World Politics
The Limits of International Cooperation on Migration:
Bilateral Readmission Agreements in the European Context
Journal: World Politics Manuscript ID: Draft Manuscript Type: Research Article Subject Category ­ Select one main category, Comparative Politics or International International Relations Relations, for the “required” field, and any applicable subcategories: Keyword: cooperation, migration, Europe, European Union Despite growing globalization in the last half century, accompanied by the creation of multilateral international economic institutions, international migration flows have been governed predominantly by unilateral state action. It is surprising, then, to observe the rise of bilateral readmission agreements, beginning in the late 1950s, and their rapid proliferation since the 1990s. These intergovernmental treaties typically require migrant sending states to accept both their own citizens, and citizens of other states who have transited their state, who find themselves in violation of the laws of the host state. We argue that the spread of bilateral Abstract: readmission agreements is directly related to state costs of managing migration unilaterally. As these costs grow, states seek to reduce their domestic costs by enlisting the efforts of sending states in migration management. We model the costs in terms of migrant flows, welfare payments, domestic political backlash to migrants, and the relative costs of managing migration flows. Evidence from the European bilateral readmission agreements is consistent with our hypotheses. The distinctive nature of migration flows, illuminated by bilateral readmission agreements, suggest that international cooperation on migration will be severely limited, despite many calls to collaborate. [email protected]
http://mc.manuscriptcentral.com/wp
Page 1 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
THE LIMITS OF COOPERATION ON INTERNATIONAL MIGRATION:
BILATERAL READMISSION AGREEMENTS IN THE EUROPEAN CONTEXT
ABSTRACT: Despite growing globalization in the last half century, accompanied by the
creation of multilateral international economic institutions, international migration flows have
been governed predominantly by unilateral state action. It is surprising, then, to observe the rise
of bilateral readmission agreements, beginning in the late 1950s, and their rapid proliferation
since the 1990s. These intergovernmental treaties typically require migrant sending states to
accept both their own citizens, and citizens of other states who have transited their state, who
find themselves in violation of the laws of the host state. We argue that the spread of bilateral
readmission agreements is directly related to state costs of managing migration unilaterally. As
these costs grow, states seek to reduce their domestic costs by enlisting the efforts of sending
states in migration management. We model the costs in terms of migrant flows, welfare
payments, domestic political backlash to migrants, and the relative costs of managing migration
flows. Evidence from the European bilateral readmission agreements is consistent with our
hypotheses. The distinctive nature of migration flows, illuminated by bilateral readmission
agreements, suggest that international cooperation on migration will be severely limited, despite
many calls to collaborate.
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 2 of 33
INTRODUCTION
Despite growing interdependence in the last half century, international migration flows have
been governed predominantly by unilateral state action. It is surprising, then, to observe the rise
of bilateral readmission agreements, beginning in the late 1950s, and their rapid proliferation
since the 1990s. These intergovernmental treaties typically require migrant sending states to
accept both their own citizens, and citizens of other states who have transited their state, who
find themselves in violation of the laws of the host state. We argue that the spread of bilateral
readmission agreements is directly related to state costs of managing migration unilaterally. As
these costs grow, states seek to reduce their domestic costs by enlisting the efforts of sending
states in migration management.
We first argue that, because migrants receive a “bundle of rights” when entering the host state,
the larger the bundle of rights, the larger the state’s economic and political costs of migration.1
Next, we observe decreasing national control over migration generated as a byproduct of treaties
that relinquish sovereignty on human rights, labor mobility, and border controls. Where states
have constrained their ability to control migration flows, they experience higher flows and higher
costs; these states are therefore more likely to sign readmission agreements. Although most
states have signed international agreements that limit their ability to control the entry and
residence of foreigners,2 European states are the most constrained by treaties involving European
integration and, hence, are the locus of most bilateral readmission agreements. Finally, because
migration flows are unique, bilateral, and non-reciprocal, agreements governing those flows will
be bilateral bargained agreements. Host states minimize their side payments to receiving states
by signing agreements only with those states that actually send migrants. The quantitative
evidence from European bilateral readmission agreements is consistent with the cost argument:
the number of migrants, the domestic political backlash to migrants, and the cost differential in
migration management explain, in part, the pattern of bilateral readmission agreements.
The analysis of readmission agreements contributes to our broader understanding of the degree
and shape of cooperative activity in the international system, as well as the impact of
international cooperation on state sovereignty. Migration flows differ from other types of
international economic flows because they are unique to specific countries, bilateral, rather than
multilateral, and non-reciprocal. These patterns favor bilateral agreements rather than
multilateral agreements when cooperation promises benefits for both parties. However, the fact
that these bilateral agreements are concentrated in Europe suggests that the benefits are nominal
in most instances, which does not augur well for the diffusion of even bilateral cooperation on
migration issues in the international system.
Our research also suggests that the sources of international cooperation in other international
economic activities – problems of collective action and/or the provision of public goods – are
absent from the voluntary migration issue area. Moreover, bilateral readmission agreements are
a form of international cooperation that reduces migration and reclaims sovereignty for states,
rather than relinquishes sovereignty. This requires us to reconsider the implications of
“international cooperation” or “global governance” in an era of globalization.
2
[email protected]
http://mc.manuscriptcentral.com/wp
Page 3 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
We begin by defining readmission agreements and the purposes they serve in the international
system. We proceed by reviewing the literature on bilateral readmission agreements. Building
on the characteristics of international migration, we then propose a model to explain the shape,
number, and third country signatories to the agreements in Europe. In the following sections, we
present our data, methodology and results of our empirical tests. In the conclusions, we place
these findings within the context of international cooperation more generally.
READMISSION AGREEMENTS DEFINED
Under customary international law, sending countries have long been expected to accept the
return of their own citizens who are in another country illegally.3 This practice is a fundamental
component of the basic concept of nationality, grounded in the individual’s right of return to his
or her own country.4 States have been reluctant to codify this customary international law
because doing so might imply that, absent a written treaty, states might not be obligated to
respect the right of return.5 Thus, readmission agreements have been the exception, rather than
the rule. Moreover, formal readmission agreements tend not only to codify the expectation that
countries will readmit their own citizens but also expand it, requiring that signatory countries
also admit stateless persons and citizens of third countries that entered the receiving country
through the sending country. Readmission agreements now generally apply to all unauthorized
migrants, or “all persons who do not, or who no longer, fulfill the conditions in force for entry to,
presence in, or residence on, the territory of the requesting state.”6 The agreements frequently
include technical provisions on procedures, data collection, and human rights protection.7
The first modern bilateral readmission agreements were signed between the late 1950s to the
mid-1960s, mostly between Western European states. The earliest agreements followed the
development of the European Steel and Coal Community (1952), the Nordic Passport Union
(1954), the European Economic Community (1957), and the Benelux Economic Union (1958).
The second wave of agreements began in the 1990s, and this wave was far larger. Between 1950
and 1990, only 14 readmission agreements were concluded; between 1990 and 2000, 124
readmission agreements were concluded. An additional 79 agreements were negotiated between
2000 and 2006.8 Of these readmission agreements, the vast majority are bilateral treaties.9 This
wave of agreements was instigated by economically developed, net-recipient states in Western
Europe and targeted less-developed source countries in Eastern Europe, Africa, and beyond.
At first, these readmission agreements were concluded between European states and third
countries and were completely outside the purview of the European Union (EU). In 1994, the
European Union adopted a model readmission agreement that would serve as guide for
individual member states as they pursued agreements with states outside the EU.10 It wasn’t until
1999, after the Tampere European Council, that the EU gained the competency to conclude
readmission agreements itself. Since then, the European Union has received mandates to
negotiate agreements with 18 countries; agreements have been concluded with 13 of those
countries while negotiations have stalled with the remaining five countries.11 So far, individual
states have continued to pursue their own agreements even while the EU has pursued Unionwide agreements.
3
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 4 of 33
Table 1 presents a summary of readmission agreements signed by European states and other
wealthy democracies. Although readmission agreements are not unique to Europe, according to
the Intergovernmental Consultations on Asylum, Refugee and Migration Policies, these treaties
represent 89 percent of all readmission agreements, the vast majority of the agreements to date.12
However, the variation among countries in Europe is significant: Ireland has signed only four
and the UK 16, whereas France has signed 68 and Italy 51. Out of all the European agreements,
multilateral agreements negotiated by the EU represent only three percent of the total. A more
detailed perusal of the agreements reveals that the signatories of these agreements vary widely.
This empirical overview illustrates the puzzles posed by the advent of bilateral readmission
agreements, concentrated in Europe: the bilateral form of the agreements, the variation among
states in number and signatories, and the timing of the agreements.
INTERNATIONAL COOPERATION AND READMISSION AGREEMENTS
What motivates states to sign these agreements? Much of the literature on readmission
agreements has focused on the normative and policy dimensions of the agreements, including the
implications for human rights.13 In research that directly addresses the determinants of
readmission agreements, Jean-Pierre Cassarino argues that geographic proximity, migration
salience and incentives are the most important factors propelling formal readmission
agreements.14 According to his analysis, geographic proximity provides the source of the
immigrants. Migration salience in the source country leads to a decline in readmission
agreements, as the country of origin is less willing to sign a readmission agreement.15 Finally,
countries of destination provide incentives to transit and/or origin countries to persuade them to
sign the agreements. The higher the level of incentives, the more likely the countries of transit
and/or origin are to sign the agreements. His argument is conditional; given geographic
proximity, readmission agreements will only occur if migration salience is low and incentives are
high.
