Analyzing votes from the presidential primaries, and census and trade data, a new report from UCLA researchers shows that two of the prevailing messages from Republican candidate Donald Trump’s presidential campaign — namely the negative effects of Mexican immigration and trade with Mexico and China on the American economy — are not significant factors in counties where he actually holds the most voter support.

The report shows that Trump’s support is concentrated in counties and commuting zones — which include urban areas as well as rural ones that share a common market — that are less likely to have significant numbers of Mexican immigrants and less likely to experience import competition from China and Mexico. Import competition typically contributes to depressed employment for low-skilled, high-pay work.

Nevertheless, economic conditions in areas where Trump enjoys his strongest support are worse than the country as a whole. In these places the average poverty rate is 15.19 percent, compared to the national average of 13.5 percent. And the average unemployment rate is 9 percent compared to the national rate of 4.9 percent.

“Our results show that, while many people in Trump-voting counties are struggling financially, trade and immigration are not to blame for those struggles,” said Raul Hinojosa-Ojeda, primary researcher and executive director of the UCLA North American Integration and Development Center, which studies transnationalism, immigration, immigrant remittance corridors and globalization. “We think it is vitally important that policymakers understand the realities in order to make decisions that have actual potential for improving economic conditions.”

Researchers used demographic and socioeconomic data from the U.S. Census Bureau’s American Community Survey from 2010 and 2014. They examined voting results from the Republican and Democratic primaries for 2,621 of the 3,007 counties in the United States.

Hinojosa-Ojeda, who is also a professor in the UCLA Cesar E. Chavez Department of Chicana and Chicano Studies, and the other researchers analyzed voting results as compared to migration patterns specific to each region, as well as exports and imports and the overall pattern of economic well-being in the areas. They applied the data both to individual counties as well as multi-county commuting zones. Results of commuting zone data closely reproduced what researchers found at the county level.

Less than 2 percent of the counties studied exhibited both very high support for Trump and a very high number of Mexican immigrants or very high exposure to Mexican or Chinese imports. In the report, a given variable, like the number of immigrants from Mexico, in a county or commuting zone was defined as “very high” if the value was higher in that area than it was in 75 percent of other counties or commuting zones.

Meanwhile, voters in areas that have very high numbers of Mexican immigrants and very high import competition are more likely to support either a Democratic candidate or a Republican candidate other than Trump, according to the report, which was co-sponsored by the UCLA Institute for Research on Labor and Education and its director, Abel Valenzuela, also a professor of Chicana and Chicano studies.

Among counties where Trump received support during the Republican primary, only 16 percent had a population of non-citizen Mexican immigrants that was higher than the national median. Among commuting zones in which Trump received above average support during the Republican primary, only 39.9 percent had an above-the-median population of non-citizen Mexican immigrants.

“Trump has also argued that two major U.S. trading partners — Mexico and China — have limited domestic consumption of U.S.-made goods and hamstrung the ability of U.S. firms to produce exportable goods,” said Maksim Wynn, public administration analyst at the Institute for Research on Labor and Education. “This creates the assumption that his support base would come from counties that import the most and export the least. Our research shows the opposite is true.”

According to calculations by the UCLA researchers based on data from WISERTrade and the U.S. Census Bureau’s 2012 economic census, Trump supporters are more likely to live in areas that have actually benefited from net exports to China and Mexico.

Researchers found that the greater impact of imports from Mexico or China in a county or commuting zone, the more likely it is that voters in those areas supported a Democratic candidate or Republican other than Trump.

“This contradicts a core tenant of Trump’s trade-policy narrative,” Hinojosa-Ojeda said. “Though many of his supporters may be marginalized workers, trade cannot generally be blamed for that status.”

The UCLA study counters previous research by the Wall Street Journal on Aug. 11, which stated that Trump won 89 of the 100 counties most affected by competition from China. UCLA researchers attribute the disparity to a limited sample size of voting counties in the Wall Street Journal analysis, as well as a reliance on larger multi-county commuting zones and a trade exposure based on national level, without an analysis of exports or migration at the county level.