The Impact of Economic Coercion on Public Opinion: The Case of US–China Currency Relations

Explaining Public Responses to External Threats

This section considers why foreign economic threats might influence public opinion. Scholars and policy makers disagree over (1) whether such threats are likely to increase or decrease support for policy change in a target state and (2) the mechanisms through which threats influence public opinion. Here, we contrast three approaches to this question, which we label interests, identity, and information. Each approach generates different hypotheses about the impact of external pressure on public opinion.

The interest-based perspective argues that individuals support policies when the personal economic benefits exceed the personal economic costs. When a foreign state threatens to retaliate against a country if it continues with its current policies, this increases the costs of maintaining the status quo. To the extent that individuals form opinions based on the material costs and benefits of a policy, external threats should reduce support for the status quo and increase support for policy change (Allen 2008; Kirshner 1997; Mack and Kahn 2000). The most prominent interest-based approach to IPE, known as “open-economy politics,” suggests that individuals whose pocketbooks are most likely to be harmed by external punishment—for example, those that are employed in sectors that are being threatened—should be most likely to change their opinions in response to foreign threats (e.g., Lake 2009).

A second perspective maintains that external threats are likely to generate a public backlash because these hostile actions will lead citizens of the targeted country to seek to “defend their group’s image” (Gruffydd-Jones 2019). Social identity theory posits that individuals’ self-esteem depends on the status of their nation, and threats from out-group members can activate a desire to protect the reputation and honor of the in-group. Thus, in the face of foreign sanctions, citizens of the target state may grow more defiant of “foreign interference, thereby strengthening…[support for the] objectionable policy’’ (Kaempfer, Lowenberg, and Mertens 2004, 30). ² This defensive reaction should be particularly strong for individuals with high levels of attachment to the in-group—in this case, individuals with high preexisting levels of national pride (de Hoog 2013; Gruffydd-Jones 2019).

A third view also maintains that external threats reduce support for policy change but through a different mechanism. In the “informational” perspective, threats of economic sanctions from a target state are consequential because they provide citizens with additional information about the consequences of the policy. Individuals often have limited knowledge about how they will be affected by policy changes. Thus, when formulating their preferences, individuals rely on informational shortcuts, or heuristics, such as endorsements from political elites (Popkin 1991). Individuals are more likely to favor a policy when informed that the policy is supported by a political party that they favor (Brader and Tucker 2012; Druckman, Peterson, and Slothuus 2013; Guisinger and Saunders 2017) or by a trusted and neutral international organization (Grieco et al. 2011; Linos 2013). While individuals respond favorably to endorsements from credible and trustworthy elites, the opposite may occur when the endorser is perceived as a hostile “outsider” with different interests or values—whether that is a national political party that one dislikes (Samuels and Zucco 2014) or a hostile foreign country (Bush and Jamal 2015; Grossman, Manekin, and Margalit 2018; Gruffydd-Jones 2019). External threats therefore provide new information to citizens about the policy preferences of foreign actors. In response, citizens are likely to infer that policy change would help the foreign country. Citizens who are skeptical of the foreign state’s intentions may also conclude that policy change would harm the home country. One key implication of this argument is that coercive threats and endorsements should have similar effects on public opinion since they carry similar information about the sending state’s preferences. This approach also expects individuals’ responses to external threats and endorsements to vary based on whether they view the external endorser in a benign or hostile light.

The Empirical Context: The US–China Exchange Rate Conflict

Chinese exchange rate policy is a useful setting for examining how economic coercion influences public opinion. Several features of this case make it particularly suitable for addressing this question. First, the United States has employed a mixture of coercive threats and diplomatic encouragement, which enables us to compare responses to each type of foreign pressure. Second, the relatively low salience of exchange rate policy is helpful because it means that respondents are less likely to enter the study with strong predefined preferences. Finally, due to strict limitations on the media, foreign threats and endorsements were not sensationalized or editorialized by domestic actors to the same degree as they might have been in more democratic contexts (see Online Appendix A). China’s controlled information environment therefore reduces the possibility that public responses to foreign threats and endorsements will be shaped by national media coverage.

In addition to the advantages of this case for research design purposes, it is an important case to study because of its real-world significance. Exchange rate policy has been a sore spot in the world’s most important bilateral relationship for the past fifteen years. The long-standing economic row between the world’s two largest economies is rooted in the massive expansion in trade between China and the United States since the early 1990s. For most of this period, China has maintained a heavily managed and “undervalued” exchange rate. ³ US policy makers have regularly complained that Beijing’s currency policy acts as a subsidy for Chinese exports that makes them more competitive in the American market and contributes to America’s trade deficit with China.

