Within the African American community there seems to be a fairly clear and consistent geographic split in the context of the 2016 Democratic primaries: Black voters further south favor Secretary Clinton by substantially larger margins than those up north. This article is not about that split, but instead focuses on a similar cleavage that exists within the Latino vote in this year’s Democratic nominating contest. However, the Latino community’s divide is not as geographically simple as the one that exists within the African American community. New York provided a clear example of this: it is much further north (and is generally seen as more ideologically liberal) than Nevada and Illinois, two states where Senator Sanders did relatively well among Latinos, yet he lost with New Yorkers of Latino descent by margins similar to those seen in Sun Belt states like Texas and Florida. If the north-south split seen in the Black vote does not apply as cleanly to the Latino vote, then what explains this result?
First, some raw data to orient this analysis. Based on exit polls from NBC News, the Latino vote broke down as follows:
Texas — 71% Clinton / 29% Sanders
Florida — 68% Clinton / 32% Sanders
*Arizona — 67% Clinton / 33% Sanders
New York — 64% Clinton / 36% Sanders
Illinois — 50% Sanders / 49% Clinton
Nevada — 53% Sanders / 45% Clinton
*Arizona did not have exit polling. However, Secretary Clinton saw two of her three largest margins of victory in the two counties in the state that are majority Latino: Santa Cruz and Yuma. Furthermore, considering how badly she performed with White voters in nearby Western states like Nevada, Utah, and Colorado, it seems very hard to believe that she could have won Arizona by nearly 20% without a substantial margin among Latino voters. As such, I assume that she won Latinos there by just about the same 2-to-1 margin indicated by exit polls in other states where she won the Latino community.
Exit polling is notoriously inaccurate, but I find these baselines fairly believable. The basic idea is that Secretary Clinton won Latinos by roughly 2-to-1 margins in Texas, Florida, Arizona, and New York, and basically split them 50-50 with Senator Sanders in Illinois and Nevada.
To restate an earlier point, there isn’t quite the geographic consistency one might have expected to see. Illinois is north of Texas and Senator Sanders won its Latinos — yet New York is even further to the north and Secretary Clinton won the Latino vote there! Secretary Clinton seems to run strong with Latinos in the southwest — but Nevada is part of that region, and Senator Sanders won the Latino community there! One might be inclined to posit some kind of Mexican American vs. Cuban American and Puerto Rican split in the Latino vote — but then how to account for Texas, whose Latinos are overwhelmingly of Mexican descent and gave Secretary Clinton her largest margin of victory among Latino voters to date?
The way to begin explaining this seemingly erratic pattern of division within the Latino vote is by looking not at geography, but at time. Specifically, time living in the United States. The cleavage within the Latino community that I believe determines any given set of Latino voters’ preference for either Secretary Clinton or Senator Sanders is how long said population has been present in the US. I refer to this as “degree of Americanization:” the longer a state’s Latino community has been in the US, the more “Americanized” it becomes — which in turn makes them more likely to favor Secretary Clinton over Senator Sanders.
Looking at the states where Secretary Clinton won Latinos (Texas, Florida, Arizona, and New York) versus those where Senator Sanders was preferred (Illinois and Nevada), the first thing that jumps out is how much more long-standing the Latino presence in Secretary Clinton’s states is. The four Clinton states have major Latino communities whose presence extends back decades — centuries in the case of Texas and Arizona! In contrast, Illinois and Nevada’s Latinos are very recent arrivals, a great many only coming to the country in the past 30 years. The major temporal dividing line, I submit, is the year 1970, when the effects of President Lyndon B. Johnson’s immigration liberalization policies began to take full effect. Latinos who were in the US before then (or whose parents were in the US before then) are predisposed to favor Secretary Clinton, while Senator Sanders does well with those who came after that year.
