A standard governance prescription is to move power closer to the people — decentralise. The theory is appealing and the African trend is strong, with Kenya's 2010 devolution among the most ambitious. But decentralisation is double-edged: it can improve accountability or simply relocate corruption to a level where local elites capture it. This module gives you both blades.
The case for decentralisation
- The decentralisation theorem (Oates, 1972) — local governments can tailor public-goods provision to local preferences, which differ across places, so decentralised provision beats uniform central provision where preferences vary and there are no strong spillovers or scale economies.
- Tiebout sorting (1956) — where people are mobile, they 'vote with their feet', choosing the jurisdiction whose tax-and-service bundle suits them; competition between jurisdictions then disciplines local governments (limited in poor settings where mobility is low).
- Yardstick competition (Besley-Case, 1995) — voters compare their local government's performance and taxes against neighbouring jurisdictions, giving incumbents an incentive to perform; the neighbour is a benchmark that pierces the information problem.
- Proximity and information — local citizens can observe and monitor local officials more easily than distant central ones, tightening accountability (the bottom-up-monitoring logic of the last module).
The case against: elite capture
Bardhan-Mookherjee: decentralisation can worsen capture
Pranab Bardhan and Dilip Mookherjee showed formally that decentralisation's effect on accountability is ambiguous because local governments may be more captured by local elites than central ones. Local political competition is often weaker, local media and civil society thinner, the local rich relatively more dominant, and local checks (audit, courts) less developed. Where these conditions hold, devolving power and money hands them to a local elite who capture them more completely than a (still-imperfect) centre would — so decentralisation relocates and can even intensify corruption rather than reducing it. Proximity cuts both ways: citizens are closer to the official, but so are the local powerful.
The evidence: genuinely mixed
Cross-country and within-country studies do not deliver a clean verdict. Some find decentralisation associated with lower corruption (Fisman-Gatti, 2002, across countries); others find federations and decentralised systems can be more corrupt (Treisman, 2007), partly because more layers mean more independent bribe-takers (the Shleifer-Vishny commons problem from module 3 applied to tiers of government). The reconciliation is that decentralisation's effect depends on conditions: it helps accountability where local political competition, local information/media, and local oversight institutions are strong; it relocates corruption to captured local elites where they are weak. Decentralisation is not good or bad for governance in the abstract — it is a structure whose effect is conditional on local accountability institutions.
Kenya's devolution
Kenya's 2010 Constitution created 47 county governments with substantial functions and a constitutionally-guaranteed share of national revenue (the equitable share). It is a natural experiment in the trade-offs of this module. The accountability gains are real: services and resources reach long-marginalised regions that the centre neglected, local priorities are better reflected, and there are now 47 sites of political competition rather than one. But the capture risks materialised too: county-level corruption (inflated tenders, ghost workers, irregular spending flagged repeatedly by the Auditor-General), thin county-level oversight and audit capacity, and local elites well-placed to capture devolved funds. Devolution did not abolish corruption; consistent with Bardhan-Mookherjee, it relocated much of it to the counties and created new sites of it — while genuinely improving the geographic reach of the state. Both things are true at once, which is exactly what the theory predicts.
Exercise
A country is debating whether to devolve health and infrastructure spending to 30 new regional governments, hoping to reduce the corruption seen at the centre. (1) State the strongest theoretical arguments that devolution will improve accountability. (2) State the elite-capture counter-argument and the conditions under which it dominates. (3) Using Shleifer-Vishny from module 3, explain a mechanism by which adding a layer of government could increase total corruption. (4) Recommend three complementary institutions that should accompany devolution to tilt the outcome toward the good case, drawing on Kenya's experience.