Center for Global Development, Working Paper 208, April Jenny C. Aker, Michael W. Klein, Stephen A. O'Connell, and Muzhe Yang (2010)

In a study of cross-border trade between Niger and Nigeria and domestic trade within Niger, Aker et al. find that ethnicity is a significant determinant of trade flows. 

Many scholars have observed the existence of trade barriers between countries.  Yet what is the root cause of such barriers?  Aker et al. examine two African countries and find some interesting and unexpected results.  Trade barriers in this case are not simply a product of tariff law and transportation costs but also ethnic groups.  In the second half of the article, Aker et al. go on to investigate if ethnic groups are important in determining domestic trade patterns as well, with some interesting findings.

The authors evaluate cross-border market integration between Niger and Nigeria with monthly price data on freely traded millet and cowpeas over 1999-2007.  These goods are chosen because they are homogenous (that is, local millet is not superior to foreign millet) and frequently traded across the border. 

Methodologically, prices in 42 markets on each side of the border are assessed. 

The data is from Niger’s Agricultural Market Information System.  The authors control for road distances, transportation costs, cellular phone coverage, rainfall, and trader, farmer, and transporter ethnicity.  They note that the border effect they identify “cannot be attributed to official trade restrictions” because both countries are members of the Economic Community of West African States, a customs union with generally free trade in food staples (6).  Aker et al. note that the boundary between the countries is not geographic, but “created by the United Kingdom and France in the wake of the 1884-85 Berlin Conference, was drawn through a region within which the Hausa ethnic group had historically supported an active trade in livestock, textiles, salt, kola nuts, and food staples” (2).  The 1906 border, 1500 kilometer long, artificially split the Hausa, Fulani, and Kanuri people into the two countries.  A panel survey conducted by Aker provides data on ethnic dispersion across the markets in the study.  Paired regression as well as regression discontinuity design is used. 

Regarding Niger and Nigeria, Aker et al. find that “a common ethnicity can diminish the effect of an international border on price differences between countries” (18). 

The authors then examine Niger alone to determine if ethnic links affect trade patterns within countries as well as internationally.  Looking at trade between Hausa and Zarma areas and controlling for a range of other explanations, they find that “The difference in ethnicity seems to create a substantial transaction cost (or a barrier to trade) at the border” (20).  Millet prices increased 26 percent upon crossing the internal ethnic border, and cowpea prices 22 percent.  The authors conclude that “The internal border effect and the international border effect are similar in magnitude. The barriers to market integration between the Hausa and Zarma regions of Niger therefore appear to be at least as great as those imposed by the international border with Nigeria. The deadweight losses in foregone internal trade may correspondingly be of a similar order of magnitude” (21).

Aker et al. suggest a range of potential explanations for the internal ethnic trade barrier.  These include: differences in public services, taxes, and trade associations between Hausa and Zarma regions; market segmentation; linguistic differences; gender; and the role of social networks in credit access. 

After evaluating these potential explanations against the data, the authors conclude that the last two are most important.  Firstly, 30 percent of Zarma traders are female but only 5 percent of Hausa traders are.  Male Hausa refusal to barter with female traders could segment the ethnic markets, leading to price differences in each.

Secondly, social networks may affect the propensity to lend across ethnic groups. 

This is particularly important because traders in Niger borrow frequently from each other.  The authors write that “In the context of imperfect credit markets, where firms and traders cannot assess the unobservable characteristics of traders and clients, social networks can play an important role in circulating information about credit histories and other attributes of relevance to potential trade partners.

Additionally, membership in the same ethnic group may offer a means of recourse for nonpayment that does not exist for transactions between members of different ethnic groups.”  For this reason, “The ethnic border within Niger may therefore reflect, at least in part, the prevalence of credit market imperfections and the resulting reliance upon borrowing and lending within ethnic groups” (25).

Overall, this paper shows that border effects exist in trade between two African countries, that ethnic identity is an important mediator of this effect, and that it is a mediator not only of the cross-border trade one would expect but also trade between ethnic groups within one country.