International Review of Law and Economics 23 (2004) 345364

Cooperation in the shadow of regulatory competition: the case of asylum legislation in Europe
Sgolne Barbou des Places a , Bruno Deffains b,
b

European University Institute, Florence, Italy, and Facult de droit et Sciences Economiques, Universit Nancy 2, France CREDES, Facult de Droit et Sciences Economiques, Universit Nancy 2, 13 Place Carnot, c.o. #26, c.o. 26-13 Place Carnot, Nancy Cedex 54 035, France Received 15 December 2002; accepted 20 July 2003

Abstract Traditional analysis considers that the granting of protection to refugees is an international public good, and thus explains both the heterogeneousness in refugee protection in Europe and the spiral that has hardened the EU Member States asylum legislation from the mid-1980s onwards as the result of free riding in the provision of the good. In contrast, the paper considers that the heterogeneousness in refugee distribution is best explained by the joint product model and that the spiral of restriction is best explained by the common pool resource model and regulatory competition theory. The paper explains, and gives empirical evidence of the emergence and development of a competitive game among the EU Member States, and shows the result and the consequence of this upon cooperative attempts among States. 2003 Elsevier Inc. All rights reserved.
JEL classication: K3 Keywords: Asylum; Refugees; Regulatory competition; Joint product; Common pool resource

1. Introduction The current establishment of a European common asylum policy raises two main questions. The rst question is: How can an equitable distribution of asylum seekers among EU

Corresponding author. Tel.: +33-3-83-19-26-52; fax: +33-3-83-19-25-33. E-mail addresses: Segolene.Barbou@iue.it (S.B. des Places), Bruno.Deffains@univ-nancy2.fr (B. Deffains).

0144-8188/$ see front matter 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.irle.2003.07.001

346

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

Member States be ensured? It is the burden sharing objective. The second question is: How can a European legislation that is fair to asylum seekers and refugees be enacted, i.e. that provides protection in accordance with the standards of the 1951 Geneva Convention on refugees? More generally, the issue at stake is the organisation of cooperation among States that, for many decades, have unilaterally decided and implemented their national asylum policy. To encompass all these issues, it is necessary to have a complete view of the asylum policies in the EU before the Treaty of Amsterdam 1997, i.e. before the EU Member States established a European asylum regime. In particular, two striking phenomena must be explained. First, the unevenness in refugee distribution and protection among EU Member States. It is crucial to determine the causes of the heterogeneousness in asylum policy and to evaluate whether, when it leads to a situation of unequal distribution in Europe, it is problematic or compatible with the development of a European asylum policy. Second, we must explain the causes, and evaluate the result and effects, of the spiral of restriction that characterises the evolution of Member States asylum legislation from the mid-1980s onwards. All of the EU Member States have introduced a wide range of procedures relating to the arrival, admission and entitlements of people wishing to claim refugee status in their territory. They have enacted restrictive legislation, preventing asylum seekers from entering into national territory and modied the procedural guarantees for the asylum application examination in order to restrict them. In addition, they have withdrawn social welfare and legal aid entitlements, limited the right to work and to education for asylum seekers and their families. Finally they have developed measures to encourage the return or repatriation of asylum seekers to their country of origin. While this evolution has been well documented in legal literature (Bouteiller-Paquet, 2001; Crpeau, 1995; Jeannin, Meneghini, Pauti, & Poupet, 1999; Joly, 1999; Noll, 2000), there is a dearth of analysis using a law and economics approach. The rare economic analysis that study the asylum issue, generally in an international perspective, assume that the provision of refugee protection is an international public good (see in particular Suhrke, 1998), and thus implicitly admit that both the heterogeneousness of asylum seekers distribution in Europe and the spiral of restriction in asylum legislation are the result of free riding in the provision of the good. Yet empirical evidence does not conrm such theoretical expectations deriving from the public good characterisation. For this reason this paper mobilises the joint product and the common pool resource models, together with regulatory competition theory, in order to explain the heterogeneousness in refugee provision and the spiral of restriction. The paper deliberately proposes a law and economics perspective, and thus stresses the consequences of legislative amendments on the choice of rational actors via their impact on the costs and benets of decisions. The paper builds upon the chronological evolution of the EU Member States asylum policy. Section 1 shows that after the Second World War, European States had generous asylum policies and welcoming legislation, and the heterogeneousness in the reception towards refugees among Member States was not considered problematic. This situation can only be explained if we consider that the good, which is to provide protection to asylum seekers and refugees, is not an international public good but a joint product, i.e. States could expect private benets from hosting and protecting asylum seekers and refugees. In addition, the joint product model permits an understanding of the formation of an alliance. But, as

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

347

Section 2 explains, by the mid-1980s, the number of asylum applications dramatically increased and there was a shift in asylum policies. States started competing to prevent asylum seekers from accessing their territory and claiming refugee status. As benets have decreased and costs increased, States have started competing for a new good: to grant as little protection as possible. The previous alliance was thus replaced by a process of regulatory competition that can be interpreted using the common pool resource model. We therefore explain the spiral of restriction, and show the excessive efforts by the EU Member States, in accordance with the common pool resource model. We then compare the theoretical explanation with empirical evidence and show how competition was started and developed. In Section 3 we evaluate the result of the competitive game and conclude that it is negative. In Section 4 we discuss the possible forms of cooperation among the EU Member States likely to replace or limit the negative competitive process. We show that the many attempts to set out cooperative schemes between the EU Member States have failed in the shadow of competition. Then we turn to the current development of the (post-Amsterdam) European asylum policy and evaluate what the conditions for efcient cooperation in the eld of asylum could be. Finally, we reach some general conclusions with regards to our central questions: What is the optimal asylum policy both for asylum seekers and Member States in the EU and how should such an optimal policy be established?

