A Matched Pairs Analysis of International Protection Outcomes in Ireland

Gerard Keogh
DOI: https://doi.org/10.48550/arXiv.1705.10131
2017-05-29
Abstract:We examine over 40,000 International Protection (IP) determinations for non-EEA nationals covering a 16 year period in Ireland. We reconfigure these individual outcomes into a set of over 23,000 matched pairs based on combination of direct matching and propensity score matching. A key feature of this approach is that it replicates the statistical features of an experimental set-up where observational data only are to hand. As a consequence we are able to identify those explanatory factors that in fact contribute to the grant of IP. This is a key innovation in the analysis of protection outcomes. We centre our study in the realm of International Relations studies on protection. We are particularly interested in whether immigration policy is a latent tool used to influence the odds of a grant of IP, specifically via the introduction of the Immigration Act 2004. Using both conditional maximum likelihood and mixed effects models we find this is not the case, this conclusion is both novel and profound in a matched pair context. On this basis we conclude there can be little justification for the perception that immigration policy is a latent tool affecting the protection process in Ireland.
Applications
What problem does this paper attempt to address?