The analysis sorts the chords according to their durability and assumes that the x-percentiles are exclusive executive agreements, where x ∈ [0, 0,1]. For example, x = 0.03 assumes that the least durable agreements of 3% are exclusive executive arrangements. It then leaves these agreements out of the analysis, runs the preferred model (5) and collects the estimated coefficient for the contract indicator and its standard error. Note that the assumption that the least durable agreements are exclusively executive agreements is extremely restrictive. In reality, some exclusive executive agreements are much more likely to last longer than agreements between Congress and the executive branch. It is therefore to be expected that this approach will distort the permanence of agreements between Congress and the executive branch upwards, making it more difficult to discern a distinction between treaty permanence and executive agreements. If it can be demonstrated that even in these restrictive cases, contracts survive management agreements, this provides particularly strong evidence of the greater durability of contracts. Finally, the record does not contain any information on the party primarily responsible for the termination of the agreement. All of the theories summarized above focus on the reliability of the United States as a negotiating partner. Agreements that have been terminated for reasons that have nothing to do with U.S. participation should therefore not be included in the analysis. Footnote 85 At the same time, the identity of the party responsible for terminating the contract is only observable if the researcher analyzes each termination individually.
Even then, identifying the person responsible is often a subjective assessment. However, the failure to observe responsibility for the breach of the agreement is unlikely to result in a significant bias in the estimates derived from the data. Bias is introduced only when an unin observed variable is related to both the variable of interest and the outcome variable. Here, this means that bias is only introduced when the other party`s probability of breaking the agreement varies between treaties and agreements between Congress and the executive branch. Footnote 86 However, as noted above, only the reliability of the United States is linked to the distinction between treaties and executive agreements. It would therefore be surprising if a partner country`s propensity to withdraw were linked to the choice of political instrument. Footnote 87 The President may enter into an international agreement on the basis of existing legislation or subject to legislation to be promulgated by Congress; and a treaty is an international agreement concluded in writing between two or more sovereign States and subject to international law, whether enshrined in a single legal act or in two or more related instruments […].