Everyone was aware that the itinerary was quite artful. It took the ambassador past every agricultural station which the American mission had established; the Senator would never be out of sight of some sign of American aid. That he would not be seeing a typical countryside disturbed none of these realistic people. (Burdick and Lederer Reference Burdick and Lederer1958, p. 244)
Foreign aid budgets are constantly under pressure (Lancaster Reference Lancaster2007; Morgenthau Reference Morgenthau1962; Spencer et al. Reference Spencer, Lindstrom, Haddad and Mulmi2011; Tendler Reference Tendler1975), and public support for foreign aid is generally low (Page and Shapiro Reference Page and Shapiro1992; van Heerde and Hudson Reference van Heerde and Hudson2010). Even when the donor country public supports the ends for which foreign aid is used, they often do not support augmenting the foreign assistance budget (e.g. Wojtowicz and Hanania Reference Wojtowicz and Hanania2017; see also Hurst, Tidwell, and Hawkins Reference Hurst, Tidwell and Hawkins2017; Milner and Tingley Reference Milner and Tingley2013; Prather Reference Prather2016; Scotto et al. Reference Scotto, Reifler, Hudson and VanHeerde-Hudson2017). Many aid agencies seek to increase support by making aid more observable.
One strategy is to brand physical manifestations of assistance with the flag, the aid agency name and logo, or other donor signifiers. Across the developing world, one finds examples of donor credit-claiming both on infrastructure projects, such as health clinics, and at trainings or workshops. As public diplomacy, branding activities ostensibly work to make recipients aware of the sources of aid (Dietrich, Mahmud, and Winters Reference Dietrich, Mahmud and Winters2018).
At home, aid agencies use images of branded aid to publicize projects to domestic audiences. Although the reasons for this effort are not consistently articulated, a common argument states that if the general public sees taxpayer dollars used to improve the lives of people in poor countries, they should support foreign aid more, yet this question has not been investigated empirically.Footnote 1 When citizens see images of branded aid, are they more likely to say it is well spent and express support for foreign assistance?
We explore these questions with an original survey experiment on British nationals. Respondents receiving branded visual treatments were more likely to identify the donor and more likely to think that aid recipients could also identify the donor. In addition, the evidence suggests that the treatments improved perceptions that aid money is “well spent.” Only among conservatives, however, did the treatments increase support for the specific aid program, and for this same group, the videos decreased overall support for foreign aid relative to a pure control group, an unexpected finding. We conclude the article with suggestions for future studies on this topic.
Seeing foreign aid in action and supporting foreign assistance
Observing a branded foreign aid intervention can provide two kinds of information to donor country citizens. First, if we assume that they view the intervention as helpful, the image can demonstrate to them that a beneficial intervention was funded using resources from their home country. Findings in psychology show that helping others and donating money cue a variety of positive emotions (e.g. Cain, Dana, and Newman Reference Cain, Dana and Newman2014; Crocker, Canevello, and Brown Reference Crocker, Canevello and Brown2017). Visual information linking a development project to the donor country could result in positive affect toward the project and toward foreign aid in general.
Second, branding can signal to an individual in a donor country that people in the aid-receiving country are aware of resources coming from the donor (Dietrich, Mahmud, and Winters Reference Dietrich, Mahmud and Winters2018). The donor country citizen might believe that it is good for people in aid-receiving countries to be aware of the donor’s generosity because it will improve public opinion toward the donor, bring about economic benefits, or reduce security threats. Inferring that branded aid benefits the donor country, the person seeing branded aid will become more supportive of aid.
In this exploratory study, we exposed donor country citizens to either unbranded or branded information about a specific aid intervention. We expect more positive reactions to foreign aid among the set of respondents exposed to branded treatments (H1).Footnote 2
Political conservatives tend to be skeptical of foreign aid for development, given their preferences for budget discipline and free markets (Milner and Tingley Reference Milner and Tingley2011, Reference Milner and Tingley2013; Paxton and Knack Reference Paxton and Knack2012), but are more likely to advocate for the instrumental use of aid in pursuit of the national interest (Fleck and Kilby Reference Fleck and Kilby2006). These differing ideological preferences exist at the level of political parties: while left-leaning parties express more positive foreign aid positions in their election manifestos (Dietrich, Milner, and Slapin Reference Dietrich, Milner and Slapin2019), conservative parties associate foreign aid with the “national interest” (Mawdsley Reference Mawdsley2017). Given that branding makes aid observable and promotes the donor image abroad, we expect positive treatment effects to be larger among conservative respondents when aid promotes the national interest (H1a).
Research design
We embedded an informational experiment in a survey conducted on a sample of self-reported British nationals registered with experimental labs at the Universities of Essex, Edinburgh, and Oxford in March/April 2016.Footnote 3 A total of 777 subjects completed the survey: 382 student subjects, 382 non-student subjects, and 13 subjects who did not identify their status.Footnote 4 Every respondent received a £3 Amazon voucher as compensation.
