�������

With election season heating up, one outcome is all but certain for state legislatures come November: There will be more winners from the far left and far right of the political spectrum, leading to more polarization across state governments.

That is the implication of a groundbreaking new Stanford study that offers the most extensive look to date at how state elections are fueling polarization in American politics. The analysis finds that at various points in the state electoral process — from filing an application to run for office through the general election — candidates who are ideologically more extreme have a distinct advantage. Moderates once had the upper hand, but that has since disappeared.

“We already know that polarization has increased markedly in state legislatures,” says Andrew Hall, a Stanford political science professor and senior fellow at the Stanford Institute for Economic Policy Research (SIEPR) who co-authored the study. “Our research shows for the first time that the rising number of ideologically extreme people running for state legislature plays a major role — and has important implications for polarization at the federal level, too, since most members of Congress start out as state legislators.”

The research, detailed in a working paper, finds that candidates who are more extreme are increasingly running for state office. They are also winning primaries and general elections more often than at any time in at least 30 years, and are often shoo-ins for victory given that 80 percent of general elections aren’t competitive. The study is based on election results for more than 84,000 candidates for state legislature from 1992 to 2020.

In addition to its overall breadth, the study is remarkable for its novel measure of ideology across all candidates — including those who lose and may never hold office. The measure uses machine learning to predict candidates’ hypothetical future roll-call votes from the campaign contributions they have received. By their calculation, Hall and his collaborators show not only that more extreme candidates are winning at the state level, but also that the entire initial pool of would-be legislators — Republican and Democrat combined — has become markedly more polarized over time. Meanwhile, fewer moderates are throwing their hats into the ring.

The significance of the study’s findings extends beyond election day, the researchers say. State legislatures are taking the lead on hot-button policy issues — among them the pandemic response, abortion, voter registration, gun ownership, and climate change. State office is also a common gateway to national office, so polarization at the state level likely portends ongoing polarization in Washington, D.C.

“As polarization — and the legislative gridlock it creates — has grown, it’s important to understand how the state legislative election system is contributing to it and to think about policies that could offset its effects,” Hall says. His co-authors were Andrew Myers, a predoctoral research fellow at SIEPR, and Cassandra Handan-Nader, a doctoral student in political science.

A dramatic shift since 2010

When it comes to unraveling America’s recent shift to political divisiveness, researchers and others have identified many contributors — including gerrymandering, mass media, the Twittersphere, incumbent lawmakers, and voter sentiment. Another potential explanation is the initial pool of candidates. In his 2019 book, , Hall makes the case that changes in who runs for office contribute as much to polarization as incumbents do.

Analyzing the role of candidate ideology is challenging. Candidates who run and have never held office don’t have legislative voting records, which are far more reliable than campaign stump speeches in evaluating ideology. Adam Bonica, a Stanford associate professor of political science, in 2014 the idea of using campaign donations as a measure of ideology for incumbent and non-incumbent candidates, and later in 2018 how polarization is rising among candidates for federal office.

Hall and his co-authors set out to build on Bonica’s ideas by looking at state elections. “State campaign finance records would also allow us to impute how people who never make it into office would have voted if they had,” Myers says. “Donors, especially sophisticated ones, have a lot of ground-level information about candidates and will contribute largely based on the candidate’s ideology.”

The researchers analyzed data on nearly 44 million campaign contributions spanning three decades. To ensure apple-to-apple comparisons of donations made to incumbent and non-incumbent candidates at the same point in the electoral process, the researchers trained their machine-learning model to consider only donations made to incumbents when they first ran for office.

Using this method, they analyzed candidate ideology at three stages in state elections: filing to run, the primary, and the general election. Among their key findings:

• Polarization of the entire pool of candidates has increased dramatically in recent decades.
• Competition for state office, meanwhile, has remained low over time. ������� 20 percent of general elections are decided by less than 20 percentage points — which, at the upper range, is not very competitive at all, Handan-Nader says. And nearly 15 percent of candidates don’t face any opponents, in either the primary or general election.
• Since 2010, there has been a marked reversal in the number of state legislative districts deemed safe for either Republicans or Democrats. Specifically, where Democrats held the majority of safe seats from 1992 through 2010, Republicans now have the largest number by far.
• More ideologically extreme candidates in state primaries have a modest advantage that has doubled since 2010. In that time, moderates have seen their once-favorable advantage in general elections shrink to nearly zero.
• In general elections that don’t involve a U.S. presidential ticket, however, moderates have a somewhat larger advantage. According to the study, this fits with a theory that many voters vote along party lines, without knowing much about the candidates. In off-cycle elections, voter turnout is lower and those who go to the polls are likely more informed about who is running.
• Ideologically extreme Republican candidates lose more share of the vote in general elections than do their Democratic counterparts on the left.

A ‘unified measure’ of extremism

The researchers say that the patterns they identify open multiple avenues for future research.

For example, the study’s measure of candidate ideology could be used to follow candidates from the state legislative pool to the federal office pool, Handan-Nader says. “We now have a unified measure of ideology that could in the future trace the path of ideological extremism at different levels of government to understand when it happens and when it gets worse,” she says.

There are policy implications, too. “We may want to explore ways to encourage more-moderate people to run for office,” Hall says. His previous work, for example, explored how rising state legislative salaries draws in moderate candidates. He says reforms to the primary system — including changes to campaign finance laws and the introduction of alternative ways of voting — could also help level the playing field between more-extreme and moderate state candidates.

“Gridlock and performative ideological posturing are major challenges to effective government,” Hall says. “Sensible government requires an adequate supply of sensible politicians. Most of our members of Congress start out as state legislators. If we want a deeper supply of pragmatic politicians, we need to start by getting more moderate people to run for and win state legislative elections.”