2. Heterogeneousness in the provision of refugee protection: an explanation using the joint product model There have always been important differences among the EU Member States in the accommodation and protection of asylum seekers and refugees. Many factors explain this unevenness in refugee distribution. Geographic proximity to the region in crisis, the asylum seekers perception of the States goals, ethnic or family ties (see Bocker & Havinga, 1998). The differences in asylum legislation also inuence the distribution of refugees and create differences in burden. However, the majority of refugee law specialists explain the uneven distribution of refugees in Europe by free riding strategies, thus assuming that the provision of protection to refugees is an international public good. In contrast, we deem that the inequitable distribution is the result of differences in legislation that come, as the joint product model explains, from the existence of private benets in providing protection to refugees. 2.1. Benets of providing protection to asylum seekers and refugees Throughout the 20th century, and more specically since the end of the Second World War, States have devoted a considerable amount of effort and resources to the task of providing refugees with international protection. Their purpose has been twofold: rst, to safeguard the lives and liberty of people whose basic rights have been threatened in their country of origin; and second to safeguard their own interests by ensuring that large-scale population movements are managed in a predictable manner and in accordance with agreed principles (UNHCR, 1997). This description can help understand the goals of States which provide refugee protection. Such provision, whether it takes the form of directly granting

348

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

asylum or the form of voluntary contributions to refugee agencies, can be conceptualised as producing two kinds of benets for States: security and altruistic benets. Security benets can be expected because once displaced beyond their national borders, asylum seekers and refugees may constitute a potential threat to other States. According to Acharya and Dewitt (1997), security comes from the ability to control entry and exit so that the core values, institutions and processes of that system are not threatened. Increased security can be regarded as the principal benet of managing the population of asylum seekers and granting asylum, because the accommodation of refugees potentially reduces the risk of refugees spreading the very conict from which they are eeing. Together with direct security benets,1 States can also expect altruistic benets. It is generally acknowledged that altruistic externality arises when the donor derives utility from knowing that less fortunate people are better off (Sudgen, 1982). In the eld of asylum, such altruistic benets result mainly from the incidence of provision for refugees on the reputation of the country. The image of a State that is protective of human rights is one of the major benets that a State can expect when it develops a generous asylum policy. Germany is a signicant example, where the right to asylum was incorporated into the Constitution in order to compensate for the shame of the Second World War acts. Suhrke (1998) also explains Western countries involvement in resettlement plans in the aftermath of the Second World War by their willingness to gain a good reputation. The analysis of the benets of regulating and granting asylum permits the discussion of the nature of the good (i.e. providing protection to asylum seekers and refugees). As they consider that security and altruistic benets are non-excludable and non-rival, the majority of specialists argue in favour of an international public good. According to Suhrke (1998), if one State admitted refugees, the others would benet from the greater international order that ensued regardless of their own admissions, so refugees provision is an international public good from which all States benet, irrespective of which country receives refugees. The consequences of this interpretation are important. Firstly, according to public good theory, cooperation produces positive-sum benets which in turn create the will to share the burden among actors as the benets of the contribution exceed the costs of the contributors. At an international level, cooperation should produce a level of provision of a valued public good which individual States cannot attain on their own. Public goods theory thus suggests that we should expect a signicant under-provision of international public goods in the absence of central taxation and enforcement authorities. The problem exists because it is in the interest of an individual member to free ride on the efforts of others. Such arguments have been applied in the NATO context to explain not only the motives for cooperation on national defence but also to demonstrate that the members of an alliance have an incentive to cheat by free riding on the common security provided by other members. Interestingly, most studies adopt a similar vision for asylum and consider that by improving the well-being of those in need, provision for refugees yields non-excludable and non-rival benets for the

1 Of course these benets have to be balanced with indirect security burden as dened by Milner (2000) that concern the impact of refugees on distributive justice and political opportunity within a given State. Refugees may threaten governments and social harmony because they create serious nancial and social burdens on the receiving nations.

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

349

community of nations. Free riding is logically expected because States rely on the payments of others by withholding some or all of their estimated expenses. Secondly, another consequence can be anticipated according to the exploitation hypothesis given by Olson and Zeckhauser (1966). According to their argument, it is ceteris paribus the larger States whose action will make more of a difference to the total common effort than the action of small States. As a result, larger States will have an incentive to contribute a disproportionate share to the overall effort. We will therefore observe a systematic tendency for exploitation of the big by the small.2 In the case of refugees, it can be anticipated that rich nations shoulder a disproportionate burden of the good for poor nations in terms of the share of the GDP allotted to provision for refugees. Last, public good theory predicts sub-optimal public good reimbursements because a contributor will not account for the spill-over benets that its contribution confers on others. 2.2. From public good theory to joint product model We now have to assess these theoretical expectations in light of empirical evidence. Is the international provision of protection for refugees a public good? The rare contributions that use the OlsonZeckhauser test show that the granting of asylum fails to conform to the pure public good model. They indicate that if one ranks the number of annual asylum applications in proportion to the size of a Member States population or GDP, one nds that the top positions, i.e. the highest burdens, are predominantly taken by smaller States (Thielemann, 2002). Using a Spearman rank-correlation test, Betts (2002) sought to ascertain whether the GDP is rank correlated with the level of burden borne. The idea is that if there is a positive and signicant correlation between the GDP and the asylum provision as a proportion of the GDP amongst the EU Member States, there is exploitation of the big by the small such that the EU asylum provision is an international public good.3 But in the 1990s, there was no empirical support for the exploitation hypothesis. Statistical insignicance suggests that the level of free riding by poorer States on wealthier States is limited and that provision for refugees cannot be described as a pure public good. The pure public good benets thus fail to explain why some of the EU Member States unilaterally provide a disproportionately higher level of asylum provision than others, despite being smaller in terms of their GDP. Far from free riding, several economically smaller States such as The Netherlands and Denmark provide proportionally high levels of asylum, while some larger States, such as Italy and France, bear far less of the burden (Betts, 2002). How can such a situation be explained? The most convincing argument is that the provision of protection is more likely to yield some private benets (reputation of a major promoter of human rights and/or security benets when the contributor is near the region of instability). In light of the assessment of private benets it makes sense to use the joint product theory developed by Sandler and Hartley (2001) whereby activities are
2 In an effort to conrm this hypothesis, a number of studies have analysed the correlation of defence spending and GDP in the NATO context. They have shown that the United States, as NATOs largest State, has contributed a disproportionate share of the burden for common defence. 3 The author assumes that the EU asylum provision is a normal good, for which the will to provide increases with the GDP.

350

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

characterised as yielding both pure public good benets for the community as well as contributor specic benets. This theory helps explain why the collective action implications, as regards the provision of protection for asylum seekers and refugees, differ from those resulting from the pure public goods scenario. Disproportionate burden sharing need not occur, so the burden cannot simply be correlated with an income measurement. If a poorer State receives a large amount of nation specic benets, such as status enhancement, then it may carry a relatively heavy burden. Furthermore, sub-optimality will be attenuated as the share of nation specic benets increases. If, for example, all derived benets are nation specic, then an efcient allocation will be achieved since the benets will match marginal costs.4 If we decide to apply the joint product model to the asylum issue, we may therefore expect an inequitable distribution of refugees in the European Union as the private benets differ from State to State. We may also understand that the country closest to a particular conict from which refugees emanate has the greatest interest in stabilising the crisis through the accommodation of refugees (e.g. the case of Austria during the Balkan crisis). There is also substantive evidence for the claim that a countrys willingness to host refugees is positively related to its more general commitment to norms such as human rights protection and solidarity (Thielemann, 2002). This is well evidenced by the fact that the Scandinavian States and The Netherlands have, for many years, provided disproportionately high levels of refugee protection. Thus, we deem that the heterogeneousness of provision of refugee protection among Member States and the correlative uneven distribution of refugees among them is the result of the difference in specic benets rather than the result of free riding. The joint product approach also explains why, in the decades following the Second World War, the game was an alliance: Member States were spontaneously contributing to the general aim of providing protection. Yet the joint product model does not offer an analytical framework that is capable of explaining the substantial changes that took place in the mid-1980: a spiral of restrictions in refugee legislation appeared. We propose to use the literature on common pool problem to explain this phenomenon.

