When international organisations provide humanitarian assistance in fragile states, they face a paradox: the aid arrives precisely because governments have failed their citizens, yet concerns persist that foreign provision might further undermine state legitimacy. New experimental evidence from Pakistan reveals a counterintuitive finding: humanitarian aid combined with clear attribution to foreign sources actually increases support for domestic governments. The mechanism is subtle: recipients interpret international presence as a positive signal about their country's institutional trajectory.
This finding arrives at a critical juncture for humanitarian policy. Over 350 million people required humanitarian assistance in 2024 – the highest level since World War II – while aid budgets face mounting political pressure due to a perceived lack of tangible benefits for donor countries. These needs increasingly concentrate in the world's most fragile settings, where weak institutions and political instability make every intervention consequential. To sustain political support for humanitarian programs, donors must demonstrate that assistance advances not only moral imperatives but also strategic interests. One key tool is soft power: the ability to shape preferences and attitudes through appeal rather than coercion. If humanitarian aid can improve recipients' perceptions of donor countries, foster values aligned with democratic governance, and influence political behaviours, it becomes more than charity – it becomes a foreign policy tool. Yet rigorous evidence on whether humanitarian aid actually achieves these soft power objectives has remained elusive.
An Experiment with Vulnerable Women in Pakistan
To test whether humanitarian aid functions as soft power, I conducted a randomised controlled trial with 2,450 extremely vulnerable women across three urban areas in Sindh, Pakistan. These women represent exactly the population that receives humanitarian assistance globally: 84% have no formal education, 93% are classified as deprived or severely deprived, and 88% face stressed, crisis, or emergency levels of food insecurity. Pakistan provides an ideal setting – it has the world's fourth-largest population requiring humanitarian assistance while navigating political instability and being categorised as highly fragile by the OECD.
The experiment cross-randomised two interventions to test whether humanitarian aid can function as a soft power tool. Two-thirds of participants received four monthly cash payments of $25 each – enough to cover a family's basic needs for one month. Separately, half of all participants received information explicitly identifying the aid's foreign source, clarifying that the Pakistani government played no role in the project. Both interventions achieve their goals: The aid significantly improves humanitarian outcomes, while the information treatment correct participants’ beliefs about the source of the funds, as shown in Figure 1, especially among those who initially believe the aid does not come from a foreign donor. This design isolates whether soft power effects come from the material assistance itself, from knowing its foreign origin, or from the combination of both.
Figure 1. Impact of Information Treatment on Participants’ Beliefs About Source of Funds
Three Key Findings
Aid and information each improve perceptions of international organisations, but their combination generates the strongest effects. Participants who received either aid alone or information alone showed 4-5 percentage point increases in approval of international organisations' work in Pakistan. But the combination generated a 7-9 percentage point increase – approximately 10% above the control group mean. This demonstrates that attribution amplifies aid's soft power potential: recipients need to connect material benefits with their foreign source to substantially improve perceptions of international actors.
The intervention generates modest shifts in cultural tolerance, but broader value changes remain limited. The combination of aid and information increased respect for people from other cultures and beliefs by 5.2 percentage points (7% of the control mean) and respect for people from different social groups by 4.1 percentage points (5% of the control mean). However, the intervention showed no effect on deeper value orientations like universalism or positive-sum thinking. Shifting fundamental values likely requires longer, more intensive interventions than four months of modest cash transfers.
Most surprisingly, the combination of aid and information about its foreign source increases support for the domestic government. When participants both received assistance and learned that international organisations provided it, their satisfaction with Pakistan's government increased by 0.125 standard deviations (Figure 2), with similar results using incentivised behavioural measures. This seems counterintuitive: Why would receiving foreign aid while knowing it came from foreigners – not your government – make you more satisfied with that government?
Figure 2. Effect of Humanitarian Aid and Information on Government Attitudes
The mechanism is neither credit misattribution nor simple life improvement. The information alone had no effect on government attitudes. The aid alone had no effect either – participants who received assistance without knowing its foreign source showed no improvement in government attitudes despite experiencing the same material benefits. And credit misattribution seems implausible: the information explicitly stated the government played no role, and participants' beliefs confirm they understood the aid came from international sources, not their government.
Instead, the mechanism operates through signalling about Pakistan's institutional trajectory. When vulnerable women learn that international organisations are actively investing in Pakistan and successfully delivering assistance, they interpret this as validation that external actors view Pakistan as having sufficient stability to merit engagement. The combination of foreign presence (signalling confidence) and actual aid delivery (demonstrating results) improved participants' views of government effectiveness and satisfaction.
Evidence supports this interpretation: the combined treatment increased participants' optimism about Pakistan's political future, particularly for women like themselves. Recipients don't credit the government with providing the aid, but they interpret international investment as validation that their country's prospects are improving – and they attribute this trajectory to the government.
What Doesn't Change
While government attitudes shifted with the combination treatment, costly political behaviours did not. The intervention showed no effects on political participation inside or outside households, nor on willingness to sign a public petition advocating for women's rights. Similarly, neither the aid nor the information alone affected political attitudes or cultural values – only the combination generated modest shifts in tolerance. These precisely-estimated null effects confirm that while humanitarian aid can shift attitudes, changing costly political behaviours is more difficult – whether due to the intervention's duration, design, or the structural constraints participants face.
Policy Implications
These findings directly address a longstanding concern among humanitarian practitioners: that foreign aid provision might undermine local governments by highlighting state failure or creating parallel service delivery systems. This study demonstrates that the opposite can occur. Rather than weakening state legitimacy, humanitarian aid combined with clear foreign attribution can actually strengthen government support – not by stealing credit, but by conveying positive signals about institutional viability that recipients attribute to their government's improving trajectory.
This has three immediate implications for humanitarian policy in fragile states. First, invest in attribution. The information intervention here was remarkably light-touch – brief messages during surveys – yet it substantially amplified aid's soft power effects. Resources spent identifying foreign sources are not wasted; they're essential for maximising impact. Second, manage expectations about value change. While modest shifts in tolerance are achievable, fundamentally reshaping worldviews requires sustained engagement beyond short-term humanitarian interventions. Third, recognise that aid's political effects may work through unexpected channels: rather than directly building or undermining government legitimacy, humanitarian assistance can influence how recipients interpret their country's future prospects.
For donors facing political pressure to justify aid budgets, these findings demonstrate that humanitarian assistance can advance strategic interests through soft power – improving perceptions of foreign actors, fostering modest value shifts, and changing support for governments in recipient countries. As humanitarian needs concentrate in increasingly fragile settings, understanding these political dynamics becomes essential not only for sustaining aid programs but also for ensuring they contribute to rather than complicate efforts to strengthen governance in vulnerable states.
About the Author
Miguel Fajardo-Steinhäuser is a PhD candidate in Economics at the London School of Economics.
His research interests are broadly in development economics, political economy, and industrial organisation. To learn more about his research, visit: https://sites.google.com/view/mfajardo-steinhauser/home
Social Media Handles:
