A recent GeoCurrents post suggested that the electoral geography of the United States has strikingly deviated from that of Latin America during the 21st century. In the U.S., population density has become one of the main influences on voting patterns, with cities favoring candidates on the left and the countryside preferring those on the right. An analysis of the 2026 presidential election in Peru revealed no such divide, as both the primate city of Lima and its metropolitan area voted strongly for the (narrowly) victorious rightwing candidate, Keiko Fujimori.
Could an urban/rural electoral divide be emerging in other Latin American countries? As Colombia recently held a presidential election, also narrowly won by the rightwing candidate, it bears examination on this score. At first glance, it seems to fit the pattern found in the United States better than that of neighboring Peru. Bogotá, by far the largest city in Colombia, gave the leftwing candidate, Ivan Cepeda, a seven-point victory, even though the province that almost surrounds it favored the right-wing candidate, Abelardo de la Espriella, by roughly same margin (see the map posted below).
In Bogotá and its larger metropolitan area, which extends into Cundinamarca Department, socio-economic class was a much better predictor of voting behavior than population density. As can be seen in the middle map of the figure posted below, which shows the densely settled “urban zone” of Bogotá proper, the northeastern neighborhoods strongly supported de la Espriella while the southwest strongly favored Cepeda. Not coincidently, the former area is the wealthiest part of Bogotá and the latter is the poorest part. As can be seen on the map to its left, the metropolitan parts of Cundinamarca had a mixed voting pattern. Here the strongest support for the left was found in the municipality of Soacha, a major city itself with a population of nearly 900,000. Not surprisingly, Soacha is described by Wikipedia as “an important industrial zone [that] is home to mostly working-class families.” Intriguingly, in the capital district of Bogotá, the area that voted most heavily for Cepeda is Sumpaz, a high-elevation, low-density locality in the district’s far south that is described by Wikipedia as “completely rural, with no city services.”
In other parts of Colombia as well, voting patterns rarely reflect the urban/rural divide. Instead, the strongest influences are socio-economic class, regionalism, and ethnicity/race (as we shall see in this and subsequent posts). As can be seen on the map posted below, all the country’s largest cities* favored the candidate who triumphed in the departments in which they are located, often with roughly similar margins. In short, in Colombia’s 2026 presidential election regional political culture figured much more prominently than rural/urban differentiation.
Not all Colombian departmental capitals, however, voted for the candidate favored by the department in which they are located. In Boyaca, for example, Tunja, the capital and largest city, supported the left while almost all other parts of the department voted for the right. But a closer look at Boyaca’s voting patterns shows that urbanism was not the decisive factor. As can be seen on the map posted below, the cities of Sogamoso and Duitama, which are almost as large as Tunja, favored the right, although narrowly in the latter case. The decisive factor in Tunja’s leftwing orientation is probably its large university. As the map also shows, the three other parts of Boyaca that voted for Cepeda are not densely populated. Two of them are reported – by AI – to have strong labor organizations associated with mining and heavy industry, whereas the third has a very large indigenous minority. Again, class and ethnicity outweigh any possible urban/rural electoral divide.
The two additional figures posted below also seem to indicate the general lack of correlation between population density and voting behavior in Colombia. The paired maps show that three of the country’s most densely populated departments, as well its three least densely populated ones, supported the leftwing candidate, Ivan Cepeda.
The maps posted above, however, might be misleading, as they examine voting patterns at the departmental level, giving them a low level of resolution. The analysis thus far has only shown that cities in Colombia’s right-leaning departments also tend to support the rightwing candidate, or at least did so in the most recent presidential election. It still remains to be seen whether rural areas in left-leaning departments also supported the leftwing candidate in the same contest. That will be the subject of the next GeoCurrents post.
If Colombia had the same urban/rural electoral divide that characterizes the United States it would lean more to the left, as it is a much more urbanized country. Colombia’s intensive urbanization is unfortunately concealed by most official statistics. The Wikipedia article “urbanization by sovereign state,” for example, ranks the U.S. as slightly more urban (83.3%) than Colombia (82.4%). But different countries define “urban” in different ways, and the U.S. Census Bureau employs a broad definition, classifying any small census-block area that has at least 5,000 inhabitants or 2,000 housing units as “urban,” even if it is far from any sizable city. Using different definitions and methodology, the UN’s “World Urbanization Prospects 2025” determined that only 35 percent of the people of the U.S. live in actual cities, as opposed to 62.5 percent in Colombia. Measuring “rural” and “urban” populations is, however, notoriously difficult, and finding general agreement on a single metric is probably impossible.
One problem is the fundamentally different spatial patterns of “urbanization” in the United States and most other countries, including Colombia. Most Colombian cities are densely populated, and there is usually a sharp divide between the city and the countryside. The United States is characterized by more of a gradient. Few of its cities are thickly settled by global standards, and density tends to gradually decline from urban cores to outer cities, to inner suburbs, to outer suburbs, to exurbs, and finally to the actual rural zone. Dispersion, moreover, often extends to the smallest towns and hamlets (or “census-designated places,” as used by Wikipedia). A tiny moribund settlement in the American West can occupy a larger area than a substantial town in Colombia or Europe.
* Excluding those located within the metro areas of larger cities.