Behind every functioning state is a bureaucracy of people deciding, daily, whether to do their jobs well, honestly, and at all. The previous courses analysed the bureau as an organisation; this module zooms into the official as an employee, applying the economics of personnel — selection, incentives, and monitoring — to the public sector. It is where state capacity is actually built or lost.
Weber's ideal type
The Weberian bureaucracy
Max Weber's ideal-type bureaucracy is defined by: recruitment and promotion by merit and qualification (not patronage); a clear hierarchy of offices; decisions governed by written, general rules rather than personal whim; impersonality (officials serve the office, not friends and kin); and full-time, salaried careers with predictable advancement. The point of these features is not bureaucratic rigidity for its own sake; it is to make administration competent, predictable, and resistant to capture — to create officials who serve the rules rather than themselves or their patrons. Evans-Rauch (last module) showed these features predict growth.
The personnel economics of the state
Finan, Olken, and Pande (2017) organise the modern evidence around three classic personnel problems applied to the state — and show the public sector has features that make each harder than in a firm:
- Selection — who applies and who is hired. The state is a huge employer whose hiring rule shapes the talent and motivation of the entire administration.
- Incentives — how effort is rewarded. Public outputs are often multidimensional and hard to measure, so high-powered pay-for-performance is risky.
- Monitoring — whether officials actually show up and do the work. Absenteeism among teachers and health workers is a first-order problem in many low-capacity states.
Selection: do higher wages attract the wrong people?
A worry about raising civil-service wages is adverse selection on motivation: pay more and you attract money-motivated applicants, crowding out the public-spirited. Dal Bó, Finan, and Rossi (2013) tested this directly in a Mexican government recruitment programme that randomised the offered wage across regions.
Higher wages attracted abler AND more pro-social applicants
The randomised higher wages drew applicants who were more able (higher IQ, more qualified, more experience) and no less — if anything more — pro-social and public-minded. There was no trade-off between competence and motivation: paying more widened and improved the pool on both dimensions. The result rebuts the 'high pay corrupts the mission' worry and supports competitive public-sector wages as a selection tool — provided recruitment is genuinely meritocratic so the better pool is actually hired, not screened by patronage.
Incentives and the multitasking trap
Performance pay is the obvious lever, and it sometimes works — Muralidharan and Sundararaman (2011) found teacher performance pay in India improved learning. But the public sector is the textbook home of the multitasking problem (Holmström-Milgrom): when a job has many dimensions and you reward only the measurable ones, effort flows to what is measured and away from what is not. Pay teachers for test scores and you may get teaching to the test, narrowed curricula, or outright cheating; pay tax officers for revenue and you may get over-aggressive assessments against the easy-to-tax. The lesson is not 'never use incentives' but 'incentives are powerful and dangerous where output is multidimensional and gameable' — use them where the measure is hard to game and aligned with the true objective, and rely more on selection and monitoring elsewhere.
Monitoring: getting people to show up
In many states the binding problem is simpler than incentives — officials are absent. Studies have documented teacher and health-worker absence rates of 25-40% in parts of South Asia and Africa. Monitoring interventions (cameras with attendance-linked pay for teachers; the broader absence literature) can work, but they run straight into the multitasking and political-economy problems: monitored staff resist, unions push back, and monitoring one dimension (attendance) doesn't ensure effort while present. Technology that makes attendance and delivery automatically visible — biometric attendance, GPS-tracked deliveries — is the current frontier, and links to the anti-corruption-technology evidence in the next module.
Exercise
A health ministry runs clinics plagued by three problems: low-skill staff, high absenteeism, and weak effort when present. A minister proposes a single fix — large performance bonuses tied to the number of patients treated. (1) Using selection, incentives, and monitoring, explain why this single instrument is unlikely to fix all three problems and could worsen some. (2) Apply the multitasking problem specifically to the 'patients treated' metric. (3) Use Dal Bó-Finan-Rossi to argue for a selection-side reform and address the 'higher pay will attract mercenaries' objection. (4) Propose a balanced reform package across the three margins and justify the mix.