Cassarino, however, does not systematically evaluate theses hypotheses. Moreover, the focus on
incentives begs the question of the circumstances that would generate an offer of incentives.
Thus we build on Cassarino’s initial insights by focusing, in part, on the factors that would cause
receiving states to offer incentives to sending states to sign readmission agreements. Implicit in
Cassarino’s argument is a bargaining framework that relies on the concept of best alternative to
negotiated agreements (BATNA).16 As states’ best alternative changes, states consider the
benefits of international agreements. However, we provide the empirical referents to understand
the state’s calculus of the changing environment. We also draw on the migration literature to
understand the pattern of flows as well as migration policy literature to elaborate why states are
constrained in handling these flows. Finally, we build on the literature on international
cooperation to explain why receiving states dominate the negotiation process.
Our argument has three parts. We first argue that, because migrants receive a “bundle of rights”
upon entry, receiving states that have finely woven social nets and democratic institutions
experience costs of migration that impel them to bargain with sending states to reduce those
costs. Second, states that have constrained their ability to control migration flows experience
4
[email protected]
http://mc.manuscriptcentral.com/wp
Page 5 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
higher flows and hence higher costs and are more likely to sign readmission agreements.
Although most states have signed international agreements that limit their ability to control the
entry and residence of foreigners,17 European states are the most constrained by treaties
involving European integration and, hence, are the locus of most bilateral readmission
agreements. Finally, we argue that, because migration flows are unique, bilateral, and nonreciprocal, agreements governing those flows will be bilateral bargained agreements. Each of
these component parts is elaborated below.
MIGRANTS’ “BUNDLE OF RIGHTS” AND ASSOCIATED COSTS
Migration differs from other international economic flows in the costs that it imposes on certain
states in the international system. It is well known that trade flows, for example, often affect the
distribution of wealth within a country, with some groups gaining from openness to the
international system and others losing. These international transactions, however, do not usually
impose costs on state governments and may, in fact, provide benefits via tariff revenues or
enhanced economic growth. Ruhs and Chang have documented the distinctive nature of
international migration, as opposed to the flow of goods, services, and money.18 Migrants enter
states with a “bundle of rights,” although the size of the bundle varies with the nature of the state.
Wealthy democracies provide migrants with the largest bundle of rights based on prior political
processes that grant civic, political, and social rights to citizens, which are applicable in many
cases to migrants. Migration thus generates costs borne by states that motivate them to remove
those that do not meet the requirements of the state (undocumented migrants) through bilateral
readmission agreements.
We argue that there are at least three types of costs borne by the host states: welfare costs,
political costs, and border control costs. These costs grow as the migration streams become
unbalanced, with net immigration from poorer states.
Welfare costs. The cost of providing social welfare benefits to immigrants may affect the
incentives for receiving states to pursue readmission agreements. There is some evidence that
generous benefits may actually attract immigrants to particular countries.19 Immigrants generally
do not utilize social welfare benefits more than citizens, controlling for other socio-economic
factors, but immigrants tend to be systematically younger, poorer, and less educated than host
state citizens, which means that they are more likely to be eligible for welfare benefits.20 Social
welfare costs thus represent a significant cost generating pressure for states to pursue
readmission agreements.
Two elements are important when calculating the welfare benefits costs of immigration. First,
how great is the overall cost of the state’s welfare expenditures? States that devote more
resources to social welfare benefits will face an even greater burden from increased immigration.
Second, how effectively can the state exclude undocumented immigrants from social welfare
benefits? States that can avoid paying for the social welfare of undocumented immigrants will
be much less concerned about restricting the flow of these immigrants.21 Excluding
undocumented immigrants from government benefits depends on both a legal ability to exclude
migrants and the practical ability to identify undocumented migrants. In Germany, for example,
5
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 6 of 33
there are strict laws that prevent undocumented migrants from accessing healthcare and
education. There are also sophisticated, linked computer databases that streamline enforcement,
and regular workplace checks. In France, on the other hand, undocumented immigrants have a
legal right to healthcare, and the courts have been reluctant to enforce laws that make being
caught without proper identification and documentation a deportation-worthy crime. Multiple
agencies are responsible for investigating employers for illegally hiring undocumented workers,
which makes enforcement cumbersome, and prosecution is rare.22 Thus, the social welfare costs
due to undocumented immigrants incurred by France would likely be greater than by Germany.
We argue that states with greater costs are more likely to pursue readmission agreements to
reduce welfare expenditures.
H1. The higher the state expenditures on welfare to immigrants, the more likely host states will
be to negotiate readmission agreements with sending states.
Political costs. There are political costs associated with an unwanted migrant population. To
citizens, states appear to be ineffective in protecting their borders. And public opinion polls
suggest that citizens prefer fewer migrants.23 Although the salience of the issue varies among
states, a certain portion of the population may organize to protest the presence of unwanted
migrants, to the detriment of the government in power. It is the organized opposition to
immigration that threatens the parties in power politically. Thus we argue,
H2. As anti-immigrant attitudes are organized into competitive political parties, host states will
be more likely to negotiate readmission agreements with sending states.
Costs of border controls. Unilateralism is the fallback of states. To offset the decline in control
over borders, states have begun to spend ever increasing amounts on border control. Offshoring
border control to sending and transit countries represents a cheaper mechanism for border
control, given the high cost of labor and the labor intensive nature of border controls. So states
will be more likely to sign readmission agreements with poorer states when border control costs
increase, and transit and sending states promise a mechanism for reducing border control costs.
H3. As the wage differential between sending and receiving states increase, host states will be
more likely to negotiate readmission agreements with sending states, to offshore the costs of
border controls.
Mitigating factors. Finally, there may be mitigating factors that reduce the costs of unwanted
migration. The state of the economy is one indicator of the ability of unwanted migrants to find
employment and of the demand by employers for more labor. If the economy is growing rapidly
and generating new jobs, there will be a larger demand for migrants than when the economy is
growing slowly and unemployment is high and/or rising. Previous research suggests that this
relationship between the supply of immigrants, and the demand for them, holds for
undocumented migrants as well. Hanson and Spilimbergo find that states decrease border
enforcement during positive shocks to sectors that employ a high number of undocumented
workers.24 They argue that states respond to private lobbying by business interests to allow the
inflow of at least some undocumented migrants. At the same time, states must balance this
against the demand by labor unions and anti-immigrant interest groups to crack down on
6
[email protected]
http://mc.manuscriptcentral.com/wp
Page 7 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
undocumented immigration. These competing forces lead to an equilibrium level of
undocumented immigration that changes in response to economic conditions within the receiving
country.25 We suggest that states will thus be less likely to pursue new enforcement and control
mechanisms, such as readmission agreements, when the state is experiencing a period of
economic expansion and full employment.
H4. As the economy expands, host states are less likely to negotiate readmission agreements.
H5. As unemployment decreases, host states are less likely to negotiate readmission agreements.
THE REDUCTION OF MIGRATION CONTROL AND THE LOCUS OF
AGREEMENTS IN EUROPE
States now confront a reduced ability to control their borders and their resident foreign
population. Before World War II, states dealt with unwanted immigration primarily through
unilateral expulsion. The post-World War II period brought with it a number of changes that
affected the ability of states to unilaterally control migration; in particular, certain states adopted
various international agreements that reduced their ability to control entry and residence of
migrants.
Global mobility alone may not represent a problem for states. It is only when states are unable to
sort and control the entry and exit of tourists, business people, students, and permanent residents
that global mobility infringes on state sovereignty. Therefore, the voluntary relinquishment of
the ability to control migration through various human rights, labor mobility, and border control
agreements affect states’ ability to sort and control migration flows.26 All states that sign these
multilateral, regional, and bilateral agreements are constrained in their ability to control
migration. However, European states, though the European Union and the European Economic
Area, have the densest network of regional agreements that constrain their ability to control
migration.
To delineate the constraints placed on states, we describe briefly each type of agreement and its
impact on state control of migration.
International human rights norms. The events of World War II prompted the international
community to establish human rights norms. The creation of the United Nations in 1945 and the
adoption of the Universal Declaration of Human Rights in 1948 mark the beginning of this era.
The international treaty that most directly deals with migration is the 1951 United Nations
Convention on Refugees.27 The principle of non-refoulement means that states must adjudicate
asylum claims from any person appearing at their border making such a claim. Although the
requirements to achieve refugee status are narrow, and narrowly interpreted by most states,
wealthy democracies have had to develop a large administrative apparatus to adjudicate these
claims. Moreover, states must either let the asylum seekers work to support themselves,
allowing some to disappear into the society, or provide subsistence for the asylum claimants
while their claims are adjudicated. The 1967 Protocol extended the refugee provisions, which
were initially limited to persons in Europe fleeing events before January 1, 1951, to the rest of
7
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 8 of 33
the world. As wealthy democracies closed their borders to labor migration, the refugee
convention forced these states to deal with an ever growing number of asylum claims, the vast
majority of which were declined. This convention has generated a substantial flow of asylum
seekers, some of whom cannot be deported, thus further constraining European states’ ability to
control migrant entry and residence.
