ethnicity and voting

Economic Disparities and Election Results in Turkey

Article-Grid, Elections, Ethnicity / By

The Republic of Türkiye (Turkey) is characterized by stark discrepancies in regional economic productivity. In 2021, nominal per capita GDP figures ranged from 153,479 Turkish Lira (17,089 US$) in Kocaeli Province, located east of Istanbul, to 26,837 Turkish Lira (2,988 US$) in Ağrı Province, located along Turkey’s eastern border. Although maps of Turkish per capita GDP by province are readily available, I made my own to highlight these regional disparities, accomplished mainly by using a two-color scheme.

The correlation between economic productivity and ethnicity in Turkey is close, at least on the low end of the spectrum. As can be seen by comparing the two posted above with that posted below, Turkey’s least economically productive provinces are all located in the primarily Kurdish southeast. The most economically productive provinces are concentrated on the other side of the country, in the northwest. This latter region includes both coastal provinces and provinces located in the northwestern quarter of the central Anatolian Plateau.

Historically, Turkey’s western coastal strip was much more productive than most of the rest of the country. Over the past several decades, however, a number of cities on the Anatolian Plateau have seen significant industrialization and rapid economic growth. Several have been deemed “Anatolian Tigers,” defined by Wikipedia as cities that “have displayed impressive growth records since the 1980s, as well as a defined breed of entrepreneurs rising in prominence and who can often be traced back to the cities in question and who generally rose from the status of small and medium enterprises.” Some of these “Tiger cities,” however, are located closer to the coast than the plateau. (The locations of the “Anatolian Tigers,” as defined by Wikipedia, are shown on one of the maps below.)

A comparison of the map of Turkish economic productivity with that of the 2023 presidential election (see the previous post) reveals some interesting connections. To illustrate these patterns more clearly, I have outlined in red the provinces that supported challenger Kemal Kılıçdaroğlu, superimposing them on the per capita GDP map.

As can be seen, a majority of the Turkey’s richer and poorer provinces supported Kılıçdaroğlu, whereas most of its mid-level provinces supported Recep Tayyip Erdoğan. Closer analysis, however, reveals that other factors are more important than per capita GDP for understanding the election results. For example, Kocaeli, Turkey’s most economically productive province, supported Erdoğan, as did most of its prosperous northwestern neighboring provinces. Support for Kılıçdaroğlu in the poor southeast, moreover, is less a reflection of economic standing than of Kurdish ethnicity, as most Kurds reject Erdoğan’s pronounced and ethnically inflected Turkish nationalism.

An interesting exception to the general patterns of anti-Erdoğan sentiments in the southeast is Şanlıurfa, which is one of Turkey’s poorest provinces; it also has, according to Wikipedia, a Kurdish majority. Yet Şanlıurfa decisively supported Erdoğan, as it had in the past several elections. But Şanlıurfa is more ethnically mixed than the Wikipedia article on it indicates; a 1996 study found that it had a Kurdish plurality but not a majority. Şanlıurfa is home to a substantial Arab community, as well as a very large refugee population. As the paired set of maps post below shows, the Arabic-speaking part of Şanlıurfa voted overwhelmingly for Erdoğan. (The pattern is markedly different, however, in Turkey’s most heavily Arabic-speaking province, Hatay, as will be examined in a later post.)

Kılıçdaroğlu’s other main region of support was the far west and southwest, covering the European portion of Turkey and the Aegean and Mediterranean coasts. Historically, this was the wealthiest part of Turkey, and it has long been the most liberal and Western-oriented part of the country. It is not surprising, therefore, that it supported the more liberal and Western-oriented candidate.

As the upper map posted above shows, Erdoğan’s main bastions of support are located in the core Anatolian Plateau, with the important exception of those areas dominated by the Alevi religious minority, and the Black Sea coast. Much of this region has seen substantial economic growth and infrastructural investment under Erdoğan’s leadership, has long been noted for its more conservative interpretations of Turkish nationalism. Many of the districts along the eastern half of the Black Sea coast had been heavily populated by ethnic Greeks before the early 1920s; when the local Greeks were expelled, Turks moved in, many of whom had themselves been expelled from Greece at the same time. It is not surprising that the descendants of such people are noted for their pronounced Turkish nationalism.

Two important provinces in the western portion of the Anatolian plateau supported Kılıçdaroğlu the 2023 election. One is Ankara, where the national capital (of the same name) is located, and the other is Eskişehir, which is noted for its large and strong universities. It is not surprising that these more cosmopolitan provinces voted against Erdoğan. Yet both provinces, and that of Istanbul as well, supported Erdoğan in the 2018 election, when his opposition was divided (see the maps below). From 2018 to 2023, Erdoğa lost ground in major urban areas, along the southern Mediterranean coast, and in the far northeast (another area with large ethnic minorities). Yet his overall vote stayed roughly the same, indicating a solidification of support over the core Anatolian region. Kılıçdaroğlu’s Alevi faith may have played a significant role here.

