Geopolitics

A recent GeoCurrents post suggested that race played a significant role in the 2026 Colombian presidential election. The racial factor is easily seen in the voting patterns of Black Colombians, of Afro-Colombians, who number between five and seven million, constituting some nine to 14 percent of Colombia’s population. As the map to the left in figure posted below shows, Afro-Colombians are heavily concentrated in the Pacific coastal region. As the map to the right reveals, this region voted heavily for the losing leftwing candidate, Ivan Cepeda. Many of the municipalities in this region are more than 80 percent Black, and almost all of them gave more than 80 percent of their votes to Cepeda. As Black Colombians have a roughly 80 percent poverty rate, it is not surprising that they favored the leftwing candidate.

The correlation between Afro-Colombian identity and leftwing voting is not, however, as close as it might appear. A lack of correspondence is seen in some parts of the racially mixed Caribbean region. As can be seen on the paired maps posted below, a few heavily Black municipalities are scattered across northern Colombia. Many of them voted strongly for Cepeda, but several favored the victorious rightwing candidate, Abelardo de la Espriella. As the next map shows, the most conservative-voting Black-majority municipalities have relatively small populations and can be considered substantially rural.

Curious about the differences between Black voting patterns in the Pacific and Caribbean regions of Colombia, I queried AI (Grok) about possible explanations. The response seems convincing, at least at a superficial level:

Pacific Afro-Colombian communities (Chocó, parts of Valle/Cauca/Nariño) tend toward higher isolation, stronger ethnic/organizational identity, and alignment with left/progressive platforms addressing territorial rights, violence, and exclusion. Caribbean Afro-Colombian populations are often more integrated into mestizo-majority coastal societies, with greater influence from traditional parties, cattle/agriculture economies, and patronage networks. Group consciousness and ethnic voting are real but weaker or more conditional there.

Of the conservative-voting Black municipalities seen on the map above, Altos del Rosario is the most intriguing. Approximately 82 percent Afro-Colombian, it gave 63 percent of its votes to de la Espriella. Altos del Rosario is noted for its high poverty level (roughly 70 percent), its meager infrastructure, and its poor public services. In its 2024 mayoral election, it narrowly selected an unconventional leftist, Juan Carlos Romo Pérez, who goes by the name “Happy.” A relatively young politician, Romo Pérez is from La Guajira Department in far northern Colombia, which has a sizable indigenous population. He represents the Movimiento Alternativo Indígena y Social (MAIS), a political organization that advocates eco-leftist policies.* On gaining office, he focused on infrastructural development and quickly gained widespread support. According to AI, “a mid-2024 survey placed Romo Pérez near the top among mayors of small municipalities in the region, [with an approval rating of] around 82–83%.” In April 2026, however, he was hit with a corruption scandal. Charged with the “improper handling of public resources,” Romo Pérez  was suspended from office for three months without pay.

I have a difficult time understanding how a community like Altos del Rosario could elect an eco-leftist mayor in 2024, give him strong approval, and then turn to a hardline conservative presidential candidate in 2026. Perhaps the corruption charges played a major role, but I suspect that the main factor is the differences between local and national politics, as mediated through patronage networks. But without further investigation, this is unfounded speculation.

Examining Black-majority municipalities in Colombia that voted for a rightwing presidential candidate made me wonder whether any Black-majority counties in the United States supported Donald Trump in 2024. As I am quite familiar with the two maps posted below, I initially doubted that any could be found. But after investigating the issue, I found 10 such counties, mostly located in Mississippi and Georgia. As can be seen on the next map posted below, these counties are characterized by both narrow Black majorities and narrow Trump victories. Heavy Republican voting by White residents, as would be expected in the rural South, along with low Black turnout, could easily explain these seeming discrepancies. Nationwide, moreover, roughly 15 percent of Black American voters nationwide opted for Trump in 2024. If replicated locally, such a voting pattern could easily swing a closely contested local election in his favor.

The most curious Black-majority Trump-voting county Issaquena County, Mississippi, which has the largest African-American majority (62 percent) in the set. But note on the map that Issaquena gave Trump only a slight plurality of its votes, 47.6 percent. Kamela Harris essential lost the county because it gave an inordinately high percentage of its votes to third-party candidates, some on the far right, some on the far left, and others unclassifiable. Such skewed election results partly reflect the county’s very small population (1,338). Only six residents of Issaquena voted for Claudia De la Cruz of the Marxist/Leninist Party for Socialism and Liberation (PSL), but that still represented almost one percent of the county’s total votes.

* Main aspects of the MAIS platform include: Defense of human rights and “Mother Earth”; Transformation of socio-economic, territorial, cultural, and political structures to achieve equity, peace, and buen vivir; Unity in diversity, ancestral indigenous wisdom, environmental harmony, participatory democracy, and ethical/transparent politics; An alternative economic model respecting different visions of development and environmental protection.

The previous GeoCurrents post on the 2026 Colombian election argued that Colombia, like most Latin American countries, does not have the stark urban/rural electoral divide that now characterizes the United States. The post showed that in Colombia’s conservative central region, cities and rural areas alike generally voted for the rightwing candidate, Abelardo de la Espriella, with the important exception of Bogotá.

Colombia’s Caribbean region, however, exhibited a different pattern. As can be seen on the first map posted below, this part of the country solidly favored the losing leftwing candidate, Ivan Cepeda. But as the second map shows, de la Espriella triumphed in many of the region’s municipalities (or local districts), often by a significant margin. Most of these right-leaning areas are in the interior portion of the region and thus have significant connections with Colombia’s conservative central highlands, although quite a few are near or even on the coast. More important, most of these conservative-voting areas are characterized by low population densities. In contrast, the largest cities of the region all favored the left. A degree of urban/rural political differentiation thus seems evident.

The urban/rural political divide in Colombia’s Caribbean region, however, is not pronounced. As can also be seen on the map posted above, the Caribbean municipalities that voted most heavily for the left actually tend to be located in more peripheral areas. A map limited to Bolivar Department (posted below) allows us to explore these patterns in more detail. As can be seen, four of the five municipalities containing the department’s largest cities were carried by Cepeda. But some of the “bluest” (rightwing-voting) as well as some of the “darkest magenta” (leftwing-voting) municipalities are sparsely populated and essentially non-urban. As a result, we can only generalize about the political orientation of the Caribbean region’s largest cities, which favored the left, although not overwhelmingly. No generalizations seem possible about the department’s lightly populated municipalities. Here, local factors tend to be far more important than population density. In particular, the Colombia’s Caribbean countryside has long been noted for its “political machines” and patron-client networks, which can largely determine election results at the local level.

Although Colombia’s Caribbean region has been tending left in recent elections, it does not have the consistently leftwing orientation that characterizes the Pacific coastal region and the southern highlands. As can be seen in the map set posted below, in 2018 five of the region’s seven departments favored the victorious rightwing candidate Ivan Duque, over Gustavo Petro of the left. Four years later, Petro swept the Caribbean region and won the national election. Conservative Colombians hoped that de la Espriella would have a better showing in the region in 2026, as he is a native son. Although born in Bogotá, de la Espriella was reared in Monterío, the capital of Cordoba Department, located roughly 31 miles (50 km) from the Caribbean Sea. He speaks with a costeño (Caribbean coastal) accent, has supposedly identified himself as “costeñísimo,” and used the slogan “costeño vota a costeño” in his campaign. This regionalist strategy was not successful.

A recent GeoCurrents post suggested that the electoral geography of the United States has strikingly deviated from that of Latin America during the 21^st 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.