3. The spiral of restriction in refugee protection: a common pool problem The content of legislation not only inuences the differences in the burden among the EU Member States but also impacts on the nature of the game. As the conditions for refugees reception and protection are not identical in the Member States, asylum seekers may choose their destination State on the basis of a preference for the country that provides better treatment: this is the asylum shopping strategy. We now propose to provide evidence
4 It is important to note that the portion of the public ingredient found in joint products depends on the ratio of excludable benets (i.e. State specic benets) to the total benets. In addition, joint products may give rise to a coordination game in which one of the Nash equilibria has all players contributing to the collective action. If the portion of jointly produced private benets is considerable enough with regard to the total outputs, contributing to the activity may even become the dominant strategy. This has implications for the forming of alliances if alliance design allows potential allies to take advantage of ally specic benets as well as excludable public benets. For example, NATO burden sharing behaviour is generally considered optimal if nation specic benets are integrated.

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

351

and explain the development of a process of regulatory competition among the EU Member States asylum legislation. 3.1. The origins of regulatory competition The rst question is: Why did a process of competition start among Member States and lead to a spiral of restriction in refugee protection? As the costs and benets are central to our analysis, we argue that the change in asylum policies comes from the change in costs and benets of providing protection. First, benets have signicantly decreased. After the fall of the Iron Curtain, altruistic benets and reputation effects have decreased. The instrumental use of the right to asylum granted to people eeing persecution from communist countries before the fall of the Iron Curtain turned out to be meaningless. In addition, all the EU Member States have faced economic difculties, recession and unemployment and the benets of hosting immigrants have dropped. Secondly, costs have increased because of changes in migration trends. Asylum seekers who were traditionally European and skilled migrants have been replaced by less skilled asylum seekers coming from other regions of the world and of different ethnic origin. The costs of integration have thus increased. But most of all, the costs have grown with the number of refugees within the territory. The 1990s witnessed an eruption of conicts that led to mass refugee movements, including some within Europe itself (ex-Yugoslavia). Table 1 highlights this major change in asylum seekers trends and the massive increase of asylum seekers in Europe. The number of asylum applications in the European Union increased six-fold from the early 1980s to the early 1990s. However, this increase has affected Member States differently. From an EU total of 5.7 million asylum applications in the period 19802000, Germany took 2.7 million alone, Austria, Belgium, France, The Netherlands, Sweden and the United Kingdom accounted for another 2.5 million while the other eight Member States dealt only with a million asylum applications. Therefore, as every examination of an asylum application implies nancial, administrative and social costs, the opportunity cost for host countries became very high (and probably superior to perceived benets) during the 1990s.5 In addition, the budgets of the government ofces in charge of asylum procedures dramatically increased. In France, for example, it increased from FF 52.9 million in 1988 to FF 142.9 million in 1990, and in Germany from DM 143.5 million in 1992 to DM 486 million in 1993. According to Velling and Woydt (1993), the average annual cost of an asylum seeker in Germany was DM 12,000 to DM 15,000 at the beginning of the 1990s, 50% of which may be accounted for by way of welfare benets and 50% by way of the costs of accommodation and administration. If we add the social and political costs that States bear when the number of refugees increases, the marginal cost of asylum policy is certainly growing. Moreover, refugees have an impact on distributive justice and they create serious nancial and social burdens on the EU States. As a consequence, the increase of costs has generated a change in behaviour. At this stage a change appeared in the game played by the States due to the emergence of a new good. The previous cooperative game has been replaced by a competition in legal procedure.
5 Following Jandl (1995) and Betts (2002), we consider that the actual number of asylum seekers is a more accurate representation of the burden borne by a given State than any attempted estimates of expenditure levels that could be derived from the available data.

352 S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

Table 1 Asylum applications in EU Member States 19802001 (UNHCR, 1999)

Austria 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 9,260 34,560 6,310 5,900 7,210 6,720 8,640 11,410 15,790 21,880 22,790 27,310 16,240 4,750 5,080 5,920 6,990 6,720 13,810 20,100 18,280 30,135 Belgium 2,730 2,290 2,910 2,910 3,650 5,300 7,640 5,980 5,080 8,110 12,960 15,170 17,650 26,880 14,350 11,420 12,430 11,790 21,970 35,780 42,690 24,549 Denmark 70 120 300 800 4,310 8,700 9,300 7,590 11,310 5,280 18,990 12,910 20,070 16,480 7,990 10,050 7,390 5,570 6,080 6,950 10,350 12,403 Finland n.a. 20 10 20 30 20 20 50 60 180 2,740 2,130 3,630 2,020 840 850 710 970 1,270 3,110 3,170 1,651 France 19,910 19,860 22,510 22,350 21,710 28,930 26,290 27,670 34,350 61,420 54,810 47,380 28,870 27,560 25,960 20,170 17,410 21,400 22,380 30,910 38,590 47,263 Germany 107,820 49,390 37,420 19,740 35,280 73,830 99,650 57,380 103,080 121,320 193,060 256,110 438,190 322,610 127,210 127,940 116,370 104,350 98,640 95,110 117,650 88,363 Greece 1,790 2,240 1,190 450 760 1,400 4,230 6,930 8,420 3,000 6,170 2,670 1,850 810 1,300 1,310 1,640 4,380 2,950 1,530 3,080 4,650 Ireland n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 30 40 90 360 420 1,180 3,880 4,630 11,090 14,800 10,324 Italy 2,130 3,640 3,140 3,040 4,560 5,420 6,480 11,030 1,240 2,120 4,830 26,470 6,040 1,650 1,790 1,730 680 1,860 11,120 33,360 14,000 9,620 Luxembourg n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a 390 260 430 1,710 2,910 630 689 The Netherlands 1,350 1,590 1,210 2,020 2,600 5,640 5,870 13,460 74,902 13,900 21,210 21,620 20,350 35,400 52,570 29,260 22,170 34,440 45,220 39,300 43,900 32,579 Portugal 1,640 600 1,120 610 380 70 280 440 330 160 80 260 690 2,090 770 450 270 300 370 310 200 192 Spain n.a. 330 2,460 1,420 1,180 2,360 2,280 2,480 4,520 4,080 8,650 8,140 11,710 12,620 11,990 5,680 4,730 4,980 6,650 6,410 7,930 9,219 Sweden n.a. 12,650 10,230 7,050 12,000 14,500 14,600 18,110 19,600 30,340 29,420 27,350 84,020 37,580 18,640 9,050 5,750 9,660 12,840 11,230 16,300 23,513 United Kingdom 2,350 2,430 4,220 4,300 2,910 4,390 4,270 4,260 4,000 11,640 26,210 44,840 24,610 22,370 32,830 43,970 29,640 32,500 46,020 71,150 75,680 70,995

