S7 Ep48: How politicians co-opt bureaucrats

How can we cut corruption in government, procurement that generates kickbacks for local politicians? We could hire officials on merit. Across the developing world, many governments now do exactly that. So why do public contracts still go to political friends?

In this week's VoxDev Talk, Sarah Brierley (LSE) tells Tim Phillips why merit-based hiring is not enough. If politicians cannot choose who gets the job, they can often control what happens next: where officials work, and whether they are promoted. This power may be sufficient to co-opt them. In Ghana, where she did much of her research, nearly half of senior local officials say contracts in their district go to firms that fund the ruling party.

The research behind this episode:

Brierley, Sarah. 2026. The Co-opted State: How Politicians' Control Over Bureaucrats' Careers Threatens Governance. Cambridge: Cambridge University Press. Open access.

To cite this episode:

Phillips, Tim, and Sarah Brierley. 2026. "How politicians co-opt bureaucrats.” VoxDev Talks (podcast).

About the guest

Sarah Brierley is Associate Professor of Comparative Politics in the Department of Government at the London School of Economics and Political Science. Her research spans the state, corruption, bureaucracy, and election campaigns, with a focus on sub-Saharan Africa. She holds a UKRI Future Leaders Fellowship for a project on political finance in Africa.

Research cited in this episode

Clientelism. Politicians hand out public resources, such as schools, jobs or contracts, in return for votes. Most research assumes that officials simply do what politicians ask. Sarah argues that officials have real control over who gets what, so politicians must first win them over.

Hiring on merit. Sarah checks whether senior local officials in Ghana are hired on merit in three ways: expert surveys, a survey of the officials themselves, and staff records. Of those hired between 2010 and 2016, 67% sat an exam. When the governing party changed after the 2008 election, the kind of people hired into professional jobs did not change. The kind of people hired into low-skilled jobs did.

Randomised response technique. A way to ask survey questions that people would rather not answer. Each person rolls a die where the interviewer cannot see it. On one number they must say "yes", on another "no", and otherwise they tell the truth. Nobody knows what any one person meant, but the odds of each roll are known, so the true share of "yes" answers can be ca