AOPSSS #1: “Political Regimes and Refugee Entries: Motivations behind Refugees and Host Governments.”

In the first session of AOPSSS, Masaaki Higashijima (Tohoku) and Yu Jin Woo (Waseda) presented their fascinating new paper “Political Regimes and Refugee Entries: Motivations behind Refugees and Host Governments.” [Slides] Loren Collingwood (UC, Riverside), Kevin Cope (UVA), and Adrian Shin (Colorado) offered excellent discussant comments, and the other session participants provided valuable feedback on a range of issues. We’re all looking forward to seeing the next version of the paper!

If you have any questions or comments for the authors, please add them below!

6 replies on “AOPSSS #1: “Political Regimes and Refugee Entries: Motivations behind Refugees and Host Governments.””

1) Inverted U shape is derived casually. I’d like to see more formal derivation (e.g. using math).
2) Regional analysis is better, though I wonder “region” is tricky to define. Instead, the authors may control summary variables of neighboring countries (e.g. average polity score).
3) Confidence intervals are wider for authoritarians than democracies. It might be because variance is heteroskedastic, not because the relationship is quadratic.

I really enjoyed the online workshop yesterday; many thanks go to Charles Crabtree, who has been working so hard to make this happen and let us be connected, two wonderful presenters Masaaki Higashijima and Yu Jin Woo, and everyone online yesterday.

My comment to the study: I was also wondering if the two processes, pull and push, can be analyzed in a combined model (I suppose this was first mentioned by one of the discussants and Yusaku Horiuchi). I would be persuaded if we can see more justification from the authors! Anyhow, it is a very nice project on the important topic and with interesting hypothesis. I hope we can read this paper in a good journal (soon!).

Many thanks for the inspiring presentation! I have three comments.
1. I wonder if government partisanship may also explain within-country variations. Although highly autocratic countries do not hold elections, moderately autocratic ones have some competitive elections. There should be ideological differences between those parties to some extent. The intuition behind this is left-leaning parties are more likely to accept a greater number of refugees, whereas right, populist parties are not (the logic is similar to that of immigration).
2. Does this help address the question “What happened to those with intermediate regime type?” This category refers to competitive authoritarianism in which partisan differences should matter to explain policy outcomes.
3. Related to this, there is also variation in ideological orientations among authoritarian governments. For example, Peru’s military regime was known to be leftist and promoted radical redistributive policies, but Chile was opposite. In short, further political variables may also matter, in addition to types of political regimes. It is worth considering this possibility to improve your theory.

Thanks Fukumoto-san for the additional comments.

(1) Yes, it may be useful to formalize our theory. In fact, Gaibulloev et al’s (2018) IO piece on regime types and terrorism introduces a simple formal model to derive a bell-curve hypothesis.

(2) At the current manuscript, we control for likely neighboring effects, including neighbors’ polity iv score, proportions of neighbors under civil war, and average numbers of refugees in neighboring countries.

(3) We compute country clustered standard errors to deal with heteroscedasticity at this point, but we’ll check about this.


Thanks for your comments and encouragements. Yes, I think your point is related to Fukumoto-san’s first point above on the possibility of formalizing the argument. Since the section of combined preferences is now very short in the paper, it would be better to combine two perspectives more tightly in that section.


Thanks for your very intriguing insights. I think ideological differences in authoritarian regimes sounds interesting and it’s possible to (roughly) measure by using ready-made datasets like DPI’s government partisanship variable. We’ll take a look at whether regime’s effects are magnified/mitigated through partisanship (or exploring the relationship between partisanship and refugee itself may produce a paper :-)).


Well, our Jackknife analyses are quite simple and we just drop each country or year one by one to see if results remain unchanged, which is often done in this sort of cross-national time-series analysis. Maybe there should be better ways..

Leave a Reply

Your email address will not be published. Required fields are marked *