The focus on selection models is a very indirect approach to identifying differences in policy instruments based on a number of sound assumptions, such as a correct specification of the model and a causal relationship between the identified models and the hypothetical motives. Without these assumptions, the observed actions can be the result of a variety of different motivations, making it impossible to deduce which instrument is the most reliable. For example, Martin`s conclusion that a high GDP correlates with the use of a contract does not necessarily mean that the contract is used because the partner country has a high GDP. As seen below, the treaty is a common footnote 59 for agreements between the United States and Western European countries. Footnote 60 On average, these countries tend to have a high GDP per capita, but they also share a number of other characteristics that could explain the results, such as a common history of Roman law and respect for legal formalism. Moreover, Martin`s findings that «high-quality» agreements increase the likelihood of contract use are consistent not only with signal theory, but also with the hypothesis proposed by Bradley and Morrison that «large» agreements are subject to selective Senate attention. Table 1 provides summary statistics. As can be seen, 5% of all agreements between 1982 and 2012 were concluded in the form of a treaty, which makes the use of the treaty exceptional. 20% of all agreements ceased to apply at some point during the observation period. The average agreement was valid for 15.26 years. Among the agreements that are no longer in force, the average shelf life is 7.3 years. LpPC values range from −17 to 17 with an average of −0.10.

On average, 50 percent of the seats in the Senate at the time the deal was signed were held by the president`s party. For 71% of the deals, the government was divided, with the White House exhausting one party and either the Senate, the House of Representatives, or both held by the other. Taken together, these figures suggest that the average deal could not have been adopted in the form of a treaty without the support of all parties, making it a potentially costly instrument. The results of international agreements can be compared in a number of dimensions. One possible measure is the degree of compliance with an agreement. However, the comparison of agreements on the basis of implementation rates has several disadvantages in this context. «Compliance» is not only difficult to define, but notoriously difficult to measure and verify in most contexts. Footnote 62 Even if it were possible to measure compliance accurately, it would nevertheless leave open the question of how compliance levels can be compared between different agreements in different thematic areas. Footnote 63 Motivated by the theoretical work discussed earlier, this article instead compares treaties and agreements between Congress and the executive branch based on the strength of the commitment involved, as measured by sustainability.

In fact, the dynamics surrounding the ex-post agreement between Congress and the executive branch are often described as if the executive branch had submitted a particular agreement to Congress and Congress would then have considered that agreement in isolation. .