Reacting to the Olive Branch: Hawks, Doves, and Public Support for Cooperation

Abstract A popular view holds that foreign policy hawks have an advantage at bringing about rapprochement with international adversaries. This idea is rooted in domestic politics: voters respond more favorably to efforts at reconciliation when their own leader has a hawkish rather than a dovish reputation. Yet, domestic reactions are only part of the equation—to succeed, rapprochement must also evoke a favorable response by the adversary. In this research note, we argue that hawks who make conciliatory gestures may face international liabilities that could offset their domestic advantages. Foreign audiences should view doves who make overtures as more sincere and should therefore be more willing to support cooperation with foreign doves than with foreign hawks. We field a pair of survey experiments to examine whether Americans respond differently when foreign hawks versus foreign doves deliver the olive branch. We find that foreign doves fare better at eliciting cooperation because they are deemed more sincere, though the prospect of military vulnerability limits how willing Americans are to respond positively even to a dove who makes a gesture. Thus, while past research has shown that hawks are better positioned domestically to initiate rapprochement, our findings suggest that they have a harder time eliciting a favorable response from the adversary.

In recent decades, scholars of international relations have embraced the idea that major policy shifts are best accomplished by unlikely leaders. Numerous studies have argued that when making peace with foreign rivals, hawkish leaders have a domestic political advantage. Empirical research has supported this conjecture, finding that domestic audiences tend to respond more favorably when hawks deliver the olive branch to a foreign rival compared to when doves do the same. 1 Yet, rapprochement ultimately requires cooperation by both sides-it succeeds only if the adversary responds in kind. Thus, to understand whether reconciliation is indeed more likely under hawkish leaders than dovish ones, one must consider the foreign reaction to the cooperative gesture. Unfortunately, while scholars have extensively investigated whether and why hawks have a domestic political advantage at pursuing rapprochement, the international dimension has drawn less attention. 2 In this research note, we examine the effect of leader reputation on how foreign audiences respond to a leader's peaceful gesture toward their country. Do hawks enjoy an international advantage at making peace, with foreign audiences more receptive to international cooperation with hawks than with doves? Or, alternatively, is the hawk's domestic advantage at initiating rapprochement accompanied by an international disadvantage at eliciting a favorable response?
We explore foreign reactions to hawkish leaders both theoretically and empirically. We first consider whether and why the same considerations that give hawks a domestic advantage at pursuing conciliation should help or harm the hawk internationally. We argue that logic suggests an international disadvantage for hawks. Individuals in the target state should be more likely to view a leader making a friendly gesture as insincere when that leader is a hawk rather than a dove; hawks who make gestures are acting counter to type, which engenders suspicion rather than trust. Thus, individuals should be more inclined to reciprocate conciliatory gestures by foreign doves compared to identical gestures by foreign hawks, though stronger gestures could narrow the gap by making hawks seem more sincere.
Evaluating hawks' potential foreign disadvantage using historical data is challenging. If, as studies have found, rapprochement is domestically riskier for doves than hawks, we might observe doves delivering the olive branch only when they have reason to expect the adversary to welcome it. Selection bias could therefore obscure differences between doves and hawks in historical situations. To sidestep selection issues and hold constant factors that could confound the effects of leader reputation and gesture strength, we fielded experiments embedded in public opinion surveys.
This approach yields potential insights on two distinct fronts. First, surveys can shed light on mass public opinion. Scholars have argued that leaders gauge public opinion when deciding whether to initiate rapprochement. 3 Likewise, the leaders targeted by friendly gestures have reasons to assess whether their own publics will support responding in kind-for example, public opinion appears to have encouraged the Eisenhower administration to reciprocate Soviet advances in the 1950s, 4 and helped buoy Israel's cooperative response to Egyptian overtures in the late 1970s. 5 Studying the public's reaction therefore helps us understand the domestic political incentives that leaders would face when deciding whether to return or rebuff a foreign enemy's peaceful overture. 6 Second, surveys of the mass public can yield insight into how leaders themselves might think about peace overtures from doves versus hawks, independent of any public pressure. As Kertzer finds, a large body of survey evidence suggests that elites and masses tend to respond remarkably similarly to external stimuli (i.e., experimental treatments); 7 this should be especially true when one considers parts of the public that resemble elites in their traits. Thus, by studying the mass public, particularly "elite-like" sections of the public, we can glean insight into whether elites might make a distinction between hawks and doves who deliver the olive branch.
We examine these questions in the context of the United States, a country that, given its numerous long-standing rivalries, could realistically be the target of a peace overture. We fielded two preregistered survey experiments on diverse samples of US adults in April and May 2021. 8 Both studies described a future scenario involving competition between the US and Russia in the Arctic. All subjects read that the Russian leader made a peaceful overture toward the US, announcing a drawdown of Russian forces in the area. We randomly varied whether the Russian leader was described as a hawk or a dove and the strength of the Russian gesture (i.e., how many military bases he closed). We then asked respondents whether the American president should reciprocate by withdrawing US forces. The two studies were nearly identical, except that in one the US was militarily stronger than Russia in the Arctic (US-Strong study) while in the other the military advantage in the Arctic was flipped (US-Weak study). 9 Thus, the US-Strong study assesses whether the dove's advantage exists when the US is in a favorable military position and drawing down US forces should be relatively popular, while the US-Weak study explores whether foreign doves have an advantage in circumstances when voters might be particularly unwilling to make the kind of risky gesture needed for reconciliation.
