How Greg Abbott Bused 100,000 Migrants Directly Into Trump’s Electoral Column

Three percentage points. In a presidential election decided by margins thin enough to fit inside a rounding error, three percentage points is not a trivial number. It is the gap, county by county, between places that received buses full of migrants from the Texas border and the places that didn’t, and it ran, uniformly, toward Donald Trump.

A new study published in Sociological Science traces the 2024 electoral consequences of one of the more unusual state-level policy experiments in recent American history. From April 2022 onward, Texas Governor Greg Abbott chartered buses and transported more than 100,000 newly arrived immigrants to six cities, each led by Democratic mayors, each with so-called sanctuary ordinances that restricted local cooperation with federal immigration officers. The destinations were Chicago, Denver, Los Angeles, New York, Philadelphia, and Washington, D.C. Abbott, asked to explain his reasoning at the Republican National Convention, said the program would “take the border to them.”

What researchers from the University of North Texas and the University of Southern California have now shown is that it did something else as well. In the counties receiving those buses, Trump performed measurably better in 2024 than in any of his two previous presidential races. The research team, led by sociologist William Scarborough, used county-level voting data across three election cycles (2016, 2020, and 2024) in a two-way fixed-effects model designed to isolate what was distinctive about the treated counties in 2024 specifically. The answer, across three different comparison groups and nine robustness checks, came back consistent: roughly 3.4 percentage points of additional Trump support that couldn’t be explained by demographics, income, education, or broader electoral trends.

That’s county-level. To understand why, the team went further.

Using individual-level exit poll data from the Associated Press’s VoteCast survey, Scarborough and his colleagues pulled apart the mechanism into its constituent pieces. Two groups drove the effect. First, swing voters: people who had voted for Biden in 2020 and lived in a busing destination were more likely to vote for Trump in 2024, pushed, the analysis suggests, by amplified concerns about crime. Among former Biden voters who expressed crime as their top issue, just 2.5 percent voted for Trump if they lived outside a busing destination. Inside a busing destination, that figure rose to nearly 10 percent. Second, Republican turnout: self-identified Republicans who had sat out the 2020 election were 10 percentage points more likely to vote in 2024 if they lived somewhere buses arrived, with immigration concern as the main driver. Among immigration-concerned Republicans in destination areas, turnout reached 60 percent. In non-destination areas, the same kind of voter turned out at less than 40 percent.

The third possible mechanism, reduced Democratic turnout, did not materialise. Democrats who voted in 2020 were no less likely to vote in 2024 regardless of where the buses went. The busing program, it seems, did not depress the other side; it energised its own.

“Texas Governor Abbott and President Trump were able to activate and mobilize anti-immigrant sentiments that had a measurable impact on the outcome of the presidential election,” said Brady, a professor at the USC Price School of Public Policy. “Given the overwhelming evidence that immigration is associated with lower crime, voters should be particularly skeptical when politicians try to falsely link immigration and crime.”

This is where the study gets into stranger territory. The places that received migrants were enormous cities, several of them among the most immigrant-dense metropolitan areas on the planet. The 100,000 migrants bused over two years represented a tiny fraction of the existing foreign-born populations in Los Angeles County, Cook County, or New York’s five boroughs. Prior research has found that the busing program had no measurable effect on local crime rates in destination cities. Immigrants, the evidence consistently shows, commit crimes at lower rates than native-born citizens. None of that seemed to matter much.

What mattered was the narrative. Media coverage and political messaging constructed, over the months that buses arrived, a public image of crisis in these cities. The perception of threat, not the actuality of it, is what the analysis traces through to the ballot box. “Our study confirms longstanding sociological work on minority threat, showing that many individuals react against the perception that they are in closer proximity to minority groups,” Scarborough said, adding that in this case that reaction expressed itself as support for candidates who broadly oppose immigration. “This was not, however, a foregone conclusion. Media coverage and political messaging framed the busing program as a crisis, significantly shaping how it was perceived by residents and raising their concerns.”

The minority threat concept has a long pedigree in sociology, rooted in Herbert Blumer’s mid-twentieth century work on how groups respond not to direct contact with other groups but to the symbolic construction of those groups in what he called the public arena. The Texas busing program, the paper argues, was an unusually clean natural experiment for testing this idea: high visibility, moderate demographic impact. The spectacle was disproportionate to the demographic reality. Which made it, from a research design perspective, an almost ideal probe of whether perception alone can move votes.

There are things the analysis cannot tell us. It cannot fully separate the effect of the buses from the broader saturation of immigration as a political issue in 2024, though the fixed-effects design does its best to control for national trends. It can’t tell us which specific news coverage was responsible, or whether the effect would have been smaller in the absence of coverage that (in some markets) amplified the arrival of buses as a recurring crisis story. The study also can’t establish that the counties tested would have stayed on their pre-2016 trajectory in the absence of the program; it can only test that they diverged from comparable counties. The parallel trends assumption holds in the pre-period, which is reassuring, but it remains an assumption.

What the study can say, with some confidence: the political impact of the busing program was probably driven less by any actual contact between destination residents and new arrivals than by the media and political environment those arrivals were embedded in. As state governments accumulate new tools for intervening in immigration policy, and as those interventions become, in themselves, objects of sustained media coverage, the research suggests that the electoral effects may come not from the policy’s substance but from its spectacle. With 2026 midterms already in view, and immigration enforcement expected to remain a dominant issue, the question of whether that dynamic repeats, and whether it can be operationalised at smaller geographic scales, seems worth watching rather carefully.

DOI / Source: https://doi.org/10.15195/v13.a11

Frequently Asked Questions