It is widely accepted in recent work in economics and political science that ethnic diversity has a negative impact on the provision of public goods such as health and education. Indeed, the conventional wisdom holds that a negative relationship is so well-established empirically that research should no longer examine whether such a relationship exists but focus instead on why it exists (Habyarimana et al. 2007). This hypothesis further has important implications for other major outcomes, including economic growth and development. Indeed, Banerjee et al. (2005) describe the notion that social divisions undermine economic progress, not just in extremis, as in the case of a civil war, but also in more normal times' as 'one of the most powerful hypotheses in political economy'.
This conventional wisdom highlights the costs of diversity and could be interpreted as providing support for polices that minimize it. Our recent article in World Development (2016), 'Ethnic Heterogeneity and Public Goods Provision in Zambia: Evidence of Subnational "Diversity Dividend' (http://www.sciencedirect.eom/sdence/article/pii/S0305750X15002405) by contrast, lends support to the view that diversity can be beneficial--not only for normative reasons, but also because, under some conditions, it can also support concrete welfare gains.
Diversity debit or diversity dividend?
We challenge the conventional wisdom on ethnic diversity on empirical grounds. We show at the sub-national level strong evidence for a positive relationship between ethnic heterogeneity and some measures of public goods provision, in particular welfare outcomes related to health and education. In other words we find a 'diversity dividend' rather than the 'diversity debit' emphasized in most research in this area.
The article draws in particular on analysis we conducted at district level for Zambia, using a new dataset compiled from administrative, budget, and survey data. These data cover a broader range of public goods outcomes than previous work, including on both budgetary outcomes (such as per capita expenditures on health) and welfare outcomes (such as school enrolment and infant mortality). We consider separate models for district-level budgetary and welfare outcomes, exploring why relationships may differ for each. In our statistical analysis, we consider a variety of alternative specifications and models.
We find a negative relationship between ethnic diversity and central government...