In the early 2000s, due to concerns that the undervaluation of China’s currency was hurting American economic interests, American policy makers began to pressure Beijing to increase the value of the RMB. This pressure has continued, with varying degrees of intensity, over the last fifteen years. Generally, the executive branch—speaking through the US Treasury—has avoided threatening language and instead has encouraged revaluation, “stress[ing] the desirability of [China] moving to a flexible market-based exchange rate regime” (US Treasury Department 2003, 7).

Similarly, sitting presidents have typically been circumspect in their language on this issue. However, as candidates, they have often directed more threatening language toward China. In 2008, then presidential candidate Barack Obama bluntly stated that the US trade deficit with China was “directly related to its manipulation of its currency’s value,” promising to increase pressure on Beijing, though he stopped short of threatening sanctions (Palmer 2008). During the 2012 presidential campaign, Republican nominee Mitt Romney promised during a televised debate to label China as a currency manipulator on “day one” of his administration (Branigan 2012). During the 2016 campaign, then candidate Donald Trump regularly blasted China for manipulating the RMB, adding that he would impose a 45 percent tariff on all Chinese products entering the United States (Haberman 2016).

Congress has also adopted a coercive approach at times. It has periodically considered legislation that would impose trade sanctions on China for the “manipulation” of its currency. For example, in 2006, Senators Chuck Schumer (D-NY) and Lindsay Graham (R-SC) introduced a bill proposing a 27.5 percent tariff on all American imports of Chinese goods unless the RMB appreciated sharply against the dollar. Two additional bills targeting China’s exchange rate policy emerged in Congress in 2010 and 2011. If passed, the legislation would have allowed US industries to employ antidumping measures and countervailing duties to defend against competition from countries with undervalued currencies (Galantucci 2015, 427). ⁴

China’s leadership, however, has not been publicly receptive to US economic coercion. For instance, in a May 2005 speech, Chinese Premier Wen Jiabao asserted that “external pressure on China to change its [currency] policies was counterproductive” (Foot and Walter 2011, 111). Five years later, responding to renewed pressure from the United States, Wen responded tersely: “Do not work to pressurize us on the renminbi rate” (Beattie, Chaffin, and Brown 2010). Along these same lines, a top Chinese economist explained to The New York Times in 2010 that “the greater the outside pressure, the more difficult it is for the Chinese government to raise the exchange rate, and the more difficult it is for the Chinese people to accept a revaluation of the Chinese currency” (Wines 2010). Both the “identity” and “informational” logics may be at play here. Consistent with the social identity perspective, top Chinese officials, such as Premier Wen Jiabao and Commerce Minister Chen Deming, often deflect US pressure about the exchange rate by stressing that this is a matter of “national sovereignty” (Foot and Walter 2011, 111; Qiao 2010). As per the informational logic, some Chinese elites claim that US efforts to get China to revalue the currency are driven by “US economic and political interests”—the implication being that appreciation would be good for the United States but not for China (Wu 2012).

A large literature examines whether US pressure has been effective in getting China to appreciate its exchange rate, but the literature has not reached a clear consensus. Some studies find that US pressure contributed to appreciation (Bowles and Wang 2016; Chao and Tung 2013; Weiss and Wichowsky 2018). Other works conclude that American pressure was either ineffective (Liu and Pauwels 2012; Frankel and Wei 2007; Steinberg 2015; Steinberg and Shih 2012) or that it reduced appreciation of the RMB (Ramírez 2013). We do not directly test whether US threats impacted Chinese policy choices but focus instead on its impact on Chinese public opinion. Our study contributes to this debate because it tests a key mechanism underlying the argument that US pressure is ineffective—that such pressure undermines public support for appreciation.

Research Design

To test the interests, identity, and information theories, we fielded a survey experiment in China that asked citizens about (1) their support for exchange rate appreciation/depreciation as well as (2) their beliefs about the consequences of appreciation. Our main treatments of interest vary the amount and type of information that respondents received about US pressure toward Chinese currency policy. One-third of respondents were randomly assigned to receive no information about US pressure. A second group received a “threat treatment,” which stated that the United States has threatened to impose tariffs on Chinese-made products if the RMB does not appreciate. The remaining third of respondents received an “endorsement treatment” that referred to a US endorsement of Chinese appreciation; this treatment contains information about US policy preferences but does not mention any US efforts to threaten China.

