SEMINAR: Methodological issues in empirical remittance modelling with panel data: evidence from immigrants to Germany

11 febbraio 2016
11 February 2016
Contatti: 
Doctoral School of Social Sciences
via Verdi 26, 38122 - Trento
Tel. 
+39 0461 283756 - 2290
Fax 
+39 0461 282335

Skype: school.socialsciences

2 PM - room 3G, Department of Economics and Management

Speaker: Giulia Bettin, Università Politecnica delle Marche

Abstract

The empirical modelling of remitting behaviour by migrants has been the object of a considerable amount of micro-level literature. Although there is a good deal of variety among the data employed for this purpose, some features are pervasive and general, such as the censored nature of international transfers. Moreover, the increasing availability of panel dataset makes it possible to explore the time dimension of remittance behaviour, but this poses the additional need to address permanent unobserved heterogeneity and state dependence.
However, these issues have not been jointly accounted for in the empirical migration literature so far. In this paper, we propose two dynamic, random-effects sample selection models: we combine the Maximum Likelihood estimators of the sample selection and double hurdle models (Heckman, 1974; Cragg, 1971) with the approach proposed by Heckman (1981) for dealing with state dependence and unobserved heterogeneity in nonlinear models. Our objective is therefore two-fold: (i) we introduce feasible estimators for a potentially large class of empirical problems; (ii) we offer novel empirical evidence in the field of remittance modelling by applying our proposed models to the German SOEP dataset.

Download