1. Introduction
How do governments handle domestic backlash when backing down in international crises? The literature on audience costs posits that if leaders act inconsistently with their previously declared stance, the public might punish them because of concerns over their competence and/or the country’s credibility and honor (Fearon, Reference Fearon1994). Skeptics have raised doubts as to the existence and importance of audience cost in determining states’ behaviors (Snyder and Borghard, Reference Snyder and Borghard2011; Trachtenberg, Reference Trachtenberg2012), and one reason for this is the problem with the empirical research done on the phenomenon (Gartzke and Lupu, Reference Gartzke and Lupu2012). Existing work on audience costs used experiments with hypothetical scenarios to investigate audience cost and its mechanisms (e.g., Tomz, Reference Tomz2007; Levy et al., Reference Levy, McKoy, Poast and Wallace2015). The lack of connection of such hypothetical scenarios to real-world settings cannot conclusively extract the real opinions of the respondents regarding the performance of the leaders in crises and, therefore, cannot definitively prove the existence of audience cost. Another problem with a majority of existing empirical work is the dependence on US-based respondents, who might not have a direct and salient connection to the disputes. This weakness reduces the generalizability of audience cost and requires replication studies done in other countries to further verify the concept.
A growing number of recent works have used real-world disputes to investigate audience cost in non-Western settings (Driscoll and Maliniak, Reference Driscoll and Maliniak2016; Quek and Johnston, Reference Quek and Johnston2018; Kohama, Quek, and Tago, Reference Kohama, Quek and Tago2024). A common topic that they also study is whether leaders can reduce the domestic backlash they face when choosing to back down in response to an international crisis. The findings are robust: leaders’ approval ratings can be improved if they effectively justify their reasons for backing down. These experiments primarily involved escalatory behavior in territorial disputes between states and ranged from demonstrative to military actions, but in no experiment was the disputed territory said to have changed hands. In other words, the country said to be currently in control of the territory in question is not described as having lost said territory.
I designed an experiment in this study involving the Senkaku/Diaoyu Islands dispute between Japan and China. What distinguishes this study from previous work on audience costs is its use of fait accompli, a unilateral revision of the territorial status quo, in the experimental scenario. In essence, Japan is described as having lost control over a disputed island. This is a qualitatively different case compared to the scenarios used in previous research. Territorial disputes are the most likely to spark interstate conflicts (Vasquez, Reference Vasquez1996; Hensel and Mitchell, Reference Hensel and Mitchell2005; Henehan and Vasquez, Reference Henehan and Vasquez2006). Therefore, it can be expected that leaders will face stronger public backlash when they back down after losing a territory, compared to other forms of crises.
I conducted an online survey experiment in Japan in December 2021. The results show that the Japanese government can reduce audience costs through four rhetorical measures: emphasizing the threat of war, the potential for casualties associated with the use of force, mediation by the United Nations (UN), or seeking international legal arbitration. Among these, countering China’s actions by suing Beijing in an international court and asking the UN to intervene proved to be the most effective. Out of the potential strategies considered, only accentuating the economic cost of conflict did not reduce audience costs in a fait accompli scenario. The paper is structured as follows. I begin by introducing the literature on fait accompli to highlight the motivation behind this study. Then, I outline the theoretical foundation and introduce our hypotheses. The third section will be devoted to explaining the experimental design. Finally, I will discuss the results of the experiment and the implications of the study.
2. Fait accompli and the cost of backing down
2.1. The implications of fait accompli
Fait accompli is defined as the unilateral action to take a limited area, with the calculation that the victim state will relent (Altman, Reference Altman2017). The logic is that, for the victim state, the risk of a full-scale war with the challenger outweighs the benefits of attempting to take back the lost territory. Therefore, they are better off accepting the loss and backing down from the crisis. States that aspire to change the territorial status quo are incentivized to choose a fait accompli strategy because the costs of such actions are lower than annexing territories by brute force (Tarar, Reference Tarar2016). Furthermore, rising revisionist powers often use fait accompli as a way to “test the water.” Through fait accompli, they can learn where the limits are for the status quo declining powers, and through this, avoid going to war earlier than expected (Hastey, Reference Hastey2020). In sum, there are two goals in territorial fait accompli: to revise the status quo through incremental changes and to avoid a full-scale war.