The interests, identity, and information theories each makes unique predictions about how the experimental treatments influence support for appreciation as well as respondents’ beliefs about the consequences of appreciation. The main prediction of the interests approach is that threats of economic punishment increase support for policy change. In this application, the theory predicts average levels of support for appreciation will be higher among those assigned to the threat treatment. The causal effect of the threat treatment should be particularly large for individuals who would be hurt by foreign punishment, which in this case refers to US restrictions on Chinese exports. Trade restrictions would be most costly for export-oriented enterprises, foreign-owned companies, and others that participate in international business ventures and global production chains (Milner 1988; Kirshner 1997; Pond 2017). Thus, the effect of the threat treatment should be strongest for individuals employed in internationally oriented firms. ⁵ Finally, the threat treatment is expected to influence individuals’ beliefs about the consequences of appreciation. Since the goal of the threat is to convince its target that policy change is necessary to avoid the imposition of material costs, the threat treatment should cause members of the target state to believe that policy change will help reduce trade tensions and that the policy change will be materially beneficial for their country. Finally, the interest-based approach expects the endorsement treatment to have no discernable effect on respondents’ preferences or beliefs because it does not include a threat to raise the material costs of noncompliance with US preferences.

The identity approach, by contrast, expects that foreign economic threats should reduce public support for policy change in the target state by triggering a defensive reaction among citizens to protect their country’s reputation. Thus, average levels of support for appreciation will be lower among those assigned to the threat treatment. The extent of individuals’ nationalist identities should moderate the impact of the threat treatment. Subjects with the strongest attachment to the nation are more likely to respond in such a defensive manner (Gruffydd-Jones 2019). Thus, threats should intensify opposition to appreciation most strongly for highly nationalistic individuals. Regarding individuals’ beliefs about appreciation, subjects in the threat treatment should be more likely to believe that appreciation will damage China’s international reputation since backing down in the face of a threat would make the country appear weak. If the identity argument is correct, the endorsement treatment should have a more limited impact on policy preferences since—in the absence of threatening China with punishment—it is less likely to enflame a nationalist response. ⁶

Finally, if threats impact public opinion in target states through an informational mechanism, the threat and endorsement treatments should reduce support for appreciation in a similar manner because both treatments carry the same information about US preferences. This approach starts from the assumption that preferences are sensitive to new information because respondents are uncertain about the consequences of appreciation. Foreign threats and encouragement directly provide citizens with new information about the foreign actor’s policy preferences. This information is likely to change peoples’ policy preferences for two reasons. First, treated subjects are more likely to expect that appreciation primarily benefits the country advocating the policy change, the United States. Second, these treatments are likely to convince some people that appreciation is bad for China’s economy. Individuals who view the United States as China’s adversary are particularly likely to respond in this manner because people tend to view relations with adversaries in zero-sum terms (Herrmann, Tetlock, and Diascro 2001; Mutz and Kim 2017; Rousseau 2002). People who hold negative opinions of the United States are more likely to believe that policies the United States advocates are bad for China. ⁷ For this reason, the informational argument expects the two treatments to have the strongest effect on preferences among individuals who have negative opinions of the sending state (the United States) or who view the US–China relationship in adversarial terms. Table 1 summarizes the expectations of each theory.

OPEN IN VIEWER

Table 1.

Overview of Theoretical Expectations.

	Interests	Identity	Information
Effect of threat on support for appreciation	Positive (especially for individuals in internationally oriented firms)	Negative (especially for individuals with strong national identities)	Negative (especially for individuals with negative opinions of United States)
Effect of endorsement on support for appreciation	None	None or weakly negative	Negative (especially for individuals with negative opinions of United States)
Beliefs about appreciation	Compliance with threats improves trade relations	Compliance with threats harms country’s reputation	Compliance with threat and endorsement benefits America’s economy and harms China’s economy

Description of the Data

In order to examine the influence of economic coercion on support for exchange rate policy changes, we fielded an online survey to Chinese citizens in February 2017. This was just weeks into President Trump’s tenure who carried his campaign’s criticism of China’s currency policy into his new administration (Holland and Lawder 2017). At the time of the survey, discussions about exchange rate policy in the Chinese media rarely focused on the role of US threats, as we show in Online Appendix A. ⁸ Consequently, our survey respondents who are not provided with any information about the US position are unlikely to have this issue in the forefront of their minds. ⁹

To recruit survey participants, we used a Chinese crowdsourcing service called KuRun, which operates in similar fashion to Amazon’s Mechanical Turk but with a narrower focus on survey research. ¹⁰ This strategy allowed us to reach a large pool of potential respondents, which would have been cost-prohibitive using traditional sampling methods. In all, we reached over 2,500 respondents from across the country. Although our sample covers all of China’s thirty-one provinces, the sampling method is not scientific. Overall, the sample is more urban, younger, and more affluent than the average citizen. Yet, this netizen-esque demographic is precisely the population the Chinese government’s online public opinion monitors are targeting in their own research (Denyer 2013). Moreover, a recent study of the emerging trend toward Internet recruitment in China shows that online convenience samples generate attitude estimates that are highly consistent with national probability samples (Li, Shi, and Zhu 2018).

One challenge with using online samples in China is the inattentiveness of respondents. To alleviate this concern, we dropped all cases where the survey was completed in less than five minutes because respondents who completed the survey at that speed were likely paying limited attention to the questions. We selected the five-minute cutoff for our main results because supplementary analyses of our data indicate that the data are much noisier below this threshold. ¹¹ Importantly, however, as shown in Online Appendix D, the main findings we present are similar when we include all observations in our analyses or when we use different thresholds for dropping observations.

Our main question of interest asked respondents whether they believed that China should increase or decrease the value of the RMB. The prompt begins by providing some basic information about the meaning of an increase or decrease in the value of a currency. ¹² The control group received the following information: “An increase in the value of the RMB relative to other currencies makes imports cheaper, whereas a decrease makes China’s exports more competitive in world markets. What do you think China should do?” Respondents then chose where to place a sliding bar, with possible responses ranging on a scale from “Decrease RMB value” to “Increase RMB value.” The scale for this variable ranges from 0 to 10, with higher (lower) values indicating more support for appreciation (depreciation).

Immediately after the main question about preferences for currency appreciation/depreciation, a series of follow-up questions asked individuals about the expected consequences of currency appreciation. We asked respondents whether they believed that increasing the value of the RMB would (1) help China avoid trade tensions with America, (2) be good for America’s economy, (3) be good for China’s economy, and whether it would (4) be good for China’s reputation abroad. For these four questions, respondents selected between five categories on a Likert-type scale, ranging from “strongly agree” (1) to “strongly disagree” (5).

We experimentally manipulated information about the position of the United States on China’s exchange rate policy. Some respondents were randomly selected to receive the “threat” treatment, which added that “America has threatened to impose taxes on its imports of Chinese-made goods if China does not increase the value of the RMB relative to the dollar.” Another randomly selected group of respondents received an “endorsement” treatment, which stated that “America has encouraged China to increase the value of the RMB relative to the dollar.” ¹³

Randomization checks indicate that the three experimental groups are observationally equivalent (see Online Appendix E). We therefore focus our analyses on simple difference-in-means comparisons between treatment groups. However, in Online Appendix F, we show that multivariate models that control for effects of other covariates produce substantively similar results to those displayed below. As described below, the survey also contains a number of questions that are useful for testing the three theories’ predictions about heterogeneous treatment effects.

The Influence of Foreign Pressure on Support for Appreciation

Figure 1 displays the mean responses to the question about the preferred amount of appreciation/depreciation, along with 95 percent confidence intervals, for the three experimental groups. The two treatment groups have lower means than the control group, indicating a stronger preference for depreciation, but these differences are relatively small, and they are not statistically significant. Thus, for the full sample, foreign endorsements and threats do not appear to strongly influence preferences in one direction or the other. However, external pressure may shape the preferences of some subsets of the population, and the next sections evaluate this possibility.

OPEN IN VIEWER

Figure 1.

Attitudes toward exchange rate appreciation. Diamonds indicate mean level of support for exchange rate appreciation for each group. Lines provide 95 percent confidence intervals surrounding the means. Higher (lower) values on this scale indicate more support for exchange rate appreciation (depreciation).

Perceptions of the United States and Support for Appreciation

The informational updating argument posits that an individual’s perceptions of the United States and US–China relations should moderate the impact of US threats and endorsements on support for currency appreciation. Two variables were used to operationalize individuals’ attitudes toward the United States. Our most basic measure asked individuals about their general opinion of America. Respondents chose where to place a needle on a five-point ordinal scale ranging from very negative (operationalized as a frowny face) to very positive (as indicated by a smiley face). Respondents were also asked about their views about China’s relationship with America, and whether the countries are best characterized as adversaries, cooperative competitors, or international partners. Individuals had to rank order the three statements, and we focus here on how they ranked the adversarial relationship. According to the information theory, which anticipates updating with the perceived preference of the endorser or threatening actor as a reference point, we expect that the two experimental treatments are more likely to reduce support for appreciation among individuals with negative opinions of the United States and those who view the United States as an adversary.

To test this hypothesis, we ran separate ordinary least squares regressions of support for appreciation on the two experimental treatments, where we split the sample based on whether or not they have negative views of the United States. ¹⁴ The upper panels of Figure 2 present the estimated average treatment effects for each subgroup. The left side shows that the treatments have virtually no influence on preferences toward exchange rate policy for respondents who have neutral or favorable opinions toward the United States. However, the upper right portion of the figure shows that people who dislike America respond strongly: each treatment reduces average support for appreciation by almost half a point (p < .05 in both cases). ¹⁵

OPEN IN VIEWER

Figure 2.

