Remittances and Working Poverty

Abstract

This article shows that the level and the predictability of remittances reduce working poverty in receiving economies through their effects on labour market dynamics. It takes advantage of the new cross-country dataset (ILO, KILM 7th edition) containing information on the share of individuals working for less than US$2 per day. To identify the main impacts, the article proposes a novel approach to deal with the endogeneity of remittances and migration. In addition, the results are robust to the possible error in measuring working poverty, to the potential attrition bias, and to the presence of various control variables.

Acknowledgements

We would like to thank the editor and two anonymous referees for the useful comments and suggestions on the article. This work was supported by the Agence Nationale de la Recherche of the French government through the programme ‘Investissements d’avenir’ ANR-10-LABX-14-01. The authors acknowledge the FERDI (Fondation pour les Etudes et Recherches sur le Developpement International) for financial and publication support. The views expressed in this article are those of the authors and should not be attributed to the institutions to which they are affiliated. Official data used in this article are available from the authors upon request.

Notes

1. This article exploits the exchange rate shocks before and after the 1997 Asian financial crisis to see how changes in remittances affect poverty. Being assumed to be exogeneous to the decision of remitting, these shocks provide a natural experiment.

2. This may arise from the fact that the impact is ambiguous if the revenue from the migrant replaces the revenue he would have procured without migrating. It may be due also to the difficult task of comparing strictly identical individuals and households. Alternative econometric strategies have been proposed to address this identification problem.

3. In contrast, remittances could enhance competitiveness and growth in the receiving economies when they are associated with an increase in capital investment in exporting industries, or through the additional trade induced by more networking activities (Fayad, 2011).

4. For more details regarding the techniques used by ILO, see ‘Working Poverty Around the World: Introducing new estimates using household survey data’, retrieved from http://kilm.ilo.org/manuscript/sectionA.asp.

5. Estimates available from the authors upon request.

6. A careful description of the control variables and the discussion on expected signs are presented in the Online Appendix.

7. We thank an anonymous referee who has suggested controlling for income inequality.

8. The bilateral migration matrix comes from the World Bank web site (http://data.worldbank.org/data-catalog/global-bilateral-migration-database). Migration data are available at a 10-year window. We therefore imputed the same migration stock value to the years composing each decade. However, it is certainly likely that with its limited within-country variation, controlling for country-fixed effects in the econometric models will reduce the potential significance of this variable.

9. Since the econometric models always control for both FDI and exports, along with country-fixed effects, we are controlling for the extent to which the remittance receiving country is also natural resource rich country or not.

10. See Rajan and Subramanian (2008) for a similar empirical strategy to deal with the endogeneity of remaining explanatory variables.

11. The results are available upon request.

12. This is a concern when the process governing the missing data on the working poor rate is related to some economic factors that simultaneously affect remittances and the dependent variable.

13. Amuedo-Dorantes and Pozo (2012) and Cox-Edwards and Rodriguez-Oreggia (2009) investigated how different remittance patterns (level and unpredictability) influence the labour supply of recipient households using Mexican household survey data with contrasting results. The former found limited evidence of the labour force participation effects of persistent remittances. According to the latter, the labour supply increases when remittances inflows are less predictable.

14. Influential observations within the set of predicted residuals have been removed to reduce the outlier bias. We calculate the studentised residuals from the remittance equation and we exclude all observations whose absolute values of residuals are greater than the conventional threshold of 2.

15. Estimates are reported with the corresponding first-stage statistics. The unpredictability of remittances, as it is computed as the residual, is therefore considered as strictly exogenous.