Since the end of World War II, the US government has, according to the Council on Foreign Relations, spent nearly $4 trillion on foreign aid, broadly referring to money that seeks to provide humanitarian relief, economic development, and/or military arms to other countries.
Despite the numerous benefits that foreign aid programs can yield (Savun and Tirone, Reference Savun and Tirone2011; Goldsmith et al., Reference Goldsmith, Horiuchi and Wood2014; Ingram, Reference Ingram2019; but see, e.g., Wright, Reference Wright2009), including economic development (e.g., Dalgaard et al., Reference Dalgaard, Hansen and Tarp2004; Karras, Reference Karras2006; Economides et al., Reference Economides, Kalyvitis and Philippopoulos2008) and the promotion of democracy (e.g., Dietrich and Wright, Reference Dietrich and Wright2015; Heinrich and Loftis, Reference Heinrich and Loftis2019; Ariotti et al., Reference Ariotti, Dietrich and Wright2022), a long-standing elite consensus in support of such programs, and their importance for US global leadership (Obama, Reference Obama2007; Lieberman and Kyl, Reference Lieberman and Kyl2015), American public support for such spending is quite low, particularly when compared to other types of government spending (Oldendick and Hendren, Reference Oldendick and Hendren2014). This holds true even when foreign affairs moves atop the political agenda.Footnote 1 And while the American mass public stands out as especially averse to foreign aid spending (e.g., Diven and Constantelos, Reference Diven and Constantelos2009), this pattern, i.e., of a lack widespread support for large public expenditure on foreign aid programs, particularly compared to other types of government spending, also appears to manifest among mass publics in other Western democracies (Henson and Lindstrom, Reference Henson and Lindstrom2013; Wood, Reference Wood2018).
One common explanation for such spending attitudes, focusing on the United States, relates to uncertainty about what “foreign aid” actually entails and ignorance, specifically gross overestimates about foreign aid's share of the federal budget (e.g., Hurst et al., Reference Hurst, Tidwell and Hawkins2017; Scotto et al., Reference Scotto, Reifler and Hudson2017). However, data from a module on the 2020 Cooperative Election Study (CES) shows that even among the approximately four in ten Americans who know that foreign aid comprises a small proportion of the federal budget (less than Medicare, Social Security, and national defense), there is more support for cutting foreign aid (38 percent) than for increasing it (12 percent). Data from the 2008–2018 General Social Survey (GSS) shows that even among the approximately one in five Americans who say they are “very interested” in international issues, far fewer people agree that we are spending “too little” on foreign aid/assistance to other countries (13 percent) than believe we are spending “too much” (60 percent). Collectively, this suggests that even if ordinary people paid more attention to international affairs and/or possessed more factually correct information about the nature of government spending, there would likely still be a dearth of public support for government allocating large sums toward foreign aid programs and assistance to other countries.
In seeking to explain American public support (or lack thereof) for foreign aid, extant scholarship tends to focus on framing and exposure to different types of information. These include exposure to information such as the extent of global wealth inequality (Nair, Reference Nair2018), and characteristics about potential recipient countries, such as whether it is a democracy (e.g., Doherty et al., Reference Doherty, Bryan, Hanania and Pajor2020), is in a developing and/or negatively stereotyped region of the world (e.g., Baker, Reference Baker2015; Blackman, Reference Blackman2018), engages in rights abuses (e.g., Heinrich and Peterson, Reference Heinrich and Peterson2020), or conversely, whether such aid would benefit the donor country on the international stage (e.g., Wood and Hoy, Reference Wood and Hoy2022; Chung et al., Reference Chung, Pechenkina and Skinner2023; Strange, Reference Strange2023). This focus on framing experiments means that we know a good deal about how exposure to different kinds of information shape attitudes toward foreign aid, but know less and lack consensus about the demographic and attitudinal factors that shape public opinion toward foreign aid, particularly relative to, for example, what is known regarding attitudes toward economic redistribution (e.g., Alesina and Glaeser, Reference Alesina and Glaeser2004; Piston, Reference Piston2018) and immigration (e.g., Valentino et al., Reference Valentino, Soroka, Iyengar, Aalberg, Duch, Fraile, Hahn, Hansen, Harell, Helbling, Jackman and Kobayashi2019; Citrin et al., Reference Citrin, Levy and Wright2023).Footnote 2
In short, extant research, both observational and experimental, has yielded important insights regarding public opinion toward foreign aid spending. What is lacking, however, is clear and comprehensive evidence regarding how and why attitudes toward government, the actor most responsible for crafting, funding, and distributing foreign aid, may matter, and specifically whether trust in one's national government, which I refer to here as political trust, matters. While existing scholarship has paid some attention to how trust matters for foreign aid support, these studies either focus on trust in other people (Bayram, Reference Bayram2017), trust in international organizations (Bayram and Graham, Reference Bayram and Graham2022), or trust in other countries (Brewer et al., Reference Brewer, Gross, Aday and Willnat2004), rather than trust in one's domestic national government.
To the extent that existing work has focused on this latter phenomenon, the empirical results, are mixed, with some studies (Chong and Gradstein, Reference Chong and Gradstein2008; Bauhr et al., Reference Bauhr, Charron and Nasiritousi2013; Bodenstein and Faust, Reference Bodenstein and Faust2017), but not all (Paxton and Knack, Reference Paxton and Knack2012) finding evidence of a statistically significant relationship between domestic political trust and public support for foreign aid expenditure. These aforementioned studies also include a measure of domestic political trust as a control variable in cross-sectional multivariate regression models, rather than as a theorized explanatory variable of interest. As such, it remains unclear as to why such variables (political trust and foreign aid attitudes) may be linked.