Attitudes toward the United States and support for currency appreciation. Circles indicate difference between treatment group and control group in support for currency appreciation. Lines indicate 95 percent confidence intervals for the treatment effects. Positive (negative) values indicate that treatment makes respondents more supportive of currency appreciation (depreciation).

The lower panels of Figure 2 utilize a different measure of attitudes toward the United States: whether individuals believe that America is China’s adversary. ¹⁶ For people who do not consider the two countries to be adversaries, US endorsements or threats about appreciation have no impact on attitudes toward appreciation. Similar to the previous results, we find that the treatments reduce support for appreciation for individuals who perceive the United States in adversarial terms, though only the threat treatment achieves statistical significance. The difference in means between the threat and endorsement treatments is not statistically significant either.

Overall, the findings in Figure 2 show, consistent with theories of heuristic decision-making, that individuals become more opposed to a policy when they learn that actors who they dislike are in favor of the policy. The fact that the endorsement and threat treatments have similar effects is consistent with the argument that threats influence policy preferences because both reveal the same underlying information about the interests of the threatening state.

Economic Interests and Support for Appreciation

The interest-based approach expects that threats of trade retaliation should increase support for appreciation for those individuals who are most likely to be harmed by foreign retaliation. We use several variables to capture individuals’ personal sensitivity to protectionist trade policies. Most directly, we measure this based on whether an individual’s employer exports any of their products. We also asked respondents whether their employer is a foreign-owned company since these businesses and their employees have a particularly strong stake in avoiding a trade war. Finally, the survey included a broader question of whether the respondents’ work “involve[s] any international partnership or services.” If threats of retaliation induce support for appreciation, people employed at export-oriented firms, foreign-invested enterprises, and those who have international business relationships should be particularly responsive to such threats.

Contrary to the interest-based theory, Figure 3 shows that individuals with international economic ties do not respond any differently to external threats than others. The upper panels compare the treatment effects for individuals who are employed in foreign-owned companies and those employed elsewhere, showing that the treatments are not statistically significant for either subgroup. The middle two panels examine whether having an international business partnership matters. We find some evidence that those without any international partnerships respond negatively to foreign threats: the threat treatment is negative and statistically significant for this group, and the endorsement treatment falls just shy of statistical significance (p = .11). And while the treatment effects for individuals with international ties have the correct positive sign, the size of the effects is relatively small and is far from conventional levels of statistical significance. We also fail to detect any statistically significant treatment effects when we split the sample between those employed at exporting firms and others, as in the bottom portion of the figure. ¹⁷ To summarize, there is no subset of the population for which threats of economic punishment increase support for appreciation, which casts doubt on the notion that people respond to threats on the basis of their economic self-interest, at least those aspects of economic interest that have been the focus of the open-economy politics approach.

OPEN IN VIEWER

Figure 3.

Economic interests and support for currency appreciation. Circles indicate difference between treatment group and control group in support for currency appreciation. Lines indicate 95 percent confidence intervals for the treatment effects. Positive (negative) values indicate that treatment makes respondents more supportive of currency appreciation (depreciation).

National Identity and Support for Appreciation

Lastly, we test the expectation that external threats are most likely to reduce support for a policy among individuals who identify particularly strongly with their nation. To test this hypothesis, we use three measures of national pride that have been used in previous research on attitudes toward international economic policy (Mansfield and Mutz 2009). The questions ask respondents how much they agreed or disagreed (on a five-point Likert-type scale) that (1) they would rather be a citizen of China than of any other country, (2) Chinese culture is superior to others, and (3) the world would be better if more people were like the Chinese. Since these questions were included prior to the exchange rate experiment, they represent respondents’ level of national pride prior to being threatened. Previous identity-based theories have hypothesized that external threats should have a stronger negative effect on support for policy change among respondents who score high on these variables (Gruffydd-Jones 2019).

The data fail to support the hypothesis that individuals who more strongly identify with the Chinese nation are more likely to respond negatively to foreign threats on the exchange rate issue. As shown in Figure 4, across all three measures of national pride, neither treatment has a statistically significant effect among the most nationalistic individuals. ¹⁸ However, the evidence indicates that our treatments have statistically significant negative effects among individuals with low levels of national pride. In two cases, the encouragement treatment reduces support for appreciation among those with low levels of national pride, and in one case, the threat treatment has this effect.

OPEN IN VIEWER

Figure 4.

National identity and support for currency appreciation. Circles indicate difference between treatment group and control group in support for currency appreciation. Lines indicate 95 percent confidence intervals for the treatment effects. Positive (negative) values indicate that treatment makes respondents more supportive of currency appreciation (depreciation).