As predicted, foreign hawks were generally seen as less sincere than foreign doves who made identical gestures. In the US-Strong study, foreign doves were significantly more likely to elicit support for reciprocation, whether their gesture was low-cost or costly. In the US-Weak study, foreign doves were advantaged only when the gesture was so costly that it reversed the local balance of power. By contrast, when Russia remained more powerful in the Arctic, reciprocation was deeply unpopular, whether the foreign leader was hawkish or dovish. In both studies, the foreign hawk's disadvantage did not diminish with costlier gestures. Finally, further analysis provided little reason to think that US political elites would react differently to our experimental vignettes than the ordinary Americans we interviewed.
Overall, our findings sow doubt about the international appeal of hawks' conciliatory gestures. Empirically, our study centers on one intrinsically interesting case-7. Kertzer 2020. 8. We preregistered our preanalysis plans with OSF (registration DOIs 10.17605/OSF.IO/W2H7K and 10.17605/OSF.IO/CRV57). 9. We fielded the US-Weak study in April 2021 and the US-Strong study in May 2021, but present them in this order for intelligibility. We detail the research process in the appendix.
the United States-but we expect that individuals in other democracies, as well as in nondemocracies, should respond to hawks with similar skepticism. 10 Several realworld examples fit the idea that the dove's foreign advantage could apply both in the US and elsewhere: dovish leaders such as the Soviet Union's Mikhail Gorbachev, West Germany's Willy Brandt, and South Korea's Kim Dae-Jung put their countries on the path to reconciliation with long-term enemies such as the US, Eastern Europe, and North Korea, while overtures by hawks like Iran's Mahmoud Ahmadinejad and Cuba's Fidel Castro (toward the US), as well as China's Deng Xiaoping (toward India), fizzled. 11 Together, our logic and findings not only indicate that leaders' reputations can have countervailing domestic and international effects but they also contribute to the broader literatures on rapprochement, trust-building, and the role of reputation in international relations. 12 Hawks vs. Doves: Who Has the International Advantage?
Protracted rivalries are costly, driving up defense spending, discouraging trade, and reducing the prospects for both formal and informal cooperation. Given these costs, rivals may consider pursuing rapprochement. Numerous studies have examined how leader reputation influences the peacemaking process, particularly how reputation affects leaders' ability to rally domestic support for peace. These studies largely conclude that "hawks"-those with a reputation for favoring conflictual and competitive policies-find it easier than "doves"-those known for preferring diplomacy and cooperation-to sell peace at home. Going against type reassures voters that the policy is sound and that the leader has moderate preferences. 13 Thus, hawks who deliver the olive branch are more likely to receive domestic support than doves, an idea supported by much empirical evidence. 14 However, reconciliation requires the target to return the friendly gesture. Thus, to understand whether hawkish leaders are indeed better positioned to achieve reconciliation, one must consider the reaction of policymakers and ordinary citizens in the target country. We assume that, on average, individuals are security seeking, preferring to spend resources on military forces (such as arms buildups and troop deployments) mainly when they think it necessary to deter aggression. Thus, individuals are more willing to reciprocate foreign gestures-relaxing their country's defenses-10. Of course, in nondemocracies, the pull of public opinion might be weaker, and leaders might diverge more strongly in their preferences and beliefs from the public; nevertheless, our theoretical logic has the potential to apply in a broad range of countries.
11. See the appendix for examples of doves' versus hawks' success at rapprochement. 12. On rapprochement and trust building, see Axelrod 1984;Kydd 2007;Osgood 1962. For a review of the literature on reputation, see Dafoe, Renshon, and Huth 2014. 13. Cowen and Sutter 1998;Cukierman and Tommasi 1998;Schultz 2005. 14. Mattes and Weeks 2019;Sigelman and Sigelman 1986;Trager and Vavreck 2011. when they think the foreign leader has peaceful rather than expansionist motives. 15 However, individuals are unsure about the foreign leader's intentions. 16 Even when the adversary makes a seemingly friendly overture, the possibility of later defection makes it risky to reciprocate. 17 How do audiences decide whether a foreign leader genuinely wants peace? Recent evidence suggests that reputations do not only matter domestically: observers also make guesses about foreign actors' intentions (and therefore, future behavior) based on country or leader reputation. 18 Following this logic, when assessing whether a foreign adversary genuinely wants peace, voters and policymakers should consider the foreign leader's reputation.