European Convention on Human Rights. At the same time that the international community was
establishing international human rights standards, the European states were drafting even more
rigorous and specific human rights policies for themselves. The Council of Europe adopted the
European Convention on the Protection of Human Rights and Fundamental Freedoms in 1950.
The Convention came into force in 1953, and the European Court of Human Rights (ECHR),
which was created as part of the Convention, began functioning in 1959.28 The Convention
represents an affirmation of the rights articulated in the Universal Declaration of Human Rights,
but it also more specifically defined these rights and created an institution for enforcement in the
ECHR. Only 2.5% of the cases that came before the court between 1959 and 1993 involved
foreigners, but these cases (and the convention on which they are based) are significant for two
reasons: (1) they explicitly limited the right of the state to expel migrants under all
circumstances, and (2) they influenced the development of legal protections for migrants within
member states.29
In the cases brought to the ECHR, immigrant plaintiffs have most frequently cited Article 3
(protection against inhumane treatment) and Article 8 (right to normal family life). The court
has ruled in favor of migrants who were primarily raised in the host country but were there
illegally. In these cases, the court has said that returning these migrants to their home country
would be inhumane, as they have no familiarity with the country of their birth. In addition, the
undocumented parents of such children cannot be expelled either, as family separation would
violate the right to normal family life.30 The court has also asserted that state security is not a
valid reason for violating the provisions of ECHR.31 In addition, the EU has adopted minimum
standards for migrant family reunification.
European Union labor mobility and human mobility agreements. The economic benefits of
regionalization in Europe were sufficient for states to sign treaties that allowed citizens of other
member states to enter and to work in the host country. Although the European Union is now the
locus of all these mobility agreements, initially there were at least three notable agreements
involving labor mobility: the Nordic Union, the Benelux Union, and the European Coal and
Steel Community (ECSC). The European Economic Community (EEC) followed these
agreements and adopted the ECSC provisions on labor mobility. The European Union
incorporated the ECSC and EEC agreements and granted additional mobility rights. While these
agreements relinquished states’ ability to screen individual workers, and later individual citizens
of member states, the agreements did not provide for unlimited residence rights for all
individuals. Thus, these mobility agreements generated a foreign population, with some
proportion that did not fulfill the requirements of residence.
European border controls. The Schengen Agreement, signed in 1985, radically changed the
ability of member states to control their own borders.32 Until this agreement, each state
maintained and controlled its national borders, and all migrants still had to pass through border
8
[email protected]
http://mc.manuscriptcentral.com/wp
Page 9 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
checks so that receiving countries could distinguish between citizens of EU member states and
third country nationals. Schengen, however, abolished all internal border control mechanisms,
which allowed for the free movement of everyone, not just Europeans. Each Schengen member
agreed to common policies and procedures for visas, asylum requests, and border controls for its
external borders.33 But once a foreigner entered a member country, she was free to move
throughout the Schengen area with no further documentation checks. So, an Algerian could
enter Spain, for example, and then travel to France with no additional paperwork or contact with
French authorities. Thus the Schengen Agreement relinquishes control over entry to the least
well policed borders, and to those countries with geographic proximity to poorer states.
There were also regionally specific changes in the international system that affected the level and
direction of migration. Western European states were confronted with the collapse of exit
controls in neighboring states as the Soviet empire disintegrated in the early 1990s. Before the
fall of the Berlin Wall in 1989, Europe’s eastern border was largely impermeable. Subsequently,
Europe’s eastern border became extremely porous, and Eastern Europeans were eager to the take
advantage of the opportunity to emigrate to the economically and politically more developed
states in Western Europe. The sending states could no longer effectively prevent citizens from
emigrating without the authoritarian tools they used throughout the Cold War, and receiving
states had little infrastructure in place to deal with the new flow.
At the same time, the political disorder in the newly democratizing countries provided an
opportunity for immigrants from all over the world to get a foothold in Europe by immigrating to
Eastern European states. The number of Chinese in Hungary, for example, rose from nearly zero
to over 10,000 in the 1990s; many of these migrants then moved on to Western Europe. During
the same period, Chinese migration to Italy increased by 260%, and similar increases occurred
across the continent.34 Thus, the international environment presented western democracies, and
European states in particular, with increased migration flows.
Each of the above mentioned treaties, multilateral and regional, makes claims on state
sovereignty – the ability to unilaterally determine which migrants may enter and which migrants
may reside. This loss of sovereignty generates an effort to reclaim control through other means –
bilateral readmission agreements. In other words, states’ best alternative to negotiated agreement
has changed, impelling them to the negotiating table.35
H6. As host states sign international and regional agreements that relinquish their ability to
manage their migration flows, they will be more likely to negotiate readmission agreements.
NEGOTIATION OF READMISSION AGREEMENTS – THE PRIVILEGING OF
BILATERALISM
Given that migration generates costs, in increasing amounts, depending on the “bundle of rights”
defined by the political and social fabric of states, why don’t states develop a multilateral system
to deal with these issues? Unlike trade flows, which tend to be multilateral and reciprocal,
migration flows tend to be bilateral, unique, and non-reciprocal. To explain the bilateral nature
9
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 10 of 33
of international cooperation on migration, we look to the distinctive migration flows in each
country.
Although migrants now move in all directions, the main flows of voluntary migrants are in the
direction of more stable and wealthier states.36 In migration research, much has been made of the
changes in transportation and communication technology that have reduced the costs of
migration and contributed to the increased movement of peoples across international borders.37
Although travel has indeed increased, the picture of “permanent” migration is more nuanced.38
“Permanent” migration is defined as movements across international borders for more than one
year. Although the number of migrants has increased dramatically, from 74.1 million in 1960 to
188 million in 2010, so too has the global population.39 In fact, migration as a proportion of the
global population has been virtually constant, increasing only from 2.7% of the global population
in 1960 to 2.8% in 2010. Moreover, every region has been affected by migration. About half of
all international population movements are among countries of the developing world and only
37% of international migrants move from the developing world to the developed world.40
However, there has been a substantial regional shift in the location of migrants, with a doubling
or tripling of migrants as a proportion of the population in wealthy democracies. In North
America, the migrant stock has risen from 6.7% of the population in 1960 to 14.2% of the
population in 2010. In Europe, the proportion has risen from 3.5% to 9.7%.41 Comparable
figures for the Organisation of Economic Co-operation and Development (OECD) countries are
4.2% and 10.9% respectively. This distributional shift reflects both the shift in the patterns of
migration as well as differential levels of population growth. These migration patterns are
reflected in the description of states as “sending,” “transit,” and “receiving” states.
Migrant streams are determined by many factors, but theories of migration suggest that migrant
networks are an important variable that helps to maintain migrant flows. In Europe, the variation
in source countries is illustrated in Table 2. The top five sending states for four of the most
important receiving states in the EU present very different, historically based migrant profiles.42
France and the United Kingdom receive migrants primarily from ex-colonial states. Sweden’s
flows reflect its refugee resettlement and asylum seeker profiles. And Germany’s migrants
reflect early and more recent guest worker program flows. Out of these top five sending states
for the four countries, 18 observations are unique. Only Turkey and China appear in more than
one country profile. These data indicate that flows are both unique and bilateral and are
representative of the types of voluntary migration flows in the international system today.
The third characteristic of these migration flows is their non-reciprocity.43 Migration flows
among states of similar wealth tend to be balanced and relatively small, as migrants move in
either direction in response to specific economic opportunities. Flows among states with high
variation in wealth and wages tend to be unidirectional. If migrant flows between states are
relatively balanced, the costs associated with migrants tend to be reciprocal. Moreover, the
leverage states have on each other to accept the return of undocumented migrants is also
reciprocal. However, when flows are unbalanced, then the costs associated with migration are
not reciprocal.
10
[email protected]
http://mc.manuscriptcentral.com/wp
Page 11 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
The readmission agreement itself is designed to serve the interests of the receiving states. The
receiving state can still accept immigrants that it wants to accept, but it reserves the right to send
back anyone who does not have the proper documentation to be in the country. The sending
state, however, has little desire to accept the return of their own citizens, and they have even less
desire to accept the return of foreigners who simply transited through their territory. Receiving
states must thus provide other incentives to the sending state in order to reach an agreement.
These incentives can include larger quotas for legal migration, expedited and streamlined
migration processes, and technical assistance for implementing the agreement.44 They may also
receive benefits not directly related to migration, such as increased foreign aid payments.45
Readmission agreements thus exemplify quid pro quo cooperation, where parties with divergent
interests can nonetheless sustain a mutually beneficial cooperative arrangement through side
payments.
Reciprocity has long been recognized as an important factor in facilitating cooperative
behavior.46 If states are able to develop “tit-for-tat” strategies, cooperative behavior can emerge.