The challenger states have been quite successful in both goals. They have used fait accompli to acquire territories much more than brute force or coercion (Altman, Reference Altman2017). Altman (Reference Altman2020) and Hastey (Reference Hastey2023) pointed out that states have engaged in this practice throughout history. In the majority of cases, the strategy worked: victim states gave up the lost territory and did not retaliate. Tarar (Reference Tarar2016) argued that war would occur only after a fait accompli when challengers, misled by incomplete information, take more than what is acceptable to their victims and leave them with no other choice but to retaliate.
The reactions of the victims determine whether wars will break out or not as a result of fait accompli. In other words, to judge whether the challengers’ fait accompli strategies have succeeded in their goal of avoiding war, we need to look at the victims’ responses. However, this aspect has been overlooked in the current literature. Previous work focused overwhelmingly on incentives for challengers while failing to explain the varied reactions of the victim states to fait accompli. A notable recent exception is Ngo (Reference Ngo2023), in which the author explained why some victim states chose to retaliate immediately and why others chose to relent and gave up. The author concluded that the decision to retaliate or not depends on the victim states’ calculation of both the values of the territories and the values of restraint. Referring to several case studies in history, Ngo explained that in cases where target territories were perceived to have vital values, as well as when restraint would put victims in a more disadvantageous bargaining position compared to fighting back, the victim states would retaliate, and wars would break out.
This study attempts to explain victim states’ behaviors by looking at whether the victim states have the option to back down after a fait accompli against them. Territories often have specific characteristics that make territorial crises more war-prone: strategic values, symbolic virtues and indivisibility (Huth, Reference Huth1999; Senese and Vasquez, Reference Senese and Vasquez2003; Wiegand and Choi, Reference Wiegand and Choi2017; Zellman, Reference Zellman2018). Therefore, leaders would be expected to have a harder time backing down from crises after losing territory compared to other types of crises. However, this factor has not yet been addressed in previous experiments on audience costs. Therefore, using this scenario in experiments might challenge the current understanding of how governments can or cannot back down.Footnote 1
2.2. The case of Senkaku/Diaoyu Islands
This study uses territorial disputes between China and Japan over the Senkaku/Diaoyu islands to test the hypotheses. This case is most appropriate for several reasons. First, among the territorial disputes between Japan and other countries, this is the dispute in which fait accompli is most likely to occur. In the Takeshima/Dokdo dispute with South Korea, Japan does not have control over the disputed territory. In the Northern Territories dispute with Russia, Japan does not face the same level and frequency of provocations as it does in the Senkaku/Diaoyu dispute. Therefore, the risk of fait accompli is considered to be much higher for the Senkaku/Diaoyu disputes.
Second, fait accompli is a realistic possibility that could happen in this dispute. For one, the Senkaku/Diaoyu Islands are among the First Island Chains, constraining China’s access to the Pacific Ocean, and the desire to gain a strategic advantage is one of the motivations for states to engage in fait accompli (Reiter and Poast, Reference Reiter and Poast2021). Furthermore, as pointed out previously, rising revisionist states such as China could use fait accompli to test the commitment of declining status quo powers (Hastey, Reference Hastey2020). A fait accompli against a close ally of the US, such as Japan, is a tool through which China can achieve this goal. Based on these theoretical foundations, many analysts consider the risk of fait accompli to be particularly high in this dispute (Pajon, Reference Pajon2012; Cronin, Kliman, and Krejsa, Reference Cronin, Kliman and Krejsa2018; Liff, Reference Liff2019; Ordaniel and Baker, Reference Ordaniel and Baker2023).
3. Methodology
3.1. Hypotheses
I argue that leaders can limit audience costs in fait accompli situations by explaining to the public the reasons for backing down.Footnote 2 The hypotheses are divided into two groups.
The first group of hypotheses concerns the public preference for peace and the cost of war. I call this group the ‘war avoidance hypotheses.’ My case, Japan, is known for its pacifism and strong anti-war sentiment (Berger, Reference Berger1993; Katzenstein and Okawara, Reference Katzenstein and Okawara1993; Katzenstein and Okawara, Reference Katzenstein and Okawara2001; Katzenstein, Reference Katzenstein2018). As a result, it is expected that when the Japanese government backs down for the stated reason of avoiding war, the respondents’ approval ratings will increase, compared to when they do not give any reason for backing down.
Losing a territory might be considered unacceptable for the domestic audience. However, if accepting the loss of territory means that they can avoid a full-scale war, this domestic backlash can be mitigated. The first hypothesis is derived as follows.
Hypothesis 1 The public approval rate will be higher when the government explains why it is backing down by claiming a threat of war from the aggressor, compared to when no reason is given.
