A General Model of Bilateral Migration Agreements
ABSTRACT Freeman (2006) suggested that auctioning immigration visas and redistributing the revenue to native residents in the host country would increase migration from low-income to high-income countries. The effect of the auctioning of immigration visas, in the Ricardian model from Findlay (1982), on the optimal level of immigration for the host country is considered. It is shown that auctioning immigration visas will lead to a positive level of immigration only if the initial wage difference between the host country and the source country is substantial. The cost of the immigration visa is more than half the earnings of the immigrant worker. Copyright © 2009 Blackwell Publishing Ltd.
- SourceAvailable from: Silke Übelmesser[Show abstract] [Hide abstract]
ABSTRACT: This paper identifies the migration policies that emerge when both the sending country and the receiving country wield power to set migration quotas, when controlling migration is costly, and when the decision how much human capital to acquire depends, among other things, on the migration policies. The paper analyzes the endogenous formation of bilateral agreements in the shape of transfers to support migration controls, and in the shape of joint arrangements regarding the migration policy and the cost-sharing of its implementation. The paper shows that in equilibrium both the sending country and the receiving country can participate in setting the migration policy, that bilateral agreements can arise as a welfare-improving mechanism, and that the sending country can gain from migration even when it does not set its preferred policy.10/2011;
A General Model of Bilateral Migration Agreements∗
Jes´ us Fern´ andez-Huertas Moraga†
IAE-CSIC, Barcelona GSE and IZA
September 29, 2008
Unilateral migration policies impose externalities on other countries. In order to try
to internalize these externalities, countries sign bilateral migration agreements. One
element of these agreements is the emphasis on enforcing migration policies: immigrant-
receiving countries agree to allow more immigrants from their emigrant-sending partner
if they cooperate in enforcing their migration policy at the border. I present a simple
theoretical model that justifies this behavior in a two-country setting with welfare
maximizing governments. These governments establish migration quotas that need to
be enforced at a cost. I prove that uncoordinated migration policies are inefficient.
Both countries can improve welfare by exchanging a more “generous” migration quota
for expenditure on enforcement policy. Contrary to what could be expected, this result
does not depend on the enforcement technology that both countries employ.
Keywords: international migration, cooperation, migration policy
JEL Classification Numbers: F22
∗I received valuable comments and suggestions from Alessandra Casella, Esther Hauk and seminar par-
ticipants at Columbia University. Of course, remaining errors are only mine.
†Instituto de An´ alisis Econ´ omico (CSIC), Campus UAB, 08193 Bellaterra (Barcelona), Spain. E-mail:
Why do countries cooperate in establishing migration policies?
immigrant-receiving and emigrant-sending countries sign migration agreements?1Clearly,
In particular, why do
these countries could have opposing interests: immigrant-receiving countries could want to
restrict immigration, while emigrant-sending countries could want to relieve their excess la-
bor supply as much as possible. In this paper, we show that cooperation among countries
with conflicting interests can be Pareto-improving since unilateral migration policies impose
externalities on other countries which can be partly internalized by migration agreements
despite a priori conflicting interests. This migration externality takes place because the
immigrant-receiving country does not take into account the welfare of its emigrant-sending
counterpart when deciding about its optimal migration policy. The result is that this optimal
migration policy tends to be inefficiently over-restrictive, thus harming the emigrant-sending
country’s welfare. There is also an externality on the other side: since restricting migration
is costly, the emigrant-sending country has no incentives to do so and therefore imposes
inefficiently high enforcement costs on the immigrant-receiving country.
Bilateral migration agreements allow to internalize this externality. One element of these
agreements is the emphasis on enforcing migration policies by which immigrant-receiving
countries agree to allow more immigrants from their emigrant-sending partner if they coop-
erate in enforcing their migration policy at the border and thereby share the costs. I present
a simple theoretical model that justifies this behavior in a two-country setting with welfare
maximizing governments. These governments establish migration quotas that need to be
enforced at a cost. The costly enforcement technology is modeled following Ethier (1986b)
original paper. I prove that unilateral migration policies are inefficient whereas both coun-
tries can improve welfare by exchanging a more “generous” migration quota for expenditure
on enforcement policy. Contrary to what could be expected, this result does not depend on
the enforcement technology that both countries employ.
The World Trade Organization (WTO) is an institution where countries can get together
and negotiate mutually beneficial trade agreements. When countries set their tariffs uni-
1The case for cooperation between immigrant-receiving countries who would unilaterally like to divert
undesired immigrant inflows to their neighbors has been studied elsewhere. For example, see Barbou des
Places and Deffains (2004) for an application of the Sandler and Hartley (2001) joint product theory to the
case of refugee distribution.
laterally, they hurt other countries because they improve their own terms of trade at the
expense of others’ terms of trade. This creates a Prisoner’s Dilemma where countries would
be better off if they all lowered their tariffs but in fact they do not have the incentive to do
so unilaterally. In order to remove this inefficiency, international cooperation is required and
this is obtained through the WTO.2A key element that explains why international coop-
eration enhances efficiency is the assumption that freer trade increases world output. This
paper shows that a similar reasoning can be applied to migration policy. However, an im-
portant difference must also be highlighted: whereas the theory of trade agreements is based
on the assumption that all participating countries benefit from higher volumes of trade, this
paper shows that migration agreements can be signed even when the immigrant-receiving
country welfare is decreasing in the magnitude of the migration flow at the same time that
the emigrant-sending country welfare is increasing in the magnitude of the migration flow.