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

353

3.2. A new good and a new game Let us consider that there is a xed number of asylum seekers and that asylum seekers are costly for asylum countries.6 In that case, there is a good for which the EU Member States compete through the designing of asylum rules: to provide as little protection as possible. It is important to note that this good is both rival and non-excludable. It is rival because lowering the number of asylum seekers enjoying protection in one country tends to increase the number of persons claiming protection in the other countries. The good is non-excludable because no State can be excluded from enacting legislation that lowers the level of protection granted, thus reducing the number of persons enjoying protection. This implies that each country is able to inuence the number of asylum applications by changing its asylum legislation: by becoming more restrictive a country deters future applicants. Therefore the identied good is an international common pool resource so that the fundamental condition for regulatory competition is satised (Klliker, 2003). We deem that this theoretical expectation is met in practice. In the mid-1980s, States were responsive to asylum seekers preferences. They were convinced that asylum seekers were comparing the legal protection offered throughout Europe and selecting the country with the most generous asylum policy as a destination State. Yet, there are many obstacles in the way of asylum shopping and not all asylum seekers choose a destination State on the basis of the expected legal treatment (see Bocker & Havinga, 1998). But, if the reality of the arbitrage among legal norms might be a weaker factor than expected, pressure on law suppliers has clearly taken the form of fear of the arbitrage (Barbou des Places, 2003). Analyses of public opinion, together with political declarations provide evidence of the conviction that protective legislation is an attractive policy. As a consequence, the real or perceived threat of huge ows of migrants entering their territories gave Member States an incentive to adapt their legislation following the example of their direct competitors. The result was the emergence of a competitive game. We are not hypothesising perfect competition but the process of regulatory competition in its dynamic sense, as rivalry among jurisdictions (Deffains, 2001). The observation of the law suppliers actions shows evidence of the States reaction function. Taking the example of France and Germany, Rotte, Vogler, and Zimmermann (1996) rst looked at national legal measures for asylum migration and their actual effects on the partner countries. They wanted to nd empirical evidence for existing interdependence between asylum migration to France and Germany, respectively, which may be inuenced by legal actions. They estimated two ordinary xed-effects panel models, the number of asylum seekers in each country being the endogenous variable. The vector of exogenous variables included several indicators for socio-economic and political positions. Potential spill-over caused by national legislation was modelled by including the legal procedures of a country in the equation for the other one. As noted by the authors, the most important nding is that the French law reforms in 1991 resulted in an increase of asylum seekers coming to Germany instead. Despite its isolated nature, the study shows
6 This proposition is not incompatible with the fact that immigration of workers in principle increases the welfare of economies, due to the greater availability of a production factor, thereby shifting the supply curve to the right. In fact, with high unemployment and structural strains on social budgets in Europe, we simply assume that the costs could exceed the benets of the reception of refugees.

354

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

that there is considerable interdependence of asylum policies in the EU. National legislation aimed at a further reduction in asylum migration contains a potential for devaluation races in asylum laws among Member States using a beggar-thy-neighbour effect in this eld. This devaluation race was in fact Europe-wide. Unilaterally, each Member State tried to develop mechanisms aimed at dissuading asylum seekers from selecting it as their future host country and thus at redirecting the ows of asylum seekers to their neighbours. Member States deliberately used national regulations as a strategic weapon in international competition in which gains in one country were at the expense of costs imposed on others. Several factors provide evidence for this beggar-thy-neighbour attitude. First, the simultaneity of law amendments (Jeannin et al., 1999), that tends to reveal the interdependence among national legislation. Second, there are striking examples of convergence within the content of the legislation. Three sets of techniques, invented in one country and then copied by the others, are particularly signicant. The rst took the form of sanctions imposed on carriers transporting improperly documented passengers (Cruz, 1995). The second example concerns the creation of international or transit zones in ports and airports. A third convergent evolution in Member States legislation was their incorporation of two complementary concepts: safe third country (Achermann & Gattiker, 1995; ECRE, 1995) and manifestly unfounded application, measures that permit the limitation of the access to the status of refugee, that justify the curtailing of the examination procedure, and that limit procedural rights and guarantees. All Member States have imported the three techniques into their national legal order. This is signicant because all these measures clearly had the effect of redirecting asylum seekers towards the competitors and/or other States. Member States have designed diversion policies, and created a general market of deection (Landgren, 1999; Noll, 2000) as a result of their containment strategy. This effect is empirically observable: when one State implements a measure that, for instance, establishes carriers liability or enhances the safe third country technique, one, or several, other Member States become the receptacle of the asylum seekers who change destination. Therefrom the chain amendments can be observed during the 1980s, and 1990s (Barbou des Places, 2003; Rotte et al., 1996).

4. The devaluation race as a consequence of regulatory competition Following the common pool resource model, the race should produce a sub-optimal outcome, and should lead to a tragedy in Hardins words. The competitive game in the eld of asylum conforms with the model and we propose, rst to capture the dilemma by a two-countries game in deterrence, then to provide empirical evidence of the negative result of competition among Member States asylum legislation. 4.1. A two-countries game in deterrence Let us consider States which have the competence and the ability to deter asylum seekers by increasing the volume of administrative procedures, designated by a (i.e. norms regulating access to the national territory, rules organising the access to legal protection and legal rights and subsidies conferred upon asylum seekers and refugees). We assume that