We thus investigate how, from the perspective of a country deciding whether to return a friendly gesture, the foreign leader's reputation influences support for reciprocation. When the foreign leader making the conciliatory gesture is a dove, their reputation reinforces the gesture and indicates that the leader truly seeks improved relations. By contrast, a hawkish reputation conflicts with the conciliatory gesture, generating suspicion rather than reassurance. Individuals may worry that a foreign hawk has expansionist intentions, and that the concession is a ploy to lull the target into a false sense of safety for later exploitation. By this logic, individuals should be more likely to think that foreign doves making peace overtures are sincere. These perceptions of sincerity, in turn, should increase support for reciprocation.
This reasoning suggests that hawkish reputations may be a double-edged sword: the domestic rewards of counter-to-type behavior may be offset by penalties at the international level. Peace initiatives could be easier for hawks to sell at home, but hawks may be less likely to invite foreign trust and thus foreign willingness to reciprocate. If so, the prospects for peacemaking by hawks would be narrower than envisioned by past research. This logic yields the following hypotheses: H1 (Dove's Advantage): Individuals are more likely to support reciprocating a gesture by a foreign dove than reciprocating the same gesture by a foreign hawk.
H1a (Sincerity 1): Individuals are more likely to believe that a foreign leader wants peace if the foreign leader is a dove rather than a hawk. 19 Our prediction that doves have an international advantage aligns with prior work by Clare, though we specify a different mechanism and expect the dove's advantage 15. Kydd 2007. 16. Yarhi-Milo 2014. 17. Schultz 2005. Dafoe, Renshon, and Huth 2014;Kertzer, Renshon, and Yarhi-Milo 2021;Renshon, Dafoe, and Huth 2018. 19. This wording deviates slightly from our preanalysis plan, which wrongly focused on the sincerity of the gesture rather than that of the leader. Our theory suggests that, for decisions about reciprocation, the key mechanism involves beliefs about whether the foreign leader is committed to peace in the long term.
to apply more universally. Clare argues that foreign doves have an advantage only when their own domestic opposition is hawkish. Under such circumstances, striking a deal with the dove appears preferable, since the dove will likely offer more concessions than a hawkish successor government. 20 In contrast, we argue that the dove's advantage does not depend on the domestic political landscape. It also does not result from expectations that doves will grant more generous deals, but rather from beliefs that doves are more likely to follow through on rapprochement in the long run.
We also explore how gesture strength, which we allow to vary independently of reputation, affects the target's perceptions. A key insight from the literature on rapprochement is that the nature of peace overtures influences their success. 21 To achieve détente, gestures must be costly enough to make the sender vulnerable to defection, separating opportunists from those with a genuine desire to reconcile. 22 Recent experimental work confirms that costlier foreign gestures increase support for cooperation. 23 We build on these insights but hypothesize that the strength of the gesture should matter more for peace overtures by foreign hawks. According to this logic, foreign doves making conciliation efforts should be seen as relatively sincere whether the gesture is costly or not. By contrast, there is more opportunity for signal strength to affect perceptions of sincerity when the leader is hawkish. Individuals should tend to doubt foreign hawks' sincerity when the gesture costs little, but unequivocally costly signals can reveal that hawks are serious about peace. We thus posit an interaction between leader reputation and signal strength, whereby strong signals by hawks do more to increase support for reciprocation (relative to weak signals) than strong signals by doves. In other words, signals by hawks have more room to be informative, helping them close the gap with doves.
H2 (Gesture Strength): As the strength of the foreign leader's gesture increases, the dove's advantage shrinks.
H2a (Sincerity 2): As the strength of the foreign leader's gesture increases, the gap in perceptions of the sincerity of doves versus hawks shrinks.
H2b (Costly Signaling): Individuals are more likely to support reciprocation of a stronger gesture.
Beyond these hypotheses, our preanalysis plan (PAP) outlined three additional sets of expectations. First, we considered another mechanism that could produce a dove's 20. Clare 2014. Clare's empirical tests use observational data, potentially exposing them to the selection bias mentioned earlier. Testing Clare's argument against ours in a survey experiment would require varying the domestic circumstances in the foreign state, which is beyond the scope of our experiments. However, considering whether doves' international advantage is even more pronounced in the presence of a hawkish opposition is a fruitful avenue for future research. 21. Kupchan 2010;Kydd 2007;Larson 1987;Mattes 2018;Osgood 1962. 22. Kydd 2007. Kertzer, Rathbun, and Rathbun 2020. advantage: individuals might prefer reciprocating gestures by doves not because they think doves are more sincere, but because they think doves are more exploitable. If so, doves' gestures would be reciprocated in the short run but rapprochement would fail in the longer run. 24 Second, we considered the possibility that nonreciprocation will prompt greater fears of foreign punishment when the foreign leader is hawkish, lending hawks rather than doves an international advantage. Third, we proposed that individual-level beliefs about the foreign threat moderate the dove's international advantage. We did not find support for any of these three expectations. To conserve space, we discuss these hypotheses and empirical tests in the appendix.

