Mr.

• In Event: The Diversity of Public Administration

Abstract

This paper analyzes how social interactions, ethnic geography, and other factors, affect the deployment and tenure of regional executives in Kenya under two presidents- Moi and Kibaki. We leverage administrative dataset on district commissioners (DCs) compiled by Carter & Hassan (2020) in a linear-in-means regression discontinuity design (RD), and Cox proportional hazard regressions. We investigate RD stability using relative treatment effects derivative. Social interactions is measured by the number of work stations a DC in a particular district has served relative to a countrywide average. Where such stations are fewer, we infer stronger social networks. Ethnic geography is proxied by a dummy for misaligned district. Misaligned districts constitute non-core political bases. We establish negligible discontinuity in the probability that a DC is non-native around the social interactions’ cutoff for Moi. Under Kibaki, this probability changes from about 88% to 96% around the cutoff. Non-native officers were likelier to govern in non-core, or smaller, districts during Moi’s era; this bias is non-existent for Kibaki. Under both regimes, long-serving DCs were co-ethnics to majority residents in the districts they governed. In terms of tenure, officers were rapidly shuffled in non-core districts. As social interactions rose (or networks weakened), Kibaki rapidly shuffled officers. Raising social interactions above the threshold lengthened officer’s tenure. We refrain from drawing suggestions for policy action.

Author

• Kefa Maunda Simiyu, University of Nairobi