On the political economy of data collection: Lessons from the unaccomplished population census (Democratic Republic of the Congo, 2006-2018)

Cyril Owen Brandt and Tom De Herdt

Type: DRC, Working Paper

Organisation: Institute of Development Policy (IOB)

Country: DRC

Date: 13/03/2019

Full summary

In 2018, twelve years after a working group began to prepare a population census in the Democratic Republic of the Congo (DRC), the census project has still not been implemented. This working paper asks, ‘How have different stakeholders shaped the census project since 2006?’


The latest census in the DRC dates back to 1984. The administrative records are in poor condition, so surveys and evaluations since are increasingly imprecise. The need for a new census is thus uncontested. The unaccomplished census, as we understand it, is a large project carried out by a complex group of national and international public and private decision-makers. In this document, we try to reconstruct the history of the unaccomplished census.


This paper analyses the project in light of low domestic capacities, a politicisation of the census, unmet funding promises by the government, and fragmented and inconsistent donor activities.


Main findings
  • Different and sometimes contradictory forces have shaped the census project over time.
  • The census project has continuously enrolled supporters despite contradictory forces and little evidence of its success.
  • There has been major tension around technology with little domestic ownership.
  • It is difficult to implement projects in which donors need the state as much as the state needs them.


  1. When working with government actors, take the value of the partnership seriously and learn how to deal with your partner.
  2. A census can only be successful if it keeps a degree of distance from the political process, yet stays close enough to garner sufficient support.
  3. If the census is too big to succeed, try a good-enough census.


Download the two page summary of this working paper: