The Effect of Geostrategic Competition on Public Attitudes to Aid

Abstract China’s rise is ushering in a new era of geostrategic contestation involving foreign aid. In many traditional OECD donors, aid policy is changing as a result. We report on a survey experiment studying the impacts of rising Chinese aid on public opinion in traditional donors. We randomly treated people with vignettes emphasising China’s rise as an aid donor in the Pacific, a region of substantial geostrategic competition. We used a large, nationally-representative sample of Australians (Australia is the largest donor to the Pacific). As expected, treating participants reduced hostility to aid and increased support for more aid focused on the Pacific. Counter to expectations, however, treatment reduced support for using aid to advance Australian interests. These findings were largely replicated in a separate experiment in New Zealand. Knowledge of Chinese competition changes support for aid, but it does not increase support for using aid as a tool of geostrategy.


Introduction
"I've changed, and I think he's changed, and it is all about China." Representative Ted Yoho speaking about his and President Trump's newfound support for foreign aid (cited in New York Times 2018).
Rising Chinese aid is a source of concern amongst political elites in traditional OECD donor countries (for example, Bolton 2018; New York Times 2018). One result has been increased attention devoted to these countries' own foreign aid (New York Times 2018). This renewed focus has not, primarily, been motivated by the nominal reason for aid giving: helping people in developing countries. Rather, interest has been driven by geostrategy.
Although attitudes to the rise of Chinese aid amongst political elites and policymakers have been well documented, to date, there has been little study on the impact of China's rise on public opinion about aid. Yet public beliefs matter. Research shows public preferences influence aid volumes Tingley 2010, 2013;Stern 1998), as well as the purpose aid, is given for (Bueno de Mesquita and Smith 2009).
In this paper, we report on the findings of two pre-registered, survey experiments conducted to test the impact of awareness of China's rise as an aid donor on public opinion. 1 The experiments involved providing participants in treatment groups with vignettes emphasising China's rise as an aid power. In doing this, we contribute to the small literature on factors shaping public preferences about aid. In particular, we add to understanding about how the public reacts to the perceived geostrategic competition born of China's rise.
The setting for the vignettes used in our experiment is the group of aiddependent small island developing states found in the central and southern Pacific. These Pacific countries (hereafter referred to simply as 'the Pacific') serve as a particularly pertinent context for our experiments, as Chinese aid to their region has been the catalyst for heightened geostrategic tensions (Citowicki 2020). China has risen rapidly as an aid donor in the Pacific and is now the third-largest donor to the region (Dayant and Pryke 2019). China's rise has elicited clear responses from traditional donors. Aid from the USA and Japan, which had been declining, is now increasing again (Dayant and Pryke 2019). Aid to the Pacific from New Zealand is also rising, partially in response to China's presence (Cabinet National Security Committee 2018). Influenced by China, Australia has launched a new aid lending facility to the region, while the United Kingdom has renewed its aid focus on the Pacific (Anders and Cornish 2018). Such geostrategic contestation is far from unique to the Pacific. However, the Pacific provides a setting where China's rise and impact on other aid donors is particularly clear.
The first of the two survey experiments involved a large, representative sample of Australians. We then replicated the findings in an independent experiment conducted in New Zealand. Australia and New Zealand were chosen because, respectively, they are the largest and second-largest aid donors to the Pacific (Dayant and Pryke 2019). 2 Both countries exist in close proximity to the region and have a heavy aid focus on the Pacific (OECD DAC 2018). Beyond the Pacific focus, both countries are fairly typical OECD donors: in 2018, the most recent year with data, each was near the OECD median in terms of aid generosity (OECD DAC 2018). 3 1 Both studies were pre-registered with the AEA RCT Registry. (Australian experiment: AEARCTR-0004249; New Zealand experiment: AEARCTR-0005055). 2 These rankings stem from data provided by Dayant and Pryke (2019) who include both traditional donors and China. Rankings come from 2017, the most recent year with data. Except where otherwise cited, all Pacific aid data in this paper and the experiments comes from Dayant and Pryke. Public support for aid in Australia and New Zealand also appears in line with that found in most OECD donors (Burkot and Wood 2017;Paxton and Knack 2012). While findings from Australia and New Zealand are not guaranteed to transfer to other OECD donors, both countries are typical enough for results to be relevant elsewhere.
In each experiment, we tested the effects of treatments involving Chinese aid in the Pacific on views about the donor's (Australia or New Zealand's) aid volume, the extent to which the donor should focus its aid on the Pacific, and the extent to which the donor should use aid to advance its own national interests, including geostrategic interests. In Australia, we also tested whether framing China's rise in a manner that was factually accurate but forceful had a greater impact on public opinion than more measured commentary.
Key findings from the experiment in Australia were thatas anticipatedboth treatments caused support for aid cuts to fall and increased desire for more aid to be focused on the Pacific. However, counter to our expectations, we did not find that telling people about China's rise as an aid donor in the Pacific increased the view that Australia's aid should be more heavily focused on advancing Australia's national interests. Rather, treatments shifted views in the opposite direction: support for focusing aid on advancing the national interest fell. At odds with the behaviour of political elites, the Australian public responded to China's rise by becoming less supportive of using aid to advance Australia's national interest.
Also counter to our expectations, we failed to find clear evidence that a more forceful framing of China's rise had a larger impact on views about aid than a measured framing did.
Other than the difference between forceful and measured framings, which we did not test, key findings were largely the same when we replicated the experiment in New Zealand.

Literature and hypotheses
Early theorists of foreign aid (most famously, Morgenthau 1962) contended aid was typically given for geostrategic reasons, or to advance donors' economic interests. In subsequent years, this theorising has been corroborated to an extent by observational studies of aid flows (Alesina and Dollar 2000;Bueno de Mesquita and Smith 2009;Dreher, Eichenauer, and Gehring 2016). Findings from recent experimental work on public opinion suggest public support for aid is motivated to a considerable extent by the national interest. Heinrich, Kobayashi, and Long (2018) found strong evidence that national interests trump moral motives for giving when perceived donor costs are high. Wood and Hoy (2018) found national interest treatments to be efficacious means of reducing support for aid cuts. And, although the findings of Hurst, Tidwell, and Hawkins (2017) are mixed, they found emphasising the national interest to be effective at increasing support for aid. Dietrich, Hyde, and Winters (2019) also found some evidence that emphasising donor benefits increases certain forms of support.
These findings led to our first three hypotheses: (H1) Informing people of China's rise as an aid donor will reduce support for aid cuts. The rationale for this hypothesis was that, in line with existing work, a rational response to the rise of a potentially hostile power in an adjacent region ought to be support for keeping aid flowing as a source of countervailing power. We focused on views about aid cuts, rather than aid increases, as studies have found views on aid cuts are more elastic to experimental treatments (Hurst et al. 2017;Scotto et al. 2017). (H2) Treating people with information on China's rise will increase the share of people wanting aid focused more on the Pacific. The rationale for H2 is similar to H1. If anything, we had stronger grounds to expect this changethe share of aid to the Pacific can be increased without an overall change in the aid budget or associated costs to the public purse. (H3) Treating people with information on China's rise will cause treated participants to become more likely to want aid focused foremost on the national interest.
The rationale for H3 is that, if public support for aid is motivated by concern for the national interest and using aid as a tool of geostrategy, information about a rising competitor should increase the desire to focus aid on the national interest. There is a clear precedent for this expectation based on recent academic work. In particular, in experimental work from Japan, Kohno et al (forthcoming) find evidence that providing information to participants about China's role as a rival aid donor reduces willingness to cut aid as a form of sanction against human-rights-violating regimes. They also provide evidence that this effect is primarily driven by participants' security concerns.
In the Australian experiment, we added one further hypothesis: (H4) Even when factually accurate, information on China's rise ought to be more effective in changing attitudes to aid when framed in a forceful manner, as opposed to more measured, balanced information.
In Australia, the tone of reporting on China's rise as an aid donor has varied considerably across media outlets. Sometimes reporting has been balanced, in other instances, it has been hyperbolic (contrast, for example, ABC News 2018; The Guardian 2018). The effects of framing are well documented (Chong and Druckman 2007) and are thought to affect attitudes to aid (Hudson and vanHeerde-Hudson 2012). This provided cause to anticipate a forceful framing of China's rise would have a greater impact on opinion.

Research Design
In May and June 2019, a socio-demographically representative sample of 2,001 voting-aged Australians received our questions as part of fortnightly Ipsos MORI online omnibus surveys. Respondents were randomly selected from a much larger pool of people who had registered to periodically participate in surveys and receive compensation in the form of shopping vouchers. Our survey questions and treatments were provided at the start of the surveys. To achieve our desired sample size, our questions were run across two rounds of consecutive surveys (people could not participate more than once). In each round, respondents were randomly allocated into one of three groups: a control group (that was provided no vignette) and two treatment groups. Randomisation was stratified by age and gender, ensuring balance across groups in both of these dimensions.
Both treatments involved vignettes containing factually accurate information as well as commentary from the Lowy Institute, an Australian foreign policy think tank. The Lowy Institute was chosen as the institute is Australia's most high-profile foreign policy think tank. The think tank is non-partisan and is a regular provider of analysis and media commentary.
Both vignettes were designed with the help of journalists. Both were nearly identical in length. The vignettes' key difference was in content. The first group received clear commentary about China's rise presented in a balanced manner. The second group received a more strident article. Both were similar in tone to types of commentary found in Australian media.
The first treatment group (referred to hereafter as the 'measured group') was provided with the following vignette: New research shows China is now an important aid donor to the Pacific

October 2018
China is now an important provider of foreign aid to the Pacific according to a new Lowy Institute study.
The Lowy Institute's work shows that in 2016, the most recent year with data, China was the fourth largest donor to Pacific Island countries, giving slightly more money than the United States, but less than Australia, New Zealand and Japan.
Commenting on the finding, Jonathan Pryke, Director of the Lowy Institute's Pacific Islands Program, said: "China is now in the top tier of aid donors to the Pacific. This will have economic and strategic ramifications. Aid can buy power. We need to keep things in perspective though. Australia still gives over five times as much aid to the Pacific as China does." 4 The second treatment group (hereafter the 'forceful group') was provided the following vignette: 4 In both the Australian and New Zealand experiments, we referred to the Pacific island aid recipient countries of interest simply as 'the Pacific'. While other aid recipients do border on the Pacific (for example, countries in Central America), in Australia and New Zealand, the term Pacific is exclusively reserved for the island states that were the focus of our experiments. Chinese aid buys power in the Pacific

October 2018
China is now one of the largest donors of foreign aid to the Pacific according to a new Lowy Institute study.
The Lowy Institute's work shows that in 2016, the most recent year with data, China gave over $100 million US dollars to Pacific Island countriesnearly twice as much as the United States gave in the same year. Since 2012, China has lent more to the Pacific than many major international organisations, including the World Bank.
Commenting on the finding, Jonathan Pryke, Director of the Lowy Institute's Pacific Islands Program, said: "China's influence as an aid donor has grown substantially in the Pacific. There's a real anxiety, now, that China will use this influence to pave the way for stronger strategic ties, including establishing permanent military bases in Australia's immediate region." Prior to receiving the treatments, respondents were asked questions about their background attributes and immediately after the vignettes, they were asked their views about foreign aid. Specifically, respondents were asked: whether Australia gives too much, the right amount, or too little aid; whether Australia should focus a greater or lesser share of its aid on the Pacific; and whether Australia's aid should primarily focus on advancing Australia's interests or helping people in developing countries. Don't know options were provided with all questions. Response orders were randomly flipped in all questions. Question order was randomised. The questions contained relevant information about current aid volumes and focus. All information in the questions was accurate. 5 Questions are included in online Appendix 1.
The New Zealand replication was run by the survey firm UMR from 26 November to 3 December 2019 and involved a socio-demographically representative sample of 1,147 New Zealanders. As in Australia, the survey was an omnibus survey and involved participants randomly selected from a large participant pool. Because of sample size constraints, only one treatment was used in New Zealand. This treatment was effectively the same as the measured treatment provided in Australia. Of the two treatments, the measured treatment was more appropriate for the New Zealand context, where a smaller media market means less variety in reporting tone, and where reporting has tended to be more of a kind with the measured treatment.
Mutatis mutandis, the same three questions were asked of New Zealand participants. The only differences stemmed from differences between the two donor countries (New Zealand focuses more aid on the Pacific, for example). The treatment and questions used in New Zealand are detailed in the online Appendices 2 and 3. 5 Information in the questions was sourced from Australian and New Zealand Governments' information on aid spending and OECD data (OECD DAC 2018).

Results and Discussion
To analyse responses, we created dummy variables, each coded one if the respondent provided the response of interest and zero otherwise, with 'don't knows' coded as missing. Table 1 provides information on the views of the control group (in the row 'Control group mean') alongside the experiment's treatment effects on responses to the three questions of interest. The table displays treatment effects in comparison to the control (columns (1), (3) and (5)) and the difference between the measured and forceful treatments (columns (2), (4) and (6)). Treatment effects come from OLS regressions without controls. (As we show in online Appendix 4, treatment and control groups are well balanced across most traits.) In online Appendix 6, we present alternative models including binary logistic and multinomial logistic regressions, as well as the effects when controls are added. Results are very similar across all models. The only difference of note is that, with controls included, the effect of the measured treatment on the national interest question is no longer statistically significant.
The treatments had the expected impact on responses to the first two questions. People became less likely to favour aid cuts, and the share of people who wanted more aid focused on the Pacific increased. The effects were statistically significant and substantively meaningful. In all three experiments, the forceful treatment shifted opinions by about 10 percentage points. These treatment effects are somewhat smaller than those found in comparable work in the USA (Hurst et al. 2017;Scotto et al. 2017) but similar to successful treatments in aid opinion work from Australia and the United Kingdom (Scotto et al. 2017;Wood 2018;Wood and Hoy 2018).
While both treatments succeeded in shifting opinion, some findings were unexpected. Although the forceful framing had a larger effect than the measured framing in all three areas, the differences were small and not statistically significant.
Even more unexpected was the direction of the treatments' effects on views about using aid to advance Australia's national interest. Figure 1 demonstrates the magnitude and direction of effects. As the chart shows, the treatments did not increase the share of Australians wanting aid focused foremost on the national interest. They had the opposite effect: support for the national interest fell. Although we do not show the findings here, the treatments had the concomitant effect of raising the share of respondents who wanted aid primarily focused on helping people in developing countries (see online Appendix 6.3). Table 2 contains the results from New Zealand. Views of the control group are shown as the control group mean. Table 2 also shows treatment effects derived in the same manner as in Australia. (As online Appendix 5 shows, the treatment and control groups were broadly balanced.) As can be seen in control group means, underlying appetites for aid cuts are lower than in Australia, while the desire for an increased Pacific focus is higher. However, when the measured treatment in Australia is compared with the treatment in New Zealand, the effects are similar. In New Zealand, the treatment has statistically significant effects on views on aid volume and focus on the Pacific. The magnitudes are similar to Australia and the effects are in the same direction. As in Australia, the treatment's effect on views  about whether aid should help advance New Zealand's interests has the opposite sign from expected. However, the effect is not statistically significant at conventional levels in New Zealand. In online Appendix 7, we show alternate regression models for the New Zealand experiment; findings are similar.

Conclusion
China's rise as an aid donor is changing the way political elites in OECD countries view development assistance. In this paper, we report on one of the first empirical studies to examine the impact of China's rise on public attitudes to aid. We found that, much like elite opinion, public views about aid are clearly influenced by China's rise. When treated with information about China, Australians and New Zealanders became less likely to want aid cut, and more likely to want their aid focused on the Pacific. In this, the public of both countries responded in a manner aligned with recent government actions. These findings also fit with existing academic work on motivations for aid giving. Yet we also found that, when they did change, public preferences in our experiments shifted against using aid to advance donor interests. This is at odds with elite behaviour, it also appears to be at odds with the findings of the one other academic study focused on China's rise and public attitudes to aid (Kohno et al. forthcoming).
One potential explanation for the difference in findings could be differing public attitudes to the security threat posed by China. In Kohno et al. (forthcoming), observed effects were highest amongst Japanese participants most concerned with the security threat to Japan posed by China. In our work, the forceful treatment, which spoke of China's rise in threatening terms, was not clearly more effective at changing views than the measured treatment. This provides some suggestive evidence that Australians and New Zealanders may be disinclined to view China as a threat to their own countries' interests. Rather, our findings raise an interesting possibility: that the public in Australia and New Zealand see China primarily as a threat to Pacific Island countries, and hence the nature of the changes we observed.
Whether this potential explanation is correct or not, will be a fruitful ground for future research. For now, we have added to the nascent literature on motivations for aid giving. We have clearly demonstrated that information on the rise of China changes public preferences about foreign aid. Yet, we have also shown that the rise of a potential geostrategic competitor does not inevitably increase the public's desire to see aid focused on the national interest.

Conflict of interest
All three of the paper's authors confirm they had no conflict of interest associated with the production of this paper.