The survey began with a set of questions unrelated to foreign aid. We then indicated that the rest of the survey would be about foreign aid and asked three initial questions to gauge respondent awareness of foreign aid.
We randomly assigned respondents to one of five experimental conditions, summarized in Table 1. The first condition was a pure control with no video treatment. The other four conditions each included a short video about UK-funded health clinics in Bangladesh.Footnote 5 All respondents in groups 2–5 were informed that they would watch a video about a “British foreign aid project in Bangladesh.” All four videos featured a 1-minute-long dramatization of the way in which prenatal services provided by a Smiling Sun Clinic calmed the anxieties of two soon-to-be-parents. In the control video treatment, the video included the Smiling Sun logo and name at the bottom of the screen but did not include any reference to UK aid or the UK.
Table 1 Summary of experimental conditions
Notes: “X” Indicates that the condition included the component.
Treatments 3–5 featured “UK aid” branding: the UK aid logo and the tagline “From the British People” along the bottom of the video. The branded video included this branding in lieu of the Smiling Sun logo and name. The highlighted video featured the branded video treatment plus a photo showing the UK aid logo on the main entrance of a clinic; it zoomed in on the logo for emphasis. Text in the highlighted video said, “When the British Government funds projects like the Smiling Sun Clinics, they often require that the project show the ‘UKAID’ logo.” The strategic video included this and further information about the strategic value of credit-claiming: “The goal is to enhance the visibility of UKAID’s projects, to advance the UK’s other foreign policy goals, and to improve the UK’s image in Bangladesh.”
Following all video treatments, we asked an initial question about aid delivery mechanisms and then asked one of our main outcome questions, “In your opinion, does the UK government do a good job of ensuring that our foreign aid is well spent by the countries that receive it?” We then asked whether the government should “spend more, the same amount, or less money to help people in poor countries,” and questions about the usefulness of foreign aid and support for the UK government. For respondents in video conditions, we inquired whether the UK should increase its support of Smiling Sun Clinics, should fund similar projects elsewhere, and whether they thought that Bangladeshis were aware of UK funding for the clinics. After demographic questions at the end of the survey, we included a manipulation check, asking respondents who were shown a video if they could name the project funder.
In the next section, we first present the manipulation checks. We then look to see if there were effects on how respondents thought about aid effectiveness and on their levels of support for the particular development project featured in the video or foreign aid in general.
We report the mean value of the outcome variables in each treatment condition. We provide statistical tests for the comparisons between treatment and control conditions individually and a joint significance test for the three branded conditions.Footnote 6 Our test statistics come from linear regressions that include indicators for all treatment conditions and subject pool indicators; we use robust standard errors to calculate p-values.
For the findings related to aid effectiveness and support for aid, we also look, in accordance with our preregistered hypotheses, for heterogeneous treatment effects based on ideology. Respondents were asked to self-identify on a five-point ideology scale that ranged from “very left” (1) to “very right” (5); respondents could place themselves anywhere on the scale in 0.1 point intervals. The median self-placement in our sample was 2.5. Respondents at the midpoint or above are coded as “conservative” and below the midpoint as “liberal.”
Table 2 Respondents by experimental condition and partisanship
Results
The first column of Table 3 presents the manipulation check. Respondents who saw any of the branded videos were more likely to absorb information about UK funding. Each treatment indicator is significant, as is a test of their joint significance. In the highlighted video condition, an additional 35% of respondents name the UK as the funder; the proportion is somewhat lower in the strategic video condition and lower still in the video that simply has a static UK aid brand at the bottom. The survey provided other opportunities for respondents to infer UK involvement (e.g. the introduction to the videos; the question reported in the second column of the table); almost half of the group that saw the control video correctly answered that the project is UK funded. Therefore, the evidence strongly suggests that branding accomplishes one of its missions: it communicates to a broader donor country audience that a development intervention has been funded using government resources.
Table 3 Awareness of UK aid
Notes: Top panel presents mean values across the four video treatment conditions. Middle panel presents the p-value for a χ 2 test of h 0: independence of treatment conditions from responses to question. Bottom panel presents regression coefficients for the treatment indicators in a regression with subject pool fixed effects (not reported); robust standard errors in parentheses. ***p < 0.01; **p < 0.05; *p < 0.10.
The second column of Table 3 shows that branding leads people to infer that individuals in aid-receiving countries are aware of the aid’s source. In the control video condition, only one of four respondents thinks that Bangladeshis are informed about the origins of the funding for the clinics. This jumps by 20 percentage points in the branded video condition and by around 20 percentage points more in the remaining two conditions.
Does seeing aid branding in action change respondents’ perceptions of foreign aid? Table 4 provides suggestive evidence that observing branding in action through the highlighted or strategic videos yields more positive assessments of how the UK uses its aid. Willingness to agree with the statement that the UK does a “good job of ensuring that aid is well spent” increased by at least a quarter of a point on a five-point scale.
