Politicians aren’t listening to the public on aid. They’re listening to the media.

In recent months, several commentators have pointed out that the dramatic cuts to aid budgets happening around the world are not a response to declines in public support.

According to a recent poll conducted in June 2026, for example, 79% of Americans agree that foreign aid can be “a good and effective part of US foreign policy”, while 72% agreed that it ‘“keeps the US safer and more secure”.

Similarly, the Development Engagement Lab’s longitudinal tracker has found that in France, Germany, the UK and the US, around 50% of people say that aid budgets should be maintained or increased and that “most people’s views remain stable over time”.

This apparent disconnect between government actions and public opinion raises an important question for international development policy, as Devex recently asked: “If the public supports foreign aid, why don’t politicians speak up?”

An over-estimation of media influence

According to our recently published research, the answer may lie in politicians over-estimating the media’s influence over public opinion.

Based on a survey of over 100 UK MPs and 2,000 members of the public, we found that many politicians believe that media coverage shapes public opinion on aid. For example, 68% believe that news coverage influences public support for the size of the UK’s overseas aid budget. This is consistent with long-standing assumptions within policy circles about the supposed influence of the “CNN effect” and media-induced “compassion fatigue”.

We also found that a politician’s belief in media influence is positively associated with the belief that the media should influence aid policy. In fact, a belief in media influence is the best and most robust predictor of thinking it is acceptable for the media to influence aid policy.

Finally, we found that UK politicians are significantly more likely than the UK public to think that the press influences public attitudes. Only 41% of the public believe that news coverage about overseas aid influences public support for the size of the UK’s overseas aid budget, compared with 68% of MPs.

These findings support the idea that aid policy is influenced by politicians’ belief in the media’s ability to change public opinion.

They also suggest that politicians may well be over-estimating the media’s influence, especially given that previous research has consistently found very limited support for the media’s influence via the “CNN effect” or “compassion fatigue”.

This means that politicians may be responding less to actual public attitudes than to what they believe those attitudes will become once shaped by media coverage.

Reframing the benefits of aid is not enough

As the co-founder of Global Citizen, Michael Sheldrick, has argued, aid budgets have become “so easy to cut” because politicians “don’t see [public] support… they see media stories that frame foreign aid as wasteful or bogged down in bureaucracy”.

This is not an isolated finding. Professor Scott’s earlier research has also shown that when aid agencies earmark a percentage of their humanitarian aid for “under-reported crises”, they appear to be correcting for a media influence that does not exist.

This insight matters because it suggests that strengthening political support for aid is not purely a communications challenge – as is often suggested. Re-framing the benefits of aid in new or alternative ways will not necessarily shift political support if politicians continue to mistake critical media coverage for public opinion.

Efforts to strengthen political support for aid should also involve correcting politicians’ overestimation of how strongly the media shapes public support. This could involve sharing more robust evidence on public attitudes towards aid and giving politicians greater confidence that public support is often more resilient than they assume.

The media does not brainwash anyone. But a belief that it does can have just as much influence as if it did. That belief, therefore, needs correcting.