One possible explanation is that individuals with low levels of nationalism have more room to become nationalistic in response to threats, whereas those with high levels of nationalism are less responsive to our treatments because they already express strong support for these policies (e.g., Hetherington and Suhay 2011). Thus, peoples’ sense of national pride appears to shape how individuals respond to external threats, albeit not in the manner we originally hypothesized.

Causal Mechanisms

The Influence of Foreign Pressure on Beliefs about Appreciation

The three theories that were outlined earlier all suggest that external threats influence preferences by changing peoples’ beliefs about policy change. However, each of the three theoretical approaches focuses on a different set of beliefs. In this section, we examine how foreign threats influence peoples’ beliefs about the consequences of a policy change. Figure 5 presents the mean responses, along with 95 percent confidence intervals, for our four questions that asked about the expected consequences of an exchange rate appreciation, with respondents split based on experimental condition.

OPEN IN VIEWER

Figure 5.

Beliefs about the consequences of currency appreciation. Diamonds indicate mean level of agreement with the listed statement about the consequences of currency appreciation, where higher values indicate stronger agreement. Lines provide 95 percent confidence intervals surrounding the means.

In the upper left segment of Figure 5, we see that the treatments have limited impact on beliefs about whether RMB appreciation would help avoid a trade war. Similarly, the upper right panel of Figure 5 shows that our experimental treatments do not influence whether people believe that an appreciation would be good for China’s reputation abroad. These findings suggest that the threats did not influence our respondents’ beliefs in the manner expected by the “interests” and “identity” approaches, though these are admittedly imperfect measures of these theories’ underlying concepts. ¹⁹

The informational theory expects the treatments to change respondents’ beliefs about two issues: whether China benefits from appreciation and whether the United States benefits from appreciation. The lower left panel of Figure 5 examines whether respondents believe that appreciation would be good for China’s economy. The mean response in the control group and threat treatment groups are similar, with means of 3.38 and 3.35, respectively. However, the mean response in the endorsement treatment (3.27) is moderately lower than the control group, and the difference is statistically significant at the 10 percent level (p = .07); the difference between the two treatment groups is not statistically significant (p = .17). This provides some evidence that foreign endorsement of a policy can create a perception that this policy works against the target state’s national interest.

The most striking differences are found in the bottom right section of Figure 1, which shows respondents’ views about whether the US economy would benefit from an appreciation of the RMB. The mean for the control group (3.33) is considerably lower than the means in the endorsement (3.62) and threat treatments (3.58). The difference-in-means between control and each treatment group is statistically significant (p < .01). The difference-in-means between the two treatments groups is not statistically distinguishable from zero.

This evidence suggests that foreign pressure can indeed influence how individuals in the target country believe a policy will work—but in the opposite direction of what the target state would hope. We find some evidence that foreign advocacy reduces the belief that a policy benefits the home country’s economy and strong evidence that such advocacy creates a perception that foreign countries will benefit economically from a policy change. The fact that the threat and endorsement treatments intensify the perception that the sending state would benefit from policy change in a similar fashion provides further support for the information-based theory of opinion formation.

Causal Mediation Analysis

In the final step of the analysis, we examine whether the changes in beliefs that we observed in the preceding section are important mechanisms through which our experimental treatments influence preferences (among some subgroups). We use causal mediation methods to test whether our four belief variables mediate the relationship between our experimental treatments and preferences for exchange rate appreciation. This approach decomposes the “total effect” of the treatments into two components. The “indirect effect” refers to the effect that is accounted for by a mediator variable, in this case our measures of beliefs. The “direct effect” is the remaining effect that consists of all other potential channels (see Imai et al. [2011] for more details).

Table 2 presents our estimates, which are based on separate models for each of the four potential mediator variables. The table also presents the results of the mediation analysis for (1) the full sample, (2) anti-American subsample, and (3) pro-American subsample. We do so because the previous section showed that attitudes toward the United States is an important moderating variable. ²⁰

OPEN IN VIEWER

Table 2.

Causal Mediation Analysis.