Research Design
To test our hypotheses, we fielded two preregistered survey experiments in the United States. We chose the US for two reasons. First, the US is a global superpower whose enduring rivalries pose a risk of large-scale military conflict and affect US relations with third states and relationships between third states. Understanding the conditions under which the US would embrace foreign peace overtures is therefore important for understanding broader patterns of international conflict and cooperation. Second, the research finding that hawks have a domestic advantage has taken place primarily in the US. 25 It is therefore useful to know whether, in the same country in which domestic hawks are known to have a peacemaking advantage, foreign hawks instead generate a reconciliation liability. Of course, future research could gain valuable insight by fielding experiments in other countries; in the meantime, studying the US provides insight into an important case while also developing an experimental approach that could be used to study the effects of foreign-leader reputation elsewhere. We administered the US-Weak study to 2,389 respondents in April 2021, and the US-Strong study to 2,186 US-based respondents in May 2021. 26 In both studies, we first measured respondents' party affiliation and foreign policy views and prescreened for attentiveness. 27 Those passing the attention checks progressed to the experiment. The vignette began: "We are going to describe a situation the US could face in the future, in the year 2027, involving Russia. Some parts of the description may seem important to you; other parts may seem unimportant." Our scenarios involve Russia because Russo-American relations fit the theoretical parameters for rapprochement between rivals. Americans have consistently ranked Russia among the US's four "greatest enemies" in recent years. 28 Moreover, Americans, even among the mass public, would likely be aware of a Russian leader's identity and reputation. 29 An Arctic scenario allows us to manipulate key features of the context, and future tensions between the US and Russia there are likely. News media report about the Russian buildup in the Arctic, and the US government acknowledges the threat. 30 Both studies randomized two key features of the scenario: the foreign leader's reputation (hawk vs. dove), and the strength of the cooperative gesture (low cost or costly). We also randomized the US president's party (Democrat or Republican) and reputation (again, hawk or dove). We did this so respondents would not draw conclusions about the US president based on the foreign leader's reputation and gesture, which could introduce confounding. 31 In general, our design mirrored Mattes and Weeks' study of the domestic hawk's advantage for comparability. 32 We first described the US president, choosing a fictional president so we could manipulate party affiliation and reputation. We held constant gender (a man) and experience (he became president in 2025 after serving five years in the Senate). We then described his hawkishness/dovishness and party affiliation. Table 1 shows the wording of these and other treatments.
We next described Russia in 2027. We noted that "Russia remains a non-democracy" and described a fictional male Russian leader who had entered office two years prior; we chose a fictional leader so we could manipulate reputation for hawkishness/ dovishness. Half of our respondents read that the Russian leader "has a reputation for favoring military solutions over diplomatic ones," and the other half that he "has a reputation for favoring diplomatic solutions over military ones." We reinforced 28. For polls on Russia, see <https://news.gallup.com/poll/1642/russia.aspx>. The US public is politically polarized regarding Russia. In our data, 57 to 58 percent of Democrats deemed Russia an "enemy" or "unfriendly," compared to 51 to 52 percent of Republicans. However, we do not think this polarization would affect our conclusions. Moreover, our scenario is situated in the future and involves a fictional Russian leader, which should reduce polarized reactions. In the appendix, we show that the dove's advantage holds among both Democrats and Republicans.
29. See <https://www.pollingreport.com/russia.htm>. Renshon, Dafoe, and Huth 2018 suggest that leader-specific reputations are especially likely when, as in our scenario, leaders have significant foreign policy authority (such as autocratic leaders) and/or security considerations are involved. One might wonder whether, within a tense rivalry, domestic audiences would ever view foreign leaders as dovish.
To check plausibility, the appendix explores the case of Gorbachev, who quickly became broadly known to the US public, was seen as more cooperative than his predecessors, and inspired greater trust than the Soviet state.
30. See, for example, Lopez 2021; Walsh 2021. The US-Russia Arctic scenario has also been used by other scholars (e.g., Gottfried and Trager 2016). Mattes and Weeks 2019 found that the hawk's domestic advantage was nearly identical whether the foreign country was Russia or China. 31. We also asked basic factual questions about the scenario to help respondents internalize information and to gauge attentiveness. 32. Unlike Mattes and Weeks 2019, we did not have a "no-gesture" condition. Our interest is in how the foreign leader's reputation influences reciprocation given a gesture, rather than how the foreign leader's reputation moderates the effect of a gesture relative to no gesture. The latter is an interesting topic for future research (though we start to explore it by studying how leader type moderates the difference between strong and weak gestures). these treatments by mentioning previous statements by the Russian president about military conflict versus peaceful diplomacy. 33 We next described the international situation. We said that "a major security concern for the US in 2027 is its very tense relationship with Russia" and that the two countries "disagree strongly over many important foreign policy issues." We then explained that "the Arctic is a key site of competition," that "40 percent of all trade is shipped through Arctic waters," and that "the area contains massive oil and gas reserves." We further reported that "the Pentagon has called the Arctic 'the first line of defense'" and that "US officials have repeatedly warned about Russian ambitions in the Arctic." To fix perceptions of military power, we said that "in 2027, the US military is overall about 30 percent stronger than the Russian military." Had we omitted information about military power, respondents might have guessed based on hawkishness/ dovishness or gesture type. Assumptions about military power, in turn, could affect willingness to reciprocate, introducing confounding.