The issue with unilateral migration flows is that receiving states have very little leverage over the
behavior of immigrant sending states, other than border controls. Therefore, as noted above, in
order to control costs associated with migration, receiving states are required to offer incentives
to sending states. However, receiving states want to sign agreements only with those states that
generate flows, to minimize the cost of the incentives. This is the primary reason that even
region-wide agreements through the auspices of the European Union have been so limited. Why
should Germany offer incentives to Tajikistan through an EU-wide agreement, when it receives
few migrants from Tajikistan?
We argue that host nations will want to sign agreements with only those countries that generate
migrant flows. Because countries have different migrant networks, they will sign agreements
with different countries. The costs generated by migration flows accrue to specific states rather
than to the system as a whole. Receiving states want to minimize their side payments and
therefore want to negotiate only those agreements necessary to halt the flows. Thus we argue
that the private nature of the externalities generates the bilateral relationship between sending
and receiving states. A multilateral agreement would require that receiving states offer side
payments high enough to entice the sending states most capable of demanding a high price for
their cooperation – even though some of the receiving states receive few migrants from those
specific sending states. Thus the overall cost of the multilateral agreement would be much
higher than the sum of the costs of many bilateral agreements.
H7. As net migrant flows from individual sending states increase, host states will be more likely
to negotiate bilateral readmission agreements with those states.
Our theory focuses on how the costs of migration have changed for states in the international
system. In other words, states’ best alternative to negotiated agreement has changed over time
due to changes in migrant access to state supplied benefits, the domestic population’s reaction to
migrants, and to the rising number of immigrants generated by adherence to treaties that
constrain states’ ability to manage migration in their own interests. However, not all states have
experienced the same degree of rising costs. As noted above, European states have relinquished
the most sovereignty over migration management due to treaties involving European integration;
11
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 12 of 33
they are also the states with the most finely woven welfare nets in the international system. That
is why we anticipate that most bilateral readmission agreements would be signed by European
states. In the best possible world, we would evaluate our hypotheses with a global data set which
would provide the greatest degree of variance on the independent and dependent variables.
However, our theory is data intensive: we need data on bilateral readmission agreements, which
are limited primarily to European states, and on bilateral migration flows, which are even more
limited.47 Therefore we evaluate our hypotheses on a subset of European states. That said, the
European data set represents substantial variation on both independent and dependent variables.
Our quantitative analysis helps to confirm the plausibility of our theory rather than being an
exhaustive evaluation of the theory.
Because European states have signed the greatest number of international and regional
agreements, we argue that they will be the most likely to sign readmission agreements.
However, our explanation differs from one strand of the literature on European integration that
suggests a “spillover” effect of integration.48 That is, one component of integration – in this case,
agreements on labor mobility and border controls – will cause states to resolve the problems
generated by the initial integration through further integration. This would suggest that the
member states of the EU would direct the institutions of the European Union itself to sign
readmission agreements in order to deport more effectively the undocumented population.
However, the history of readmission agreements at the level of the EU undermines this
interpretation of spillover. The EU, as noted above, did draft model readmission agreements that
were taken up by individual states. And the EU received authority to negotiate readmission
agreements on behalf of all members of the EU for some 18 states. However, only 13 of those
negotiations have come to fruition – and constitute only three percent of all European
readmission agreements. European states, including states that are not EU members
(Switzerland and Norway, notably), are the main actors in negotiating readmission agreements.
A QUANTITATIVE TEST: DATA AND METHODS
We test the hypotheses laid out above with data from nine European states: Belgium, Denmark,
Finland, Germany, Italy, the Netherlands, Norway, Spain, and Sweden. Our data are limited to
these nine states because historical data for bilateral migration flows are only available for these
nine countries. We analyze the data using the Cox Proportional Hazards model, which is a
survival model. This method of analysis is appropriate because it allows us to find the
probability that a state will conclude a readmission agreement with a particular country as time
passes, accounting for changes in the independent variables. So, for example, we expect that as
the number of immigrants from a particular country increases, the risk that the receiving country
will sign an agreement will increase as time goes on. The Cox Proportional Hazards model
allows us to test such claims.49
The Risk Period. In survival analysis, one must identify the period of time in which there is a
risk that the event of interest might occur; in this case, the “event” is the conclusion of a
readmission agreement. Outside this risk period, we assume that there is no risk of the event
occurring. We identify the risk period here as beginning with the very first international human
rights treaties (the Universal Declaration of Human Rights and the European Convention on
12
[email protected]
http://mc.manuscriptcentral.com/wp
Page 13 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Human Rights) and the establishment of the first European institutions, specifically the Nordic
Passport Union, the Benelux Economic Union, the European Steel and Coal Community, and the
European Economic Community, in the early 1950s. Indeed, the very first readmission
agreements were signed in 1952 between a small number of European states. The risk period
continues through the present time.
The earliest date for which we were able to obtain data on source country-specific migrant flows
for the widest number of countries is 1980. The data on migrant flows end between 2000 and
2004. This means that the data are both left- and right-censored. It is left-censored because
countries were at risk of signing an agreement before 1980, and it is right-censored because
countries may have signed agreements after 2004; in fact, they may continue to sign agreements
into the future. Although this censoring is not ideal, the data still allow us to evaluate most of
our hypotheses.
The Data. The unit of analysis in this study is the dyad-year. Each dyad includes a destination
country (one of the nine European states mentioned above) and a source country. Only relevant
dyads are included in the dataset; we define relevant dyads as ones in which at least one person
from the source country has migrated to the destination country during at least one of the years in
the dataset. This is a low bar to clear, but the record of readmission agreements shows that some
destination countries do sign readmission agreements with states that send only a small number
of migrants. Italy, for example, has signed a readmission agreement with Mexico, despite the
fact that the greatest number of migrants that Mexico ever sent to Italy in a single year was 207,
and in many years, it sent none. France has also signed a many agreements with countries that
send only a small number of migrants; for example, it has an agreement with Dominica despite
the fact that the small Caribbean nation sent an average of only nine migrants per year in the
years that it sent any migrants it all. Countries for which not a single migrant was recorded
during the entire time period are thus not included. We assume that countries that have no
exchange of migrants are not at risk for signing a readmission agreement. The dyadic dataset
allows us to test an important part of our theory, which is that readmission agreements with a
particular sending country become more likely as the flow of migrants from the sending state
increases. The drawback of this approach is that bilateral migration flow data are very limited,
restricting both the number of receiving states we can include in the study as well as the time
span. Appendix 1 provides the overall summary statistics for the data we analyze.50
The Dependent Variable. Data on the signing of readmission agreements comes from the
MIREM project at the European University Institute. Once a receiving country has concluded an
agreement with a sending country, this is marked as a “failure,” which in a survival analysis
means that the event of interest has occurred.
The Independent Variables. We test our hypothesis about the effect of international/regional
agreements by including a dummy variable for signatories of the Schengen Agreement, which
was drafted during the time frame covered by our analysis. Ideally, we would like to test the
effect of each agreement that we think was important in restricting state sovereignty over
migration control: the UN Convention on Refugees (and the 1970 Protocol), the European
Convention on Human Rights, the ECSC, the European Community, the Benelux Economic
13
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 14 of 33
Union, and the Nordic Union. However, many of these are collinear during the time period of
the analysis, so we limit our focus to the effect of the Schengen Agreement.
We measure the welfare cost of immigrants by multiplying the per capita cost of government
social welfare programs by the total migrant stock within the receiving country, giving the sum
of the per capita welfare costs of migrants. We then divide this number by the total population,
which gives the welfare costs of migrants per capita (this accounts for the fact the larger
countries likely host a larger number of migrants).51 While these data accurately capture how
much the states spend on social welfare, it does not necessarily capture differences in the level of
access that immigrants have to government services. Our calculation assumes that migrants in
different states have the same level of access to government welfare programs, when in fact
some states may restrict access to these benefits only to citizens. However, there is likely to be a
correlation between higher welfare expenditures and a greater absolute welfare expenditure on
migrants, even if restrictions on access can prevent migrants from capturing some benefits.
We expect that the political costs of migration will grow when the political salience of migration
is high, and this provides an incentive to receiving states to sign readmission agreements. We
use the percent of the vote won by right-wing parties as a measure of the political salience of
migration. Growing support for right-wing parties indicates that the public thinks there are too
many immigrants. Immigration restriction is a significant part of the platforms of right-wing
parties across Europe, and anti-immigrant sentiment is a key determinant of support for the
parties.52
We expect that states will try to “offshore” the expense of border control through readmission
agreements when the costs of border control increase in the receiving state in comparison the
costs of control in the sending state. We estimate this cost differential by subtracting the sending
state’s GDP per capita, adjusted for purchasing power parity from the receiving state’s GDP per
capita, adjusted for purchasing power parity.53
We use the growth rate of real GDP per capita and unemployment rates as measures of factors
that may mitigate demand for readmission agreements.54 The ILO unemployment data come
primarily from official employment office records, although the data from Finland and Italy
come from the ILO’s Labor Force Survey. Where possible, registered unemployment rates are
used; total coverage figures are used for Finland and Italy. As the growth rate of real GDP per
capita increases, we expect that demand for readmission agreements should decrease. When
unemployment decreases, we expect that the demand for readmission agreements should
decrease.
