The literature on the “fiscal contract” proposes that states trade public service provision in exchange for tax payments and other forms of compliance from their citizens (Bates and Lien Reference Bates and Lien1985; Levi Reference Levi1988; Moore Reference Moore, Bräutigam, Fjeldstad and Moore2008; Timmons Reference Timmons2005). In aid-dependent developing countries, however, the state may not be the actor providing the basic services and public goods that citizens use, and this may have implications for the strength of the fiscal contract. In recent years, international donors and non-governmental organizations have placed increasing emphasis on prominently branding the development interventions that they fund – they want the citizens of aid-receiving countries to know that projects are funded by an ostensibly benevolent foreign government (DFID 2012; Sacks Reference Sacks2011; USAID 2012). While the goal of this project branding is to improve attitudes toward the donor countries among recipient country citizens, one potential negative externality of the branding is that it might interrupt the virtuous circle in which government performance leads to legitimation of the government among citizens and to citizen compliance with the government in terms of tax payments and other behaviors (Sacks Reference Sacks2011, Reference Sacks2012).
The idea that non-state provision of social services might undermine state legitimacy is widespread in the literature on non-governmental organizations in the developing world (Brass Reference Brass2010). While non-governmental organizations can provide services that benefit poor people in developing countries, they may do so at the expense of state legitimacy (Fowler Reference Fowler1991), particularly in those cases where developing country governments have explicitly tried to legitimize their rule through promises of service provision (Bratton Reference Bratton1989). Gubser (Reference Gubser2002: 141) bluntly suggests that states might be “embarrassed” by other actors taking on government responsibilities. Whaites (Reference Whaites1998) argues, from a normative perspective, that non-governmental organizations have an obligation to try to build capacity within the state to offset this risk. Lake (Reference Lake2010) argues that the obvious presence of foreign actors in the state-building process can be self-defeating.
In this paper, we collect individual-level evidence on whether or not knowledge about foreign-funded development interventions undermines citizens’ confidence in the state. We do so by using a survey experiment conducted on Indian citizens in which we describe a project in the health sector and vary the information that we reveal about the source of project funding. India is an appropriate case to study because the penetration of foreign aid donors is not so great that citizens will automatically assume that the project is foreign-funded and yet is large enough that citizens will not have trouble understanding the prompt. Overall, we find null results that, like existing results from observational studies (Sacks Reference Sacks2012), call into question the extent to which foreign aid is harming the fiscal contract.
HYPOTHESIS
If citizens are aware that foreign actors are providing the goods and services that national governments are supposed to provide, this may undermine confidence in the government. Because core functions are being outsourced to contractors or non-governmental organizations from other countries, citizens may view their government as ineffective or irrelevant. On the other hand, Sacks (Reference Sacks2012) argues and finds observational evidence for the viewpoint that successful government management of foreign aid actors (at either a national or local level) may serve to demonstrate government capacity to citizens and thereby increase legitimacy. We test the initial hypothesis.
METHODOLOGY
We used Amazon's Mechanical Turk to recruit a convenience sample of 1,400 respondents from India in August 2012.Footnote 1 Potential subjects were offered US$0.15 to “take a 10-minute survey in which we ask your opinions about a healthcare program in India.” Respondents were required to be in India and have a 95% or higher approval rate for previous work done on Mechanical Turk. The survey was in English, requiring an intermediate degree of proficiency to complete.
Recruiting through Mechanical Turk means that our sample is likely to differ from a random sample of Indian respondents in several ways. As compared to the average Indian, the average respondent in our sample is more likely to be young, highly educated, living in an urban area, and proficient in English. From this sample, we can generalize to a young Indian elite that is likely to be relatively politically engaged and also relatively likely to observe evidence of foreign sponsorship of development interventions through indirect means (e.g. news broadcasts).Footnote 2
After a series of 11 demographic and background questions (including four questions about the severity of the HIV/AIDS problem in India), respondents were presented with a vignette about an HIV/AIDS prevention program. In the control condition, the vignette read as follows:
-
The Samastha project is a comprehensive HIV/AIDS program, involving prevention, care, support, and treatment, in the state of Karnataka, a state that exhibits high rates of HIV infection. The project's goal is to reduce transmission and the impact of HIV in selected districts, with a focus on rural areas. The Samastha project provides system strengthening at the state level and services in 13 high-HIV/AIDS prevalence districts in Karnataka.
-
Since the beginning of the project in 2006:
-
– The program increased coverage of care for children affected by HIV and AIDS to about 54.5% of the estimated 33,000 infected children in Karnataka.
-
– Among Karnataka's estimated 1.2 million pregnant women annually, the proportion who received counseling and testing for HIV increased from 20.1% to more than 67%.
-
Across four treatment conditions, we inserted a sentence in the vignette (following the first sentence) that described the funding and implementation of the project. The four treatments read as follows:
-
1. (Foreign Funder (U.S.)). The program is supported by the government of the United States through the United States Agency for International Development (USAID).
-
2. (Foreign Funder (U.S.) and Local Implementer). The program is supported by the government of the United States through the United States Agency for International Development (USAID), and is administered by the Karnataka State AIDS Prevention Society and local NGOs.
-
3. (Foreign Funder (U.S.) and Foreign Implementer). The program is supported by the government of the United States through the United States Agency for International Development (USAID) and is administered by international NGOs.
-
4. (Foreign Funder (Canada)). The program is supported by the government of Canada through the Canadian International Development Agency (CIDA).