Do Swedish-Americans Vote for Democrats? National Origins and Voting Behavior in the United States

Article-Grid, Cultural Geography, Elections, North America / By

Percent Scandinavian US MapIn responding to a recent GeoCurrents post comparing electoral geography in Minnesota and northern California, commentator Barzai makes some important points about ethnicity and national background. As he notes, people of Scandinavian and German descent are a much more significant factor in Minnesota than in California. More importantly, he argues that the concept of a monolithic “White” population is challenged by such differences in national origin.

These are significant points, but the actual patterns of voting behavior and national origin in the U.S. are more complicated than one might think. It is certainly true that people of Scandinavian background are disproportionally concentrated in Democratic-voting Minnesota, as the first map indicates. Some other areas of Scandinavian settlement are also relatively liberal, such as Western Washington. Utah forms an intriguing exception, as it is a highly conservative state with a substantial Scandinavian population, but in this case conversion into the politically conservative Mormon Church (Church of Jesus Christ and Latter Day Saints) has been a more influential factor. But more serious problems in the linkage between Scandinavian Voting Minnesota MapScandinavian heritage and Democratic-Party-voting are also encountered, First, even in Minnesota the percentage of the population that claims Scandinavian background is not all that high. Second, some of the most Democratic-Party-voting counties in Minnesota and neighboring states do not seem to be heavily Scandinavian. To show this, I have outlined several mostly rural counties in western Minnesota and eastern North and South Dakota that voted for Barack Obama in 2012 and that generally swing left in national elections. Yet neither here nor in the generally Democratic-voting areas of southwestern Wisconsin, northeastern Iowa, and northwestern Illinois does one find—according to this map—Scandinavian concentrations.

All of this follows, however, only if this map of “Scandinavian-Americans” is reasonably accurate. But it may not be; so much intermarriage has occurred since the wave of northern European migration washed over the Midwest that such numbers are difficult to establish. Individual maps Nordic Americans Mapshowing the concentrations of Swedish-, Norwegian-, and Danish-Americans, derived from 1990 census data show substantially different patterns. (I have also included a map of Finnish-Americans; although Finland is not technically a Scandinavian country, it is a fellow Nordic country.) Here the Swedish map and especially the map showing Norwegian-Americans do show major settlement in many of the more left-voting rural counties of the upper Midwest. But they also show concentrations in more conservative areas, such as North Dakota, western Iowa, and Nebraska, the latter being an especially Republican-oriented place. Danish-Americans actually tend to be concentrated in conservative areas, although again much of this is linked to heavy Danish recruitment by the Mormon Church. Northeastern Minnesota—the most reliably Democratic-voting White, non-metropolitan part of the country—is in contrast heavily Swedish and especially Finnish. But the same ethnic mix is found in the Upper Peninsula of Michigan, which has many habitually Republican-voting counties.

German Americans MapGerman ancestry is much more widespread in the United States than Scandinavian ancestry, as can be seen by comparing the key to the map posted to the left with the keys of the maps posted above. German settlement was pronounced across most areas of the Midwest, but there seem to be few if any correlations between the strength of German settlement and modern voting behavior.

Dutch Americans mapProbably the best example of a linkage between contemporary voting and national background among White Americans pertains to people of Dutch ancestry. Several counties that were heavily settled by the Dutch are still noted for the their strongly conservative political orientations. This is true both in northwestern Iowa and in Western Michigan, two areas that stand out on the map of Dutch ancestry. The more specific factor here, however, is again religion, as these areas are also bulwarks of Dutch Reformed churches, which have in general retained their conservative orientations much more in parts of the United States than in the Netherlands. As the Wikipedia describes Ottawa County, Michigan (the largest city in which is Holland):

The Christian Reformed Church in North America was the biggest Christian denomination in the county with 67 churches and 33, 700 members, followed by the Reformed Church in America with 37 congregations and 33,300 members. …

Ottawa County is a stronghold of the Republican Party. The last Democratic Party candidate to carry the county was George B. McClellan in 1864.

1964 US Presidential Election MapBoth Ottawa County and a cluster of counties in northwestern Iowa clearly stand out on the 1964 U.S. presidential election map. In this pivotal contest, these counties supported the staunchly conservative candidate Barry Goldwater, unlike most other parts of Iowa and Michigan. What I find especially interesting in this map, however, is the fact that a number of counties in northeastern Illinois—which is now a generally Democratic-voting area—also supported Goldwater.

Such patterns obviously call for more research, but I must now move on to other issues.