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

355

the greater the a the higher the bureaucratic costs to manage the system of reception and protection of refugees. Restrictive and non-liberal asylum policies imply costs related to border controls and strict asylum application examination. Note that a restrictive asylum policy is more costly than a liberal one. As a consequence, if a State prefers more a, it will be less attractive to asylum seekers but the costs per asylum seeker will be higher. Suppose moreover that asylum seekers have to nd refuge in one of the competing States: when a country chooses a, it does not take into account the fact that its policy imposes costs on other States. In other words, an increase in a imposes a negative external cost on the other States. The consequence is that each country will opt for a too restrictive asylum policy and supra-national coordination may be justied. To demonstrate this phenomenon, imagine a world divided into two countries, 1 and 2. For the reasons presented above, the countries spend resources on legal procedures in order to process asylum applications.7 Denote aj A as the level of legal procedures chosen by country j, where A denotes the set of possible legislations. a can also be interpreted as the costs of legal procedures. Suppose that the welfare (of citizens) of countries 1 and 2 can be represented by V1 (a1 , a2 ) and V2 (a1 , a2 ), respectively. Thus, the level of procedures chosen by country 2 has an impact on the welfare of country 1 (and vice versa). Following Marceau (1997), we suppose that Vj (aj , ak ) is strictly concave in aj and that (Vj /ak ) < 0, j = k. This implies that the legal procedures chosen by the States are perfect substitutes. For this reason, we can observe in Fig. 1 that an indifference curve for country 1 (2) is necessarily concave to the horizontal (vertical) axis. Moreover, the fact that procedures taking place in country k affect negatively the welfare of jurisdiction j can be rationalised by the following argument: a more restrictive legislation in k can hurt (in terms of direct or indirect security burden) j because refugees nd it relatively more protable to settle in j. The problem of country 1 is maxa1 V1 (a1 , a2 ). The rst- and second-order conditions of this maximisation problem are V1 (a1 , a2 )/a1 = 0 and 2 V1 (a1 , a2 )/2 a1 < 0, respectively. This denes a reaction function for country 1: a1 = f1 (a2 ). For country 2, the analogue to the problem of country 1 leads to a reaction function a2 = f2 (a1 ). A Nash N N N N N N equilibrium is then a pair (a1 , a2 ) such that a1 = f1 (a2 ) and a2 = f2 (a1 ). This is depicted in Fig. 1 in the (a1 , a2 ) space where the reaction curves are negatively sloped. The N N Nash equilibrium pair (a1 , a2 ) lies at the intersection of the two reaction curves and is denoted by N. This simple case shows that the Nash equilibrium level of the actions is higher than the Pareto efcient level. In the context of asylum legislation, this means that the Nash equilibrium level of legal procedures could be higher than the Pareto efcient one. This can be seen in Fig. 1, where the countries indifference curves through the Nash equilibrium are depicted (denoted by u1 and u2 , respectively). Taking into account the shape of the indifference curves, it is possible to observe that all the points in the hatched area to the southwest of N would improve the situations of both countries. Hence, the Nash equilibrium is not Pareto efcient: it entails legislation that is too restrictive.

7 The costs of legal procedures have to be distinguished from the costs of accommodation of refugees. These last costs could be high but they concern only accepted refugees in a Member State. As we are interested in asylum seekers we focus on the administrative costs of considering the applications.

356

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

Fig. 1. Nash equilibrium.

The argument behind the result of legislation which is too restrictive is that a State, when choosing its level of procedures, trades off its own benets (reduced number of refugees and increased indirect security) and costs (forgone resources, perhaps reduced direct security), but does not take into account the fact that this has a negative impact on the welfare of the other States. There are costs to other countries that are not internalised. Finally, according to the model, the countries, in the absence of global regulation (or collective action) earn less than they potentially could if resources were properly managed. 4.2. Evidence of a race to the bottom Legal scholars have dwelt on the constant and convergent restriction of legal rights conferred upon asylum seekers in the EU Member States. They have described a new asylum regime that reects the emergence of a new paradigm, from a regime implementing a selective but integrative policy of access and full status recognition paired with full social rights, to one which maximises exclusion, undermines status and rights and emphasises short-term stay for refugees (Joly, 1999). This description indicates the existence of a race to the bottom. First, the actions developed by Member States have jeopardised the security of potential and actual asylum seekers (UNHCR, 1997). Through the use of the safe third country and manifestly unfounded application techniques, asylum seekers are bounced back and forth between countries until one nally accepts to host them. Lengthy procedures are also problematic because before their asylum application is examined and a status granted

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

357

or refused, asylum seekers are left in a legal limbo, sometimes without subsidies. This situation creates difculties for asylum seekers and for the host State. Second, in comparing States goal (to provide as little protection as possible) with the instruments used to achieve it (restrictive procedures), one discovers that the result is sub-optimal. Although the array of restrictive measures may have slowed the inow of asylum seekers, it has certainly failed to stop it. Moreover, in attempting to limit the number of asylum seekers arriving and remaining on their territory, these States have actually damaged their own interests (UNHCR, 1997). The costs of bureaucracy have proven considerable and States, by putting the emphasis on migration control and border protection, have used a very high quantity of human resources (customs, police, civil servants in charge of asylum application examination). The systems developed to reduce costs have in fact generated new costs. In addition, during the competitive process, States have frequently amended their legislation, sometimes every year. We must underline the costs of modifying legislation (costs of legislative procedure, time spent by Members of Parliament and civil servants involved, costs of implementation of the new law, etc.). A third, and problematic, aspect of regulatory competition is the (non) compliance of certain legislation with international obligations. All Member States are bound by the provisions of the Geneva Convention, the 1967 New York Protocol and the European Convention on Human Rights (see Bouteiller-Paquet, 2001). In order to redirect asylum seekers towards other countries, States have enacted measures which are not arguably in full conformity with international norms (Crpeau, 1997). These potential encroachments upon the spirit of the international provisions are costly for States international reputations because part of their international image and benets gained from hosting refugees is to be able to tout themselves as protectors of human rights. In addition, when low-income countries close their borders to refugees, they tend to justify their actions by referring to the precedents set by the more afuent States. Last, the settings of containment measures can be challenged before national courts and important case-law indicates that States face internal criticism. Finally, the procedures set out by States have created side effects. The restrictive asylum practices introduced by many of the industrialised States have converted what was a relatively visible and quantiable ow of asylum seekers into a covert movement of irregular migrants that is even more difcult for States to count and control (UNHCR, 1997). There is also widespread agreement that in many cases, keeping asylum seekers out has had the side effect of allowing for the development of networks of migrant smugglers (see Ghosh, 1998; Salt & Hogarth, 2000). As a result, States are obliged to reinforce procedures and to allocate more human resources in order to ght against smuggling networks, and abuses of their asylum systems, which constitute signicant, indirect, bureaucratic costs.