	Benefit American Economy			Benefit Chinese Economy
	Endorsement	Threat		Endorsement	Threat
	All Respondents			All Respondents
Indirect	−.079**	−.070**	Indirect	−.091*	−.021
Direct	−.049	−.066	Direct	−.038	−.115
Total	−.128	−.136	Total	−.129	−.137
	Anti-American			Anti-American
Indirect	−.059**	−.059**	Indirect	−.137*	−.115
Direct	−.346	−.338	Direct	−.269	−.282
Total	−.405*	−.397*	Total	−.407*	−.398*
	Pro-American			Pro-American
Indirect	−.098**	−.080**	Indirect	−.053	.027
Direct	.102	.079	Direct	.057	−.029
Total	.004	−.001	Total	.003	−.002
	Improve China’s Reputation			Improve Trade Relations
	Endorsement	Threat		Endorsement	Threat
	All Respondents			All Respondents
Indirect	−.016	−.041	Indirect	−.007	.022
Direct	−.106	−.095	Direct	−.117	−.164
Total	−.122	−.137	Total	−.125	−.142
	Anti-American			Anti-American
Indirect	.061	−.093	Indirect	.019	.021
Direct	−.460**	−.305	Direct	−.425**	−.419**
Total	−.399*	−.398*	Total	−.406*	−.397*
	Pro-American			Pro-American
Indirect	−.053	−.016	Indirect	−.016	.022
Direct	.062	.014	Direct	.025	−.031
Total	.009	−.002	Total	.009	−.009

Note: Cell entries provide the total, direct, and indirect effect of the experimental treatments across different mediator variables. Estimation is based on Hicks and Tingley (2011).

*p < .1.

**p < .05.

First consider the results for the full sample. As per our previous results, we find that each experimental treatment reduces overall levels of support for appreciation by about 0.13, but this total effect is not statistically significant. Importantly, however, beliefs about the US economy are a statistically significant mediator for both treatments, and this indirect effect is around half the size of the total effect. This implies that both treatments create a belief that appreciation benefits the American economy, which leads to a reduction in support for appreciation. We also find that a belief that appreciation hurts China’s economy is a statistically significant mediator for the endorsement treatment (p < .10). ²¹ The other two mediators have small and statistically insignificant indirect effects. For the full sample, then, even though the experimental treatments do not create a strong shift in average opinions, they do reduce support for appreciation by altering beliefs about which countries are likely to benefit from a policy change. ²²

The similarities and differences between the pro- and anti-US subgroups are informative. It is unsurprising that a belief that the United States benefits from appreciation is an important mediating variable for individuals with negative opinions about the United States. Interestingly, a belief that the US economy benefits from appreciation has a similarly sized and significant indirect effect among people with positive opinions about the United States. Online Appendix I probes this relationship more deeply. It shows that the threat and endorsement treatments have a similar effect on the belief that the United States benefits from appreciation for pro-American and anti-American respondents (see Online Appendix Tables I2 and I3). Moreover, believing that appreciation is good for the United States has a similarly sized and significant negative association with support for appreciation across the two subgroups (see Online Appendix Tables I4 and I5). Thus, the distinct responses of pro- and anti-American individuals to the experimental treatments are not driven by differences between the two groups’ desires to hurt the US economy.

Instead, the sharpest difference between these two groups lies in the importance of beliefs about the Chinese economy. The perception that appreciation is bad for China has a much larger indirect effect among respondents who dislike the United States than those with positive opinions about the United States. As we show in Online Appendix I, both groups oppose policies that they perceive as being bad for China’s economy. The key source of this difference is that the treatments cause anti-American individuals, but not those with pro-American views, to believe that appreciation is bad for China.

The results of these mediation models help us understand why external threats have different effects on policy preferences based on peoples’ views of the sending state. The treatments lead both groups of people to update their beliefs about whether the United States will benefit from appreciation—presumably because this is an obvious implication of foreign threats and endorsements. The larger total effect among the anti-American subgroup arises because foreign threats lead these individuals to conclude that China’s economy will be hurt by appreciation—a finding that is consistent with the information updating theory. By contrast, threats do not lead people with more benign views of the United States to conclude that policy change will hurt their own country, likely because they are less likely to view the bilateral relationship in zero-sum terms.

On the other hand, perceptions about how appreciation benefits the two countries’ economies cannot fully explain why pro- and anti-American individuals respond differently. The direct effect for the anti-American subgroup remains substantial, which implies that there are other important mechanisms that our analyses have not fully modeled. For instance, those who view the United States as a helpful partner may simply want to maintain good relations with this country, while individuals who view the United States as a hostile adversary may be less concerned or even happy about seeing the two countries’ relationship sour. The evidence shows that informational updating is one important mechanism through which external threats alter policy preferences in this issue area. At the same time, more traditional mechanisms, such as those emphasized by social identity theory, are likely operative as well.

Summary and Conclusions

In this article, we use the case of US–China currency relations to test three competing theories of the effect of foreign economic threats on public opinion. The leading approach in IPE, known as open-economy politics, argues that individuals’ economic self-interest, as determined by variables such as their industry of employment, influences preferences on foreign economic policy. This approach would therefore expect the public—especially those in internationally oriented industries—to become more supportive of exchange rate appreciation when a country’s trading partners threaten to impose tariffs on the home country. Our survey-based evidence finds no support for these hypotheses: US pressure did not make any group of individuals, including those employed in trade-dependent parts of the economy, significantly more supportive of exchange rate appreciation. It is possible that economic interests do shape citizens’ reactions to economic threats, but in this case, they do not appear to work in the manner expected by standard scholarship in IPE.