The real project on which the vignette is based was jointly sponsored by the United States and Canada and jointly administered by international and local NGOs; therefore, the vignettes do not involve direct deception, although each treatment omits some information.
Immediately following the vignette, respondents were asked a series of questions about their impressions of the project. They were then asked about government performance on the issue of HIV/AIDS and asked to provide favorability ratings of several national institutions and several foreign countries. Specifically, respondents were asked whether they had a very favorable, favorable, unfavorable, or very unfavorable opinion of (1) Prime Minister Manmohan Singh and (2) the Lok Sabha (the lower house of parliament in India), and they were asked (3) to “assess the job that the national government is doing in addressing the problem of HIV/AIDS across India” and (4) to “rate the job that your state government is doing in addressing the problem of HIV/AIDS within (the respondent's state).” For these latter two questions, they could again choose one of four answers (very poor, poor, good, and very good). Respondents also were asked, “How much do you think that foreign aid from other countries contributes to the provision of social services in India?” and could answer on a four-point scale (nothing; a little; some; a lot). If branded aid undermines government legitimacy, we expect respondents in the treatment condition to rate government institutions and performance more negatively.
RESULTS
Before looking at the results with regard to legitimacy, we show in the first column of Table 1 that respondents in the treatment conditions were more likely to say that foreign aid contributes to the provision of social services. Comparing the three conditions where foreign funding is mentioned to the control condition, the average response to this question moves 0.13 units (p < 0.01). This significant treatment effect is evidence that the experimental manipulations conveyed the intended information about the foreign funding of development interventions and respondents absorbed this information.Footnote 3
Table 1 Average Treatment Effects for Four Treatment Conditions
Note: Standard errors in parentheses; 95-percent confidence intervals in brackets. p-values are for two-tailed t-tests of H0: No difference between the treatment and control groups
For the two follow-up questions referring to the prime minister and the legislature, the estimated treatment effects are in the direction predicted by our main hypothesis: when respondents hear about foreign funding and implementation of a large-scale social service project, they express less favorable opinions about Prime Minister Singh and the Lok Sabha. The differences, however, are not significant across any of the four treatment conditions, nor are they significant when all four conditions are combined (b = −0.07, p < 0.22 for the prime minister; b = −0.07, p < 0.17 for the legislature) or when the three U.S. conditions are combined (b = −0.06, p < 0.29 for the prime minister; b = −0.08, p < 0.14 for the legislature).
For the other two follow-up questions, which asked respondents to more generally assess the performance of the national and state governments in the sector where the project is occurring, we estimate treatment effects in the opposite direction. Across the three U.S. treatments, respondents give more credit to both the national and state governments after hearing about a U.S.-funded project. These results are not statistically significant, either individually or when combined (b = 0.06, p < 0.18 for the national government; b = 0.02, p < 0.68 for the state government). The direction of these estimated treatment effects resemble the correlations found in Sacks (Reference Sacks2012), where citizens think more highly of their government when they also think that foreign actors are involved in social service provision. In contrast, when respondents hear about Canadian funding, the ratings of the national and state governments appear lower as compared to the control condition (although not significantly so).
We tried two strategies to reduce the level of noise in our analysis. First, we estimated the treatment effects using linear regression models in which we included controls for respondents’ reported gender, age, education, income and exposure to the news. The inclusion of covariates only marginally changed the estimated treatment effects and produced nearly identical p-values (results not reported). We also estimated treatment effects within relevant subgroups of respondents: those who report reading the newspaper or watching news daily, those who scored in the top-third of an index of political activity, and those who ranked HIV/AIDS as one of the most important issues facing India. In all cases, we expected to find stronger treatment effects. However, for none of these subsets of respondents did the estimated average treatment become significant at conventional levels (results not reported).
SUMMARY
Scholars and development practitioners have expressed concern that citizens in developing countries might lose confidence in the abilities or raison d’être of their own government if they see foreign actors funding and implementing social services. A decrease in legitimacy risks reducing the government's ability to collect tax revenues and otherwise extract compliance from citizens (Levi and Sacks Reference Levi and Sacks2009; Sacks Reference Sacks2012; Weber Reference Weber1968). As foreign aid donors have increasingly emphasized the branding of development interventions, a loss of government legitimacy might be an unintended negative externality.
Results from a survey experiment in which citizens are told about a health intervention and randomly assigned to learn that it is either foreign-funded or do not provide evidence that information about foreign funding has such effects. Overall, the results are null, and the point estimates are in two different directions: particular government institutions (i.e., the sitting prime minister and the legislature) suffer from lower approval ratings in the treatment conditions, whereas overall government performance is assessed more positively.
The null results provide experimental evidence that, like the findings in Sacks (Reference Sacks2012), suggests that foreign aid is not interfering with the development of a fiscal contract between citizens and the states. Some of our point estimates are positive, like those in Sacks (Reference Sacks2012), but they are estimated with too much uncertainty for us to say that foreign-funded service provision is actually improving citizen attitudes toward the state. Our results nonetheless still serve as a rejoinder to those who worry that state development is undermined by the presence of foreign actors.
We are not willing to go so far as to say that these results mean that foreign aid does not have a deleterious effect on state capacity. The presence of foreign aid projects may still hinder the development and retention of capacity within the state. But in terms of finding evidence that the presence of foreign aid projects is interfering with the amount of legitimacy accorded to the state by its citizens, we have found little.
SUPPLEMENTARY MATERIALS
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/XPS.2014.31
