Two‐Way Migration between Similar Countries

• Date: 01 January 2017
• Author: Jens Wrona,Udo Kreickemeier
• Published date: 01 January 2017
• DOI: http://doi.org/10.1111/twec.12377

Two-Way Migration between Similar

Countries

Udo Kreickemeier

1

and Jens Wrona

2

1

Faculty of Business and Economics, TU Dresden, Dresden, Germany and

2

D€

usseldorf Institute for

Competition Economics (DICE), University of D€

usseldorf, D€

usseldorf, Germany

1. INTRODUCTION

IN this paper, we develop a simple model to explain two-way migration of high-skilled

individuals between developed countries. While the phenomenon of two-way migration

has received little attention in the theoretical literature, it is quantitatively important, in partic-

ular for high-skilled individuals migrating between high-income countries. Table 1, which is

based on data from Docquier et al. (2008), shows for country pairs within the EU15 and the

OECD, respectively, the share of bilateral migration that can be characterised as two-way.

The share is measured by the index of bilateral balance in migration (Biswas and McHardy,

2005), which for each country pair (i,j) is given by Bij 2 minðEmij;Emji Þ=ðEmij þEmjiÞ,

with Emij as the stock of emigrants from country iresiding in country j.

1

The numbers in

Table 1 are the average values of the index for the respective country group, in a given year

and skill group. The data show that the share of two-way migration is highest for high-skill

individuals, that it has grown over time, and that it is higher within the homogeneous group

of EU15 countries than in the more heterogeneous group of OECD countries.

Focusing on high-skilled (tertiary educated) individuals, Figure 1 gives a more disaggre-

gated view at the level of country pairs for the EU15 (below the diagonal) and the OECD

(above the diagonal).

2

The figure confirms that a lot of high-skilled migration between EU15

countries is two-way in nature, while this is true to a lesser extent for the larger and more

heterogeneous group of OECD countries.

We are grateful to Florian Baumann, Herbert Br€

ucker, Ron Davies, Giovanni Facchini, Andreas Hoefele,

Wilhelm Kohler, Mario Larch, Oliver Lorz, Nina Neubecker, Hillel Rapoport and Daniel Sturm, to semi-

nar participants at Tinbergen Institute, WHU Koblenz, the University of Mainz, and Bond University,

and to participants of the Norface Conference ‘Migration: Global Development, New Frontiers’, the

ETSG Annual Conference, the G€

ottingen Workshop on International Economics, the T€

ubingen confer-

ence on ‘Worker-Specific Effects of Globalisation’, the TEMPO Conference on International Migration,

the Research Committee in International Economics of the German Economic Association, the RGS

Doctoral Conference in Economics, the RIEF Doctoral Meetings in International Trade and International

Finance and the Spring Meeting of Young Economists (SMYE) for helpful comments and discussions.

We thank Dominique Bruhn, Till Nikolka, Leopold Schiele, Simone Schotte and Sebastian Schwarz for

excellent research assistance.

1

The construction of the index is directly analogous to the well-known Grubel–Lloyd index measuring

intra-industry trade, that is two-way trade in goods within the same industry.

2

The figure plots (15 914)/2 =105 country pairs from the set of EU15 countries and (30 929)/

2=435 country pairs from the sample of OECD countries. Chile, Estonia, Israel and Slovenia are omit-

ted, as data on two-way migration are not available for these countries in Docquier et al. (2008). Note

that country pairs are ordered such that for the set of EU15 countries the net-emigration country appears

first, while for the set of OECD countries the net-immigration country is named first. Hence, the strict

separation in above and below the 45 degree line.

©2016 John Wiley & Sons Ltd

182

The World Economy (2017)

doi: 10.1111/twec.12377

The World Economy

Despite this regularity, there is, of course, an incidence of substantial two-way migration

for specific country pairs that are part of the OECD but not part of the EU15. Taking Canada

and the USA as another prominent example of rather similar countries, we observe substantial

high-skilled migration in both directions, with the share of two-way migration being 0.5 for

the year 2000.

3

The key challenge in explaining two-way migration of similar (highly skilled) individuals

within a group of similar (high-income) countries –rather than one-way migration from low-

income to high-income countries –lies in the fact that country differences cannot be expected

to play a central role. The model we develop in the main part of this paper therefore uses the

assumption that countries are identical in all respects (this assumption is relaxed later on). In

both countries, there is a continuum of workers with differing abilities, which are private

knowledge. The production technology, borrowed from Kremer (1993), exhibits complemen-

tarities between the skill levels of individual workers, and profit-maximising firms therefore

aim for hiring workers of identical skill. Migration is costly, and the cost is the same for all

individuals. High-skilled individuals from both countries self-select into emigration to separate

themselves from low-skilled co-workers at home. Firms can distinguish natives and immi-

grants, which allows them to form more efficient matches, leading to larger gross wage

premia for skilled workers.

The welfare effects of migration in our model are stark: in the laissez-faire equilibrium, all

individuals are worse off than in autarky. We show that this result is due to a negativ e migra-

tion externality which leads to too much migration in equilibrium. We also show that for suf-

ficiently low migration cost the level of migration chosen by an omniscient social planner is

strictly positive (but of course lower than in the laissez-faire equilibrium), since the existence

of migrants as a distinct group of individuals enables firms to match workers of more similar

expected skill. While aggregate gains from migration exist in the social planner equilibrium,

the distributional effects are strong: all migrants gain relative to autarky, while all natives are

worse off. These distributional effects are mitigated if the social optimum is implemented via

a migration tax, since in this case the possibility of redistributing part of the gains to

non-migrants exists.

The framework we develop in the main section of our paper is deliberately stylised to

bring out the basic mechanism driving two-way migration in our model and its welfare impli-

cations in the most transparent way possible. Due to its simplicity, the basic version of the

TABLE 1

Index of Bilateral Balance in Migration for EU15 and OECD Countries

High Skill Medium Skill Low Skill Total

EU15

1990 0.61 0.53 0.20 0.35

2000 0.64 0.51 0.28 0.48

OECD

1990 0.23 0.22 0.14 0.19

2000 0.28 0.26 0.13 0.22

3

See Schmitt and Soubeyran (2006) and the references cited therein for additional anecdotal evidence

on the balance in migration flows between Canada and the USA.

©2016 John Wiley & Sons Ltd

TWO-WAY MIGRATION BETWEEN SIMILAR COUNTRIES 183