Kananga tax agency boosted revenue by letting neighborhood chiefs collect

Kananga, a Congolese provincial capital, needed to collect property tax across a fast-growing city of roughly 45,162 registered properties, using state agents with no special local knowledge of individual neighborhoods -- who actually lived at a given address, what a property was really worth, and who had the ability to pay right now versus who was genuinely struggling.

Economists Pablo Balán, Augustin Bergeron, Gabriel Tourek and Jonathan Weigel ran a randomized policy experiment testing this trade-off directly: some Kananga neighborhoods kept ordinary state agents collecting property tax, while others delegated collection to city chiefs -- traditional local authority figures with no formal state capacity, but deep, specific knowledge of who in their neighborhood could and would pay.

Let neighborhood chiefs collect the tax directly, since they already knew each household's means, capturing the gain without building new information systems.

Chiefs knew which households could pay, so they targeted visits efficiently.

State agents lacked this local knowledge, so training couldn't close the gap.

Delegation cost nothing extra because chiefs' knowledge already existed.

Hybrid test showed information, not persuasion, drove the compliance gain.

When your workforce lacks on-the-ground information, delegate execution to whoever already has it, even if they're outside your formal hierarchy.

2 percentage points, a 44% increase in revenue over agent-led collection in the same period. Chiefs also collected roughly 2 percentage points more in bribes than state agents did -- payments made in lieu of the official tax rather than on top of it -- yet third-party verification found chiefs were actually more accurate at assessing property liability and more likely to correctly exempt elderly and disabled residents as Congolese law required, and the study found no evidence chief-led collection undermined citizens' trust in government or tax morale.