Another prominent approach suggests that public responses to external threats are not based on peoples’ interests but instead based on their identities. If so, threats should reduce support for policy change, and the magnitude of this effect should vary based on an individual’s degree of national pride. Our evidence provides some support for this argument. We find that Chinese survey respondents with low levels of nationalism grow more opposed to exchange rate appreciation when informed about foreign threats, which is consistent with some variants of this theory (though contrary to our initial expectations). On the other hand, the evidence suggests that foreign endorsements, which are less likely to elicit a nationalistic backlash, had similar effects on public opinion as did threats, which suggests that the social identity approach is incomplete at best.

The evidence most strongly supports the informational updating theory. This argument maintains that foreign threats reduce support for economic policy change in the target country because these threats lead citizens of that country to infer that changing policy will benefit the threatening state at the expense of the home country. US threats and encouragement had a similar effect, leading Chinese individuals to believe that appreciation benefited the United States. Mediation analysis showed that this belief is an important channel that reduced support for exchange rate appreciation. Moreover, US pressure also increased the degree to which respondents with negative opinions about the United States believed that appreciation would be bad for China, helping to explain why this subgroup had a particularly strong negative reaction to external pressure. In sum, the evidence suggests that external threats create a perception that policy reform would benefit foreign countries at the expense of the home country, which undermines support for policy change among some segments of the population.

Of course, our findings should not be overstated. While the United States and China comprise the world’s most important bilateral relationship, it is also one that is more easily depicted as competitive and, perhaps, adversarial. Our analysis, therefore, cannot tell us whether external threats and endorsements affect public opinion in the same way when the two countries in question are on friendlier terms. Indeed, other studies find that external pressure only generates a public backlash when the sending state is viewed as an unfriendly state (Grossman, Manekin, and Margalit 2018; Gruffydd-Jones 2019). Thus, it may be the case that the Chinese public would be more likely to support accommodation when threatened by a state viewed as a friend of China. Public opinion backlashes may also be less likely on issues that are not plausibly viewed in zero-sum terms and for policy areas where citizens already have strong preexisting opinions. While our findings do not indicate that foreign pressure always has counterproductive effects on public opinion, the evidence suggests that publicly pressuring a foreign country to change course on competitive economic issues can do more harm than good when it comes to public opinion in the target state.

While keeping these caveats in mind, our findings have a number of important implications for theories of international relations and political behavior. At the broadest level, this analysis highlights the important ways in which international politics influences individual public opinion—what Kertzer and Tingley (2018) refer to as the “first image reversed” approach to international relations. Our evidence also complements other recent work that finds that external pressure reduces public support for policy change for issues such as foreign security policy (Grossman, Manekin, and Margalit 2018) and human rights (Gruffydd-Jones 2019; Bush and Jamal 2015). Thus, one contribution of the study is to show that public opinion “backlashes” also operate in the economic realm. Another contribution is to elaborate on a novel informational mechanism that can contribute to this backlash effect.

This research also contributes new insights to policy debates about the effectiveness of economic coercion as a foreign policy strategy. Our evidence casts some doubt on the view that coercive economic threats—and even endorsements—by geopolitical rivals contribute to policy change. To be sure, economic policy is not driven solely by mass public opinion, especially in an authoritarian regime such as China. On the other hand, policy makers are unlikely to entirely ignore public opinion. To the extent that leaders care about public opinion, economic coercion is likely to make it less rather than more, likely that governments will change policy. Indeed, China’s leadership has itself repeatedly claimed that their citizens’ response to foreign pressure makes it more difficult for them to appreciate the exchange rate (Foot and Walter 2011, 119). Alternatively, if and when China’s leaders feel compelled to cave in the face of foreign pressure, our findings suggest that doing so will cost them support from some of their domestic audiences.

These domestic political dynamics may also help explain why the Trump administration’s coercive diplomacy has not been more successful in extracting policy concessions abroad. As a recent article in the Washington Post explains: “One reason for worry about things spinning out of control, trade experts say, is Trump’s apparent belief that he can use threats to coax concessions out of allies…most of the world’s major powers have rejected his demands” (Long and Mufson 2018). Public opinion in these countries is likely part of the reason why US efforts have yielded few successes so far. When it comes to getting a foreign government to adopt a preferred policy reform, the most effective strategy may be to communicate preferences privately or, perhaps, do nothing at all.

Supplemental Material

Supplemental Material

Supplemental Material