I also consider the cases in which the victim government explains in detail their cost-benefit calculation. In particular, whether or not the public is more amenable to backing down when they receive information about expected casualties. Previous literature on casualty sensitivity pointed out that the public is less likely to support the use of forces if the action leads to great loss of life (Reifler, Gelpi, and Feaver, Reference Reifler, Gelpi and Feaver2005; Komiya, Reference Komiya2019; Horiuchi and Tago, Reference Horiuchi and Tago2022). Thus, it is anticipated that the Japanese public would accept their government’s lack of action, owing to the potential loss of lives involved in reclaiming the lost territories. We have the second hypothesis:
Hypothesis 2 The public approval rate will be higher when the government explains its reason for backing down by emphasizing heavy casualties and the high costs of taking back the territory, compared to when no reason is given.
In territorial disputes with Japan and other countries, China often uses economic coercion as a means through which it can force the other side to de-escalate (Zarichsen, Reference Zarichsen2015; Vesaki, Reference Vesaki2019). Therefore, it is reasonable to expect that in a potential crisis over the Senkaku/Diaoyu Islands, Beijing would threaten to impose economic sanctions – even informally under the form of import/export restrictions – should the Japanese government signal its resolve. Such sanctions might have a great impact on bilateral trade relations, which is a factor that the Japanese public is very interested in (Tanaka, Tago, and Gleditsch, Reference Tanaka, Tago and Gleditsch2017; Kohama, Quek, and Tago, Reference Kohama, Quek and Tago2024). It is anticipated that the public would support the government’s efforts to de-escalate if faced with economic sanctions that heighten the financial burdens of conflicts.
Hypothesis 3 The approval rate of the public will be higher when the government explains its reason for backing down by emphasizing the threat of economic sanctions from China, compared to when no reason is given.
The second group of hypotheses, which I call the ‘limited retaliation’ hypotheses, is based on the argument by Cho (Reference Cho2018) for the effect of provocation. Cho argued that the governments of victim states when provoked, can face two constraints from the public: the first is the increasing audience costs if they do not react to the provocation, and the second is the increasing support for retaliation. To mitigate this effect, victim states must retaliate to some extent, although not resorting to the use of force.
The first of such retaliatory measures is to bring the matter to court. This is a tactic that Japan has used before in its territorial disputes with South Korea concerning the Takeshima/Dokdo islands. The government could also decide to use the international court as a political cover (Allee and Huth, Reference Allee and Huth2006a; Allee and Huth, Reference Allee and Huth2006b). Moreover, the public generally responds positively to settlements declared by an international court. This is either because they perceive that these judicial bodies have legitimacy over the issue or because they extend their beliefs in their domestic judicial system to international bodies (Voeten, Reference Voeten2013). From this literature, I derive the following hypothesis:
Hypothesis 4 The approval rate of the public will be higher when the government justifies backing down by declaring that it will bring China to an international court, compared to when no reason is given.
It is also necessary to take into account the role of international organizations (IOs). Previous studies have shown that, to some extent, IOs can play a mediating role in interstate conflicts. The degree of their success would depend on several factors: the level of institutionalization within the mediating IO (Zhang and Qi, Reference Zhang and Qi2022), the level of accountability of the involved regimes (Chiba and Fang, Reference Chiba and Fang2014), and the economic leverage of the IOs (Karreth, Reference Karreth2018). The United Nations is highly institutionalized and has a reputation for mediating conflicts between countries (Iji, Reference Iji2017; Schiff, Reference Schiff2022; Iji, Reference Iji2022). Past studies also showed that the Japanese public has a high level of trust in the United Nations as a force for preserving international peace and security (Ogata, Reference Ogata1983; Ogata, Reference Ogata1987). This is also evident through the high public support for the more active participation of the Self-Defense Forces in UN Peacekeeping Operations (Cabinet Office, 2023). From this rich body of literature and empirical evidence, I hypothesize that the Japanese populace would support a government-led de-escalation if it is endorsed by UN initiatives.
Hypothesis 5 The approval rate of the public will be higher when the government justifies backing down by calling for the mediation of the conflict by the United Nations, compared to when no reason is given.
3.2. Experimental design
The experiment was conducted online in December 2021. The survey was made on the Qualtrics platform and distributed through Nikkei Research. The data excluded those who did not pass the attention check, resulting in a total of 1,905 respondents, all of whom were of Japanese nationality and eligible to vote.Footnote 3 All respondents were asked for their consent and were informed that they had the option to exit at any point during the survey and that if they did so, their responses would not be recorded.