5. European cooperation and efciency at solving the race to the bottom problem The game is not static and institutional solutions generally emerge to solve the collective action problem. More precisely when the interaction among the users of a common pool resource is indenitely repeated, the outcome does not have to be a tragedy but might be optimal as demonstrated by the folk theorem (Gibbons, 1992). When transaction costs are

358

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

low, such self-emerging institutions imply community building, information gathering or sanctioning devices. In practice, different forms of cooperation have emerged in the eld of asylum legislation but the many obstacles that are met can explain the recently increasing role of action at the European level. We propose to rst explain the development and result of the spontaneous cooperation of the early 1990s. Then we evaluate the post-Amsterdam evolution and the progressive establishment of a centralised cooperation. 5.1. Failed cooperation in the shadow of competition Numerous norms enacted in the early 1990s show evidence of a spontaneous and decentralised cooperation among the EU Member States. They reect a common will to achieve two entangled goals: to limit the asylum shopping phenomenon that generates regulatory competition and to substitute burden sharing mechanisms (i.e. collective action) for unilateral strategic behaviours. The Schengen Implementation Convention (SIC) and the 1990 Dublin Convention that determined the state responsible for examining the asylum applications lodged in one of the Member States of the European Community were the rst moves towards limiting the competitive process. The goal of the SIC and Dublin Convention was to establish mechanisms which would replace the asylum seekers opportunity to choose his/her destination country with a legal, State-oriented determination. To this aim, the Dublin Convention laid down criteria that determined which contracting party would be held responsible for processing the asylum application. By restricting a refugees ability to choose a destination State, States aimed at suppressing asylum shopping and thus the fuel for competition among legal norms. During the same period, Member States started negotiating and discussing burden sharing mechanisms (see Barbou des Places, 2002; Noll, 2000). Since 1992, they have adopted various measures, pushed by active and overburdened States (mainly Germany and Sweden) that try to establish procedures or mechanisms favouring an equitable distribution of asylum seekers. It is clear that the burden-sharing objective can be understood as a strategy supported by over-burdened States in order to rid themselves of an excessively large number of protection seekers, but it can also be viewed, instead, as a means to maintain and restore the States admission capacity in the long term (Hailbronner, 2000; Harvey, 1998). Burden sharing plans, particularly in the eld of asylum, were praised because they are an insurance mechanism (Suhrke, 1998) that can help States increase or maintain the protection granted to asylum seekers. The evaluation of these cooperation schemes is not easy. Some scholars stress the fact that Member States have used common norms to legitimise the shift towards restrictive asylum legislation. Referring to the two-level game metaphor, Vink (2001) concludes that governments have pursued restrictive policies that are perhaps not very different from those that they would have pursued in the absence of European cooperation. Although they could avoid accusation more easily by strategically proting from the European playing eld, especially by shifting the blame onto Europe for the tightening of their asylum policies. By acting at an international level, States have avoided certain national judicial constraints and have got around opposition from parliamentarians and NGOs supporting migrants rights (Guiraudon, 2000). Thus, in terms of costs and benets (mainly the costs induced by the

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

359

loss of a good reputation and tensions within constituencies), collective action can appear to be slightly positive. Yet, our analysis is that cooperation failed because the expected results were not met. The Dublin Convention has not suppressed asylum shopping and faces erce criticism: lengthy procedures for determining which State is responsible for processing the asylum claim, unclear criteria, and long periods of uncertainty, all appear to be problems endemic to the system. From the point of view of costs, the Dublin system of responsibility allocation and transfer of asylum seekers was costly because it has increased human and technical resources requirements. Moreover, indirect costs of clandestinity have increased since asylum seekers prefer to disappear rather than be transferred from one country to another. Finally the Dublin system has not curtailed asylum shopping. Concerning the burden sharing aim, the results are not positive either. States never agreed on a common binding rule concerning uniform admission, return, or rights granted to displaced persons. While the early 1990s proposals were ambitious and suggested systems of people sharing, i.e. mandatory distribution of asylum seekers among Member States, the measures that were nally adopted have abandoned the system of allocation of people. They establish a very light and non-binding system. Logically, the unevenness in the burden borne by Member States did not diminish during the 1990s. In fact, the Dublin mechanism does not provide Member States with any more protection against inequitable distribution. Hailbronner (2000) rightly concludes that the Schengen and Dublin Conventions have established a special type of burden shifting rather than burden sharing (see Lavenex, 1999). After more than a decade of cooperative attempts, it is impossible to attest that a shift from costly and inefcient asylum policies to a coordinated and satisfactory collective action has occurred. This result has several explanations. First, there is a problem of collective action insofar as cooperation benets are endangered by the remainder of competition opportunities. Indeed, Member States face a paradox. On the one hand, they have incentives to organise burden sharing schemes at the European level because it ows from their belonging to the EU (and progressive loss of competence with regard to border controls) that they progressively lose possibilities to insulate themselves against inows of asylum seekers entering their territory. At the same time, they are losing capacity to compete because each time a common norm is enacted another part of asylum law is pacied and removed from the ambit of interstate competition (Noll, 2000). In other words, unilateral actions and competing strategies become more and more costly as they have an impact on other elds of European integration (Barbou des Places, 2002). On the other hand, Member States have not transferred all competence to the EC. To a certain extent, they retain the capacity to insulate themselves and/or to successfully compete. In the absence of a harmonisation process, they may compete with their neighbours by maintaining different degrees of legal protection. Countries that receive few asylum seekersgenerally countries offering the lowest level of protectionmight be dissuaded from participating in cooperation if they believe that unilateral action is less costly than mandatory burden sharing schemes. Secondly, States have disagreed about how to deal with the collective action problem. The intersection of collective action problems and problems of heterogeneous interests is difcult because both call for solution strategies that are partly contradictory. In order to solve a problem of collective action, it is useful to extend the range of cooperators so that fewer and fewer actors can potentially free ride. In contrast, in order to solve problems of