Furthermore, these studies measure the concept of domestic political trust in different ways, often with single survey items, rather than with multiple questions comprising a validated multi-item scale. This empirical approach, i.e., employing single-item survey measures that only partially capture aspects of a well-validated measure of political trust (Hetherington, Reference Hetherington2005) can, by increasing the likelihood of possible measurement error, lead to researchers failing to find, or potentially underestimating, by a non-trivial degree, statistical relationships between concepts of interest (e.g., Ansolabehere et al., Reference Ansolabehere, Rodden and Snyder2008).
One study that goes beyond including a single-item measure of domestic political trust as a control variable is Heinrich et al. (Reference Heinrich, Kobayashi and Lawson2021). Using a variety of data, including surveys from the United States (MTurk in 2018) and the United Kingdom (Dynata in 2019), these authors examine the relationship between populism and mass support for foreign aid. As part of their broader exploration of how populism shapes foreign aid attitudes, the main focus of their paper, they find evidence of a statistically significant relationship between “anti-elitism,” one component part of populism, which they define as perceiving elites to be “corrupt, dishonest, and self-serving,” and thus not representing a clearly defined public will, and support for foreign aid spending. This is related to the concept of domestic political trust, which reflects general attitudes toward government and the political system but is also empirically distinct from it (Geurkink et al., Reference Geurkink, Zaslove, Sluiter and Jacobs2020). Indeed, it is possible for one to think that government is ineffective and inefficient, without believing that a small elite are nefariously defying a clear and apparent public will. Furthermore, several of the questions that (Heinrich et al., Reference Heinrich, Kobayashi and Lawson2021, 1050) employ to measure “anti-elitism” are very similar to those that are used to measure external political efficacy (Geurkink et al., Reference Geurkink, Zaslove, Sluiter and Jacobs2020, 254). In short, Heinrich et al.'s work, while informative and valuable, provides stronger evidence to suggest that populism and foreign aid attitudes are linked, but yields more suggestive, rather than definitive, evidence regarding the relationship between domestic political trust and attitudes toward foreign aid spending.
Here, I build upon these existing studies, including Heinrich et al. (Reference Heinrich, Kobayashi and Lawson2021)'s finding of a statistically significant relationship between “anti-elitism” and foreign aid attitudes, but also diverge from past work in several ways. One is by focusing specifically on political trust, rather than populism writ large, as my main explanatory variable of interest. A second is by using an explicit measure of political trust (in my primary analyses) that has been widely employed in existing studies, the 4-item American National Election Studies (ANES) scale. A third is by employing a theoretical framework that focuses more on cost–benefit trade-offs in the public mind, one that is consistent with existing scholarship on the mass-level policy consequences of political trust (e.g., Hetherington and Globetti, Reference Hetherington and Globetti2002; Macdonald, Reference Macdonald2021). Finally, I also use nationally representative survey data (from the ANES and GSS) and employ both cross-sectional and panel survey designs to test this relationship. Overall, existing work provides some evidence to tentatively suggest that domestic political trust may shape public opinion toward foreign aid spending. Here, I seek to test this more comprehensively and definitively.
In addition to providing a more comprehensive and definitive series of tests regarding how domestic political trust shapes attitudes toward foreign aid spending, which is the main objective of this paper, I also seek, by examining a wide range of individual-level explanatory factors, e.g., demographics, political identities, values, social group sentiments, along with political trust, to further advance our understanding of the individual-level factors that systematically shape of public opinion toward foreign aid. This reflects an additional contribution to the literature given that there is less consensus regarding the individual-level factors that systematically and meaningfully shape attitudes toward foreign aid spending, relative to, for example, other policies in the domain of foreign affairs such as international trade (e.g., Mutz, Reference Mutz2021), and the use of military force (e.g., Berinsky, Reference Berinsky2009).
I use a combination of cross-sectional and panel survey data from the United States to (1) thoroughly test how political trust matters, and (2) to advance a more comprehensive model of public opinion toward foreign aid spending. Overall, I find that political trust is associated, to a substantively significant degree, with support for foreign aid spending. This relationship is robust to a variety of model specifications and is present during both peacetime and wartime. I attribute this to the perceived imbalance of costs and benefits that foreign aid entails, i.e., that it imposes a financial burden on the domestic population while conferring financial benefits upon foreigners. Additionally, I find that race, ideology, core political values, general isolationist sentiment, evaluations of the national economy, and feelings toward poor people also meaningfully shape public attitudes toward foreign aid.
Overall, these findings advance collective knowledge regarding American public opinion toward foreign aid, a widely employed strategy in international diplomacy and statecraft (e.g., Schraeder et al., Reference Schraeder, Hook and Taylor1998; Lancaster, Reference Lancaster2008; Van der Veen, Reference Van der Veen2011). These findings also underscore the political relevance of political trust (Zmerli and Van der Meer, Reference Zmerli and Van der Meer2017) and suggest that consistently low levels of political trust may help to explain consistently low levels of support for foreign aid among the American mass public.