The two studies differed in the local balance of military power, however. In the US-Strong study, subjects learned that "in the Arctic itself, the US is about 30 percent stronger than Russia." By describing the US as having greater local military power than Russia, we aimed for a scenario in which Americans would be relatively He has often said that military force is essential to protecting Russian national security. He has frequently declared that he will use force when necessary. He is known for saying that "the only way to achieve peace is to be ready for war." … has a reputation for favoring diplomatic solutions over military ones. He has often said that military force is not the answer to protecting Russian national security. He has frequently declared that he believes in diplomacy and negotiations and will use military force only as a last resort. He is known for saying that "the only way to achieve peace is to act peacefully." FOREIGN LEADER CONCILIATORY GESTURE Low-Cost* Costly* … declares that Russia is closing 1 of its 7 (9) military bases in the Artic and that he is willing to consider closing additional Russian bases in the area.
… declares that Russia is closing 5 (7) of its 7 (9) military bases in the Arctic and that he is willing to consider closing additional Russian bases in the area.
*Treatments in parentheses denote the US-Weak study, which differed from the US-Strong study only in the balance of local military power.
33. We focus on words rather than previous actions to avoid leaking information about other threats, which might influence willingness to cooperate. likely to support reciprocating the Russian gesture; otherwise, we reasoned, we might encounter floor effects. 34 Thus, in the US-Strong study, both Russia and the US "have a major military presence in the Arctic," Russia "has seven military bases in the Arctic," and the US "has nine military bases in the Arctic." We described both countries' bases as housing "troops, icebreakers, aircraft, missiles, and nuclear-armed submarines." The US-Weak study was nearly identical, except that the military balance in the Arctic was flipped: Russia was 30 percent more powerful than the US in the Arctic, with nine military bases to the US's seven. Thus, the US-Weak study represents a scenario in which Americans might be particularly reluctant to reciprocate.
Next, in both studies, we provided information about the Russian leader's conciliatory gesture. Subjects read that the Russian president "makes a speech about foreign policy. He announces that he hopes to find common ground with the US on the Arctic as well as on other disputed issues." Respondents were then randomly assigned to one of two treatment conditions. In the "low-cost gesture" treatment in the US-Strong study, the foreign leader "declares that Russia is closing one of its seven military bases in the Arctic and that he is willing to consider closing additional Russian bases in the area." (In the US-Weak study, Russia closed one of nine bases.) This treatment is designed to move beyond classic "cheap talk" in offering a tangible concession, but closing one of seven (or nine) bases carries relatively few risks. It is thus a comparatively low-cost but not empty gesture.
By contrast, in the "costly gesture" treatment in the US-Strong study the Russian president "declares that Russia is closing five of its seven military bases in the Arctic and that he is willing to consider closing additional Russian bases in the area." (In the US-Weak study, Russia closed seven of nine bases.) Unilaterally dismantling all but two Russian military installations significantly weakens Russia's position and exposes Russia to the risk of the US exploiting this gesture to cement local dominance. This gesture thus matches how past literature conceptualizes effective peace overtures.
After a bullet-point summary, respondents read that the US president "is considering whether to pursue a more cooperative or conflictual policy toward Russia. He is deciding what to do about US bases in the Arctic." We then asked whether he should "build additional US bases in the Arctic," "keep the same number of US bases in the Arctic," or "reduce the number of US bases in the Arctic." 35 As a follow-up, depending on whether respondents said the president should increase or decrease US military bases, we asked how many bases should be built/closed. We then probed mechanisms and measured demographics and characteristics such as political interest and religiosity.
34. We realized the potential importance of a US power advantage only after fielding the US-Weak study. Appendix A describes our research process. 35. We also considered an alternative dependent variable: approval of the US president. Unfortunately, subjects could assume that the US president elicited a costly versus a low-cost foreign gesture. If so, differences in approval by gesture strength could be due to the strength of the foreign gesture leaking information about the president's competence, generating confounding.

Results
How did subjects respond to the Russian leader's gesture? We begin by considering this question from the perspective of our public samples as a whole; a later section investigates whether these findings would be likely to differ among political elites.
We first examine whether doves have an international advantage at delivering the olive branch (H1). To test this core hypothesis, we compared support for reciprocating the cooperative gesture when the foreign leader was a hawk to support for reciprocating the identical gesture by a foreign dove. 36 For ease of interpretation, we created a dichotomous dependent variable, reciprocate, coded as 100 if a respondent supported reducing US bases in the Arctic, and 0 if they preferred to keep or increase the number of US bases. Thus, effects can be interpreted as the percentage increase in support for reciprocation when the foreign leader is a dove, as opposed to a hawk. We examined the dove's international advantage hypothesis (H1) in both of our experimental scenarios-that is, both when the US was relatively more powerful than Russia in the Arctic and when Russia dominated locally. Following our preanalysis plan, we first focus on the low-cost gesture scenario (closure of one Russian base).
