approximately 3
1
2
years to ratify; and those with an average GDP per capita of $20,000 (e.g.,
Australia in the mid-1990s) took approximately 2
1
2
years to ratify, according to Model 1.
18
States with more NGOs per capita are also substantially more likely to commit to the FCCC
(p < .01). Countries with only two NGOs for every 10,000 people (e.g., Equatorial Guinea
circa 2000) took approximately 3
1
2
years to ratify; those with 14 NGOs per 10,000 people
(e.g., Malta circa 1996) took about 2
1
2
years to ratify; and those with 28 NGOs per 10,000
people (e.g., Luxembourg circa 2002) took approximately 1
1
2
years to ratify, all else equal.
Universality is greater than 1 and significant at p < .05, suggesting that states are more
likely to ratify the FCCC when other governments are also doing so. An increase in this
variable from 40% having ratified (the sample mean) to 50% decreases a state’s time until
ratification by over one year. There is, however, no evidence that being an EU member had
an independent effect on states’ propensity to ratify the FCCC.
Model 2 adds one variable to Model 1: CO
2
per Capita * Annex 1. The purpose of
this variable is to test whether there is an interactive effect of greenhouse gas emissions and
being an Annex 1 party on ratification of the FCCC. Although the FCCC commitments
of Annex 1 countries were not binding, to the extent that those countries did state their
“aim of returning individually or jointly to their 1990 levels,” one might expect the extent of
compliant behavior to have some bearing on Annex 1 countries’ decision whether and when
to ratify the FCCC. However, as Model 2 indicates, there is no evidence that this is the case.
The interaction term does not appear to belong in the model.
– Table 1 About Here –
As a robustness check, Model 3 adds two additional controls to Model 1: Original FCCC
18
All predictions in this paper hold the other independent variables at their mean values. All predictions
are based on the predicted ‘Years Until Failure’ of an exponential distribution survival model, which seem
a more intuitive unit of prediction than the hazard ratios the Cox model generates. The predictions of the
Cox Model are available upon request.
18
Convention