Aid cuts and NGO administration

In January 2025, the US Government indefinitely paused or cancelled almost all funding for USAID, and revoked existing contracts with most international development organisations around the world. These rollbacks coincide with substantial cuts to the UK’s overseas aid budget. While data on the impacts of this disruption are presently lacking (Abdel Ghany et al. 2026), independent estimates suggest that the resulting healthcare gaps will increase global annual deaths by 2.3 to 5.6 million, predominantly in the world’s poorest countries (Kenny and Sandefur 2025; Cavalcanti et al. 2025). There are expected to be large declines in access to basic education (Oxfam 2026), and significant, though harder to quantify, impacts on election integrity and anti-corruption programmes (Uberti and Jackson 2020). This is a historic rupture: before January 2025, no major foreign aid donor had ever fully exited the global aid ecosystem. This pre-analysis plan registers anticipated analysis of an online and phone survey with officials working at registered NGOs in Malawi, to understand the ongoing effects of this disruption on their capacity to provide services. The survey gathers rich descriptive evidence on the impact of the cuts, and we use two conjoint choice experiments to gauge preferences over the policies that donors and the Malawian government could pursue in response. A core theoretical motivation of the experimental components of this study is that NGOs are not neutral observers of the policy choices facing their governments and donors, they are organizational actors with identifiable interests that shape their preferences over the aid architecture. As implementing agents, NGOs depend on funding flows for organizational survival and have strong incentives to protect the programs and sectors in which they operate. As advocates, they have reputational and relational stakes in the welfare of the communities they serve. These dual roles generate predictable preferences over replacement aid architecture design that may diverge significantly from the assumptions embedded in donor-designed funding packages. Collectively, this data will help us understand the impact of aid cuts on a range of pertinent outcomes, and add the voice of frontline staff to ongoing negotiations about funding packages. The survey we focus on in this plan was also designed to allow comparison with a separate set of surveys with public health officials and citizens conducted in summer 2025. These supplemental surveys were conducted among 1,140 citizens living in proximity to 54 public health facilities and 160 public health officials working in the same 54 facilities. In the hypotheses below, we refer to the main survey with NGOs as the “NGO survey.” The prior surveys we refer to as the “citizens survey” and the “officials survey”. The pre-analysis plan for the citizens survey and officials survey is available at osf.io.