The Refugee’s Dilemma: Evidence from Jewish Migration out of Nazi Germany

Abstract

We estimate the push and pull factors involved in the outmigration of Jews facing persecution in Nazi Germany from 1933 to 1941. Our empirical investigation makes use of a unique individual-level dataset that records the migration history of the Jewish community in Germany over the period. Our analysis highlights new channels, specific to violent contexts, through which social networks affect the decision to flee. We estimate a structural model of migration where individuals base their own migration decision on the observation of persecution and migration among their peers. Identification rests on exogenous variations in local push and pull factors across peers who live in different cities of residence. Then, we perform various experiments of counterfactual history to quantify how migration restrictions in destination countries affected the fate of Jews. For example, removing work restrictions for refugees in the recipient countries after the Nuremberg Laws (of 1935) would have led to an increase in Jewish migration out of Germany in the range of 12 to 20%, and a reduction in mortality due to prevented deportations in the range of 6 to 10%.

New York, Abu Dhabi, London or Stay at Home? Using a Cross-Nested Logit Model to Identify Complex Substitution Patterns in Migration

“…(2) provides the results of an NL model in which all foreign destinations are included in the same nest (Buggle et al, 2019;Monras, 2020). The model introduces a natural distinction between stayers and movers, but assumes that IIA holds across all foreign destinations.…”

“…First, intended stayers and intended movers are very different, and foreign destinations are therefore likely to be more correlated with each other than with the domestic destination. This has motivated the use of separate nests for the domestic location and the foreign potential locations in recent studies (Buggle et al, 2019;Monras, 2020). Second, some foreign destinations will be more correlated among themselves compared with others.…”

“…1 While deviations from IIAleading to what are known as multilateral resistances -have been extensively examined in the trade literature van Wincoop, 2003, 2004), very few contributions have dealt in depth with this issue in migration studies. Some recent studies account for unobserved differences between home and foreign destinations, assuming that the IIA assumption holds across foreign destinations (Ortega and Peri, 2013;Buggle et al, 2019;Monras, 2020). A few other innovative works account for a violation of IIA across foreign destinations, but use a somewhat arbitrary partitioning of countries (Bertoli and Fernández-Huertas Moraga, 2015;Bredtmann et al, 2017).…”

“…For the first time, we implement it in the context of international migration. In line with Ortega and Peri (2013), Buggle et al (2019) and Monras (2020), our proposed specification first distinguishes between the home and foreign destinations in order to capture unobserved differences between the two types of location -or between "stayers" and "movers." In addition, we also identify a set of overlapping subgroups of foreign destinations defined along several dimensions that are usually perceived as relevant in migration literature (e.g., level of economic development, quality of institutions, geographic location, language spoken, and contiguity).…”

“…Ortega and Peri (2013) developed a microfounded gravity approach in which the stochastic component of the movers shares a common component across foreign destinations. 3 More recently, Buggle et al (2019) used individual data to identify factors influencing the emigration decisions and destination choices of Jewish refugees during the 1930s in Nazi Germany. They developed an explicit nested logit approach along the lines of McFadden (1978), allowing them to identify the factors inducing Jewish individuals to leave Germany.…”