1. Why political trust matters for foreign aid support
At first glance, it might seem unlikely that trust in the national government, referred to here as political trust, would matter in shaping public opinion toward foreign aid spending. Indeed, this would jibe with majority of work on this subject, the policy consequences of political trust, focusing on domestic issues (Devine, Reference Devine2024). Notable exceptions to this approach include studies focusing on [US] public opinion toward military force during the 9/11-era War on Terror and the Afghanistan and Iraq wars (Hetherington and Husser, Reference Hetherington and Husser2012), immigration (Macdonald, Reference Macdonald2021), and free trade (Macdonald, Reference Macdonald2024), although all of these policies arguably have roots in domestic affairs as well, particularly the latter two issues of immigration and trade.Footnote 3 In short, the vast majority of existing scholarship (but see, e.g., Heinrich et al., Reference Heinrich, Kobayashi and Lawson2021) has (1) focused exclusively on domestic issues, or (2) focused on how political trust matters for issues that arguably straddle both domestic and international politics.Footnote 4
While foreign aid is doled out for a variety of reasons, both self-interested and altruistic other-regarding in nature (Bueno de Mesquita and Smith, Reference Bueno de Mesquita and Smith2009; Bearce and Tirone, Reference Bearce and Tirone2010; Heinrich, Reference Heinrich2013; Chiba and Heinrich, Reference Chiba and Heinrich2019), and can ultimately serve the ends of both donor and recipient countries (e.g., McLean, Reference McLean2015; Bermeo and Leblang, Reference Bermeo and Leblang2015; Wang, Reference Wang2016; Heinrich et al., Reference Heinrich, Kobayashi and Long2018; Blair et al., Reference Blair, Marty and Roessler2022), I argue that ordinary mass publics are unlikely to view it that way. Indeed, while the benefits of US foreign aid may be clear and tangible to economic, military, and/or political elites, they are unlikely to be similarly apparent to the masses, the large majority of whom are only moderately political informed, at best (e.g., Delli Carpini and Keeter, Reference Delli Carpini and Keeter1997), particularly regarding international politics and global affairs.Footnote 5 In contrast, government spending in areas dealing with, for example, environmental protection and infrastructure may entail higher taxes and/or ultimately prove to be ineffective wastes of money, but can also yield clear, tangible benefits such as, for example, reduced pollution and fewer potholes, to the domestic [US] population.
In contrast, there is no such clear tangible benefit regarding foreign aid programs. Such spending, by definition, disproportionately benefits foreigners. As noted by then-budget director Mick Mulvaney in 2017, during the Trump administration's ultimately unsuccessful push to drastically cut US foreign aid, “the overriding message is fairly straightforward: less money overseas means more money spent here” (Shepardson, Reference Shepardson2017). More than two decades earlier in 1992, another right-wing populist candidate, Pat Buchanan, stated that “we simply do not want to fight other people's wars or use the tax dollars of our citizens to pay other nations’ debts” (Lauter, Reference Lauter1992). More recently in 2023, left-wing (on certain issues) populist candidate Robert F. Kennedy Jr. contrasted the “$113 billion committed to Ukraine” with the “57 % of Americans who can't put their hand on $1,000 if they have an emergency” (DeArment, Reference DeArment2023), implying that US tax dollars would be better served if they were spent domestically rather than abroad.
Although foreign aid can help donor countries to advance their various policy objectives, government spending on this program is fundamentally centered around US tax dollars are being spent to benefit a non-US population. In other words, the costs of foreign aid are borne by the domestic US population, while the benefits of such government largess are, I argue, in the minds of ordinary citizens, conferred upon a non-domestic population. This is not to say that such aid never flows to a US ally or a group toward whom most people would view sympathetically, but foreign aid is inherently different from domestic spending programs, e.g., on infrastructure, health, childcare, education, the environment, etc., in that the perceived cost–benefit balance is fundamentally tilted away from the domestic population. In other words, foreign aid spending entails costs and risks for the mass public, e.g., potentially wasteful and ineffective spending that comes without any clear, tangible benefits to offset them.
Regardless of the merits of US foreign aid and assistance to other countries, these messages echo a potentially powerful argument against it, that foreign aid imposes costs upon us and confers benefits upon them (Kinder and Kam, Reference Kinder and Kam2009). Furthermore, foreign aid could be ineffective in achieving its goals, end up entrenching dictators rather than promoting democracy, and/or be taken advantage of by corrupt actors rather than flowing the intended target group (e.g., Kono and Montinola, Reference Kono and Montinola2013; Bermeo, Reference Bermeo2016; DiLorenzo, Reference DiLorenzo2018). While every government policy has a non-zero chance of being ineffective and/or detrimental, foreign aid stands out in that it does not, I argue, in most ordinary people's minds, counterbalance this with potential benefits to the domestic population.
This is not to say that the American mass public, nor mass publics anywhere, are universally opposed to foreign aid. Indeed, myriad survey data shows that they are not. Rather, I am arguing that the nature of this policy entails, for ordinary people, a cost–benefit imbalance that can “activate” political trust (Hetherington, Reference Hetherington2005). By this, I mean make it “matter” for people's decision calculus when they are considering their support for said policy. This dynamic has been demonstrated widely, including for affirmative action (Hetherington and Globetti, Reference Hetherington and Globetti2002), gun control (Ryan et al., Reference Ryan, Andrews, Goodwin and Krupnikov2022; Hansen and Seppälä, Reference Hansen and Seppälä2023), economic redistribution (Rudolph and Evans, Reference Rudolph and Evans2005; Macdonald, Reference Macdonald2020), and health insurance reforms that entail “more government” (Hetherington, Reference Hetherington2005; Hetherington and Rudolph, Reference Hetherington and Rudolph2015). Political trust has also been shown to shape US public opinion toward immigration (Macdonald, Reference Macdonald2021), free trade (Macdonald, Reference Macdonald2024), and the use of military force (Hetherington and Husser, Reference Hetherington and Husser2012). Here, I seek to extend this to foreign aid, something that has not been clearly demonstrated by extant research. What all of these aforementioned policies have in common, unlike for example, receiving cash benefits from the government, such as the stimulus checks that most Americans received during the Covid-19 pandemic, or universal old-age pension programs such as Social Security and Medicare, is that they entail risks and costs, i.e., they may have detrimental societal impacts, without conferring clear and tangible benefits to wide swaths of the population. As such, it is natural for people to exhibit some degree of skepticism about supporting such policies. One thing that can help them to overcome such skepticism is whether they trust government, the actor responsible for enacting and implementing said policies. In other words, people's support for policies like foreign aid should be determined, in part, by how much they trust and confidence they have in government.Footnote 6
In sum, my argument is that foreign aid is a policy that should “activate” political trust. When policies entail costs and risks but confer few clear, tangible benefits, people are likely to be skeptical about giving government the flexibility and leeway, via public support, to enact and implement said policies. However, when people are more trusting of government, they will be more likely to give government the “benefit of the doubt” and support an increased role for government, even when the policy seems unlikely, on its face, to clearly and directly benefit themselves, their social groups, and/or their country. I articulate my formal hypothesis as follows and discuss my associated research design below.
Hypothesis:
Political trust is positively associated with support for foreign aid spending.
2. Data and methods
My main data source is the Cumulative American National Election Studies (ANES), a nationally representative survey of the American mass public that has long been recognized as a “gold standard.” The ANES also has valid measures of support for foreign aid spending and of political trust, as well as a rich battery of theoretically appropriate control variables. Unfortunately, the ANES does not ask about foreign aid spending in every year and has not done so for several election years. As such, I use data from the 1996, 2000, 2004, and 2008 ANES. I chose these years for several reasons. One is because, at the time of this writing, 2008 was the last year that the ANES asked about foreign aid spending. The second is because these election years include periods of peace (1996 and 2000) and war (2004 and 2008), and span both two-term Democratic (Bill Clinton) and Republican administrations (George W. Bush). Pooling data from these four election years also yields a sufficient sample size to validly test my hypotheses.
I supplement these pooled cross-sectional analyses of the 1996–2008 ANES with panel data from 1994–1996 and 2002–2004 ANES to demonstrate that political trust appears to drive attitudes toward foreign aid, rather than the reverse, and panel data from three GSS panel studies spanning 2006–2014 to show that the results hold even with an extremely strict test that accounts for individual fixed effects (dummy variables for each survey respondent) as a means of better accounting for unobserved pre-adult socialization factors. In the following sections, I discuss the data for main (1996–2008 ANES) analyses in detail. I discuss the data for my supplemental analyses (panel data from the 1994–1996/2002–2004 ANES and the combined 2006–2014 GSS) in later sections.Footnote 7
2.1 Dependent variable
My dependent variable (for my 1996–2008 ANES analyses) is support for foreign aid spending. This question is asked of ANES respondents as follows: should federal spending on foreign aid be increased, decreased, or kept about the same?. I code responses as follows (1 = decrease; 2 = same; 3 = increased; mean = 1.61). Given that my dependent variable is ordinal in nature, I employ an ordered probit regression model for my main analyses.
2.2 Main independent variable
My main independent variable (for my 1996–2008 ANES analyses) is trust in the national government, which I refer to as political trust. This has commonly been measured in the ANES with four questions, all of which are intended to capture people's general attitudes toward government rather than views toward any one political figure or specific branch of government. The four questions are as follows: (1) how much of the time do you think you can trust the government in Washington do what is right?, (2) would you say the government is pretty much run by a few big interests looking out for themselves or that it is run for the benefit of all the people?, (3) do you think that people in the government waste a lot of money we pay in taxes, waste some of it, or don't waste very much of it?, and (4) do you think that quite a few of the people running the government are crooked, not very many are, or do you think hardly any of them are crooked?. Consistent with past scholarship, I sum responses to these four questions into an additive index (α = 0.637; mean = 0.350) and re-scale it to range between 0 and 1, with lower (higher) values indicating lower (higher) levels of trust in the national government.Footnote 8
2.3 Control variables
For my main (1996–2008 ANES) analyses, I account for a theoretically appropriate battery of control variables, i.e., individual-level demographic and attitudinal factors that may correlate with both political trust and attitudes toward foreign aid spending. To identify such variables, I draw on past studies (employing observational research designs) that have identified various correlates of public opinion toward foreign aid spending (e.g., Kam and Kinder, Reference Kam and Kinder2007; Van Heerde and Hudson, Reference Van Heerde and Hudson2010; Paxton and Knack, Reference Paxton and Knack2012; Bauhr et al., Reference Bauhr, Charron and Nasiritousi2013; Kim, Reference Kim2013; Bayram, Reference Bayram2017; Alvarez et al., Reference Alvarez, Boussalis, Merolla and Peiffer2018; Heinrich et al., Reference Heinrich, Kobayashi and Lawson2021; Prather, Reference Prather2024). I also draw upon additional [observational] studies that focus on the determinants of mass attitudes toward foreign policy issues more broadly (e.g., Hurwitz and Peffley, Reference Hurwitz and Peffley1987; Brewer et al., Reference Brewer, Gross, Aday and Willnat2004; Hetherington and Husser, Reference Hetherington and Husser2012; Rathbun et al., Reference Rathbun, Kertzer, Reifler, Goren and Scotto2016). Collectively, these illustrative (but not exhaustive) studies, along with the variables available in the ANES, lead me to control for a combination of demographic and attitudinal factors.
To account for people's socialization and life experiences, I control for the following demographics: age (in years), gender (female versus male), race (White, non-Hispanic versus not), whether both of a respondents’ parents were born in the United States (yes versus no), and Census region of residence (Northeast versus Midwest versus South versus West). To help account for socio-economic status, I control for formal education (4-college degree or higher versus not), home ownership (yes versus no), marital status (married versus not), and household income (upper tercile versus not). To help account for exposure to news about national and/or international affairs, I also control for whether people report reading a newspaper every day (yes versus no).
In terms of attitudinal variables that may correlate with both political trust and foreign aid support, I control for partisanship (7pt; strong Democrat → strong Republican) and ideology (7pt; extremely liberal → extremely conservative) to help account for ordinary Americans’ general left-right identities and long-standing orientations to politics. Beyond these, I also control for the core political values of traditionalism (4-item index; α = 0.619) and egalitarianism (6-item index; α = 0.696). Traditionalism is coded so that higher values reflect a stronger desire for conformity, aversion to new ideas and cultures, and the need to guard against external threats. Egalitarianism is coded so that higher values reflect a stronger desire for cooperation, less rigid group hierarchies, and the need to ameliorate social and economic inequities. In addition to these core values, I also control for the general foreign policy orientation of isolationism (a binary measure), and social trust (can't be too careful versus most people can be trusted). This helps to ensure that my measure of political trust is not simply capturing a willingness to trust other countries and/or other people. I also account for feelings toward poor people (cold → warm), as this may reflect a salient group, in ordinary Americans’ minds who are the likely intended target group of foreign aid spending. In other words, people who feel more favorably toward the poor should be more likely to support policies intended to benefit such individuals (Piston, Reference Piston2018).
Additionally, I account for people's perceptions of the national economy in the past year (5pt; much worse → much better) and of their personal financial situations in the past year (5pt; much worse → much better), as people may be more politically trustful, favor “more government,” and exhibit less antipathy toward foreigners during economic “good times” (e.g., Hetherington and Rudolph, Reference Hetherington and Rudolph2008; Heinrich et al., Reference Heinrich, Kobayashi and Bryant2016; Rueda, Reference Rueda2018; Wlezien and Soroka, Reference Wlezien and Soroka2021; Hopkins et al., Reference Hopkins, Margalit and Solodoch2024). Finally, I account for year fixed effects (dummy variables for 1996 versus 2000 versus 2004 versus 2008). These help to account for factors such as real-world economic conditions, partisan control of the federal government, and whether the United States was at war. I re-scale these variables (except for age and the fixed effects) to either be dichotomous (0 versus 1), or continuous (ranging 0–1).Footnote 9
3. Main results
I present my main results in Table 1. These are pooled cross-sectional analyses, using an ordered probit regression model, of the 1996–2008 ANES. Overall, the results show that political trust (in the US national government) is positively associated, to a large and statistically significant degree, with support for federal spending on foreign aid. This is consistent with my hypothesized expectations. Indeed, the only independent variable with a larger z-statistic (coefficient/standard error) than political trust is isolationism (z = 8.20 for political trust versus z = −9.60 for isolationism). Beyond these two largest (in terms of magnitude) variables, I also find that race, ideology, core political values, feelings toward poor people, and evaluations of the national economy in the past year are meaningfully and significantly associated with support for foreign aid spending. These findings are theoretically sensible and suggest that this regression model is appropriately specified.
Political trust and support for spending on foreign aid, 1996–2008
Note: Dependent variable is support for federal spending on foreign aid (1 = decreased; 2 = same; 3 = increased). All independent variables either range between 0 and 1 or are dichotomous (0 versus 1), except for age. Ordered probit coefficients with robust standard errors in parentheses.
Source is the 1996–2008 ANES.
***p < 0.01, **p < 0.05, *p < 0.1, two-tailed test.
Given that these are ordered probit coefficients, the results in Table 1 are not directly interpretable. As such, I graphically illustrate, in Figure 1, my main relationship of interest (between domestic political trust and US public opinion toward foreign aid spending). I do so by plotting the predicted probabilities of each ordinal outcome (1 = decrease; 2 = same; 3 = increase) from Table 1 across the observed range (0–1) of political trust, holding all of the control variables constant at their observed values (Hanmer and Kalkan, Reference Hanmer and Kalkan2013).
Political trust and attitudes toward spending on foreign aid, 1996–2008. Note: Figure shows the predicted probability of supporting decreased versus kept about the same versus increased federal spending (1–3) on foreign aid across the observed range of political trust. Based on the ordered probit regression model in the first column of Table 1. All control variables are held constant at their observed values. The small black bars represent a histogram showing the distribution of political trust. The thick lines represent the predicted probabilities for each ordinal outcome. The thin dashed lines represent the 95 percent confidence intervals (robust standard errors, two-tailed test).
The results in Figure 1 show that a shift from the minimum to maximum level of trust in the national government (0 → 1) is associated with an approximately 0.32 decrease in the probability of supporting decreased federal spending on foreign aid, from 0.63 to 0.31. This is both statistically and substantively significant, i.e., these results show not only that the coefficients are different from zero, but also that domestic political trust does indeed appear to be a powerful driver of mass opinion toward foreign aid spending.Footnote 10
The results in Figure 1 also show that lower support for cutting spending on foreign aid (the modal outcome in the ANES sample) comes along with greater support for both maintaining spending levels (the second most common value) and increasing federal spending on foreign aid (the least common choice of survey respondents). On average, holding the other control variables constant at their observed values, a shift in political trust from its minimum to its maximum (0 → 1) is associated with a 0.16 increase in the probability of keeping foreign aid spending “about the same,” from 0.33 to 0.49, and a 0.15 increase in the probability of favoring increased foreign aid spending, from 0.05 to 0.20. These results are also both statistically and substantively significant.Footnote 11
In the Supplementary Appendix (see Tables B1–B4), I run additional models to show that the main results are similar if I restrict the sample to non-Hispanic Whites and control for ethnocentrism (Kam and Kinder, Reference Kam and Kinder2007), that the results hold if I include a measure of authoritarianism (Hetherington and Suhay, Reference Hetherington and Suhay2011), that the results hold when controlling for feelings toward Muslims (Sides and Gross, Reference Sides and Gross2013), and when I use an alternative measure of political trust, a feeling thermometer rating (cold → warm) of the federal government.Footnote 12
3.1 Heterogeneity in the relationship
My theoretical argument, which draws on the logic of Hetherington and Globetti (Reference Hetherington and Globetti2002), and various subsequent related work on how political trust shapes mass opinion across a range of policies (see Devine, Reference Devine2024, for a review), is that foreign aid spending is the type of policy that should, as a result of an imbalance, in the public mind, between who receives the benefits/largess and who bears the costs/risks of foreign aid spending, serve to “activate” political trust. By this, I mean make attitudes toward government, the actor who I argue is, in ordinary people's minds, most responsible for crafting and implementing foreign aid programs, relevant when ordinary people are considering their attitudes toward said policy.
However, this line of work also notes that political trust should not matter equally for everyone. For example, political trust matters little for the attitudes of people who receive the benefits of a public policy; they will be likely to support it even if they are generally distrustful of government. However, for people who bear the costs of said policy and/or do not receive the benefits, they need additional reasons to overcome their skepticism and support said policy; one such factor is being trustful of government. The logic is similar for liberals (conservatives) being asked to support conservative (liberal) policies, or Democrats (Republicans) being asked to support a policy supported by Republican (Democratic) Party elites, i.e., political trust matters more (less) for policy support when people are (not) asked to make material, ideological, and/or partisan “sacrifices” by supporting a policy that does not yield clear (broadly defined) benefits.
Building on this idea, it seems reasonable to assume that people who hold isolationist views toward foreign policy, meaning that they are less inclined to support the United States taking an active role in world affairs (e.g., Hurwitz and Peffley, 1987), will have fewer reasons to support foreign aid spending that their non-isolationist counterparts, who presumably would derive some benefits, i.e., the policy would comport with their general political orientations. In other words, people who simply believe that the United States would generally be better off if we turned inward (a non-trivial minority of the mass public), may need to be especially trustful of government in order to support foreign aid spending. For people who eschew isolationism, it seems plausible that political trust would matter less in terms of their support for government spending on foreign aid.
I test this in Table 2 by regressing support for foreign aid spending (1–3) on an interaction between political trust (ranges 0–1) and isolationist sentiment (0 versus 1), along with the same control variables as in Table 1. If political trust matters (in terms of shaping attitudes toward foreign aid spending) more for those who exhibit isolationist sentiment than for those who do not, then we would observe a positive and statistically significant interaction term. Overall, the results in Table 2 show that political trust matters for both isolationist and non-isolationist Americans, but also that its “effect” is stronger for the former group. This is demonstrated by the positive and significant coefficient for political trust (when “isolationist” is at a value of “0”), as well as by the positive and statistically significant interaction term. In short, political trust appears to matter broadly in terms of shaping public opinion toward foreign aid spending.Footnote 13
Political trust, isolationism, and support for spending on foreign aid, 1996–2008
Note: Dependent variable is support for federal spending on foreign aid (1 = decreased; 2 = same; 3 = increased). Political trust ranges 0–1; isolationist sentiment is dichotomous (0 versus 1). Includes the same control variables as in column 1 of Table 1. Ordered probit coefficients with robust standard errors in parentheses.
Source is the 1996–2008 ANES.
***p < 0.01, **p < 0.05, *p < 0.1, two-tailed test.
3.2 Assessing directionality
Thus far I have shown evidence of a positive and substantively significant relationship between political trust and public support for federal spending on foreign aid (Table 1). I have also shown that this relationship is present among both the minority of Americans who hold isolationist foreign policy orientations and among the majority who eschew isolationism, although political trust is more meaningful, i.e., it is larger in magnitude among the former group of Americans (Table 2). These results come from pooled cross-sectional regressions, in which my variables of interest are all measured contemporaneously. As such, it is difficult to establish the direction of “causality,” i.e., whether, as I have argued, political trust drives attitudes toward foreign aid spending, or whether people adjust their level of trust in the national government to justify their support (or lack thereof) for foreign aid.
While existing work has shown, via the use of panel data, that the “causal arrow” appears to run from political trust to policy preferences, rather than the reverse (Hetherington and Globetti, Reference Hetherington and Globetti2002; Hetherington and Husser, Reference Hetherington and Husser2012; Macdonald, Reference Macdonald2021, Reference Macdonald2024), it is still prudent to test this empirically where possible. I do so in Table 3 with panel data (the same respondents interviewed at multiple points in time) from the 1994–1996 and 2002–2004 ANES. Employing the same measures of political trust and foreign aid spending opinion as in my main results (Table 1), I use a series of cross-lagged ordinary least squares (OLS) regression models to test whether past values of political trust significantly predict future values of foreign aid spending support, holding past values of foreign aid spending support constant, or whether the reverse is more likely to hold true.
Cross-lagged relationship between political trust and support for spending on foreign aid, 1994–1996 and 2002–2004
Note: Shows the cross-lagged relationship (t1 = 1994/2002; t2 = 1996/2004) between political trust (as a dependent variable in 1996/2004; as an independent variable in 1994/2002) and support for foreign aid spending (as a dependent variable in 1996/2004; as an independent variable in 1994/2002). Political trust ranges 0–1 (low → high), support for foreign aid spending ranges 0–1 (decreased → same → increased). OLS coefficients with robust standard errors in parentheses.
Sources are the 1994–1996 and 2002–2004 ANES panels.
***p < 0.01, **p < 0.05, *p < 0.1, two-tailed test.
Overall, the results in Table 3 show that political trust significantly predicts support for foreign aid spending, even when controlling for past values of foreign aid spending attitudes. This is evidenced (in columns 2 and 4) by the statistically significant and substantively large coefficients for political trust. In contrast, there is little evidence to suggest that the reverse hold true, i.e., that people meaningfully adjust their degree of trust in the national government to bring them in line with their attitudes toward foreign aid spending. This is evidenced (in columns 1 and 3) by the substantively small and marginally significant (at best) coefficients for attitudes toward foreign aid. In short, the results in Table 3 show that political trust drives mass opinion toward foreign aid spending, rather than the reverse.
3.3 Accounting for unobserved factors
Thus far I have shown that political trust is associated, to a positive and substantively significant degree, with support for government spending on foreign aid. I have also taken pains to ensure that this relationship is robust. By controlling for a battery of theoretically appropriate variables that may correlate with both political trust and attitudes toward foreign aid spending, and by employing panel data to address “reverse causality,” I can also assuage some concerns about endogeneity and reduce the likelihood that my findings are simply a “false positive.” However, my research design is observational in nature and thus cannot account for certain unobserved variables such as pre-adult socialization processes, underlying the relationship between political trust and policy preferences (e.g., Peyton, Reference Peyton2020).
To better address such concerns, I employ three-wave panel data from the GSS. By using three waves (people interviewed at three different points in time), I am able to control for individual-level fixed effects, i.e., a dummy variable for each survey respondent. This is an exceptionally powerful research design (Allison, Reference Allison2009) because it is able to account for stable, unobserved factors such as people's pre-adult socialization experiences, that may correlate with both political trust and attitudes toward foreign aid spending. While not a panacea, it is, by examining only within-person variation, a much stronger test than a cross-sectional analysis (e.g., Lee and Mutz, Reference Lee and Mutz2019).
I conduct such a test in Table 4. I do so by combining data from the three GSS panel studies (2006–2008–2010, 2008–2010–2012, and 2010–2012–2014).Footnote 14 Pooling across these GSS panels and restricting my sample to respondents who answered questions about foreign aid spending and political trust in all three waves, results in a valid sample size of 6,984 (2,328 respondents × 3 interviews). This permits me to examine within-person variation (t1 → t2 → t3) and to include fixed effects (dummy variables) for each of these 2,328 individuals. While this cannot account for every possible factor that may correlate with both political trust and foreign aid opinion, it does provide an exceptionally strong test of this relationship.
Individual fixed effects model of political trust and support for spending on foreign aid/Assistance to other countries, 2006–2014
Note: Shows the relationship between political trust (ranges 0–1) and support for cutting spending on foreign aid/assistance to other countries (0 = no; 1 = yes). Model also controls for individual fixed effects (a dummy variable for each survey respondent). These variables and the constant term are not displayed here. OLS coefficients from a linear probability model with robust standard errors clustered by individual in parentheses.
Sources are three GSS panel studies spanning 2006–2014 (06-08-10; 08-10-12; 10-12-14).
***p < 0.01, **p < 0.05, *p < 0.1, two-tailed test.
The GSS uses two questions measure attitudes about foreign aid, my dependent variable of interest. The first is posed to respondents as follows are we spending too much, too little, or about the right amount on foreign aid?. The second is posed to respondents as follows are we spending too much, too little, or about the right amount on assistance to other countries?. I combine responses to these questions into a single measure of foreign aid opinion and re-code responses to be dichotomous (0 = too much/right amount; 1 = too little).Footnote 15 Given this coding scheme, where higher values indicate support for cutting foreign aid, the coefficient for “political trust” should be negative, meaning that as trust in government increases (decreases), people should be less (more) likely to agree that we are spending “too much” on foreign aid/assistance to other countries.Footnote 16
Unfortunately, the GSS lacks the same questions as the ANES to measure political trust, but it does have an appropriate substitute that asks about confidence in different institutions, and one that has been used by past work to measure the concept of political trust (e.g., Catterberg and Moreno, Reference Catterberg and Moreno2006). To measure political trust in the GSS, which is conceptualized as general attitudes toward government as a whole, I combine two questions into an index. These questions are posed to respondents as follows (the text in brackets varies), as part of a broader battery about confidence in different institutions As far as the people running these institutions [Executive branch of the federal government; Congress] are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them?. The valid responses to both of these questions range from 1 to 3 (1 = hardly any; 2 = only some; 3 = a great deal) and thus the additive index ranges from 2 to 6; I re-scale this variable to range between 0 and 1, with lower (higher) values indicating lower (higher) levels of domestic political trust.
My “control” variables in this model are individual-level fixed effects, meaning a dummy variable to represent each of the 2,328 GSS respondents in this dataset. As previously discussed, this is an exceptionally powerful control variable and thus a strict test of the relationship between political trust and attitudes toward foreign aid. I also cluster my standard errors by respondent to account for repeated observations (each person was surveyed three times) among units of analysis (Cameron and Miller, Reference Cameron and Miller2015).
Overall, the results in Table 4 show that, on average, if a person's level political trust shifted from the minimum observed value to the maximum (from 0 to 1), they would be approximately 15.3 percentage points less likely to agree that we are spending “too much” on foreign aid/assistance to other countries, with a decrease from 0.73 to 0.58.Footnote 17
In short, the results in Table 4 are robust to an exceptionally strict statistical test. They also manifest even when employing what I view as a less ideal measure of political trust (compared to the 4-item ANES scale). Overall, this further demonstrate the robustness and validity of my main findings, yielding additional evidence of a substantively significant relationship between political trust and public opinion toward foreign aid spending.
