Smoothed population size estimation for an appraisal of human right violations in Colombia in the period 1985-2018

Prof. Alessio FARCOMENI
Professor of Statistics - University of Rome Tor Vergata

Abstract
We estimate the extent of conflict-related human right violations (homicides, forced recruitments, kidnappings) in Colombia between 1985 and 2018; also at age-class, department, gender, and year level.
To do so, first we introduce general principles for conditional population size estimation and show their theoretical properties.
We then propose Ridge and fused-Ridge penalised Chao and Zelterman estimators. The fused-Ridge type penalties are particularly apt at pooling information across time and space, after specification of an opportune adjacency matrix. Penalty parameters are selected through an empirical Focused Information Criterion strategy. A simulation study illustrates how penalised estimators outperform competitors in terms of mean squared error and stability, both at the marginal and conditional level.
Our data analysis indicates that in the Colombian conflict events were somehow polarised in the North-West of the country, peaking in the period 2000-2005; with a final toll of about 6.5-hundred thousand conflict-related homicides, sixty thousand kidnapped, and twenty thousand forcefully recruited. After the signing of peace agreements in 2016 there was a slight resurgence in killings and, more broadly, in human rights violations; even if the conflict can still be considered to be on the way to being resolved. Notably, human right violations occurred with lower than expected regard for age; with a particularly high excess mortality of infants, which is a strong indication of the brutality of the conflict. Homicides impacted males about ten times more often than females, a common figure in conflict contexts.  Women were on the other hand impacted more than expected by forced recruitments.
Finally, in order to provide minor evidence of the numerical discrepancies between a war and a genocide, we analyse recent data on the Israeli's aggression on Gaza in the period between October 2023 and  June 2024; identifying substantial differences in terms of gender-stratified and infant conflict-related mortalities with respect to the Colombian case.