Share

NBER Working Paper 16196, July. Rachel Glennerster, Edward Miguel, and Alexander Rothenberg (2010).

This study finds that ethnic diversity does not undermine public goods provision in Sierra Leone.

In some contexts, ethnic diversity seems to obstruct economic and political development, empirically speaking.  Studies have found that in diverse countries, economic growth is slower and fewer public goods are provided.  Development economists theorize that diverse societies are more loosely organized and therefore lack the monitoring and enforcement capabilities necessary to prevent free riders from taking advantage of public goods.  This debate is especially important in sub-Saharan Africa, one of the poorest and most ethno-linguistically diverse areas in the world.

Glennerster, Miguel, and Rothenberg investigate this problem in Sierra Leone, with attention to that country’s ethnic diversity, collective action, public goods provision, and social capital outcomes.  Sierra Leone is an extremely poor country, with a purchasing power parity per capita GDP of $806, life expectancy of 41.8 years, and adult literacy rate of 34.8%.  It was mired in a civil war from 1991 to 2002 and ranked last in the world on the Human Development Index in 2005.  There are 18 major ethnic groups in Sierra Leone.  The authors reason that given the post-civil war context, “Ethnic appeals and divides are salient in national politics in Sierra Leone, making it a reasonable setting to test the thesis that ethnic divisions stifle local public service delivery and economic development.” (1).

Of course, there is a significant endogeneity problem in what economists call the “residential sorting of individuals.”  Individuals of the same ethnic group likely move to locations where they can live together, leading to more solidarity, social monitoring, and public goods provision.  This effect is likely especially strong in a post-war context in which many individuals fled violence.  One of the main contributions of this paper is a new econometric technique to estimate the relationship between ethnic diversity and collective action.  The authors employ ethnic composition data from the Sierra Leone Population Census of 1963 as instrumental variables for present day diversity.  Their analysis suggests that in 1963, residential sorting was not yet correlated with public goods provision, making this measure a valid instrument.  As discussed in a previous brief, this type of technique greatly reduces the problem of endogeneity and represents “a methodological advance over most of the empirical ethnic diversity literature” (1). 

Ethnic residential data is drawn from the 2007 National Public Services household survey.  Collective action and social capital data are drawn from the 2005 and 2007 National Public Services Surveys.  These measures include, for example: road maintenance; school construction; community meeting attendance; membership in women’s associations, youth groups, religious groups, credit groups, etc.; control of community disputes such as theft or physical attack; and trust of community members, chiefs, outsiders, and members of parliament.  The authors also surveyed all Sierra Leonean paramount chiefs in 2008 to assess their political strength and ability to limit free riding in public goods provision.

The authors’ findings are quite remarkable.  Once the tendency for co-ethnic habitation has been controlled for, “local ethnic diversity is not associated with worse local public goods or collective action outcomes in Sierra Leone” (2).  This result was robust to many specifications, controls, techniques, and measures of collective action, including road maintenance, group membership, trust, and school staffing and funding.  Local chief strength did not affect public goods.  The authors report that they “with high levels of confidence can rule out that ethnic diversity has even a moderate adverse impact on local outcomes” (2).

These results provide support for suggestions that “despite the leading role of ethnic appeals in national politics, ethnic divisions have been much less damaging in Sierra Leone than in many of its African neighbors, and in particular were not a leading factor in the recent 1991-2002 civil war.”  The authors note that the Revolutionary United Front rebels “targeted people from all ethnic groups, and statistical analysis of documented human rights violations shows that no ethnic group was disproportionately victimized. There is also no evidence that civilian abuse was worse when armed factions and communities belonged to different ethnic groups… Ethnic grievances were not rallying cries during the war and all major fighting sides were explicitly multi-ethnic” (2-3).

After presenting their rather remarkable results, the authors explore the question of what historical and institutional phenomena might have produced Sierra Leone’s ethnic amity compared to other negative outcomes in the region.  Why is diversity not a contentious issue in Sierra Leone compared to much of Africa?  The authors point to the historical legacy of strong Paramount Chiefs empowered by the British and who continue to hold great power in communities to this day.  Strong local leaders “help overcome the classic free-rider problem in local public goods provision,” they report.  Additionally, though Sierra Leonean politics mobilized along Krio versus non-Krio lines in the 18th and 19th centuries, post-independence politics may have swung away from clashes between the two main ethnic groups, the Mende and Temne, because dictator Siaka Stevens belonged to neither of these groups.  Finally, the common national language of Krio may have contributed to a single national identity (although this was not the case in Rwanda and Somalia).

They suggest that unlike ethnic politics in Tanzania and Zambia, “a leading explanation for Sierra Leone’s relatively good inter-ethnic cooperation is the presence of strong traditional local authorities that help overcome the classic free-rider problem in local public goods provision” (3).  Specifically, “In contrast to Tanzania, the high level of interethnic cooperation in Sierra Leone is not the result of a “modernizing” approach that dismantled chiefdom authorities and replaced them with elected local institutions. Unlike in Zambia, successful local collective action across diverse ethnic groups is maintained in Sierra Leone even when the groups are rivals for political power on the national stage” (33).

One caveat is that this paper does not suggest that ethnic divisions are totally unimportant in Sierra Leone.  Statistical analysis of post-war residential choices show that Sierra Leoneans do prefer to reside with co-ethnic individuals.  Some studies show modest co-ethnic voting patterns.  Yet overall, this paper provides a positive view of diversity and collective action.  The authors conclude that “While it is difficult – and potentially unwise – to draw general conclusions about how to achieve inter-ethnic cooperation in a continent as diverse as Africa, Sierra Leone provides evidence that ethnic differences can be highly salient in some aspects of life and yet not undermine local public goods provision, an encouraging message for other diverse societies. It also provides a stark counterexample to the view that underdevelopment in Africa is inextricably connected to tribal conflict” (33).