Table 4 Perceptions of UK aid quality
Notes: Top panel presents means for the five experimental conditions. Middle panel presents p-value for a χ 2 test of h 0: independence of treatment conditions from responses. Bottom panel presents regression coefficients for treatment indicators in a regression with subject pool fixed effects; control video is the omitted condition; robust standard errors in parentheses. ***p < 0.01; **p < 0.05; *p < 0.10.
Does seeing branding in action change respondents’ support for foreign aid? In Table 5, we look at two outcomes related to the development intervention in the video and one related to overall levels of foreign assistance. In the combined sample, we see small and insignificant effects of the branded videos for each of the three outcomes. In each case, however, we also see that there is lower support for foreign aid among conservatives in the control conditions (i.e., the pure control condition and the control video condition). Among conservative respondents, exposure to the highlighted branding condition made them about a quarter point more likely (on a five-point scale) to say that the UK should increase its support for the Smiling Sun Clinics in Bangladesh and about 15 percentage points more likely to say that the UK should support similar projects in other countries. Oddly, liberals exposed to the highlighted video are less supportive of allocating assistance to the Smiling Sun Clinics in Bangladesh relative to liberals who saw the control video. If liberals are critical of using aid to promote the donor’s image, the highlighted video treatment might evoke negative feelings toward aid, particularly absent explanation for the UK aid logo on a health clinic in Bangladesh.
Table 5 Support for foreign aid
Notes: See notes for previous table.
The third outcome variable in Table 5 presents some surprises and suggests that information about aid moves respondent opinions in a negative direction. Respondents in a pure control group that saw no video were more likely to support foreign aid spending overall; this is particularly pronounced among conservatives.
Our central hypotheses were preregistered, and we provide joint significance tests on the treatment indicators; yet there is a potential for false discoveries because of the multiple treatment indicators and the clustering of outcome variables. The Holm–Bonferroni adjustment creates a stricter standard for declaring findings statistically significant.Footnote 7 In Table 3, we examine three treatment indicators across two outcome variables from the same family: given the small standard errors, all findings continue to obtain conventional levels of statistical significance after the correction. In Table 4, we have a single outcome variable with four treatment indicators examined across three subsets of the data. In Table 5, we have three outcome variables with either three or four treatment indicators, each examined across three subsets of the data. All adjusted p-values for both Tables 4 and 5 fall above the standard critical value of p < 0.05. The results that we report should be treated as preliminary.
Conclusion
Branding development interventions can convey information to multiple audiences. On the one hand, branding speaks to those who benefit from the project, ostensibly in an attempt to improve attitudes toward donor countries. At the same time, branding shows political elites and taxpayers in the donor country how their resources are used. Whether these efforts work as expected is largely an untested proposition.
We evaluate UK citizens’ reactions to foreign assistance branding. This research is exploratory and will inform future research. Our findings contribute knowledge to research that asks whether existing attitudes about aid can be changed (e.g. Nair Reference Nair2018; Scotto et al. Reference Scotto, Reifler, Hudson and VanHeerde-Hudson2017). First, we show that branding used in communications with donor publics succeeds in conveying information to respondents: subjects who watched a video featuring branded aid were more likely to know that the project had been funded by the UK. In addition, these respondents were more likely to infer awareness of the donor among Bangladeshis. Secondly, our findings suggest that branding changes attitudes toward foreign aid: it improves perceptions that foreign aid is well spent. Because of the small sample size that we had, we hope other researchers will work to replicate this finding.
Our study also contributes to research on how ideology shapes support for foreign aid (e.g. Fleck and Kilby Reference Fleck and Kilby2006; Mawdsley Reference Mawdsley2017). We find that branding increases support among conservative respondents for expanding the intervention featured in the video. This finding supports recent evidence suggesting that conservative and left-leaning parties may value aid for different reasons: while the left supports aid for more altruistic, development-oriented reasons, conservatives support foreign aid when it pursues the national interest, for example, by promoting the image of the donor abroad. Future research could include mediator questions to test these claims.
Further, our result that branding improves perceptions about the UK’s ability to ensure that aid is well spent directly relates to the literature on credit claiming, which contends that legislators reap electoral gains for allocating benefits and also for taking credit for their actions (Stein and Bickers Reference Stein and Bickers1994). More recently, research has found that credit-claiming messages improve perceptions of the ability of legislators to deliver spending and projects to their districts (e.g. Grimmer, Messing, and Westwood Reference Grimmer, Messing and Westwood2012). Our findings show that the UK government’s credit claiming for development interventions can be successful: it increases information about who funds the intervention and improves perceptions about the UK’s ability to deliver foreign aid effectively. It also appears to increase support for aid on the domestic front, although not uniformly.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/XPS.2019.12.
Author ORCIDs
Susan Hyde, 0000-0003-4079-598X, Matt Winters, 0000-0002-3693-9461