360

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

interest heterogeneousness, it is useful to limit the range of cooperators. Heterogeneousness can be contained through the exclusion of dissenters. In fact, the best combination is achieved by a viable balance between inclusiveness and exclusiveness. Cooperation is relatively difcult to realise if the minimum sized coalition is fairly large and if cooperation makes non-cooperation more attractive to outsiders (Schelling, 1978). If all States prefer all other States to cooperate but each State individually prefers to defect, non-cooperation is the dominant strategy and the collective equilibrium outcome conforms to the prediction of the regulatory competition game. Unsurprisingly, as spontaneous and decentralised cooperation proved to be inefcient, States have favoured centralised action to organise cooperation. The Treaty of Amsterdam 1997, demonstrates the Member States intention to establish real and sustainable cooperation. 5.2. Towards centralised action in the post-Amsterdam era In the Treaty of Amsterdam, States have transferred important competence to the European Union level on several matters concerning asylum and refugee law. Since 1999, the European institutions have a very important legislative activity and they enact numerous measures in the eld of asylum. Although the process of setting the European asylum policy is still very young and in progress, we interpret the current legal developments as clues of a new tendency, whereby Member States try to set out a soft but institutionalised and centralised cooperation aiming at replacing the previous regulatory competition. One of the lessons taught by the 1990s cooperative schemes is that cooperation can be efcient only if it tries simultaneously to tackle the causes of regulatory competition and to organise centralised burden sharing. On the rst point, an evolution is already perceptible. The European institutions have enacted or proposed important norms (directives) that all lay down minimum standards. They deal with the reception of the applicants for asylum in the Member States, with the asylum procedures, with the qualication and status of refugees and with temporary protection granted to displaced persons. These norms all purport to approximate the national legislation with the explicit aim to suppress or reduce the asylum shopping strategy. Harmonisation corresponds thus to a rst step of cooperation, that Noll calls norm sharing (Noll, 2000). In addition, by accepting harmonisation and laying down minimum standards, Member States agree on a limitation to their competitive and strategic de-regulation capacities. By approximating their legislation and recognising equivalent rights to asylum seekers, they pacify the harmonised areas from the ambit of regulatory competition. The European cooperation differs from the 1990s system for a second reason. It aims at establishing centralised burden sharing. In the frame of the Freedom, Security and Justice Area developments, the Council has received competence to enact measures promoting a balance of efforts between Member States in receiving refugees and displaced persons. Since 2000, two important norms have been enacted. The rst one is the creation of a European Refugee Fund in September 20008 and the second if the Directive of 20 July 2001 on minimum standards for giving temporary protection in the event of mass inux of displaced
8

Decision 2000/56 of September 18, 2000, OJ 6/10/2000.

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

361

persons and on minimum measures promoting a balance of efforts between Members States in receiving such persons.9 The European Refugee Fund Decision favours a scal burden sharing system by creating mechanisms of nancial compensation to burdened States. The Fund will nance actions relating to conditions for reception, integration of the persons in the host society and repatriation. Progressively, the subsidies will be allocated to the State bearing the major costs of hosting protection seekers. Despite its limited budget, the Fund represents a substantive change insofar as States have agreed to institutional nancial solidarity. As for the Temporary Protection Directive, it is an attempt to set out a people sharing mechanism under the Councils supervision. Both norms represent a signicant attempt to deal collectively and comprehensively with the asylum dilemma by acting on the causes of regulatory competition (minimum standards), and by promoting complementary measures of centralised burden sharing, both for people sharing and for scal burden sharing. A third difference with the 1990s cooperation is the role of European institutions. In particular, the Commission plays a central role, as it proposes and promotes the development of new European legal instruments (harmonisation directives in particular). In addition, it plays a signicant role in comparing national legal asylum systems, identifying and promoting the adoption at the European level of what it considers to be the best practices with regards to providing protection to asylum seekers and refugees. Clearly, the current setting of a European asylum policy has important limitations. Harmonisation is limited to the adoption of minimum standards leaving space for competition among States. In addition, while the Treaty calls for the adoption of burden sharing instruments, there is no time limit for the adoption of these measures, which is another aw in the system. Yet, despite the limitations, the current legal developments demonstrate the States will to cooperate. Regulatory competition is unlikely to disappear but it might be progressively framed by the action of the European institutions and could be absorbed by a political competition for best norms inside the European political debate. In particular, we must note the enactment by the Commission of a communication that promotes the use of the Open Method of Coordination (OMC) in the eld of asylum.10 This new governance method is interesting insofar as it may complement the legislative works. If the OMC is put in practice, States will have to agree on common objectives and guidelines, establish quantitative and qualitative indicators and benchmarks. Then European guidelines will have to be translated into national policies and States will have to organise periodic monitoring, evaluation and peer review. The Commission will play an important role in the OMC mechanism as it will be able to stimulate open and public debate on the national asylum policies, select and promote best practices and oblige States to justify their individual actions. The Commission will be able to make extensive use of the shame leverage in order to dissuade States from implementing externalising measures that shift the asylum burden onto the other Member States. We can thus observe that, progressively, a European asylum policy is taking shape, that may substitute organised, institutionalised and centralised cooperation, for negative regulatory competition.
Directive 2001/55, OJ 7/8/2001. Communication from the Commission to the Council and the European Parliament on the common asylum policy, introducing an open coordination method, COM (2001) 710 nal.
10 9

362

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

6. Conclusion The law and economics perspective adopted in this paper permits an understanding of the very intense legislative activity in the eld of asylum in Europe from the mid-1980s onwards, and the generalised shift from generous national asylum policies to restrictive and deterrent asylum legislation. It also explains the difculties met by the EU Member States in setting out efcient cooperative schemes in the 1990s. The paper nally raises the issue of the efciency of the European asylum policy as it is currently set out. To analyse these questions, we have focused on the nature of the good and departed from traditional analysis that characterises the provision of protection to asylum seekers and refugees as an international public good. We rst showed that the provision of protection after the Second World War presented the characteristics of a joint product. As States expected private i.e. nation specic benets (both security and altruistic benets), the provision of protection was both spontaneous and heterogeneous in Europe. Then, by the mid-1980s, as the costs of asylum seekers reception dramatically increased, the nature of the good changed and States started competing for a common pool resource that was: to provide as little protection as possible. We theoretically explained and gave empirical evidence of the emergence of a competitive regulatory game that took the form of a spiral of restrictions in the EU Member States asylum legislation. Positive analysis conrmed that the result is a race to the bottom. The issue now at stake is the organisation of cooperation among the EU Member States. The risk of a race to the bottom is generally seen as a reason which may justify centralised decision making, including measures of harmonisation. So, can we expect that the current establishment of the European asylum policy will lead to an efcient result and is it likely to solve the destructive competition problem? We showed that the spontaneous cooperation of the 1990s failed to limit negative regulatory competition. That is why the EU Member States have chosen a rst option to solve the problem: centralised cooperation. They transferred important competence to the European institutions in the Amsterdam treaty and since that time, European institutions have set out various mechanisms that aim to limit and frame competition. However, another option to deal with the race to the bottom problem merits further exploration. After having considered the paper, it may be interesting to focus on the nature of the good. While current centralised cooperation organised by States aims to control the result of the game, a normative proposal argues instead in favour of changing the nature of the game. The idea would be to identify conditions that would allow for regulatory competition to produce a positive result. We build upon the American debate on Delaware and regulatory competition in the eld of corporate law. Pro-competition scholars such as Romano (1985, 1987, 1993) show that competition does not necessarily lead to a race to the bottom and can instead produce a climb to the top. As for the EU, it seems interesting to take into consideration Reichs view (1992) that a competition for better rules is possible, i.e. competition that does not fall below a certain level. In the eld of asylum, such a change is difcult to operate, but we deem that reection is worthwhile, that would evaluate how the perception of the good by the actors could be modied in order to change the game.