We argue that states will be more likely to conclude a readmission agreement the larger the
inflow of unwanted migrants. However, there are no accurate data on annual inflows of
undocumented migrants from particular source countries; any data that do exist are very broad
and can only provide a rough estimate. We thus use the flow of documented migrants as a proxy
for unwanted migrants. We assume that larger documented flows likely correspond with larger
undocumented flows. The data come from International Migration Flows to and from Selected
Countries: The 2005 Revision, produced by the United Nations Department of Economic and
Social Affairs, Population Division.55
14
[email protected]
http://mc.manuscriptcentral.com/wp
Page 15 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Control variables. The variation in the number of readmission agreements signed by European
states may in part be due to bureaucratic inertia. Once one agreement has been signed, it
becomes easier to sign another agreement. The state then has a model for future agreements, and
it also has experience in negotiating agreements, which reduces the cost of each future
agreement. The process of concluding agreements may generate individual bureaucrats and even
departments with an interest in pursuing more agreements as a way to expand their sphere of
policy control. This effect may be especially strong when the incentives needed to induce the
sending country to sign an agreement are low. France, for example has signed an extraordinary
number of readmission agreements, and it has even pursued agreements with far-flung countries
(for example, in Latin America) that export a relatively small number of migrants to France. The
effect of bureaucratic inertia is probably small, and it certainly wouldn’t override broader state
interests. However, it does seem reasonable that the more agreements that a state has signed, the
more agreements it is likely to sign in the future. Bureaucratic inertia is measured simply by
adding up the total number of readmission agreements signed by the receiving country up until a
given year.
Lastly, we control for geographic distance between the sending and receiving countries.56 We
expect that states that are geographically farther apart are less likely to sign agreements with
each other. The effect of distance may be overwhelmed by the effect of migrant flows, but
distance may have a distinct effect, particularly since migration from geographically proximate
states may be harder to control than migration from distant states. Persistent migrants from
nearby states will likely have an easier time overcoming barriers that receiving states erect to
control migration. Thus, receiving states may find that other tools, apart from readmission
agreements, are sufficiently effective for controlling migration from states far away but
insufficient for controlling migration from states nearby.
QUANTITATIVE TEST: RESULTS
The results of the analysis are displayed in Table 3. First, signing the Schengen Agreement has a
strong and significant effect, as expected. Surrendering sovereignty over border control by
creating a common border with Schengen member states is followed by an increase in
readmission agreements. In an alternate model specification, Schengen implementation was
measured rather than the signing of the agreement. In this specification, agreement
implementation did not have a significant effect, and the model actually suggested that the effect
is negative. This makes sense; the period of time between signing the Schengen Agreement and
the date it goes into effect is very long. The original signatories of Schengen in 1985 gave
themselves an entire decade to prepare themselves for the transition, which included
standardizing admission requirements, streamlining procedures, and, it appears, preparing for
negative externalities by pursuing bilateral readmission agreements with states outside the
Schengen Area. By the time the agreement actually went into effect, the receiving countries had
already signed the majority of the agreements that they intended to sign. States thus likely
anticipated the changes that the Schengen Agreement was going to bring, and they planned
accordingly.
15
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 16 of 33
The domestic welfare cost of migrants does not have the expected effect, but these are not
significant results.57 These results can be interpreted in two ways. First, it could be that
economic costs in the form of welfare benefits for immigrants do not, in fact, move migration
policy. If so, this is a surprising result; migration policy scholars and observers have argued that
higher welfare costs of immigration are a key determinant of negative public opinion toward
migrants.58 If our result is accurate, then this casts doubt on existing theories of migration policy.
But there is a second explanation; it could be that our measure does not accurately capture the
welfare costs of immigrants. This may be because countries restrict access to social welfare
benefits so that migrants cannot enjoy them. Thus, rising immigration rates alone may not
increase social welfare costs. To determine which explanation is correct, further research on
immigrant access to government welfare across countries and across time is needed. However,
the fact that there are no readily available data on immigrant access to government welfare is,
itself, a challenge to the theory that welfare costs drive restrictive migration policy. If we, as
researchers, can’t assess the true welfare costs of immigrants, how could policy makers (or the
public) assess this cost? If costs are opaque, then they are unlikely to drive policy. It is possible
that the perception of costs could drive policy; this might instead point towards political factors
as far more important than government expenses. Only further research will allow us to untangle
these competing explanations.
The political cost of unwanted migration, as measured by support for right-wing parties, is
strongly significant and in the direction expected. As electoral support for right-wing parties
grows, the more likely a country is to pursue restrictive migration policy through readmission
agreements.
As expected, the greater the differential in GDP per capita between receiving and sending states,
the more likely receiving states will be to conclude a readmission agreement, and this is a
significant result. This suggests that receiving states may be using readmission agreements as a
way to reduce border control costs by offshoring this duty to states where the costs of control are
much lower.
Unemployment has a positive and significant effect on the likelihood of signing an agreement, as
expected. Countries experiencing low unemployment are more to demand more immigrants and
thus are less likely to seek out readmission agreements to control immigration, and vice versa.
Similarly, when economic growth increases and demand for immigrants increases, states are less
likely to sign readmission agreements. This effect, however, appears to be relatively small and it
is not statistically significant.
The number of immigrants to the receiving country has a strong, significant effect on whether or
not two countries conclude a readmission agreement. States do not pursue agreements with all
states; they focus on ones with which there is a significant migrant flow. This explains both why
receiving states pursue different numbers of agreements and why they pursue them with different
sending states. In addition, this may explain why bilateral readmission agreements remain more
popular than multilateral agreements. Since receiving states do not share the same migration
profiles, they do not seek to conclude readmission agreements with the same sending states.
16
[email protected]
http://mc.manuscriptcentral.com/wp
Page 17 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Bureaucratic inertia, as measured by the total number of agreements signed by a country, has the
expected effect and it is significant. The more agreements a country has signed, the more
agreements it is expected to sign in the future. We suspect this is perhaps due to the
entrepreneurial efforts of the individuals and agencies tasked with concluding and managing
readmission agreements. In addition, familiarity and experience with readmission agreements
may reduce the cost of drafting and concluding new agreements, which would reduce the
incentives needed to induce a receiving state to pursue a readmission agreement.
Lastly, geographic distance between sending and receiving states has a significant effect in the
expected direction. States that are geographically farther away from each other are less likely to
conclude agreements with each other. This relationship remains significant even though migrant
flows are included in the model, which suggests that distance isn’t important simply because
geographically proximate sending states send a larger number of migrants to receiving states
than do more distant sending states. This could signify that receiving states are better able to
control the flow of migrants from more far-flung states using tools other than readmission
agreements. These other, more unilateral tools may be less effective on closer neighbors,
creating an incentive to turn to readmission agreements. In addition, it may be cost effective to
offshore border control in geographically proximate states but not in more distant states.
Beyond basic conclusions about the direction and significance of effects, interpreting log hazard
ratios in a Cox Proportional Hazards Model is not always intuitive. In their most simple form,
hazard ratio indicates that a unit increase in the independent variable signifies a β increase in the
probability that the event of interest (the signing of a readmission agreement) will occur in the
next time period (the next year).59 In the results presented in Table 3, the logged hazard rates are
presented, so that a coefficient above zero indicates a positive relationship, while a number less
€
than zero indicates a negative relationship.
The interpretation of the coefficients is complicated by that fact that diagnostic tests revealed
non-proportional hazards in the original model. This means that the effect of the independent
variables varies across time, while the model assumes that the effects do not vary with time.60 To
solve this problem, we interacted affected variables with a transformation of time (in this case,
the natural log of time) and included the interactions in the model.61 We included interactions
with the GDP Growth variable and the Distance variable, and this solved the non-proportional
hazards problem. Unfortunately, this complicates the interpretation of the constituent
coefficients. The constituent coefficients now must be interpreted as the effect only during the
first year of the analysis (1980), and not the effect of the variables over the entire time period.
Following methods suggested by Amanda Licht, we first calculated the point in time where the
effect of the variables changes sign.62 Distance between countries has time-varying effects.
During the period of our analysis, states are more likely to sign agreements with geographically
proximate states. However, after 33 years, the direction of this relationship changes. We suspect
that this is because the number of geographically proximate states is quite limited. Over time,
receiving states that continue to conclude readmission agreements are reaching farther and
farther afield to find partners. 63
17
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 18 of 33
Qualitatively, the pattern of European readmission agreements supports our theoretical
framework. Figure 1 plots signed agreements over time for the nine receiving states that have
concluded the greatest number of readmission agreements. Although the process of European
integration began shortly after World War II, alongside of the human rights agreements that
constrained state sovereignty, readmission agreements were sparse in the initial period.