The experiment began with a series of questions about demographic attributes and socioeconomic status. Respondents were then asked to answer several questions regarding their political knowledge, including a multiple choice question in which they chose the countries they thought to pose a security threat to Japan, as well as a self-assessment of political ideology.
The respondents were randomly assigned to one of six groups, one control group, and five treatment groups, each treatment group aimed at testing for one of the hypotheses mentioned above.Footnote 4 All respondents were asked to read a hypothetical scenario on the Senkaku Islands dispute between Japan and China. The scenario involved Beijing’s occupation of a maritime feature in the Senkaku Islands, as well as how the Japanese government initially responded to the crisis with a tough stance. Later, they were stated to have backed down from that tough position. Respondents assigned to the control group would proceed directly to the next part of the survey, without receiving any information on why the Japanese government decided to back down.Footnote 5 Respondents in treatment groups would read additional information on the reasons for backing down.Footnote 6
After reading the scenario, respondents were asked to rate the responses of the Japanese government to the crisis. Responses were recorded on a 5-point Likert scale from ‘strongly approve’ to ‘strongly disapprove’.
4. Results
To determine the treatment effects, this paper uses the two-tailed Welch t-tests of the raw approval score on a five-point scale.Footnote 7
Figure 1 summarizes the difference in the approval rate of the Japanese government for each of the treatment groups compared to the control group. In general, backing down is not a well-supported course of action for governments in international crises, regardless of the reasons they give. However, as shown below, there are ways for the government to earn more public sympathy and reduce the audience cost that they face. Among the hypotheses, only H3 on the threat of economic costs from China was not supported. I elaborate on the results of the treatments below.
Approval rating percentage for backing down.
4.1. The war threat from China (H1)
Compared to the control group, the approval rating of the Japanese government when backing down is justified by a threat of war from Beijing is 15 percentage points higher. On average, the approval rating (5 points Likert scale from –2 to 2) for the government of the respondents in this treatment group is –0.17, higher than the rating of –0.52 in the control group (P = 0.003).
This statistically significant result supports our H1. This result suggests that, to some extent, the Japanese public sympathizes with the government backing down from a tough stance in crisis, to avoid a full-blown war with China. This is consistent with the logic of fait accompli: the public could consider the cost of going to war with China to be much higher than the cost of losing a feature in the Senkaku Islands, and would understand why the government would back down from trying to take it back.
4.2. Heavy casualties (H2)
In the second treatment group, the approval rating for the Japanese government is higher by 17 percentage points compared to the control condition. The mean approval rating of the respondents in this group is –0.07, higher than the –0.52 rating in the control group (P < 0.001).
The statistically significant result shows that the Japanese public is considerate when judging the government’s responses to China’s provocation. The public would show more tolerance towards the government when they are made aware of the potential heavy casualties of taking back the lost territory.
4.3. Economic costs (H3)
The approval rate for Japanese governments in this group is largely not different from that in the control group. In particular, the approval rate increases by one percentage point. The means approval rating for this group is –0.45, compared to –0.52 in the control group (P = 0.50).
This statistically insignificant result shows that emphasizing the threat of economic sanctions would not help the government reduce audience costs. One possible explanation for this outcome is that in a previous incident, China imposed restrictions on the export of rare earth, a key ingredient in the production of semiconductor and automotive parts, to Japan (Bradsher, Reference Bradsher2010). Initially, this weaponization of interdependence led to key industries in Japan suffering due to the shortage of their crucial ingredients; later, diplomatic, economic and commercial efforts from the Japanese government rendered these restrictions ineffective (Vesaki, Reference Vesaki2019; Fickling, Reference Fickling2021). With this precedent, emphasizing the threat of economic sanctions from China would not help the Japanese government justify backing down because the Japanese public may be confident in the country’s ability to deal with sanctions as in the past.
Another possible explanation for this result might be that the loss of territory means the stake in the crisis is high, and therefore, the public might consider economic costs not to be as important as other factors (such as war or casualties). The public might be willing to bear the economic costs of conflicts in exchange for a possible return of the territory.
4.4. Bringing China to international court (H4)
Compared to the control group, the approval rating for the Japanese government in this group increased by 21 percentage points. The average rating is 0.08, compared to –0.52 in the control group (P < 0.001).
