The dominant narrative of the post-colonial African state is that newly installed leaders adopted "neo-patrimonial" hiring practices, using state positions to consolidate their power. Given the salience of ethnicity, and rulers' ties to certain ethnic groups, these practices often resulted in ethnic bias. Further, ethnic imbalances in access to state resources fuelled political instability. A salient assumption within this narrative is that under colonial rule, civil service hiring was meritocratic, and that any ethnic favoritism was a result of groups' differential exposure to education. We digitize fine-grained, bureaucrat-level data from the entire period of Britain's colonial rule of the Gold Coast (later Ghana) and show that this assumption is incorrect. Using contemporary census and voter registration data to uniquely code each colonial bureaucrats' likely ethnicity, we demonstrate significant ethnic bias in hiring throughout Ghana's colonial period. We argue that this bias resulted from patrimonial practices by groups that gained early entry to state positions. Using exam results data from potential civil servants from the colonial period, we further demonstrate that hiring was not meritocratic. Our results imply the need to re-evaluate claims about post-colonial politics, suggesting that many leaders did not inherit an ethnically-balanced and meritocratic civil service upon which to build. Instead, the seeds of state capture had already been sown.