When supporters of Senator Sanders insisted that he would win or tie with Latinos in New York, their assertions never rang true for me and I predicted that Secretary Clinton would get roughly the same 2-to-1 margin she enjoyed with Latinos in places like Florida. The reason I believed this goes back to the history of New York’s Latinos. New York’s Puerto Rican community has roots extending back to World War II, or even earlier. A great many of the other Latinos in the state are of South American descent, many of whom also have been present since before the 1970s. Florida is a similar case: it’s Latino community is heavily Cuban and Puerto Rican, both of whom arrived in the US in the 1950s and 1960s. It is worth mentioning that the growing number of Puerto Ricans there is fueled by growth in more recent years, however this still conforms to my theory of “Americanization” because they are moving to the state from within the US, not from overseas. Like other states on the US-Mexico border, Texas and Arizona’s Latino presence extends back not only to WWII, but all the way to the 19th century!
Turning to the causal mechanism for this split, I believe it to have both (1.) a compositional and (2.) an ideological component.
(1.) By “compositional,” I mean the demographic make up of the Latino community itself within each states. “Older” (or more “Americanized”) Latino communities feature more participation by older Latino voters, which greatly dilutes Senator Sanders’ strength with the youth vote.
In states where Latinos are relatively recent arrivals, like Illinois and Nevada, it is far more common for a lot more of the older adult population to be undocumented (because they are recent arrivals they have not yet had the time to try and gain lawful status). These Latino immigrants, by virtue of having come to the US so recently, are also far more likely to have missed the amnesty program offered in the 1980s, which further reduces the chances that they will be able to vote. The upshot is that the Latino vote in states where Senator Sanders won among Latinos is disproportionately younger — the children of the undocumented persons who themselves cannot vote.
In contrast, the states where Secretary Clinton won with Latinos are more likely to have larger numbers of older Latinos who have managed to secure lawful status and voting righta. This is partly because those who are immigrants are more likely to have been around for long enough to go through the process of getting citizenship, and are more likely to have been in the US for programs like President Reagan’s offer of amnesty. Furthermore, a greater proportion of these older Latinos are themselves US-born, which confers automatic citizenship upon them. In this way the Latino communities of these places is more “Americanized:” their mix of older and younger voters looks more like that of their neighbors, while states where Latinos preferred Senator Sanders have disproportionately young Latino voters.
(2.) There is more conjecture behind this part of the analysis, but I believe that another factor that influences how Latinos vote in each states’ Democratic primary is the ideology of the community itself. Latinos who have been in the US longer are more likely have fully adopted American political culture (or become more “Americanized”), including the US’s long-standing aversion to left wing ideologies. More importantly, Latinos who have been in the US for longer (or were born in this country) are less likely to be hanging on to the old-country socialism of their Latin American ancestors. Socialism is not as widely reviled in the US today as it was 30 years ago, but it remains FAR less popular here than in most of Central and South America. Latinos who came from those areas more recently (and their first generation children) are more likely to be ideologically comfortable with Senator Sanders’ democratic socialism. Latinos who are second or third generation (or are from family with even more long-standing roots in the US) are significantly more hostile to socialistic ideas, much like their American neighbors. Generational wealth almost certainly also plays a part in the ideological split between newer Latinos and more “Americanized” ones: undocumented Latinos are more likely to be poor, as are their children, compared to member of the Latino community who are non-immigrants or arrived earlier. Senator Sanders’ proposal to raise taxes to pay for more government services sounds great if your income is so low that you probably won’t be paying the new taxes; it sounds less great when you’re earning more and thus are more likely to be the one paying for the new social programs.
Looking forward to the three remaining states with the largest Latino populations, California, New Mexico, and New Jersey, I believe that the first two will see the Latino vote go to Secretary Clinton, while New Jersey’s Latinos break a bit more toward Senator Sanders. Like the two states on the US-Mexico border that have voted so far, California and New Mexico have large numbers of Latinos who have deep roots (extending to well before the 1970s) in the US, and this degree of “Americanization” will lead them to back Secretary Clinton. New Jersey is a less clear case. Its Latino population is newer, but how much of that is driven by foreign-born immigrants versus people moving between the states will be important. If it’s the former, then Senator Sanders should carry the Latino vote there. If the latter, then Secretary Clinton will get the state’s Latino voters.