non-flip counties — matched on urbanicity, census region, and county income — the file shows a quiet, specific pattern. Assessments of the present ran nearly parallel for nine straight years: flip counties sat about 0.05 of a ladder rung below their matched comparison in 2014–2016, roughly where they had always sat. Assessments of the future did not. The "hope gap" — how much higher people place their life in five years than today — grew steadily everywhere as the recovery matured, but it grew more slowly in the flip counties, and the shortfall peaked around 2014, two years before the election. Not misery. Foreclosed expectation.

What the chart holds stillSame kind of place, different slope

Flip counties are not a random slice of America: they are disproportionately small-town and rural, concentrated in the Midwest, middling in income (their average 2016 Republican margin was 11.3 points — they did not merely tip, they lurched). So the gray line is not the national average. It is a comparison group built from the nearly three thousand counties that never flipped, re-weighted inside 33 cells — urbanicity band (metro, small city, rural) by census region by county-income tercile — to carry exactly the flip counties' mix. This is coarsened stratification, the descriptive journalist's version of matching, not a causal design; it holds still the obvious confounders and nothing more.

Against that yardstick, the top panel is the dog that didn't bark. Current-life ratings in flip counties recovered from the crisis on the same schedule as everywhere comparable, drifting upward from about 6.9 to 7.0 on the ladder. The gap between the lines in 2014–2016 averaged 0.05 rungs — versus 0.045 back in 2008–2010. Nine years of macroeconomic drama, and the distance between flip America and matched America in the present tense moved by less than a hundredth of a rung per year.

The bottom panel moves. In 2009–2010 flip-county residents expected their lives five years out to be about as much better as matched residents did: the hope-gap difference was a rounding error (−0.016 on average through 2010). Then matched counties' expectations climbed with the recovery and flip counties' expectations stalled. By 2014 the raw difference reached −0.105 ladder rungs — flip residents at 0.37, matched residents at 0.47 — the widest of the series. The standard-of-living item tells the same story in a different grammar: everyone's "getting better" index rose sharply after 2011, but flip counties trailed by −0.061 in 2016 on the −1-to-+1 scale, double their early shortfall. Both are trajectory measures; both sag while the level measure holds.

And worry? Nothing. Within each survey regime, flip-county residents reported worry at essentially the matched rate — 26 percent versus 27 percent in 2016. Whatever was happening in these places, it did not register as day-to-day distress. It registered as a discount on the future.

The regression is where the story gets its edge. Once you compare flip-county residents with demographically similar people in similar non-flip counties, their ratings of the present actually gained ground after 2008 — the flip × 2016 coefficient on the current ladder is +0.09, statistically solid. Their expectations went the other way: the hope-gap coefficients are negative in every single year, bottoming at −0.13 in 2014 (clearly distinguishable from zero) before narrowing to −0.06 by 2016 (no longer so). Present brightening, future dimming, within the same kind of person in the same kind of place. That scissors pattern is the file's genuinely new contribution to a stale argument: it is hard to square with a story of pure material immiseration, and uncomfortably easy to square with one about expectations that stopped being believed.

Timing matters as much as sign. The expectations shortfall is not a 2016 phenomenon — it is a 2011-to-2014 phenomenon that had already begun to close before the campaign. The flip counties felt it first; by the time the rest of the country was arguing about them, the worst of the sag was behind them.

The composition checkIs it just who lives there?

Counties change residents, and flip counties skew white and non-college — exactly the groups whose national mood shifted most. So restrict both groups to white respondents without a four-year degree and re-run the comparison. The early-2010s sag survives: in 2014, white non-college flip residents' hope gap trailed their matched counterparts by −0.064. But the late-period deficit does not: averaged over 2014–2016 the restricted difference is −0.007 — effectively zero — because flip-county whites without degrees caught back up in 2015 and 2016. Read plainly: the mid-decade expectations slump in flip counties was real and visible within this fixed demographic slice, while the portion that lingered into 2015–16 in the aggregate numbers partly reflects who lives in these counties rather than something extra happening to each resident. Honest accounting cuts both ways.

Reading honestlyWhat this does — and doesn't — show

• It shows that in the counties that would later flip, measures of expected wellbeing (the hope gap, the future ladder, the standard-of-living trend item) lagged a matched comparison through the early 2010s, peaking around 2014, while measures of current wellbeing stayed parallel — and, person-for-person, slightly improved. The divergence preceded the flip; it did not follow it.
• It does not show that dimmed expectations caused anyone's vote. Counties were not assigned to flip, and this file records no individual's ballot. County vote context is an attribute joined to interviews by geography; "residents of counties that later flipped" is the only honest subject for every sentence here.
• It does not adjudicate economic anxiety versus status threat. A sagging hope gap is consistent with both — expectations can be foreclosed by a shuttered plant or by a felt loss of standing. What the data rule out, in these counties, is the cartoon of acute present-tense misery: life today was rated fine, futures were marked down, and daily worry never budged.
• The magnitudes are modest. A tenth of a ladder rung at the 2014 peak is a faint signal by individual standards — detectable only because the samples are enormous. Faint is not fake: it is the difference between a place where expectations climbed with the recovery and a place where they idled.
• Composition is part of the story. Within white non-college residents the mid-decade sag appears and then closes; some of the aggregate late-period shortfall reflects demographic mix. The cleanest claim is about the early 2010s, and that is where this article stakes it.

The op-ed war over the flip counties has mostly argued about what their 2016 vote revealed. Three million interviews suggest the more interesting fact sits earlier: between roughly 2011 and 2014, in places that looked unremarkable by every present-tense measure, the future quietly got marked down — and the people living there said so, out loud, to a stranger on the phone, years before anyone thought to ask why.