In the US-Strong study, consistent with our expectations, respondents made a large distinction between foreign hawks and doves. About 30 percent of respondents supported reciprocating the conciliatory gesture when the foreign leader was a hawk (Figure 1, top). When the foreign leader was a dove, in contrast, support for reciprocation swelled by sixteen percentage points, an effect that is both substantively large and statistically significant. We further explored whether the foreign leader's reputation affected the number of bases respondents thought should be closed or built in the Arctic: the preferred number of bases was 0.37 smaller for doves than hawks (p < .05). These findings persist when we control for the US president's party and hawkishness/dovishness and demographic/attitudinal variables as specified in the PAP. Moreover, our estimates remain the same if we include only subjects who passed at least 85 percent of our post-treatment attention checks.
An important question, however, is whether these findings generalize to situations in which voters are likely to be more skeptical of reciprocation. The US-Weak study, accordingly, described a scenario in which Russia was more powerful in the Artic than the US (nine bases for Russia, and seven for the US) and remained more powerful even when it closed one base.
Here, average support for reciprocation was much lower than in the US-Strong study. Moreover, counter to our expectations, respondents did not distinguish between foreign hawks and doves. About 17 percent of respondents supported 36. Balance tests reveal that in both studies, random assignment of foreign hawk/dove was successful with respect to both the other randomized treatments and demographic/attitudinal variables. For simplicity, our main model therefore averages across the other treatments and demographic/attitudinal variables. See the appendix for tables for all analyses reported here and supplementary analyses controlling for other treatments and demographic/attitudinal controls. reciprocating a foreign hawk's conciliatory gesture, a figure that grew by less than three percentage points when the foreign leader was a dove (Figure 1, bottom). (This effect is substantively small and statistically insignificant.) We further found that the preferred number of bases, our alternate dependent variable, was virtually identical across the two treatment groups. These (non)findings persist when we control for the other treatments and demographic/attitudinal variables. Moreover, the estimates are nearly identical even when excluding respondents who passed less than 85 percent of the post-treatment checks. Thus, this nonfinding is unlikely to be due to inattentiveness.
In sum, Americans seem reluctant to return low-cost gestures when the US is in a position of weakness, a reticence that dovish foreign leaders cannot overcome. It is interesting to consider what our findings might imply when the shoe is on the other foot, and the US is trying to convince a foreign rival to make peace. If our findings generalize to other states, the US, because of its military dominance, may find it difficult to persuade rivals to relax their defenses. Thus, our findings suggest a peacemaking analog to the "Goliath's curse": strong states might find it difficult not only to compel states to do their bidding, 37 but also to persuade rivals to make peace.

Perceptions of Sincerity
Next, we evaluate our predictions about the sincerity mechanism. To assess whether voters see foreign doves as more sincere than foreign hawks, we asked whether subjects agreed or disagreed that the Russian president "wants peace with the US." To facilitate interpretation, we again created a binary variable coded 100 if the respondent agreed and 0 otherwise. To test H1a we calculated the effect of a dovish foreign leader on the percentage of respondents who agreed that the Russian president wants peace. According to our PAP, we again focused on the low-cost-gesture condition. In the US-Strong study, voters were a massive thirty-one percentage points more likely to think a dovish Russian leader sincerely wanted peace, compared to a hawkish one. To see how perceptions of sincerity are related to support for reciprocation, we regressed support for reciprocation on our three randomized treatments, the sincerity mediator, and the attitudinal and demographic control variables. We found that perceiving the foreign leader to be sincere led to a sixteen-point increase in support for reciprocation, a result that was highly significant. 38 In the US-Weak study, we likewise found that dovish foreign leaders were perceived as more sincere (again, a thirty-one-point effect). Here, however, perceptions of sincerity did not appear to reassure the public that reciprocation would be wise, which could explain why a dovish foreign leader had no effect in the US-Weak study. Perceptions of sincerity increased support for reciprocation by only five percentage points (p < .08). This finding underscores the impression that Americans were hesitant to have fewer bases in the Arctic than Russia.

Gesture Strength
Finally, we investigate whether the strength of the conciliatory gesture influences whether voters react differently to foreign hawks versus doves. We preregistered the hypothesis that the gap between foreign doves and hawks should shrink when the gesture is more costly (H2). According to our PAP, we compare the dove's international advantage when Russia makes a low-cost overture versus when Russia's gesture is costly. We use the same preregistered procedures as for H1.
We first assess the effect of gesture strength in the US-Strong study. As predicted by H2b, the overall level of support for reciprocation was higher when the foreign gesture was costly: 45 percent of subjects supported closing US bases, as opposed to 37 percent when the gesture was low-cost.
