John Hood: Don’t buy myth about voter turnout

RALEIGH — Republicans in much of the country enjoyed significant success during the 2021 election cycle. They elected a GOP governor in Virginia, almost elected one in New Jersey, won several important judicial races in Pennsylvania, and even elected their first Republican candidate to local office in Seattle (city attorney) since the 1980s.

In most of these places, voter turnout was up from previous off-year cycles — a fact that politicos, in particular, ought to take to heart.

Too many have bought into the widespread assumption that lower turnout tends to benefit Republicans and higher turnout tends to benefit Democrats. This assumption reflects little more than political folklore. It’s not proposition supported by empirical evidence.

For me, the folk tale dissipated many years ago when I took a look at the turnouts and outcomes for U.S. Senate races in North Carolina and discovered no consistent relationship between the two variables — even though I’d been assured repeatedly, by operatives in both parties, that Republicans liked to run negative ads against their opponents not so much to swing undecided voters their way but simply to turn voters off so they wouldn’t show up, thus giving the GOP an edge by Election Day.

All I did, admittedly, was run a simple test for correlation across data from a single category of races in a single state. More sophisticated models are needed to test the broader proposition that low turnouts help Republicans, models that include lots of other variables from multiple jurisdictions and then attempt to adjust for them. Scholars have now published dozens of such rigorous studies. While the results aren’t unanimous, they tend to undermine the conventional wisdom.

Daron Shaw of the University of Texas and John Petrocik of the University of Missouri summarized much of this research in their 2020 book The Turnout Myth. “Seventy years of survey data and election outcomes suggest that turnout has no systematic partisan consequences,” they concluded. Among other findings, they discovered that “the candidate preference of those most likely to vote and those least likely to vote is almost indistinguishable.”

It’s not that a pro-Democratic tilt to higher turnout was a ridiculous notion, by the way. On average, Republicans tended to be more likely to vote than Democrats. And in North Carolina, especially, there were way more registered Democrats than registered Republicans. It seemed plausible, then, that efforts to register voters and make it easier for them to cast ballots — the introduction of early voting, for example — would net more Democratic than Republican votes, therefore flipping competitive races to the blue column.

Still, plausible isn’t provable. It turned out that this pervasive turnout myth was based on overly simplistic analysis. For one thing, that big Democratic edge in voter registration has been shrinking since the 1980s. For another, even if Democratic-leaning folks outnumber Republican-leaning folks in the pool of available non-voters, that doesn’t necessarily mean making it easier to register and vote will have comparable effects on these two groups. In other words, everything else being equal, perhaps the Republican-leaning bystanders will be more likely to respond to your policy changes and join the electorate. (Yes, Democrats are more likely to vote early, but the vast majority of them were going to vote, anyway. That’s redistributive, not additive.)

The more you read studies about political behavior, the quirkier it looks. For example, a new paper in the Journal of Elections, Public Opinion and Parties examined the turnout effects of sending postcards to voters in state legislative districts. The authors found that the postcards likely reduced rather than increased turnout in those districts, perhaps by pulling the attention of low-propensity voters away from marquee campaigns towards legislative races that didn’t excite them as much.

Elections can certainly turn on turnout differentials between the two major party coalitions, differentials that may reflect tactical choices or issue salience at the time. But overall turnout is not predictive of who wins elections. Please discard this myth and adjust accordingly.