Most theoretical models of migration coincide in concluding that the free movement of
factors contributes to a better allocation of resources at the world level, even when one ab-
stracts from fairness considerations (Findlay (1982)). In most cases, the upper estimate of
these efficiency gains is notably superior to the efficiency gains that can be expected from,
for example, free trade. For example, Hamilton and Whalley (1984) crudely estimated (us-
ing data from 1977) that the efficiency gains from totally removing immigration controls
could get to double world GNP. In a more recent paper, Rodrik (2002) argues that “...lib-
eralizing cross-border labor movements can be expected to yield benefits that are roughly
25 times larger than those that would accrue from the traditional agenda focusing on goods
and capital flows3!” Why are these immense efficiency gains not obtained through interna-
tional cooperation? The typical explanation (Hatton (2007)) is that the movement of people
has opposing effects on immigrant-receiving and emigrant-sending countries. Immigrant-
receiving countries tend to ask for lower migration whereas emigrant-sending countries tend
to ask for freer migration, at least in terms of low-skill migrants.
As a result, the economics literature has typically studied migration policies as a unilat-
eral phenomenon. For example, Ethier (1986b) and Ethier (1986a) use the crime-theoretic
analysis of Becker (1968) to analyze the effects of different policies aimed at reducing illegal
2Bagwell and Staiger (2003) provide a detailed discussion.
3A more modest estimate by the World Bank (2006) finds that “a rise in migration from developing
countries sufficient to raise the labor force of high-income countries by 3 percent” would yield gains 13 per
cent higher than the gains to be obtained from global trade reforms as proposed in the Doha round.
immigration. Bond and Chen (1987) extend Ethier’s analysis to a two-country model with
capital mobility but they do not allow a policy response of the emigrant-sending country
to the migration policy of the immigrant-receiving country4. The same can be said about
Woodland and Yoshida (2006) contribution, who add emigrants risk preferences to the model.
Guzman, Haslag, and Orrenius (2008) allow a response to the migration policy (on border
enforcement) of the immigrant-receiving country. However, the response does not come from
the emigrant-sending country but from the smugglers, who can partly nullify the intended
effects of border enforcement. Finally, Schiff (2007) analyzes the relative merits of common
migration policy options and proposals, such as permanent migration programs, guest-worker
programs and Mode IV in the GATS (General Agreement on Tariffs and Services).
On the contrary, Bandyopadhyay and Bandyopadhyay (1998) extension of Bond and Chen
(1987) model is more similar to the one in this paper since they consider a policy response of
the emigrant sending country. In their case, this policy consists of imposing restrictions on
capital inflows and it can render the border enforcement policy of the immigrant-receiving
country partly ineffective. Dula, Kahana, and Lecker (2006) also take the policy options of
emigrant-sending countries into account. They advance an original proposal to address the
migration externality. They claim that immigrant-receiving countries could save in border
enforcement by financing relatively more those emigrant-sending governments who would
make a bigger effort in avoiding the exit of illegal emigrants from their country, thereby
creating competition among emigrant-sending countries for the funds of the immigrant-
receiving country. This kind of auction for development aid has not been formally established
yet. A third framework that also considers both the emigrant-sending and the immigrant-
receiving country policies is proposed by Stark, Casarico, Devillanova, and Uebelmesser
(2007). In the presence of a human capital externality in the emigrant-sending country
that makes a certain level of emigration welfare improving by generating a brain gain but
additional levels welfare inferior by creating a brain drain, they show that there is scope for
migration agreements. There is an important difference with my paper since these migration
4The issue of the relationship between labor and capital mobility and optimal policies to maximize
welfare under different scenarios has a longer tradition in the literature. The classical reference in this area
is Ramaswami (1968). He used MacDougall (1960) framework to show how allowing for migration and taxing
migrants is preferred to exporting and taxing capital in a neoclassical model with two factors of production.
Calvo and Wellisz (1983) showed how the institutional restrictions (inability to discriminate labor) were key
in Ramaswami (1968) result so that there was no need to import labor in order to obtain the same outcome.
agreements are only beneficial when both countries’ welfare levels are decreasing in the
magnitude of the migration flow, that is, when the preferences of both countries are aligned,
as Hatton (2007) suggests. In this paper, it will be shown that the scope for migration
agreements remains in the absence of human capital externalities and even when one of
the countries would favor larger migration flows whereas the other benefits from smaller
The public finance literature has also addressed the issue of cooperation in migration
policies in the context of regional migration. From this literature, the most relevant result
for the purposes of this paper is that of Myers (1990), who shows that, under free migra-
tion, decentralized policies are enough to achieve efficiency because countries (regions in his
case) internalize the consequences of their policies on their neighbors through their effect on
migration flows. The immediate consequence is that the establishment of migration controls
may preclude an efficient solution since decentralized policies will then create externalities
on other countries (regions). For example, Casella (2005) shows how there are situations in
which countries (regions) can individually choose to set migration barriers optimally, thus
preventing externalities from being internalized through the effect of other policies (redistri-
bution policies in her model) on migration flows. In those situations, both migratory policies
and internal policies must be coordinated in order to achieve efficiency. The difference with
my approach is that the source of the externality in Casella (2005) is not the migratory
policy itself but the existence of technological spillovers.
As of 2004, there were at least 176 bilateral agreements on migration issues.5What is the
economic justification behind all of these? One useful starting point to address this question
is to incorporate the arguments that are actually given for signing bilateral migration agree-
ments. According to the background paper for the joint IOM/World Bank/WTO Trade
and Migration Seminar, IOM/World Bank/WTO (2004), the reasons why migrant-receiving
countries sign these agreements are:
• Combatting irregular migration.
• Responding to labor market needs of temporary or permanent nature.
• Promoting economic links with sending countries.
5The number refers only to agreements in which at least one OECD member is involved (OECD (2004)).