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

363

Acknowledgements Valuable comments by two anonymous referees are gratefully acknowledged.

References
Acharya, A., & Dewitt, D. (1997). Fiscal burden-sharing. In C. Hathaway (Ed.), Reconceiving international refugee law (p. 111). Nijhoff. Achermann, A., & Gattiker, M. (1995). Safe third countries: European developments. International Journal of Refugee Law, 7, 19. Barbou des Places, S. (2002). Burden sharing in the eld of asylum. Legal motivations and implications of a regional approach (pp. 134). Working Paper of the European University Institute, Florence, RSC No. 2002/66. Barbou des Places, S. (2003). Evolution of asylum legislation in the eld of asylum. Insights from regulatory competition theory (pp. 137). Working Paper of the European University Institute, Florence, RSC No. 2003/16. Betts, A. (2002). Public Goods Theory and Refugee Provision: The Role of the Joint-Product Model in Burden-Sharing Theory, Paper presented at the UACES workshop on Internal and External Dimensions of EU Burden-Sharing, London School of Economics, 27 April 2002, on le with authors. Bocker, A., & Havinga, T. (1998). Asylum migration to the European Union: Patterns of origin and destination. Nijmegen: Institute for the Sociology of Law. Published by the Ofce for Ofcial Publication of the European Communities. Bouteiller-Paquet, D. (2001). LEurope et le droit dasile. LHarmattan. Crpeau, F. (1995). Droit dasile, de lhospitalit aux contr les migratoires. Bruylant: Collection de Droit o International. Cruz, A. (1995). Shifting responsibility, carriers liability in the Member States of the European Union and North America. Trentham Book. Deffains, B. (2001). Competition between legal systems, a comparative law and economics perspective. In Deffains, B., & Kirat, T. (Eds.), Law and economics in civil law countries. JAI Press. ECRE (European Council on Refuges and Exiles). (1995). Safe third countries: Myths and realities. London, available at www.ecre.org. Ghosh, B. (1998). Huddled masses and uncertain shores: Insights into irregular migration. The Hague: International Organization for Migration, Martinus Nijhoff. Gibbons, R. (1992). Game theory for applied economists. Princeton, NJ: Princeton University Press. Guiraudon, V. (2000). European integration and migration policy: Vertical policy-making as venue shopping. Journal of Common Market Studies, 38, 251271. Hailbronner, K. (2000). Immigration and asylum law and policy of the European Union. Dordrecht: Kluwer. Harvey, C. J. (1998). The European regulation of asylum: Constructing a model of regional solidarity. European Public Law, 4, 561. Jandl, M. (1995). Structure and Costs of the Asylum Systems in Seven European Countries, International Centre for Migration Policy Development, September 1995. Jeannin, L., Meneghini, M., Pauti, C., & Poupet, R. (1999). Le droit dasile en Europe. Etude compare. LHarmattan. Joly, D. (1999). A new asylum regime in Europe. In F. Nicholson & P. Twomey (Eds.), Refugee, rights and realities. Evolving international concepts and regimes (pp. 336356). Cambridge: Cambridge University Press. Klliker, A. (2003, June). Competing for international economic commons: Towards a collective goods theory of regulatory competition. Paper presented at the Robert Schuman Centre and Journal of Public Policy Workshop, Europeanisation, Global Capital, and Regulatory Competition, Florence, on le with the author. Landgren, K. (1999). Deecting international protection by treaty: Bilateral and multilateral accords on extradition, re-admission and the inadmissibility of asylum request. UNHCRs New Issues in Refugee research. Working Paper No. 10, available at www.unhcr.ch. Lavenex, S. (1999). Safe third countries, extending the EU asylum and immigration policies to central and eastern Europe. Central and European University Press.

364

S.B. des Places, B. Deffains / International Review of Law and Economics 23 (2004) 345364

Marceau, N. (1997). Competition in crime deterrence. Canadian Journal of Economics, 30, 844854. Milner, J. (2000). Sharing the security burden: Towards the convergence of refugee protection and state security. QEH Working Paper. Noll, G. (2000). Negotiating asylum. The EU acquis, extraterritorial protection and the common market of deection. Nijhoff. Olson, M., & Zeckhauser, R. (1966). An economic theory of alliances. Review of Economics and Statistics, 48, 266279. Reich, N. (1992). Competition between legal orders: A new paradigm of EC law? Common Market Law Review, 29, 861896. Romano, R. (1985). Law as a product: Some pieces of the incorporation puzzle. Journal of Law, Economics, and Organization, 1, 225283. Romano, R. (1987). The state competition debate in corporate law. Cardozo Law Review, 8, 709. Romano, R. (1993). The genius of American corporate law. AEI Press. Rotte, R., Vogler, M., & Zimmermann, K. F. (1996). Asylum migration and policy coordination in Europe (pp. 131). Discussion Papers, Munchener Wirtschaftswissenschaftliche Beitrage, 96-11. Salt, J., & Hogarth, J. (2000). Migrant trafcking and human smuggling in Europe: A review of the evidence with case-studies from Hungary, Poland and Ukraine. Geneva: International Organization for Migration. Sandler, T., & Hartley, K. (2001). Economics of alliances: The lessons for collective action. Journal of Economic Literature, 39, 869896. Schelling, T. (1978). Micromotives and macrobehavior. New York: Norton. Sudgen, R. (1982). On the economics of philanthropy. Economic Journal, 92, 341350. Suhrke, A. (1998). Burden-sharing during refugee emergencies: The logic of collective versus national action. Journal of Refugee Studies, 11, 396. Thielemann, E. (2002, January). Between interests and norms: Explaining burden-sharing in the European Union. Paper presented for the UACES Workshop on European Burden-Sharing and Forced Migration, London School of Economics, on le with the author. UNHCR. (1997). The state of the worlds refugees, 199798: A humanitarian agenda. Oxford: Oxford University Press. UNHCR. (1999). Refugees and others of concerns to UNHCR: Statistical overview. United Nations High Commissioner for Refugees, Geneva, July 2000, available at www.unhcr.ch. Velling, J., & Woydt, M. (1993). Die Migrationspolitiken in ausgewhlten Industrielndern-ein synoptischer Vergleich, Mannheim: ZEW-Dokumentation. Vink, M. (2001, November). The limited Europeanization of domestic asylum policy: EU governments and two-level games. Paper presented at the rst YEN Research Meeting on Europeanisation, Workshop IV Europeanisation of Domestic Policies, Siena, on le with the authors.