As noted earlier, the bilateral readmission agreements are concentrated in two periods, the early
1960s subsequent to the signing of the Treaty of Rome, which established free movement of
labor among the original European Economic Community (EEC) member states, and the period
after 1985 which saw the signing of the Schengen Agreement among the original participants
(Germany, France, and the Benelux countries) and the Treaty of Maastricht (1992) which
extended the right of movement from workers to all citizens of member states. The plots show
spikes in signing around the times when these regional agreements were concluded. Neither the
United Kingdom nor Ireland joined to the passport free travel zone, retaining sovereign control
over borders and obviating the need for bilateral readmission agreements – and hence these states
are low on the list of signers. Italy, having been a net emigration country until the 1980s,
delayed negotiating bilateral readmission agreements until it joined the Schengen Agreement in
1990; it then quickly became one of the most prolific signers of bilateral readmission
agreements.
Switzerland is an interesting case because it is not a member of the European Union. However,
beginning in 1994, it negotiated a series of bilateral agreements with the EU that ultimately
brought Switzerland into the labor and mobility clauses of the European Union as well as
accession to the Schengen Agreement in 2004. In anticipation of non-reciprocal flows that
would follow, Switzerland negotiated a series of bilateral readmission agreements with EU
member states – the earliest with the EU candidate countries of Central Europe – as well as with
third countries (MIREM). Switzerland follows the pattern of European states. Where national
controls are constrained and flows are non-reciprocal, the costs to the receiving state increase,
leading those states to offer incentives to take back country nationals as well as transit migrants.
CONCLUSION
We have argued that migration flows are distinctive from other types of international economic
flows, generating a distinctive type of cooperation in the international system. The patterns
privilege bilateral agreements and are negotiated primarily in those cases where the bundle of
migrant rights is large and states lack the ability to effectively control their borders. Receiving
states must offer incentives to reach agreement with sending states; the size of the incentives and
the likelihood of agreement increase with increased flows and associated costs.
Migration is distinctive from other types of international economic flows both because of the
bundle of rights that generates costs for receiving states and because of the unidirectional flow of
most migration in the international system today. Migration therefore does not present issues of
collective action or the provision of collective goods that underlie many of the international
agreements and organizations that populate the international economic system today.
18
[email protected]
http://mc.manuscriptcentral.com/wp
Page 19 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Bilateral readmission agreements are akin to the recent spate of bilateral investment treaties.
Bilateral investment treaties, as described by Elkins, et al. are agreements reached between poor
nations and wealthy nations in an effort to attract foreign direct investment (FDI) from wealthy
states, by promising to treat foreign capital on terms identical to domestic capital. These treaties,
too, are bilateral and based on non-reciprocal flows.64 Yet that incentive structure differs,
because poor states are attempting to attract foreign direct investment and the agreements are
driven by a competitive dynamic among these states. Moreover, this represents only a small
portion of all foreign direct investment flows – around 70 percent of all foreign direct investment
continues to flow among wealthy OECD countries. And FDI among wealthy OECD countries is
governed by informal reciprocity among these states. In our case, bilateral readmission
agreements are driven by the wealthy states in order to reverse the flow of migrants. Reciprocal
migration flows among wealthy European states are, of course, governed by mobility agreements
among EU member states and include some European non-EU member states.
The proliferation of readmission agreements in European democracies since the 1990s, and the
continuing variation in the propensity of countries to sign these agreements, is an interesting
development for those interested in international cooperation on migration. The evidence we
have presented suggests that readmission agreements are a way for states to bilaterally reassert
their sovereignty over migration within an institutional framework that has increasingly limited
state sovereignty in favor of supranational authority.65 For European states, subjugating
themselves to a growing European authority, pooling resources, and coordinating policy have
provided many benefits, economically and diplomatically. However, this increasing
Europeanization is not without drawbacks. The readmission agreement has become one of the
tools that European states use to address the negative externalities of Europeanization. It is only
here where the costs of migration are sufficiently high that states are impelled to enter into
international agreements. Elsewhere, the smaller bundle of rights and/or the ability to adequately
control migration reduce incentives for states to enter into cooperative relationships.
One should take care, though, in drawing conclusions from increasing cooperation on
readmission about cooperation in other areas of migration. Migration does not exhibit the
characteristics of other international economic flows that generate incentives for cooperation. It
is unlikely then, that cooperation on readmission will grow into cooperation on other migrationrelated issues.
Moreover, readmission agreements are primarily concerned with restricting immigration, not
facilitating it. International cooperation and global governance, as expressed in bilateral and
multilateral treaties, tend to carry with them the notion of shared or relinquished sovereignty as
well as greater openness to the international system. Bilateral readmission agreements represent
a new instrument, wielded primarily by European states, to regain sovereignty in an issue area
upon which regional agreements have encroached. This form of “cooperation” should compel us
to rethink how we understand global governance in the contemporary era.
19
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 20 of 33
References
Achermann, Alberto, and Mario Gattiker. 1995. “Safe Third Countries: European
Developments.” International Journal of Refugee Law 7, no. 1: 19-38.
Axelrod, Robert. 1984. The Evolution of Cooperation. New York: Basic Books.
Betts, Alexander. 2006. “Towards a Mediterranean Solution? Implications for the Region of
Origin.” International Journal of Refugee Law 18, no. 3-4: 652-676.
Betz, Timm, and Cali Mortenson Ellis. 2009. “Stay at Home Migrants? The Strategic Effects
of Readmission Treaties on International Migration.” Paper presented at the annual
meeting of the American Political Science Association, Toronto, September 3-6.
Bond, Eric W., and Tain-Jy Chen. 1987. “The Welfare Effects of Illegal Immigration.” Journal
of International Economics 23: 315-328.
Bonnén, Preben, and Michael Søsted. 2003. “The Origin, Development and Perspectives of
Nordic Co-operation in a New and Enlarged European Union.” Özp [Austrian Journal of
Political Science] 32, no. 1: 19-32.
Bouteillet-Paquet, Daphné. 2003. “Passing the Buck: A Critical Analysis of the Readmission
Policy Implemented by the European Union and Its Member States.” European Journal
of Migration and Law 5: 359-377.
Box-Steffensmeier, Janet M., and Bradford S. Jones. 2004. Event History Modeling. A Guide for
Social Scientists. New York: Cambridge University Press.
Brücker, Herbert, Gil S. Epstein, Barry McCormick, Gilles Saint-Paul, Alessandra Venturini, and
Klaus Zimmermann. 2001. Managing Migration in the European Welfare State. At
http://www.frdb.org/upload/file/paper1_23jun01.pdf, accessed June 14, 2012.
Cassarino, Jean-Pierre. 2010. “Dealing with Unbalanced Reciprocities: Cooperation on
Readmission and Implications.” In Jean-Pierre Cassarino, ed., Unbalanced Reciprocities.
Cooperation on Readmission in the Euro-Mediterranean Area. Washington, DC: Middle
East Institute. At http://www.mei.edu/content/unbalanced-reciprocities-cooperationreadmission-euro-mediterranean-area, accessed June 14, 2012.
Castles, Stephen, and Mark J. Miller. 2009. The Age of Migration. International Population
Movements in the Modern World, 4th Edition. New York: Guilford Press.
Coleman, Nils. 2009. European Readmission Policy. Boston, MA: Martinus Nijhoff Publishers.
Collinson, Sarah. 1996. “‘Safe Third Countries’ and ‘Readmission’: The Development of an
Asylum ‘Buffer Zone’ in Europe.” Transactions of the Institute of British Geographers
21, no. 1: 76-90.
20
[email protected]
http://mc.manuscriptcentral.com/wp
Page 21 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
De Georgi, Giacomo, and Michelle Pellizzari. 2006. “Welfare Migration in Europe and the Cost
of a Harmonised Social Assistance.” IZA Discussion Paper No. 2094. Bonn, Germany:
Institute for the Study of Labor.
Doomernik, Jeroen, and Michael Jandl. 2008. Modes of Migration Regulation and Control in
Europe. Amsterdam: Amsterdam University Press.
Djajic, Slobodan. 1987. “Illegal Aliens, Unemployment and Immigration Policy.” Journal of
Development Economics 25: 235-249.
Duerden, Martin. 2009. What are Hazard Ratios? What is…? Series. Hayward Medical
Communications. At
http://www.medicine.ox.ac.uk/bandolier/painres/download/whatis/what_are_haz_ratios.p
df, accessed June 14, 2012.
Elkins, Zachary, Andrew T. Guzman, and Beth A. Simmons. 2006. “Competing for Capital: The
Diffusion of Bilateral Investment Treaties, 1960-2000.” International Organization 60,
no. 4: 811-846.
Ellerman, Antje. 2008. “The Limits of Unilateral Migration Control: Deportation and Inter-state
Cooperation.” Government and Opposition 43, no. 2: 168-189.
Eurobarometer. At http://ec.europa.eu/public_opinion/index_en.htm, accessed June 14, 2012.
European Union. 1991. Report from the Ministers Responsible for Immigration to the European
Council Meeting in Maastricht on Immigration and Asylum Policy. SN 4038/91 (WGI
930), December 3. Brussels: European Union.