4. Conclusion and political implications
Political trust and its dynamics have long been of interest to the study of political systems (Easton, Reference Easton1975; Citrin and Stoker, Reference Citrin and Stoker2018). It has also been shown to matter for mass politics in numerous ways (Wright and Winters, Reference Wright and Winters2010). Here, I have shown that it also has important consequences for public opinion toward foreign aid spending. These findings not only underscore the importance of political trust in shaping mass opinion (Hetherington and Rudolph, Reference Hetherington and Rudolph2015; Devine, Reference Devine2024), but also suggest that public attitudes toward foreign aid are not simply reflections of elite rhetoric (Kertzer and Zeitzoff, Reference Kertzer and Zeitzoff2017). If they were, then we would observe much higher support for foreign aid spending, consistent with a long-standing post-World War II elite consensus, albeit certainly not a unanimous one, in favor of foreign aid. Accordingly, low political trust can help to explain why mass support has long remained low, relative to other types of spending, among the American mass public. While public opinion on foreign aid spending is clearly shaped by stereotypes of potential recipient states (e.g., Baker, Reference Baker2015; Heinrich and Kobayashi, Reference Heinrich and Kobayashi2020), and the predispositions that underlie ordinary people's general willingness to favor (or disfavor) people different from oneself (e.g., Kam and Kinder, Reference Kam and Kinder2007; Prather, Reference Prather2024), I have shown that it is also meaningfully shaped by the degree to which people exhibit confidence and trust in their national government.
The real-world relevance of these findings can be illustrated with data from the 2022 CES, which fortuitously asked questions about both political trust (in the national government) and public attitudes toward extending foreign aid to Ukraine. These data show that among the majority of Americans (just over six in ten) who report having “none at all” or “not very much” trust in the federal government when it comes to handling the nation's problems, 45.1 percent favor sending food, medicine, and other aid to affected countries [by Russia's invasion of Ukraine] and just 32.1 percent favor providing arms to Ukraine. The analogous numbers among the minority of Americans (just under four in ten) who report having “a fair amount” or “a great deal” of trust in the federal government when it comes to handling the nation's problems were 58.7 percent (aid to affected countries) and 51.0 percent (arms to Ukraine). While public opinion may not decisively determine whether major world powers such as the United States provide additional aid to Ukraine, a country facing, at the time of this writing, a daunting Russian military invasion, higher public support, signaling a friendlier “pro-aid” domestic environment, would almost certainly make it easier to achieve this legislatively (Milner and Tingley, Reference Milner and Tingley2016). Moreover, a lack of strong public support for aiding Ukraine economically and/or militarily, which, I argue, is driven partially by low levels of political trust, could provide ammunition to politicians who oppose the continuation and/or expansion of such aid.
Alternatively, the successful provision of foreign aid and/or positive media framing of such spending can potentially improve politicians’ standing among the voting public (e.g., Heinrich and Kobayashi, Reference Heinrich and Kobayashi2020; Tobin et al., Reference Tobin, Schneider and Leblang2022). Rather than low and declining levels of political trust fueling greater public opposition to foreign aid spending, successful (be it real or perceived) aid provision could potentially bolster political trust (Hetherington, Reference Hetherington1998; Hetherington and Rudolph, Reference Hetherington and Rudolph2008), which, as demonstrated here, can increase mass support for spending on foreign aid, and afford national governments greater leeway and flexibility to use the economic, humanitarian, and/or military aid as a means of achieving their various foreign policy objectives.
In terms of future work, it would be worthwhile to examine whether political trust matters less for policy support when spending is discussed in specific rather than general terms (e.g., Jacoby, Reference Jacoby2000), given that people tend to be more supportive of foreign aid spending when it is portrayed in way that emphasizes specific goals such as disaster relief, food and medical assistance, and aid for women's education in needy countries rather than presented as “foreign” aid in the abstract (Wojtowicz and Hanania, Reference Wojtowicz and Hanania2017).
It would be beneficial for future work to extend this analysis to other donor states beyond the United States, including both Western-aligned multi-party democracies such as Germany, France, Japan, and Australia, as well as authoritarian regimes such as China and Russia, the difficulties of measuring public opinion in such contexts notwithstanding (e.g., Chia, Reference Chia2014). Future work would also do well to consider when political trust is a stronger driver of attitudes toward foreign aid spending. For example, it would be valuable to test whether political trust matters more, in terms of shaping political attitudes, during economic downturns or when a country is experiencing an influx of refugees. In such instances, the idea of sending tax dollars overseas rather than spending them “at home” may seem “riskier” and less appealing and thus potentially more likely to “activate” political trust among the mass public. This latter phenomenon is especially relevant in Europe, where the Russian invasion of Ukraine has drastically upended the politics of migration, security, and alliances (Bergmann et al., Reference Bergmann, Toygür and Svendsen2023). Such questions could benefit from a future cross-national study, e.g., variation in potential conditioning factors, both over time and across countries.
This study has focused on the American case, which is important to do given the uniquely powerful role that the United States plays in global politics. It is also important to note that the data I examine here (ANES and GSS) spans 1994–2014. This predates the 2016 election of Donald Trump, a far less pro-democracy internationalist figure than other early 21st-century US presidents (George W. Bush, Barack Obama, and Joe Biden). Given increased right-wing isolationist sentiment during the Trump era, a notable departure from recent decades, it is important to test, in future work, whether the relationship between domestic political trust and foreign aid spending now has a strong partisan component.
The US-led post-World War II global order, of which foreign aid is an important component, depends, in part, on domestic political forces (e.g., Milner and Tingley, Reference Milner and Tingley2010; Greene and Licht, Reference Greene and Licht2018; Heinrich et al., Reference Heinrich, Kobayashi and Long2018; Dietrich et al., Reference Dietrich, Milner and Slapin2020). One such force is public opinion. As such, the findings in this paper suggest that we should also pay greater attention to the factors that underlie mass opinion regarding foreign aid spending. Here I have demonstrated, via analyses of cross-sectional and panel survey data in the United States, that one such important factor is the extent to which ordinary people exhibit confidence and trust in their national government.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2024.70.
To obtain replication material for this article, https://doi.org/10.7910/DVN/JPRYWS
Open access
