The Limits of Government Outsourcing: Property Tax Re-Assessments in India (Duflo, Hanna, Olken & Tandon, 2026)

Summary

"The Limits of Government Outsourcing: Property Tax Re-Assessments in India," by Esther Duflo (MIT & University of Zurich), Rema Hanna (Harvard Kennedy School), Benjamin A. Olken (MIT), and Shreya Tandon (Cornerstone Research), is NBER Working Paper 35610 (August 2026). It adds a direct answer to a question the wiki's assessment feasibility objection treats mostly as a methodology problem: when a government outsources land/property assessment to private firms, who actually does it better?

The Chennai Experiment

The Greater Chennai Corporation sought to reassess roughly 280,000 properties flagged as likely undervalued by an earlier drone-based GIS survey — a rapidly growing city where many small properties had been converted to larger, more valuable multi-story or mixed-use buildings without a corresponding tax reassessment. Rather than reassess with existing staff alone, the government experimented with outsourcing some of the work to private firms. The authors worked with the government to randomize which of the city's 199 wards were reassessed by regular government tax inspectors versus by outsourced private contractors, and — among outsourced wards — whether firms were paid a fixed fee per property or an incentive fee tied to the revenue increase found. Within each ward, individual streets were further randomized into reassessment waves or a control group.

The Finding: Government Outperformed Private Contractors

Properties reassessed by government inspectors saw valuations rise 84% on average, compared to 54% for privately-assessed properties (p<0.01) — despite the government dedicating roughly half as many total man-days to the task (449 versus 808). Revenue raised per property inspected was 87% higher for government inspectors. To rule out the possibility that government assessors were simply over-valuing properties, the authors hired independent third-party surveyors — whose findings were not shared with either the government or the private firms — and found government assessments more accurate, not less, and not systematically inflated.

Why: Procurement Rigidity, Not Incentive Design

The authors test and rule out two standard explanations. First, government assessors did not appear to hold a private-information advantage about which properties were most undervalued — checked against an independent drone survey neither side could see. Second, weak incentives were not the driver: wards randomized to incentive-fee private contracts performed no differently from fixed-fee contracts on revenue, accuracy, or effort. Instead, the paper traces the gap to procurement and workforce quality: government inspectors averaged 11 years of property-measurement experience versus 1.4 years for private-firm workers, and 43% of outsourced inspectors lacked a college degree versus 15% of government workers. More than 20 firms initially expressed interest in bidding, but the government's lowest-bid procurement rules — intended to limit corruption by removing bureaucratic discretion — pushed the final contract price about 75% below the average initial bid, and firms that declined to bid reported wanting 2–10x what the government offered. Only 3 of an intended 6 firms were ultimately contracted, and — vindicating the hesitant firms — the project was abruptly cancelled after citizen protests and an approaching election, with private firms paid 16 months late. Government inspectors also had legal authority to enter properties that could not be easily transferred to private firms. The authors read this as evidence that "the limits to government outsourcing may go beyond multitasking issues" — the classic Hart, Shleifer & Vishny (1997) framing — implicating procurement design and worker skill directly.

Relation to the Georgist Case

This is a genuinely new dimension for the wiki's assessment-quality coverage: not whether land or property value can be measured accurately (the assessment objection's usual frame), but who should do the measuring when a government's own capacity is stretched. The finding cuts against a common implementation instinct — that outsourcing to the private sector improves administrative capacity — and instead identifies a specific, fixable mechanism (rigid lowest-bid procurement discouraging skilled firms from bidding) rather than a general claim that government is inherently better. It is also indirectly relevant to the wiki's assessment-corruption objection: the same anti-corruption procurement rule (no discretion to favor higher-quality, higher-priced bidders) that is meant to limit bribery here produced a different failure mode — driving away the more competent firms entirely.

Nuances and Limits

• A single city, an interrupted experiment. The reassessment project was cancelled after 8 weeks for political reasons unrelated to the study design, so results reflect only the first wave of the planned rollout.
• Government outperformance is not unconditional. The authors note their result echoes Behaghel, Crépon & Gurgand (2014) on job-placement services, but caution this is one domain, not a general claim that government assessors reliably outperform private contractors — the paper's own diagnosis is procurement-design-specific, not an argument against outsourcing per se.
• A-claim. Full text (introduction, experimental design, results and mechanism sections) read directly from the NBER working paper PDF, not reconstructed from an abstract.

Bears On

• Objection: Land value can't be assessed accurately — adds a who-assesses dimension distinct from the existing methodology-focused evidence.
• Objection: LVT enables corrupt assessors — a case where an anti-corruption procurement rule (no discretion in bidder selection) produced a different failure mode: driving away skilled contractors.
• Research: Bergeron, Tourek & Weigel: DRC Property Tax — a companion developing-country property-tax-administration field experiment, on rate-setting rather than outsourcing.
• Benefit: Property tax raises welfare in developing countries — the outcome page this administrative-capacity evidence bears on.

See Also

• Objection: Land Value Can't Be Assessed Accurately
• Objection: LVT Enables Corrupt Assessors and Will Be Gamed
• Bergeron, Tourek & Weigel: DRC Property Tax
• Tourek: DRC Progressive Property Tax
• Franzsen & McCluskey: Property Tax in Africa
• India MoHUA Value Capture Financing Framework
• Mass Appraisal Methods

Sources

1. Esther Duflo, Rema Hanna, Benjamin A. Olken & Shreya Tandon (2026), "The Limits of Government Outsourcing: Property Tax Re-Assessments in India," NBER Working Paper 35610, August 2026. RCT registry: socialscienceregistry.org/trials/9965. nber.org/papers/w35610 — full text (introduction, experimental design, results and mechanism sections) read directly from the working-paper PDF, 2026-09-02 — used for the Chennai experimental design, the 84%/54% valuation-increase finding, the 87% revenue-per-property gap, the independent third-party accuracy check, the worker-experience/education comparison, and the procurement-rigidity mechanism (A-claim; full text read and cross-checked against a second independent extraction of the same PDF).