______. 1996. “Council Recommendation of 30 November 1994 Concerning a
Specimen Bilateral Readmission Agreement between a Member State and a Third
Country.” Doc. 396Y0919(07). Official Journal C274: 20-24.
______. 1999a. Summary Report of Member States’ Contributions on Readmission
Questions. Doc. 7609/1/99, May 3. Brussels: European Union.
______. 1999b. Strategic Committee on Immigration, Frontiers, and Asylum. Doc.
13888/99, December 7. Brussels: European Union.
______. 2002. Proposal for a Return Action Programme. Doc. 14673/02, November 25.
Brussels: European Union.
______. 2005. Readmission Agreements. Memo 05/351, October 5. Brussels: European Union.
______. 2009. The Schengen Area and Cooperation. Brussels: European Union. At
21
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 22 of 33
http://europa.eu/legislation_summaries/justice_freedom_security/, accessed June 14,
2012.
Groenendijk, Kees. 1980. “The Working Group on Legal Aid for Immigrants: A Public Interest
Law Organization in the Netherlands.” In Erhard Blankenburg, ed., Innovations in the Legal
Services. Cambridge, MA: Anton Main, 165-176.
Guiraudon, Virginie, and Gallya Lahav. 2000. “A Reappraisal of the State Sovereignty Debate:
The Case of Migration Control.” Comparative Political Studies 33, no. 2: 163-195.
Haas, Ernst. 1958. The Uniting of Europe. Stanford: Stanford University Press.
Hailbronner, Kay. 1997. “Readmission Agreements and the Obligations on States under Public
International Law to Readmit their Own and Foreign Nationals.” Zeitschrift für ausländisches
öffentliches Recht und Völkerrecht [Journal of Comparative Public Law and International
Law] 57: 1-49.
Hainmueller, J. and Michael J. Hiscox. 2007. “Educated preferences: Explaining attitudes
toward immigration in Europe.” International Organization 61(2): 399-442.
Gordon Hanson, Kenneth F. Scheve and Matthew J. Slaughter. 2007. “Public Finance and
Individual Preferences over Globalization Strategies.” Economics and Politics 19(1): 133.
Hanson, Gordon H., and Antonio Spilimbergo. 2001. “Political Economy, Sectoral Shocks, and
Border Enforcement.” The Canadian Journal of Economics 34, no. 3: 612-638.
Hatton, Timothy. 2007. “Should we have a WTO for International Migration?” Economic
Policy, 339-383.
Heston, Alan, Robert Summers, and Bettina Aten. 2009. Penn World Table Version 6.3. Center
for International Comparisons of Production, Income and Prices at the University of
Pennsylvania.
IGC (Intergovernmental Consultations for Asylum, Refugee and Migration Policies in Europe,
North America and Australia). 2002. IGC Report on Readmission Agreements. Geneva,
Switzerland: IGC.
______. 2006. IGC Report on Readmission Agreements. Geneva: IGC.
ILO. 2010. LABORSTA Labour Statistics Database. Table 3A Unemployment, general level
(Thousands). Geneva: International Labor Organization.
IOM. 2007. Supplemental Materials. Prepared for the International Dialogue on Migration
Intersessional Workshop on Free Movement of Persons in Regional Integration Processes
in Geneva, Switzerland, June 18-19.
22
[email protected]
http://mc.manuscriptcentral.com/wp
Page 23 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Jacobson, David. 1996. Rights Across Borders: Immigration and the Decline of Citizenship.
Baltimore: Johns Hopkins University Press.
Joppke, Christian. 1998. Why Liberal States Accept Unwanted Immigration. World
Politics 52, no. 2: 266-293.
Katzenstein, Peter. 1996. The Culture of National Security: Norms and Identity in World
Politics. New York: Columbia University Press.
Keohane, Robert O. 1986. “Reciprocity in International Relations.” International Organization
40(1): 1-27.
Laczko, Frank. 2003. Europe Attracts More Migrants from China. Migration Information
Source. Washington, DC: Migration Information Source. At
http://www.migrationinformation.org/Feature/display.cfm?ID=144, accessed June 14,
2012.
Lavenex, Sandra. 1998. “‘Passing the Buck’: European Union Refugee Policies towards Central
and Eastern Europe.” Journal of Refugee Studies 11, no. 2: 126-145.
Lavenex, Sandra and Emek M. Uçarer, eds. 2002. Migration and the Externalities of European
Integration. Lanham, MA: Lexington Books.
Licht, Amanda A. 2011. “Change Comes with Time: Substantive Interpretation of NonProportional Hazards in Event History Analysis.” Political Analysis 19, no. 2: 227-243.
Lubbers, Marcel, Mérove Gijsberts, and Peer Scheepers. 2003. “Extreme Right-Wing Voting in
Western Europe.” European Journal of Political Research 41, no. 3: 345-378.
Mayda, A.M. 2006. “Who is against immigration. A cross-country investigation of individual
attitudes towards immigrants.” Review of Economics and Statistics.
Migration Information Source. At http://migrationinformation.org, accessed June 14, 2012.
MIREM. 2009. European University Institute, Robert Schuman Centre for Advanced
Studies. At http://www.mirem.eu, accessed June 14, 2012.
Myers, Gordon M., and Yorgos Y. Papageorgiou. 2000. “Immigration Control and the Welfare
State.” Journal of Public Economics 75, no. 2: 183-207.
OECD. 2005. SOCX 1980-2005 Aggregated Data. At
http://stats.oecd.org/Index.aspx?datasetcode=SOCX_AGG, accessed June 14, 2012.
Odell, John S. 2000. Negotiating the World Economy. Ithaca, NY: Cornell University Press.
23
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 24 of 33
Ruhs, Martin and Ha-Joon Chang. 2004. “The Ethics of Labor Immigration Policy.”
International Organization 58(1): 69-102.
Sassen, Saskia. 1996. Losing Control? New York: Columbia University Press.
Soysal, Yasemin. 1994. Limits of Citizenship. Chicago: Chicago University Press.
Steinberger, Helmut. 1985. “Reference to the Case-Law of the Organs of the Convention of
Human Rights Before National Courts: Communication on the Courts of the Federal
Republic of Germany.” In Proceedings of the Sixth International Colloquy about the
European Convention on Human Rights. Boston, MA: Martin Nijhoff.
UNDP (United Nations Development Program). 2009. Human Development Report 2009.
Overcoming Barriers: Human Mobility and Development. New York: UNDP.
World Bank. 2011. World Development Indicators. Washington, DC: World Bank.
1
The phrase “bundle of rights” was first used by Ruhs and Chang 2004.
This is described as “self-limited sovereignty” by Christian Joppke 1998.
3
Migrants can either enter without proper documents or overstay the terms of their entry.
4
Hailbronner 1997.
5
Coleman 2009.
6
This text is common to a number of European Community readmission agreements, including ones with Albania,
Hong Kong, Macao, and Sri Lanka (Coleman 2009).
7
EU 2005.
8
Statistics from the IGC (2002) and the IGC (2006) are not consistent but do support the dramatic rise in readmission agreements after 1990.
9
IGC 2002; 2006.
10
EU 1996.
11
EU 2005; 2011.
12
IGC 2006.
13
Betts 2006, Bouteillet-Paquet 2003, Ellerman 2008. There is also a large literature on their implications for
refugee and asylum policy (Collinson 1996; Lavenex 1998; Achermann and Gattiker 1995).
14
Betts 2006 implies that states are motivated by their logistical inability to unilaterally prevent migrants from
crossing their borders. For example, in the early 1990s, approximately 100,000 to 200,000 undocumented migrants
were arriving each year to Europe from Northern African countries, and 65,000 of these were sub-Saharan Africans
in transit. This, argues Betts, was the impetus for the proliferation of readmission agreements between European
states and North African states.
15
Cassarino 2010 gives the example of France signing many readmission agreements with Latin American countries
and Spain’s inability to do so, given the salience of Latin American migration to Spain.
16
See for example, Odell 2000.
17
This is described as “self-limited sovereignty” by Christian Joppke 1998.
18
Ruhs and Chang 2004.
19
De Giorgi and Pellizari 2006.
20
Brücker et al. 2001.
21
In fact, investing state resources in control may be more costly than allowing a completely unfettered flow of
undocumented immigrants (Myers and Papageorgiou 2000).
22
Doomernik and Jandl 2008.
23
See, for example, various Eurobarometer polls.
24
Hanson and Spilimbergo 2001.
25
Bond and Chen 1987; Djajic 1985.
2
24
[email protected]
http://mc.manuscriptcentral.com/wp
Page 25 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
26
Sassen 1996. Also see Betts 2006 and Footnote 14.
Some scholars have argued that the ability of states to unilaterally control immigration, using any means
necessary, has declined due to the globalization of liberal norms and the development of an international human rights infrastructure (Katzenstein 1996; Sassen 1996; Jacobson 1996; Soysal 1994). The direct effect of
international human rights norms and international agreements on national migration policies is unclear, and the
early agreements were careful to not infringe upon the ability of the state to control migration (Guiraudon and Lahav
2000). The effect of human rights regimes, particularly during the early years, should not be overestimated.
