Note: Percentages of the total 1 value are displayed for dichotomous variables, while sample standard deviation is displayed next to continuous variables.
The descriptive statistics, presented in Table \ref{table:descstats}, show that maritime insurgencies are similar to other insurgencies on ten of eleven explanatory variables. They only significantly differ on ethnic fractionalization – they tend to be less fractionalized. Insurgencies in less fractionalized countries are associated with decreased duration, so this difference should bias against my expectations.
These summary statistics confirm that green-water and smuggling are common to maritime insurgency. 89.5% of the 19 maritime insurgencies smuggled and 73.7% of maritime insurgencies operated in the green-water littoral area. The low frequency of pirate, underwater, and blue-water operations may indicate that these are costly or risky strategies.
Finally, 26 non-maritime “coastal” insurgencies occurred near their national coastlines. I identified these insurgencies using geo-located event data\cite{ralph_sundberg_introducing_2013} within 10 kilometers of the coast with the statistical program r. I then coded “coastal” insurgencies as insurgencies that were within 10 kilometers of the coastline, but were not maritime insurgencies. Proximity to the water suggests these groups could have been maritime insurgencies in different circumstances. I will probe the link between maritime operations and duration with this variable.
Quantitative Results
I find strong support for my theory of maritime insurgency resilience based on this analysis. The cross-country regressions presented in Table \ref{table:regtable} confirm hypotheses 1-4: maritime insurgencies, maritime smuggling, and green-water operations are associated with increased duration. Those explanatory variables are both statistically significant and substantively meaningful with a coefficient similar to population size – an accepted factor associated with insurgency duration. The greater significance of smuggling shows its importance to maritime insurgency. Coastal insurgencies were not significantly associated suggesting no relationship between coastal insurgencies and duration. There was not a statistically significant relationship between duration and green-water insurgency as expected by Hypothesis 4.