A new UCLA-led study challenges the popular perception that ethnic diversity is to blame for sectarian conflicts in Iraq and Northern Ireland, recent tensions in Tibet, and ethnic violence in post-election Kenya.
"Countries that are ethnically diverse do not experience more conflict than their more homogenous counterparts," said Andreas Wimmer, the study's lead author and a UCLA professor of sociology. "Rather, conflict breaks out when large segments of the population are excluded from access to government because of their ethnicity."
In fact, a country that excludes 80 percent or more of its population on the basis of ethnicity is more than three times as likely to have a civil war as a wholly inclusive country, Wimmer and his team calculate.
The effects of ethnic exclusion proved to be as important as substantial differences in a country's per capita income, a well-recognized risk factor for civil war.
"If you want peace in countries with ethnic conflict, you have to rearrange government to include real power-sharing with all ethnic groups," said Lars-Erik Cederman, study co-author and a former assistant professor of political science at UCLA who now serves as a professor of international conflict research at the Swiss Federal Institute of Technology (ETH) in Zurich. "Eliminating barriers to political participation may sound like common sense, but this research for the first time pinpoints the dramatic risk of failing to do so."
The findings appear in the April issue of the American Sociological Review.
Wimmer and Cederman led a team of social scientists, including UCLA political science graduate student Brian Min, that spent close to three years building a dataset of ethnic power relations in 155 countries from 1946 to 2005, based on the expert advice of nearly 100 country specialists from universities across the world.
Area specialists were asked to identify the politically relevant ethnic groups within a country for each year since 1945 and then estimate the extent of each group's access to political power on a six-point scale, ranging from a total monopoly on power to being powerless and discriminated against. An ethnic group was considered to be excluded if its members were absent from the highest levels of regional and national government.
For a handful of countries — both very homogenous ones, such as Korea, and very heterogeneous ones, such as Tanzania — ethnicity was not politically relevant at all. In the rest of the countries, the risk of armed conflict rose in proportion to the degree of ethnic exclusion.
"The odds of having a war — rather than peace — increase by a factor of 1.12 for every additional 10 percent of the population excluded from central government power," Wimmer said.
Particularly at risk, according to the researchers, are so-called ethnocracies — countries in which an ethnic minority representing 20 percent or less of the population monopolizes power. Examples include Iraq under Saddam Hussein, Rwanda under the Tutsi-dominated government, Sudan since 1954, Syria in recent history, Liberia until the outbreak of civil war in 1980 and Rhodesia before the white minority government was overthrown.
In addition to various ethnic power configurations derived from this dataset, the study tested 10 other potential risk factors for political violence, including economic development, ethnic diversity, population size, geography, degree of democratization and even the presence of oil reserves — the last on the theory that rebels might be inspired to take up arms against the government when control of lucrative natural resources is at stake.
Per capita gross domestic product proved to be as important a risk factor as ethnic exclusion. In line with previous research, the study found that lowering national per capita annual income from $13,000 to $6,000 more than doubles the risk of armed conflict.
Other important risks include a large population size and a political system somewhere between dictatorship and democracy, sometimes called "anocracy." Examples of anocracies include present-day Pakistan, Sri Lanka, Nigeria and Armenia.
Of the 10 additional risk factors explored, the least significant proved to be political instability (as measured by whether a country had changed its level of democratization in the previous three years), a country's geography and its degree of ethnic diversity.
"It's not that people of different ethnic backgrounds can't get along because they have different cultures or creeds," Min said. "It's that political exclusion along ethnic lines stirs up trouble."
With a 98-percent exclusion rate, Liberia between 1946 and 1980 had the highest rate of ethnic exclusion, followed by Rhodesia (now Zimbabwe), with 97 percent between 1965 and 1979, and Nepal, with 95 percent between 1946 and 1950.
"All but one of these countries did, in fact, see serious armed conflict during these periods, and they are part of a pattern that holds across cases, continents and time periods," Wimmer said. "They are good examples of the causal forces that operate in many less extreme cases of exclusion as well. You need to have a good understanding of these general mechanisms if you want to design policy that can be adapted to specific cases."
With a current exclusion rate of about 85 percent, Syria, Sudan and Rwanda tied as the study's most exclusionary countries today.
Since Sudan's and Rwanda's ethnic conflicts have been in the news for years, their vulnerability did not surprise researchers, but Syria's situation defied their expectations.
"Syria hasn't had a conflict since 1982, so it looks very stable today, especially compared to its neighbors Turkey, Lebanon and Iraq, but our general findings suggest that there are tensions that could rise to the surface in the future," Wimmer said.
On the other hand, the researchers cautioned that their model did not capture all the risk factors equally well. Pakistan's relatively low exclusion rate of 11 percent would bode well for the country's future prospects for peace.
"You might see civil war in that country for other reasons, but they would be unrelated to ethnic exclusion," Cederman said.
As much as the findings point toward the benefits of eliminating political marginalization of ethnic groups, they show that inclusion alone is no panacea for peace. This became clear when the researchers looked at the number of ethnic groups with representation in a government. Increasing the number of ethnic elites that share power also increased the risk of violence, the study found.
"Overall, more inclusion still lowers the risk of conflict, despite the potential for violent infighting among power-sharing partners," Wimmer said.
Similarly, democratization is not necessarily a guarantee of peace. Moving from autocracy to anocracy, the study found, might well increase the likelihood of violent escalation of conflict, because protest movements are no longer as effectively and ruthlessly repressed as in a dictatorship.
"Once a democracy has become routinized, you can expect stability, but that can take years, if not decades," Wimmer said. "The process of democratization can stir up ethnic conflict."
The research was funded by the UCLA International Institute, the university's social science division and the Swiss National Science Foundation.