These results demonstrate a high level of support among the Japanese public for using international judicial bodies to settle territorial disputes. It also reflects a strong belief of the Japanese public in the legal advantage that Japan has in disputes. With strong legal grounds, the probability of a judgment in favor of Japan is higher. The public might also view the act of bringing China to court as a naming and shaming method by directly pointing out the names of the wrongdoers.
4.5. UN mediation of conflicts (H5)
Compared to the control group, the approval rating of the respondents in this group is 14 percentage points higher. In particular, the average approval rating in this group is –0.14, compared to –0.52 in the control group (P < 0.001).
This result reflects the moderately high level of trust of the Japanese public in the capabilities of the United Nations to mediate conflicts. To a greater extent, it shows that the response of the international community plays a role in the bilateral crisis negotiation process. The active participation of the United Nations in trying to mitigate the effects of conflicts will help lessen the audience cost and give governments more flexibility in their responses.
5. Discussions and conclusion
The main finding of the study is that after a fait accompli occurs, there are ways for the victim state to mitigate the potential audience cost, allowing them to back down and de-escalate the tension. The results show that for the Japanese public, the threat of war, power asymmetry, third-party mediation, and adjudicating via international court help reduce audience costs. The only potentially ineffective rhetorical strategy for reducing audience costs under these conditions and in this type of crisis scenario was emphasizing the threat of economic sanctions from China.
The findings of this study support similar results of Quek and Johnston (Reference Quek and Johnston2018) and Kohama et al. (Reference Kohama, Quek and Tago2024) regarding how leaders can manage the audience cost faced by backing down after a territorial crisis. However, the significantly lower support rate for backing down across groups in this article, compared to previous works, supports the interpretation that the use of fait accompli, a more severe crisis involving the loss of territory, could produce different results compared to the less severe actions presented in the scenarios used in previous studies.Footnote 8 My results also show that the economic costs of the conflict might not play a major factor in influencing the Japanese public’s reaction to an international crisis, as previously suggested.Footnote 9 One potential explanation is the difference in the type of crisis used in the experimental scenario. The public may react differently to different levels of provocation. When the crisis involves the Chinese use of drones over the disputed territory, the Japanese public might still consider, to a certain extent, the economic cost of getting into a conflict with China and thus support their government to back down. However, when the territory is lost, the economic consequences are not considered important to the public.
This paper advances the literature on fait accompli. Potentially, the results can partially explain why some fait accompli led to war and why others did not. Backing down and accepting the loss of territory is not a popular decision among the public. Therefore, in cases where the government cannot persuade the public that they have good reasons to back down, they might be forced to carry out the tough stance they have committed. In cases where fait accompli did not lead to war, the government might have, to some extent, successfully managed public sentiment and avoided the pitfall of tying their hands. By analyzing audience cost, one particular form of cost that the victim might face in a fait accompli situation, this paper complements the current literature focusing on the costs of the challenger.
Another implication of this result is the settlement of territorial disputes through international courts. Among the justifications for backing down, this reason is particularly well-received by the Japanese public. This positive result supports the greater use of international judiciary bodies in the resolution of territorial disputes.
Future research can benefit from an improved experimental design. Due to the technical limitations of the platform used to create the survey, I was only able to randomly assign each scenario once. In other words, the respondents received only one explanation from the government. In reality, governments often give multiple explanations for their actions. Another factor that I did not include was the involvement of the United States. With the US–Japan alliance, it is unlikely that in the case of China conducting a fait accompli, Japan will act alone, without assistance from the United States.
The results of the experiment also call for replication studies to be performed in different contexts. Specifically, it would be theoretically informative to see the similarities and differences in the obtained results in Japan and other countries in territorial disputes with China, such as Vietnam, the Philippines, or India. By conducting further mirror experiments, the generalizability of the results might be improved, and knowledge of the fait accompli can be advanced. It will also potentially have a positive impact on crisis management between countries with territorial disputes.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1468109924000239
Data availability statement
The replication data and code will be provided on Harvard Dataverse if the paper is accepted https://doi.org/10.7910/DVN/9KAKDE.
Acknowledgements
The author thanks Tago Atsushi, Shinomoto Sou, the editor, and the reviewers for their helpful feedback. The author is also profoundly thankful to Huyen Le for her steadfast encouragement and support from beginning to end.
Funding statement
This research was supported by the JSPS KAKENHI Grant Number 22H00050.
Competing interests
There is no conflict of interest to declare.
Ethics approval statement
This research has been approved by the Ethics Review Committee on Research with Human Subjects, Waseda University, application number: 2021-240.
Open access
