they are already doing. Concerning the different types of agreements (rules of war is the
omitted category), arms limitation and commodity agreements both have positive
coefficients that are statistically significant (although also only at the 90% confidence
level). The coefficients for the other types of agreements are not statistically significant.
Neither is the coefficient for Major power proposer, which is not reported as it is not
statistically significant for any of the models. Whether or not a major power makes the
first proposer does not seem to affect the length of negotiations.
Looking at only the security related agreements, the signs on all of the
coefficients are the same as for the larger set of data. However, only the coefficients for
NGO proposer and Arms limitation remain statistically significant. Whether the
agreement specifies duration no longer has an effect
cooperation of the agreement.
Finally, to better test Hypothesis 3, the variable Duration is used instead of
Duration limit. The results are reported in Table 3. First, there is a much smaller number
of observations, as there are only 26 agreements that specify duration. Second, the
coefficient for the duration of the agreement is in the opposite direction of what was
expected, indicating that the longer the duration of the agreement, the shorter the duration
of negotiations will be. However, this coefficient is not statistically significant. Third,
the coefficients for the number of charter members and the depth of cooperation are both
positive and statistically significant.
To summarize these findings, negotiations for multilateral agreements are
expected to be longer when an NGO makes the first proposal and when the agreement
requires a deep level of cooperation. Agreements that are concerned with arms
11
This is not surprising as there are only 6 of the 83 security related agreements specify duration.
19
Convention