Segregation and the quality of government

Alberto Alesina, Ekaterina Zhuravskaya 15 September 2008



Racial, ethnic, and religious tensions are common around the world. Some states collapse under their weight; in less extreme cases, states’ functions are negatively affected. Economists have recognised the importance of this factor in explaining poor economic policymaking, low quality of government, and delays in development. For instance, a widely accepted explanation of Sub Saharan states’ failures is the linguistic and ethnic fragmentation of many countries of this region (Easterly and Levine 1997). In turn, the fragmentation is often a result of absurd border left by former colonisers.

The researchers who have documented the cost of ethnic fragmentation could not take into account the geographic distribution of different groups within a country. Indeed, the commonly used index of ethnic fragmentation depends only on the size of each group in a country and does not take into account whether groups are completely integrated or totally separated geographically. In reality, the degree of segregation makes a big difference and examining its consequences has important policy implications. For example, in the case of the urban economic literature on US cities, segregation, in addition to the racial composition of cities, is deemed extremely important and is widely studied. Lack of data had prevented comparable research in a cross-country setting.

In a recent paper (Alesina and Zhuravskaya 2008) we filled this gap by painfully collecting data on the ethnic and religious distribution of groups within countries at the regional (sub-national) level. We built an index of segregation for about 90 countries. The index varies dramatically across countries. Some highly fragmented countries are very segregated, meaning that regions within these countries are a lot more homogeneous than the country as a whole. For example, the national-level linguistic fragmentation index in Nigeria (0.4) is twice as large as the largest regional-level fragmentation index (0.2 in the South-West region). Other highly fragmented countries are not at all segregated. In these countries, ethnic and religious groups are uniformly distributed across the country. For instance, Bolivia’s ethnic national fragmentation index is equal to 0.7 and its regional fragmentation indices range from 0.6 to 0.7.

Segregation and the quality of government

Our first finding is that more segregated countries have much lower quality of government. To be precise, consider two hypothetical countries with the same distribution of ethnic groups; say three groups each comprising one third of the population. Suppose that in country A the three groups are totally separated geographically and in country B they are uniformly distributed across the national territory. Country B has a superior quality of government, measured by what are now standard cross-country governance indicators, i.e., the World Bank’s indices of the rule of law, control of corruption, regulatory quality, government effectiveness, and voice and accountability.

This correlation, of course, could be explained with causality going from the quality of government to segregation. Poor governance may lead people to choose to live with their own group, as it may be easier to self-organise within a homogeneous local community to provide basic public goods when government fails to provide them.

In order to study how segregation affects governance, we focus on the exogenous variation in segregation driven by the composition of major groups in bordering countries. More specifically, we construct an index of predicted segregation based on the idea that if the home country has a group that is also present in a neighbouring country, this group is likely to be concentrated near the border of the two countries. On the contrary, if the home country has a group not present in any of the neighbouring countries, the group is likely to be distributed uniformly.

Using variation in the predicted segregation index, we find that higher segregation in terms of ethnicity and language leads to lower quality of government. Moreover, the negative effect of ethnic and linguistic segregation is especially large in democracies.

Segregated politics

How can we explain this finding? First, in more segregated countries, ethnic politics is more prevalent in two senses. One is that groups that dislike each other cannot agree on a common set of rules and policies leading to a deterioration of the polity. The other is that a dominant group can more easily target the minorities when they are geographically segregated. Second, ethnic voting is likely to be more easily organised in segregated countries, leading to a selection of politicians more interested in pursuing ethnic-based policies than the common good. One can also find an optimistic message in these results: groups that live together rather than separately learn to respect each other and tolerate differences better.

Finally, a word on religion. We did not find any result on religion fragmentation or segregation, i.e. the latter seems uncorrelated with quality of government. Most likely the explanation is that religious affiliation can be imposed on people by repressive measures. In fact, the worst governments often restrict religious freedom or affiliation, so more religiously homogeneous countries may be such purely by brute force. Forced homogenisation of racial and ethnic characteristics is more difficult, though not impossible.


Alesina, Alberto F., and Ekaterina Zhuravskaya (2008) “Segregation and the Quality of Government in a Cross-section of Countries”, CEPR Discussion Paper No. 6943

Easterly, William, and Ross Levine, "Africa’s Growth Tragedy: Policies and Ethnic Divisions," Quarterly Journal of Economics, November 1997, 112:1203-125



Topics:  Politics and economics

Tags:  development, segregation, ethnic fractionalisation


Very good analysis and a thank you to the authors for researching this very important subject of fragmentation in a national sense. I noted that research (to my knowledge) in terms of understanding Lincoln's famous phrase "A house divided against itself cannot stand" is limited although the explanatory power of this phase is quite powerful. It makes sense, that if a nation (or any organization or large group) is deeply fragmented, then it will not be able to accomplish group-wide goals effectively.

I like the use of nuanced and theoretically justified measures of indexes of fragmentation, showing correlation with standards of living and other quality of life measures.

A couple thoughts: 1. An index of fragmentation can be differing goals among any number of groups, not just ethnic or religious, have the authors looked into groups based on socio-economic status, or by attitudes towards important political or economic issues (such as, for example, attitudes towards foreign policy)? Further, overall, fragmentation in an organizational setting, such as for a manufacturing firm (labor verses management, for example)

2. Further, the degree is disagreement on issues is seems important, so if one group "strongly believes" in, for example, free trade while another group "strongly believes" in restrictions on trade, the outcome of free trade is more likely to be contentious and perhaps not what either group wants. (direct contradiction of group goals verses agreement and/or non-contradiction of group goals, combined with strength of belief in these group goals)

3. Directly measuring "contention" -- Have the authors looked into Robert Putnam's (also at Harvard) work on Social Capital? It seems to also follow a "degree of fragmentation" although Putnam (I don't think) uses this term -- Putnam defines Social Capital broadly as "trust" declines with more groups (higher fragmentation) across a society.

4. Lastly (phew! but the research is very interesting) I wonder if there is any way to measure "mission creep" to more divergent goals among groups. So in year 1, group A and group B believe in free trade (example) but by year 5, group A does not believe in free trade while group B does. is there any theory regarding changing inter-group goals? (leading to higher Indexes of Fragmentation over time)

Nathaniel Ropes Professor of Political Economy, Harvard University; and Research Fellow, CEPR

Professor of Economics at the Paris School of Economics (EHESS) and the New Economic School in Moscow