AUDIENCE COSTS---advocating the agreement creates them. Those determine U.S. policy AND flip domestic support.

Audience cost is based on the theory that the audience cares more about consistency between a government’s words and deeds than the substance or results of foreign policy. Thus, the argument works independently of interest group the preferences or those of the public. Of course, it is important for decision makers to determine the extent to which the interests of the political leadership and those of interest groups and the public coincide or disagree on policy issues, but it is not germane to the audience cost argument per se. Audience cost is an additional factor that affects decisions of the political leadership, independently of the policy preferences of social groups. The domestic audience punishes its government for backing down from a public threat even if this drawback would be in its interest if the threat had not taken place in the first instance (Clare, 2007; Snyder and Borghard, 2011). One reason for this reaction can be that the audience cares about a government’s competence or a nation’s reputation or honor (Clare, 2007; Smith, 1998; Tomz, 2007). When a government backs down from a public threat, the audience could see a government’s inconsistency between words and deeds as incompetence or damaging for the national honor. Because it stands to reason that voters tend to evaluate the national honor more than interest groups and an attentive public that already have well-defined and specific interests, audience cost tend to be higher when the audience comprises more voters of the general public. This argument is supplemented by Chaudoin (2014) who proves that if the audience has strong preferences over policy choices themselves, they care less about the consistency of the leadership’s words and deeds but more about the final policy. Thus, the less clear the audience’s policy preferences are, the higher the audience costs. Yet, Levy (2012: 286) criticizes that public concerns about national honor are more affected when the adversary has initiated a threat. In this case, a government would tie the hands of the adversary more than its own by threatening publicly. Furthermore, Baum (2004: 608) emphasizes that rational voters would recognize that backing down is a wise decision in some situations and therefore a loss of national reputation is acceptable. It could also be argued that the audience sees threats just as cheap talk. Snyder and Borghard (2011: 440) emphasize that there is evidence that domestic audiences care about national honor, but little evidence that they care independently about empty threats. In the US-South Korea beef dispute, for instance, concerns about national honor seemed to play a minor role compared to the substance of the policy issue, namely, the concerns about public health.

Alternative explanation

These critiques raise the question of whether audience costs can be explained without referring to national honor. The following proposed explanation provides an alternative causal mechanism for audience costs and thus avoids the aforementioned criticism. Instead of referring to the audience’s focus on possible damage to the national honor, the following causal mechanism is based on agenda setting, issue framing, and asymmetric information (for interaction effects of audience costs with other forms of bargaining power see Janusch, 2018). Governments can shape public opinion by setting the agenda and framing issues in favorable ways, inter alia because they have an informational advantage in foreign policy over their domestic audiences (Shapiro and Jacobs, 2002). This is valid especially when international negotiations lack transparency and access to reliable information is limited to governmental actors. The executive branch, in particular, has a special informational advantage, because it is often the only actor that participates in negotiations at the international and domestic level (Shapiro and Jacobs, 2002). Thus, a government can influence the audience’s preferences by defining the conditions under which international cooperation seems to be in the national interest or not (see also Chaudoin, 2016). Should the foreign counterpart not fulfill these conditions, a government can declare an international cooperation to be counter to national interest and threaten to break off the negotiations. To back down from such a threat and to accept the demands of the foreign country would be to approve an agreement that has just been rejected as counter to the national interest, which would invite punishment from the audience.

Because interest groups and (to a lesser extent) the attentive public usually have more information and better-defined interests than the general public (Saunders, 2015), audience costs will tend to be higher, the more the audience consists of the general public rather than just interest groups and the attentive public. A government can affect the size of audience costs in two ways: agenda setting and issue framing.

First, it can increase or decrease the salience of an issue on the public agenda, for example, by going public or private (Baum, 2004; Dijkstra, 2012; Hill, 1998). A government can shape the public agenda by introducing new issues, emphasizing or de-emphasizing existing issues, or discarding issues (Tallberg, 2003). Public threats on the part of a government transmitted by the mass media can make an issue more salient (see also Baum, 2004; Slantchev, 2006). This so-called first-level agenda setting will increase the potential size of audience costs because the audience comprises more voters of the general public. Potential size means the maximum size of audience costs a government can generate when issue salience remains stable. It does not describe the audience costs that a government would suffer if it should back down, but rather the maximum audience costs a government could potentially incur by making public threats or demands, assuming a given composition of the audience. Thus, the greater a government’s ability to set the public agenda, the more it can increase the issue salience and thus affect the audience composition and the potential size of the audience costs (see Figure 2).