Counter to H2, however, costlier gestures did not produce a significantly smaller dove's advantage: the effect of a dovish foreign leader was sixteen percentage points 38. The appendix shows a full mediation analysis. The sincerity mechanism explains approximately onethird of the total effect of our treatment. Note that causal mediation analysis, like our estimate of the effect of sincerity on support for reciprocation, relies on untestable assumptions, including that the observed values of the mediators are independent of the subject's treatment status and independent of all pretreatment confounders (Imai, Keele, and Tingley 2010). when the gesture was low-cost, and twelve points when the gesture was costly (Figure 2). While the direction of this effect is consistent with our hypothesis, it is not statistically significant. Likewise, when we use the alternate dependent variable measuring the desired number of US bases in the Arctic, the difference between doves and hawks is not significant (the estimated effect of foreign dove actually increased slightly with costlier gestures; see the appendix). Furthermore, counter to H2a, the gap in perceptions of sincerity for foreign doves versus hawks was similar for costly and low-cost gestures (gaps of 28 versus 31 points). Thus, counter to our expectations, a stronger gesture does not significantly close the gap between hawks and doves.
We next move to the US-Weak study. Once again, as predicted by H2b, average support for reciprocation was substantially stronger when the foreign gesture was costly: 44 percent of subjects supported closing US bases, as opposed to only 18 percent when the gesture was low cost. Moreover, our results remained at odds with H2 ( Figure 2). The effect of a dovish foreign leader was actually larger when the gesture was costly (6 versus 3 percentage points), though the difference was not statistically significant. As the appendix shows, we found similar patterns with the alternate dependent variable. 39 Interestingly, we found a statistically significant dove's advantage, that is, support for H1, when the US started in a position of weakness but the Russian gesture was costly ( Figure 2). Why did the dove's advantage materialize for a costly but not a low-cost gesture? We suspect that it has to do with the balance of power and how the combination of the Russian leader's gesture and US reciprocation would change that balance. Given the US position of vulnerability at the outset of the US-Weak scenario, a low-cost Russian gesture meant that closing any US bases would reify the US's relative weakness. Voters were extremely unlikely to support reciprocation in this scenario even when they thought the leader was sincere. 40 Put differently, the unattractiveness of the US remaining in a position of local military vulnerability seems to generate a floor effect that squashes the dove's advantage.
When the gesture was costly, however, the picture changed. Now, as one would expect, perceptions of sincerity were strongly correlated with support for reciprocation: subjects who saw the gesture as completely sincere were much more likely (by twenty-two percentage points) to support reciprocation than those who saw the gesture as completely insincere. 41 We suspect that this is because the costly gesture would reverse the local military balance, removing the universal aversion to reciprocation and allowing the dove's advantage to emerge. 39. Because there was no dove's advantage in the US-Weak study when the gesture was low cost, there was no room for the (non)effect to shrink and thus to find support for H2.
40. Recall that although voters perceived doves as more sincere (by thirty-one points), perceptions of sincerity only weakly predicted support for reciprocation. Support for reciprocation was also very low in this scenario (18 percent). 41. At the same time, the gap in perceptions of sincerity shrank, lending support to H2a: dovish leaders increased perceptions of sincerity by twenty-one percentage points, significantly less than when the gesture was low-cost (thirty-one percentage points).
Overall, combining insights from our two studies, we conclude that doves fare better at eliciting reciprocation than hawks, but that balance-of-power considerations limit doves' international advantage. When the US would remain in a position of weakness despite a friendly foreign gesture, the foreign leader's reputation did not affect willingness to reciprocate-support for cooperation was always overwhelmingly low. But when the US was in a position of strength (either at the outset of the scenario, or because the Russian leader's gesture reversed the balance of power), respondents strongly preferred reciprocating foreign doves' gestures over hawks'. This suggests that, when the prospects for rapprochement are not too grim, doves are better positioned to kick-start the peace process. Moreover, we found no evidence that the hawk's disadvantage diminished when the leader made a costlier gesture. While we cannot exclude the possibility of gestures so costly that they would allow foreign hawks to overcome their international disadvantage, our findings suggest that, at best, hawks' concessions would need to be substantial.

Would the Results Have Differed Among Political Elites?
Studying the mass public is important in its own right because public opinion creates political incentives for leaders who are contemplating peace. However, it is also  important to ask whether the elites who ultimately make policy would have reacted similarly to our experiment, even if they were insulated from public pressure. Drawing on a wealth of past experimental evidence, Kertzer finds that elites and ordinary members of the public react very similarly to experimental treatments. 42 Nonetheless, as Kertzer points out, ordinary voters differ from political elites in traits such as age, gender, education, and income. Moreover, they presumably are less interested in politics than the elites who devote their lives to leadership roles. One might therefore wonder whether any of these five "elitelike traits" moderate our treatment effects. We used two different approaches to assess whether our results hinge on the demographic and attitudinal composition of our sample. 43 First, we interacted the five elitelike traits with our experimental treatments. With few exceptions, those who scored high on these traits reacted no differently from regular voters, consistent with Kertzer's finding about the remarkable similarity of elite and mass reactions to experimental treatments. Second, we carried out the hawk/dove analyses on subsets of individuals who most closely resemble political elites, using a variety of definitions of political elites. Our conclusions do not change when we consider subjects in these elitelike subsets only. 44 Of course, it remains possible that political elites differ from ordinary voters in ways not captured by these traits; for example, perhaps political elites are aware of the idea that hawks have a domestic advantage, and would therefore be more sympathetic to cooperation with hawks. Nonetheless, we do not see reason to think that, had we fielded our experiment on US political elites, their reactions would have differed from those of ordinary voters.
