That is a newly published piece by George Hawley, Social Science Quarterly, not yet available on-line as far as I can find:
I test the hypothesis that immigration status itself is a predictor of Democratic Party affiliation and vote choice, even controlling for other attributes. I further test whether having immigrant parents and grandparents has a similar effect. Method.To examine these questions, I created single- and multilevel models of party affiliation and vote choice using the 2016 Cooperative Congressional Election Study. Results. Even after controlling for a myriad of individual and contextual attributes, immigration status was a statistically significant and substantively important predictor of Democratic affiliation. This was also true of the children and grandchildren of immigrants, but this effect weakened over multiple generations. Conclusion. Immigration status itself appears to be an important determinant of voting patterns, which is highly consequential, given the large and growing foreign-born population in the United States.
Perhaps this explains some small part of American politics in recent times.
For the pointer I thank D.