Working paper · September 2026 draft
Roving Tax Bureaucrats: Bureaucratic Quality, Tenure, and the Costs of Rotation
Renhao Ye and Nico Ravanilla
In brief
Governments rotate tax officials to prevent corruption, but rotation also cuts short the local learning that makes capable bureaucrats effective. Using Philippine tax administration records from 1999 to 2007 and election-period transfer bans that exogenously extend postings past the three-year limit, this paper shows bureaucratic quality and tenure are complements: extended tenure yields disproportionately larger revenue gains under higher-quality officers.
Abstract
Bureaucratic rotation can limit corruption by preventing officials from becoming embedded in the jurisdictions they administer. But rotation may also interrupt the local learning that makes capable bureaucrats become effective. We study this trade-off in the Philippine tax administration, combining assignment records with district fiscal data from 1999 to 2007. We construct a leave-one-out measure of bureaucratic quality from officers' performance in prior postings. Officer quality predicts revenue across assignments, but its returns emerge only with tenure and operate through voluntary remittances rather than enforcement-intensive collections. We then exploit election-period transfer bans that exogenously extend assignments beyond a three-year tenure limit. These extensions generate larger revenue gains under higher-quality officers. The results show that bureaucratic quality and tenure are complements in the production of fiscal capacity. Personnel rules designed to constrain bureaucratic discretion can weaken state capacity when they ignore how differences in bureaucratic quality interact with time in post.
Research annotationQuestion, variables, design, findings, and mechanism at a glance▾
- Research question
- Does limiting bureaucratic tenure through rotation reduce the returns to bureaucratic quality?
- Independent variable (X)
- Two treatments. First, a predetermined leave-one-out measure of a revenue district officer's quality, built from performance residuals in prior postings only (earlier years, other districts), in the style of teacher value-added. Second, exogenously extended tenure: an officer 'overstaying' the three-year statutory limit, instrumented by exposure to election-period transfer bans, interacted with quality.
- Dependent variable (Y)
- Log district tax revenue, decomposed into voluntary remittances, assessment-based collections, and delinquent collections, plus revenue interacted with local dynastic entrenchment.
- Identification strategy
- Bureau of Internal Revenue assignment records matched to district fiscal data, 1999 to 2007, with district and year fixed effects, time-varying controls (potential revenue, transfers, expenditure, revenue goal), and standard errors clustered by district. Portability is tested by regressing revenue at horizons zero to two on quality measured entirely from other postings, complemented by a stacked event study around high-quality arrivals. The tenure test is 2SLS: overstay and its interaction with quality are instrumented by election-period transfer-ban exposure, whose identifying variation is the intersection of two administratively independent clocks, the nationally fixed election calendar and an anniversary set by an assignment decision three years earlier, with year fixed effects absorbing common election-year shocks. First-stage joint F statistics run 34 to 45 on the overstay margin.
- Main findings
- Quality is portable: a one-unit increase in officer quality predicts 11.1 percent higher revenue one year after arrival and 14.5 percent two years after (both p < .01), with no effect in the arrival year itself. The event study shows no differential pre-trends and roughly 16.3 percent higher revenue after a high-quality arrival (p = .024). The gains run entirely through voluntary remittances, whose coefficients mirror total revenue, with no detectable effect on assessment-based or delinquent collections. Returns persist under dynastic entrenchment, with no significant interaction. In the 2SLS, instrumented overstay interacted with quality is 0.277 at the one-year horizon (p < .10), implying a one-unit increase in quality raises the revenue effect of an induced overstay by about 32 percent.
- Mechanism
- Quality operates through the compliance margin rather than enforcement: better officers raise the taxes remitted voluntarily, consistent with administration built on credibility and local knowledge. Because that knowledge takes time to accumulate, the returns to talent emerge only with tenure, making quality and time in post complements.
- Why it matters
- Identifies a cost of the standard anti-corruption prescription: rotation rules that treat all bureaucrats alike truncate the returns to the most capable ones, so personnel policies designed to constrain discretion can weaken fiscal capacity when they ignore the quality-tenure complementarity.
- Speaks to
- state capacity and tax administrationbureaucrat quality and the personnel economics of the staterotation and anti-corruption policyfiscal capacity and developmentPhilippine political economy
Cite
@misc{ye_roving_bureaucrats,
title = {Roving Tax Bureaucrats: Bureaucratic Quality, Tenure, and the Costs of Rotation},
author = {Renhao Ye and Nico Ravanilla},
howpublished = {Working paper},
}