The determinants of international migration: Unbundling the role of economic, political and social institutions

• Published date: 01 June 2020
• Author: Imran Arif
• Date: 01 June 2020
• DOI: http://doi.org/10.1111/twec.12889

World Econ. 2020;43:1699–1729. wileyonlinelibrary.com/journal/twec

|

1699

© 2019 John Wiley & Sons Ltd

Received: 30 November 2018

|

Revised: 16 August 2019

|

Accepted: 1 October 2019

DOI: 10.1111/twec.12889

ORIGINAL ARTICLE

The determinants of international migration:

Unbundling the role of economic, political and social

institutions

ImranArif

Department of Economics, Appalachian State University, Boone, NC, USA

KEYWORDS

culture, democracy, freedom, gravity model, migration

1

|

INTRODUCTION

Potential migrants may decide to leave their origin country for a variety of reasons including eco-

nomic, political, demographic and social conditions in the origin and destination countries. They may

decide to migrate if these conditions improve in the destination country—pull factors—or if these

conditions deteriorate in the origin country—push factors. Potential migrants compare economic,

psychological and emotional benefits and costs of migration1; they decide to migrate whenever their

private economic and non-monetary benefits exceed costs. If policymakers understand what motivates

migration, they can make suitable policies to encourage brain gain and discourage brain drain.

Economic researchers have been identifying economic and non-economic factors motivating inter-

national migration; helping policymakers to understand the implications of migration; and helping

them to make domestic policies that would have a direct or indirect effect on migration. Some re-

searchers have found the role of socioeconomic and political factors of migration.2 For example,

Uebelmesser, Geis, and Werding (2013), Nejad and Young (2016), and Migali (2018) find that both

economic and political institutions are significant pull factors for potential migrants. Our paper further

expands upon the insights of these studies by investigating whether emigrants self-select based on

economic, political and social institutions.

Our paper adds to the existing literature in at least three ways. First, we add social dimension as a

determinant of international migration and compare the size of its significance with that of economic

and political institutions. Earlier research includes either social dimension (White & Buehler, 2018)

or economic and political dimensions (Nejad & Young, 2016) but did not include all three variables

together. Second, we separate the pull and push factors of migration. Nejad and Young (2016) con-

centrate on pull factors of economic and political institutions, and White and Buehler (2018) focus

1 Please see Falck et al. (2018).

2 Please see Baudassé et al. (2018) for a survey of this literature.

1700

|

ARIF

on the bilateral social factors. Our paper improves upon both these studies by separating the pull and

push factors of migration, as both pull and push factors are crucial in migration decision, and it will be

interesting to evaluate whether pull factors are more important than push factors and vice versa. Third,

we explore the pull and push factors of migration based on the human capital of migrants. We expect

that the education level influences the sensitivity of potential migrants to these pull and push factors.

We use the most widely used indicators of economic, political and social institutions in our analy-

sis. For economic institutions, we use the Freedom of the World index from the Fraser Institute. The

Freedom of the World index indicates the quality of economic institutions in a country (Gwartney,

Lawson, Hall, & Murphy, 2018). For political institutions, we use the well-known Polity IV index. In

addition, we use checks and balances from Keefer and Stasavage (2003) as an indicator of political

constraint. For social institutions, we follow White and Buehler (2018) and use survival versus self-ex-

pression values and traditional versus secular-rational authority from Inglehart and Baker (2000).

Our empirical strategy includes two types of modified gravity models. Our first empirical strategy

includes converting the above three variables into ratios. We consider this approach because of the

bilateral nature of migration data and unilateral nature of the above three variables. Specifically, we

convert the economic, political and social variables into ratios between the destination and origin's

corresponding indicators. These ratios allow us to interpret coefficient values as elasticities, enabling

us to quantify and compare the effect of economic, political and social institutions on migration.3

These elasticities show the change in migration flows due to a 1% change in destination indicator,

helping us understand the effect of economic, political and social institutions; all else equal, we expect

migrants move towards destinations with better economic, political and social institutions.

Although we can observe the effects of economic, political and social institutions on migration

using the above strategy, we cannot isolate the pull factors of migration from the push factors using

the above approach. Both pull and push factors are essential components of migration. Migrants may

be moving towards destinations with improved economic, political and social conditions or leaving

origin countries with deteriorated economic, political and social conditions. The ratio-scale model

does not separate these pull and push effects. To isolate pull and push effects, we employ our second

empirical strategy. This strategy includes using the dyadic migration flow as a dependent variable and

monadic economic, political and social institutions as main explanatory variables. In addition, we test

the pull and push factors of migration using the Alesina, Harnoss, and Rapoport (2016) two-period

panel data model. Furthermore, we test the robustness of our cross-sectional results by using Hofstede's

six dimensions of culture and by splitting our sample based on the income levels of the origin and

destination countries.4

We test the effect of economic, political and social institutions on migration using bilateral mi-

gration data of 103 origin and destination countries. Our migration data come from Artuc, Docquier,

Özden, and Parsons (2015), who present global bilateral migration flows by gender and education

levels during the period 1990–2000, allowing us to test the effect of economic, political and social

institutions on three types of migration flows: total, college and non-college migration. We expect that

potential educated migrants may be more sensitive to the origin and destination countries' economic,

political and social institutions.

We use the Poisson pseudo-maximum likelihood (PPML) estimator as our preferred estimation

method. There are several advantages to using PPML in this context. First, PPML accounts for zero

observations, which increases the number of observations by about 100%. Zero migration flows may

3 See Ashby (2010).

4 These robustness tests are presented in the Appendix S1.