28
Guiraudon and Lahav 2000.
29
Guiraudon and Lahav 2000.
30
The precedents are based on the following cases: Berrehab v. Netherlands (ECHR, June 21, 1988); Moustaquim v.
Belgium (ECHR, February 18, 1991); Djeroud v. France (ECHR, January 23, 1991); Lamguindaz v. United
Kingdom (ECHR, June 28, 1993); Beldjoudi v. France (ECHR, March 26, 1992). See Steinberger 1985.
31
Guiraudon and Lahav 2000.
32
The Schengen agreement was controversial when it was first drafted in 1985, and only five countries originally
signed on: France, Germany, Belgium, Luxembourg, and the Netherlands. In addition, the agreement did not take
effect until ten years later, in 1995. In the years since, however, the Schengen area has gradually grown to include 25 countries, including every member of the European Union except the United Kingdom and Ireland (Romania,
Bulgaria, and Cyprus are not yet members of Schengen, but they are set to become members in the future; Norway, Lichtenstein, and Switzerland are Schengen members even though they are not EU members.
33
EU 2009.
34
Laczko 2003.
35
This is consistent with Betts’ 2006 claim that increasing migration caused states to sign bilateral readmission
agreements. However, we model the costs that states incur that changes their best alternative to negotiated
agreement as well as the signatories to the agreements, which illuminates the fact that European states signed
readmission agreements whereas other receiving states failed to do so.
36
UNDP 2009.
37
Castles and Miller 2009, for example.
38
UNDP 2009.
39
UNDP 2009, 21.
40
UNDP 2009, 21.
41
UNDP 2009, 30.
42
Some European states experience higher gross flows from neighboring European states but, as noted above, those
flows tend to be reciprocal so that net flows are much smaller (see Migration Information Source).
43
Hatton 2007.
44
EU 2005.
45
Betz and Ellis 2009. A 1991 report to the European Council from the Ministers responsible for immigration
suggested that increasing cooperation on a wide range of economic, social, financial, and political issues, as well as assistance for addressing the root causes of migration, should be conditional on readmission agreements (EU 1991;
Coleman 2009).
46
Axelrod 1984, Keohane 1986.
47
Our data for bilateral readmission agreements come from MIREM, which provides only data on European states.
The data on bilateral migration flows is from the United Nations Population Division and is limited to a small number of wealthy democracies beginning in 1980. The data on bilateral readmission agreements from the IGC
(Intergovernmental Consultations on Asylum) are proprietary information of the member states and are only
available as summary statistics. 48
Haas 1958.
49
Box-Steffensmeier and Jones 2004.
50
Note that the total number of agreements reported for each country in Appendix 1 is different than in Table 1
because Appendix 1 only reports agreements since 1980 (when the dataset starts), while Table 1 reports all
agreements signed.
51
Data on government welfare expenditures come from the OECD 2010. Data on population and total migrant
stock come from the World Bank; since migrant stock data are only available in five-year intervals, data points for
intervening years are interpolated.
52
Lubbers, Gijberts, and Scheepers 2002.
27
25
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 26 of 33
53
World Bank 2011.
The growth rate of real GDP per capita data come from Heston, Summers, and Aten 2009; the unemployment rate
data come from the ILO 2010.
55
There is some difficulty in comparing data among destination countries because the data are reported differently.
Some countries (Finland, Italy, Spain, and Sweden) only report the combined inflow of foreigners and citizens of the
destination country from a particular country. Citizen inflows, however, are not truly immigrants, since they are
already citizens of the destination country. Other countries (Belgium and the Netherlands) only report data on
foreign flows, not citizen flows. Our solution is to use the data on the combined citizen and foreign inflows
wherever possible and the foreign inflow figures for Belgium and the Netherlands. While the data do not represent
the exact same concept, the correlation between foreign inflows and foreign and citizen inflows is about 90%.
56
Cassarino 2010.
57
We tried numerous alternative measures of welfare costs, including welfare expenditures per capita, welfare
expenditures as a percent of GDP, welfare expenditures interacted with migrant stock, welfare expenditures
interacted with migrant flows, and various other specifications. None of the alternatives delivered significant
results, and most were not in the expected direction.
58
See Hanson, et al. 2007, Hainmueller and Hiscox 2007 and Mayda 2006.
59
Duerden 2009.
60
We also tested our data using the Weibull model, which allows for the effect of independent variables to vary
over time. However, this model presents another problem; it assumes that the effect of the variables is either
increasing or decreasing monotonically over time, which is an equally unrealistic assumption (Box-Steffensmeier
and Jones 2004). Nevertheless, in this case the substantive results from the Weibull model are very similar to those
of the Cox model. We display the Cox results because the model has the fewest restrictions.
61
Box-Steffensmeier and Jones 2004.
62
Licht 2011.
63
GDP Growth is statistically insignificant but requires the time variant correction (which is also statistically
insignificant). In the case of GDP Growth, the point is at about t=9, or nine years after the beginning of our
analysis. GDP growth has a negative effect, albeit insignificant effect, on the probability of signing agreements at
the beginning of our analysis and then, by 1989, the effect changes direction. This means that somewhere around
1989, higher economic growth no longer reduces the likelihood that a receiving country will sign a readmission
agreement; in fact, high growth rates begin to make it more likely that a country will sign an agreement. We suspect
that this time-variant effect might be due the broader patterns in the business cycle that our particular time frame
captures. Between 1984 and 1993, the average annual GDP growth rate (per capita) in our sample of receiving
countries dropped from 3.5% to -1.4%. By 1995, the average growth rate had recovered, rising to 3.4%. Since
these economic trends affected all of the receiving states in our sample, our results might be skewed. A larger
sample, over a longer period of time, might provide more robust, significant results.
64
Elkins, Guzman, and Simmons 2006.
65
Ellerman 2008.
54
26
[email protected]
http://mc.manuscriptcentral.com/wp
Page 27 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Table 1: Number of Readmission Agreements (2002)
State
Number of Agreements
EU countries prior to 2004
Ireland
4
Portugal
10
Finland
11
United Kingdom
16
Greece
19
Sweden
20
Denmark
21
Luxembourg
23
Netherlands
26
Belgium
26
Austria
28
Spain
31
Germany
35
Italy
51
France
68
Non-EU European countries
Iceland
8
Norway
22
Switzerland
45
European countries’ mean
25.7
Other wealthy democracies
New Zealand
0
United States
2
Australia
4
Canada
12
Non-European countries’ mean
4.5
Source: MIREM; IGC 2002
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[email protected]
http://mc.manuscriptcentral.com/wp
Page 28 of 33
Page 29 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
Table 2: Top Non-EU Migrant Admissions 2003 (Selected EU Countries)
FRANCE
GERMANY
UK
SWEDEN
Algeria
Turkey
Pakistan
Iraq
Morocco
Russia
India
Thailand
Tunisia
Ukraine
South Africa
Serbia
Turkey
United States
Nigeria
China
Congo
China
Afghanistan (2006)
Somalia
Source: Migration Information Source
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3: Cox Model of Readmission Agreements
Readmission Agreement
Log Hazard Ratios
(Standard Errors)
Schengen Signatory
1.343*
(.727)
Per Capita Welfare Costs of Migrant Stock
-.001
(.535)
Support for Right-Wing Parties
.036**
(.016)
Difference in GDP between Sending and
.045***
Receiving States
(.009)
GDP Growth Rate
-.495
(.764)
Unemployment Rate
.114*
(.061)
Migrant Inflows (Logged)
.176***
(.056)
Total Number of Agreements Signed
.075*
(.042)
Distance between Capital Cities
-5.272***
(in thousands of kilometers)
(1.364)
GDP Growth Rate*time (Logged)
.226
(.570)
Distance*time (Logged)
1.495***
(.460)
N
24,712
Log pseudolikelihood
-464.860
Clustered by receiving country. Uses Efron method of ties. *p<.10, **p<.05, ***p <.01
[email protected]
http://mc.manuscriptcentral.com/wp
Page 30 of 33
Page 31 of 33
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
[email protected]
http://mc.manuscriptcentral.com/wp
World Politics
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Number of Readmission Agreements Signed by Receiving States 320x213mm (72 x 72 DPI) [email protected]
http://mc.manuscriptcentral.com/wp
Page 32 of 33
Page 33 of 33
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
World Politics
APPENDIX 1: Summary Statistics
Variables
Observations Mean
Readmission
Agreements
Migrant Inflows
33,063
N/A
33,060
709
Standard Minimum
Deviation
N/A
0
(32,952 no events)
6,962.27 0
GDP Growth Rate
33,063
2.08%
2.35
-8.95%
7.64%
Unemployment Rate 33,064
8.47%
4.47
1.3%
20.4%
Per Capita Welfare
Costs of Migrant
Stock
Support for RightWing Parties
Kilometers between
Capital Cities
32, 223
$348.21
257.96
$24.59
$1,091.59
32,881
3.72%
5.99
0%
26.3%
32,882
5878 km
3518
0 km
19,835 km
[email protected]
http://mc.manuscriptcentral.com/wp
Maximum
1
(111 events)
455,075