But a broad new study co-authored by MIT political scientist Daniel Posner shows that the picture is more complicated than this: There is some favoritism in African politics, but it is inconsistent and limited in nature.
The study looks at a range of goods distributed in six African countries since the 1960s to see if political patronage networks do form along ethnic lines, with regard to things such as education, natural resources and health benefits. The answer is that it varies in extent, from country to country, over time, and depending on which kinds of benefits are at stake.
“Distributive politics is about the sum total of favoritism meted out by the government vis-à-vis various constituencies,” says Posner, a professor of political science at MIT. “But what we find is that the pattern of favoritism that has been so painstakingly identified for one outcome often does not hold on other outcomes.”
Five public goods
Posner and his co-author, political scientist Eric Kramon of UCLA, lay out their results in a new paper based on the research, just published in the June issue of the journal Perspectives on Politics.
The researchers looked at five public goods for which government action has significant responsibility: Medical care to prevent infant mortality, childhood vaccinations, household electrification, access to water and educational attainment. They used data from the Demographic and Health Surveys — a global survey of developing countries, chiefly funded by the U.S. Agency for International Development — that gathers statistics at five-year intervals.
In Kenya, for instance, there is no evidence of tribal favoritism as it pertains to the provision of household electricity, vaccinations or infant mortality — but there is significant ethnic favoritism when it comes to provision of primary education and access to clean water. Being from the same ethnic group as the president, for instance, leads to a 5 percent increase in primary school completion rates.
In Zambia, ethnic favoritism has no impact on education, vaccinations or water issues, but was associated with a 5 percent increase in household electrification. And in Malawi, favoritism seems to run more strongly across the board: People from the president’s ethnic group are about 4 percent more likely to have completed primary school, have a 1 percent lower infant mortality rate, and receive more vaccinations.
The researchers, as Posner describes their strategy, “compiled information about lots of different outcomes that were all plausibly the products of government decision-making. And we look at groups that did have a co-ethnic [leader] in power at some times, but not at other times, and then compare their fortunes when they did and did not have a co-ethnic president.” In Kenya, for example, which has had three presidential regimes since 1963, the study tracks the shift in the distribution of goods relative to the changing ethnicities of its presidents.
Other scholars in the field have praised the study’s contributions. “This paper provides an important reminder that the inferences we draw about why governments do what they do depend heavily on the quality of data available for study,” says Brian Min, a political scientist at the University of Michigan. “What’s exciting is that novel technologies and data collection strategies are giving citizens and scholars a greater ability to monitor and study the investments made by their governments than ever before.”
Why only some goods?
Posner describes the findings as a “cautionary tale” for researchers who might be inclined to draw blanket conclusions about favoritism based on the distribution of a single good. Instead, he suggests, empirical research in this area needs to draw out the specific reasons why some governmental actions are linked to ethnic favoritism, while others are not.
For instance, Posner notes, there could be geographic and practical factors limiting the distribution of goods along ethnic lines: Citizens from a political leader’s ethnic group might live in places where it is difficult to extend electrical power or supply more water, but might be relatively easy to provide them with more schools or medical clinics.
“Or certain goods might be controlled by ministries that are more susceptible to political manipulation,” Posner says. “It might also be that certain groups have greater needs than others” at particular times.
And, of course, it could be that some leaders have greater or lesser senses of distributive justice pertaining to all ethnic groups, not just their own.
“With respect to why you see favoritism for some goods and not others, the answer is probably going to be specific to each case,” Posner says. “But you can’t even begin to ask that question unless you recognize that these outcomes can vary across goods. In this respect we see the paper as an agenda-setting piece.”