Of course, our experiments cannot tell us whether elites would give hawks a disadvantage in real-world decision making. On the one hand, perhaps elites would be less likely than voters to rely on heuristics such as reputation, given their access to much more detailed intelligence assessments. On the other hand, perhaps elites would have stronger priors about the foreign leader's reputation, which could amplify the effect of leader reputation. Future research should evaluate the extent to which the results uncovered in our experiments generalize to real-world elite decision making.

Conclusion
This research note complements existing work on hawks' domestic political advantage at peacemaking by examining how the target of rapprochement responds to conciliatory gestures by hawks versus doves. Logically, the same dynamic that gives hawks a domestic edge at making peace might create an international liability. A hawk's counter-to-type behavior may sow doubt among foreign audiences about 42. Kertzer 2020. 43. We did not anticipate these analyses in our PAP. 44. The appendix describes the analyses for H1 in detail. Analyses of the other hypotheses are available on request. the sincerity of the leader's desire for peace, making voters and policymakers abroad less willing to reciprocate overtures by hawks compared to identical overtures by doves. We further hypothesized that the dove's international advantage should diminish with costlier signals.
To test these conjectures, we fielded two survey experiments on large samples of US respondents. In line with our expectations, foreign doves who made a peace gesture were viewed as more sincere than hawks who pursued an identical effort. We also generally found evidence of doves' international advantage, except when recipients of the olive branch were already very unlikely to support rapprochement. When the US started out in a position of local military superiority or when the Russian gesture gave the US the upper hand, respondents showed significantly greater willingness to reciprocate a foreign dove's overture compared to a hawk's. When Russia was locally more powerful and the Russian gesture did not change this power balance, foreign-leader reputation did not matter. Under these circumstances, public willingness to draw down US forces was so low that not even the prospect of dealing with a Russian leader who seemed sincerely committed to peace convinced respondents to take a chance. Further analysis suggested that these conclusions would be unlikely to change had we fielded our experiments on the kinds of political elites who ultimately make policy.
Our research thus shows that the hawk's advantage at rapprochement may be more limited than previous work implies. The hawk's advantage at garnering support from domestic constituents for the initiation of rapprochement does not appear to be matched by an advantage at eliciting support for reciprocation from foreign audiences. Instead, the adversary is more likely to favor reciprocating a dove's overture, though balance-of-power concerns limit foreign willingness to respond positively even to a dove's gesture. Furthermore, hawks may find it difficult, if not impossible, to compensate for the sincerity gap by making costlier gestures.
Our study also has implications for the literatures on trust building among adversaries and the role of reputation in foreign policy. While previous work emphasizes the character of successful conciliatory gestures, 45 we highlight how leader reputation can influence whether a peace overture is perceived as credible. Recent research on crisis bargaining finds that a leader's past behavior shapes expectations about future behavior. 46 Here we add to these insights by showing that a foreign leader's reputation for hawkishness/dovishness can affect preferences about rapprochement.
Finally, our study raises many questions for future research, including how individuals come to see foreign leaders as either hawks or doves, how power imbalances influence public support for rapprochement, and whether the dynamics uncovered in our US-based experiments replicate in other countries.
A particularly interesting avenue for future research concerns the interaction between domestic and foreign leader reputations. If hawks have a domestic political 45. Kydd 2007;Osgood 1962. 46. Renshon, Dafoe, andHuth 2018. advantage but a foreign disadvantage, then the constellation of leader reputations most likely to promote rapprochement could depend on the barriers to peace in each country. 47 For example, if the barriers to peace are high in country A but low in country B, then the best combination might be a hawk in country A and a dove in country B. One high-profile historical case arguably fits this pattern: the Cold War ended as a result of the efforts of President Reagan (a hawk; high domestic barriers) and Mikhail Gorbachev (a dove; lower domestic barriers, given the USSR's economic decline). At the same time, how interactions between hawks and doves unfold might also depend on how individuals in the target state interpret the responses of audiences in the state that made the initial gesture. 48 For example, individuals may prefer to reciprocate a gesture by a foreign dove but may worry about the prospects for peace if they see the foreign dove encounter criticism at home. Alternatively, individuals who worry about engaging with a foreign hawk may feel reassured if the hawk receives strong backing at home. These possibilities suggest that a systematic theoretical and empirical examination of interaction effects between domestic and foreign leader types, and domestic and international audience responses, could shed light on the conditions most conducive for rapprochement. In the meantime, our findings suggest that hawkish leaders may not be the unproblematic emissaries of peace that one might hope.

Data Availability Statement
Replication files for this research note may be found at <https://doi.org/10.7910/ DVN/BFRBXF>.