Benn Miguel Posner 2010

From Projecting Power

"Political Competition and Ethnic Identification in Africa" by Benn Eifert, Edward Miguel, and Daniel N. Posner[edit]

Introduction and Research Findings[edit]

The significance of ethnic identities in Africa, especially in social and political contexts, is an ongoing source of debate. One view suggests that ethnic identities are innate and deeply-rooted; "intrinsically part of who people are." Another view holds that ethnic identity is functional, serving as a tool for mobilization and coalition building in the competitive struggle for resources and power. Eifert, Miguel, and Posner analyzed survey data which supported the latter view of the salience of ethnic identity.

The researchers hypothesize that "if ethnic identities are tools that people use to get access to political power, then they are likely to be rendered most salient when political power is at stake— that is, at election time." Second, the researchers hypothesize that the role of ethnic identity will become more important when elections are competitive. To test these assertions, they used survey data on social identity from 22 rounds across ten African nations. They reached the following conclusions from the survey data: 1)In African countries, the strength of ethnic identification changes significantly over time. 2)The changes are related to how close the survey is conducted relative to a presidential election. This effect is also affected by the competitiveness of the election. 3)When the presidential election is very competitive, the chance that the survey responder will identify themselves by ethnic terms increases by 1.8 percent points with every month that it is closer to the election.

The researchers suggest that two mechanisms might account for these findings although they emphasize that their findings do not lend themselves to a certain explanatory mechanism. The first is that people could be mobilized by politicians who "play the ethnic card." The second is that voter's recognize the tendency to distribute resources based off of ethnic lines.

The researchers also wanted to assess which identities are displaced when ethnic identity rises in importance. They looked at four other categories of social identity: ethnicity, class, religion, and gender. The researchers found that increasing the importance of ethnic identity in proximity to an election decreases the importance of class identity.The researchers also wanted to understand what types of people are more likely to identify in ethnic terms. They found strong evidence to support that people who work in modern sectors of the economy are more likely to identify in ethnic terms than people in the traditional sector of the economy.

The researchers also made novel methodological contributions through using a multinomial logit empirical methodology that permits inference about the factors associated with the salience of multiple dimensions of social identity, and through using repeated country-level observations.

Data and Methodology[edit]

The researchers used data collected in the Afrobarometer, which is " a multicountry survey project that employs standardized questionnaires to probe citizens’ attitudes in new African democracies." The researchers note that the data may be biased as it is self-reported but explain that the confidential nature of the survey lessens concerns. The researchers also question the generalization of the findings and caution that the findings might not represent Africa as a whole.

The Salience of Ethnic (and Other) Identities[edit]

The researchers' findings support the notion that ethnic identity is context dependent. The researchers also voice a concern that sampling variation accounts for the changes but the Afrobarometer uses a large, nationally representative sample. They also question whether the surveys are intentionally conducted before competitive elections but find little evidence to suggest a relationship between survey timing and electoral cycles.

The Political Sources of Ethnic Identification[edit]

Competitiveness of a presidential election is a possible explanation for variations in the tendency of survey respondents to identify in ethnic terms. The authors provide the possible hypothesis that "the strength of ethnic identity is dependent on the competitiveness of elections" and have provided the followings analyses:

Initial evaluation of the data collected from the study reveals that neither proximity nor competitiveness has an impact on ethnic identifications
The aforementioned analysis is not reflective of the data, which revealed that variation within countries was generally limited, and even completely identical in four out of the ten countries studies.
This prompted Eifert et. al to redirect their focus to emphasizing the interaction term between the proximity and competitiveness variables.
In doing so, data observations reveal that coefficients regarding the proximity and the interaction term are statistically significant.

Political Competition and Other Social Identities[edit]

The authors evaluated the observations of main dependent variable, which were answers to the question "With which group do you feel you belong to first and foremost?". To depict this process Eifret et. al have elected to utilized a modified multinomial discrete choice empirical framework, that was used in prior in the study. This resulted in one of the more crucial findings that the probability of respondents who identify themselves through class/occupation or gender decreases as more competitive elections draw near. This is significant because it additionally supports the finding that identification through ethnicity increases the more competitive an upcoming election is.

Beyond Aggregate Effects[edit]

Eifert et. al have examined four alternative explanations for why people identify based on ethnicity in Africa:

1. "Modernization will disconnect individuals from traditional concepts, such as viewing themselves through ethnic lens, in favor for class/occupational identities."
-This is disproven because it would suggest that more traditional communities would have stronger ethnic attachments, which is not reflected in data.Also, additional researchers have provided the insight that urbanization has a higher chance of deepening ethnic ties, as economic, political, and social advancement tend to exploit these group memberships.
2. "Ethnic groups that are large to form significant political coalitions consist of more individuals who are willing to identify through that ethnic group."
Alternative Hypothesis formulated by authors: "Smaller ethnic groups are more likely to identify their membership due to unwillingness to being dominated by larger group"
-Both of these hypotheses are disproven based on data, which reveals that size not only does not impact willingness of identification to a certain ethnic group, but it also does not impact any other category of social identity.
3."Supporters of current political groups in power are not as likely to be connected to their ethnic identities, while supporters of oppositional political groups are more likely to identify through ethnic ties due to political partisanship."
-This hypothesis is disproven by data collected through survey study, and similar to previous hypotheses, other categories of social identity are not impacted by stance on current political groups in power.
4. "Media portrayals of African conflicts consisting of armies of unemployed young men may make young African men more prone to identify with their ethnic groups."
-The data reveals that young men are not more likely to identify through ethnicity than any other gender/age group in the study.

When considering all of this, it is important to to note that the data of this study consists of just 10 countries, limiting the generation of meaningful results for triple interactions.

Discussion and Conclusion[edit]

Ultimately, findings reveals that ethnicity plays a significant role in terms of political power. Possible explanations for specific role of ethnicity, such as for whom it impacts politically consist of:

1. Political Elites: Politicians are better rewarded when they invoke peoples’ ethnicity to mobilize support and retain and receive greater political power. Efforts for ethnic mobilization tends to be more prioritized leading up to elections.
2. Citizens: Since elections reflect a time where voting citizens are in control of determining who gets to govern resources, they are more mindful of potential candidates who share similar identities, specifically ethnicity-based.

Due to limitations of study, there are no study-based conclusions, but an educated conclusions provided by the authors is that these two theories are complementary, not competitive.