Does Policy Harmonisation Work?


The EU's Role in Regulating Migration Flows





Eiko R. Thielemann

Department of Government/European Institute

London School of Economics






Keywords: public policy effectiveness, asylum, migration pull factors, deterrence, burden sharing, collective action problems, international co-operation, European Union


Public policy making on asylum takes place in an environment of intense public scrutiny, strong institutional constraints and international collective action problems.  By assessing the relative importance of key pull factors of international migration, this article explains why, even when controlling for their differences in size, some states receive a much larger number of asylum seekers than others.  The analysis of 20 OECD countries for the period 1985-1999 further shows that some of the most high profile public policy measures—safe third country provisions, dispersal and voucher schemes—aimed, at least in part, at deterring unwanted migration and at addressing the highly unequal distribution of asylum burdens have often been ineffective.  This is because the key determinants of an asylum seeker’s choice of host country are historical, economic and reputational factors that largely lie beyond the reach of asylum policy makers.  Finally, the paper argues that the effectiveness of unilateral policy measures will be further undermined by multilateral attempts to harmonise restrictive policies and that current efforts such as those by the European Union will consolidate, rather than effectively address, existing disparities in the distribution of asylum burdens.








Paper prepared for presentation at the 8th European Union Studies Association (EUSA) Biennial International Conference, March 27-29, 2003, Nashville, Tennessee



1. Introduction[1]


In an increasingly interdependent world, rising numbers of asylum seekers and their highly unequal distribution across countries have meant that forced migration is now regarded as one of the key challenges facing nation states today.[2]  This challenge is made even greater by the fact that one state’s policy decisions on the relative leniency or restrictiveness of its asylum regime will create externalities for other states and can thus lead to strained relations between states.[3]  As a consequence, forced migration has also come to be seen as a crucial challenge for international policy coordination, leading, for example, to rapid advances in the efforts of the European Union to provide for solutions in this area. 


Policy makers charged with finding an appropriate response to these challenges have been faced with two key questions: First, why have some states been faced with a much higher number of asylum applications than other states.  And second, what public policy measures can effectively influence the number of asylum seekers that a state receives?  From a national perspective, the most frequent response to the first question has been to argue that if states’ asylum burden is disproportionate, then these countries’ asylum procedures are probably too lenient and their welfare provisions too generous in international comparison.  By increasing the restrictiveness of their asylum policy, states will be able to redress the inequitable distribution of burdens, raising concerns in some quarters about a possible race to the bottom of protection standards.  However, there has so far been no academic attempt to use relevant theoretical models developed in the field of economic migration (Ranis and Fei, 1969; Harris and Todaro 1970; Borjas 1990; Massey et al. 1993) to systematically analyse pattern of asylum flows in order to establish the importance of policy and other historical, economic or political migration pull factors that can explain why asylum seekers apply in a particular country.


Regarding the second question, on the capacity of public policy in this area, there is still little consensus as to whether liberal states can control unwanted migration (Freeman 1994).  The ‘transnationalist’ strand of the literature (Sassen 1996; Jacobson 1995; Soysal 1994) emphasises systemic constraints that undermine the capacity of states to assert effective control in this area.  In contrast, the more ‘state-centrist’ strand of the literature (Holzer, Schneider and Widmer 2000; Guiraudon and Lahav 2000; Joppke 1997; 1998) argues that states have found new ways to regulate migration in an era of increasing interdependence, which enables them to retain much of their regulatory capacity in this area, even to the extent that their measures have undermined some of the more liberal aspects of the international migration regime. 


Largely missing from the literature have been quantitative studies that systematically analyse empirical evidence across time and space[4] and which might offer more conclusive answers about the determinants of asylum seekers’ choice of destination country and the effectiveness of public policy in regulating asylum flows.  In an attempt to fill this gap, this paper will proceed as follows:  After a short overview of recent public policy responses, an analysis of the theoretical literature on migration will identify theoretically informed pull factors.  From this the paper will generate a number of hypotheses.  The next part develops and explains the model which is subsequently used to test the hypotheses empirically based on data for 20 OECD countries over the period 1986-1999.  The final section discusses the empirical results which call into question some widely held assumptions about the underlying reasons for the unequal distribution of 'asylum burdens' and the effectiveness of unilateral and multilateral deterrence measures.



2. Setting the Scene: Forced Migration and Public Policy


Since the mid-1980s, the issue of immigration and asylum has gained considerable prominence in OECD countries.  The combination of heightened migration pressure and reduced willingness to accept inward migration, has pushed the issue towards the top of the political agenda.  As economic and political uncertainties increased in the 1990s, public opinion (often encouraged by electioneering politicians and a xenophobic media)[5] became more and more wary about inward migration, which in turn produced more pressure on politicians for "decisive" action in this area.  The important distinction between economic and forced migration often threatened to be lost in the process.  Although states are generally free to decide on the number of economic migrants they are willing to accept, in the area of forced migration international obligations such as the Geneva Convention on Refugees impose important obligations on states. However, this is not to dispute the fact that, with the door to legal immigration shut in most states since the early 1970s, a significant number of economic migrants have taken the 'asylum route' as it often has constituted the only remaining avenue for third country nationals to legally settle in one of the OECD countries.  In the 1990s, asylum applications therefore became a primary concern for policy makers in all OECD states.  Figure 1 shows that the number of asylum applications filed in the developed world increased significantly in the late 1980s and early 1990.


However, policy-makers are not just concerned about the growth in the absolute numbers of asylum applications, over which they have only limited influence given a volatile international system and their international obligations.  They are also concerned about the relative distribution of asylum applications among states, in particular when they feel that the policy measures adopted in neighbouring states are at least in part responsible for its won asylum burden.  When analysing the development of asylum applications across OECD countries, it soon becomes clear that the distribution of asylum applications has been highly unbalanced.  Public attention was drawn to this in when in 1992 Germany received over 438.000 asylum applications, which constituted 66 percent of all applications registered in the territory which now makes up European Union.[6] 


However, a focus on absolute figures might well be misleading.  When using the more meaningful measure of relative burdens, i.e. one which takes account of differences in reception capacity,[7] the unevenness in distribution becomes even clearer (see Table 1).  It can be shown that since the mid 1980s some European countries, most notably Switzerland Sweden and Germany, have borne a much higher relative (per capita) burden than the EU average.  This inequitable balance of burdens has constituted a considerable domestic challenge in some countries.  It also led to tensions between some OECD countries, particularly within the EU, as there was a feeling in some quarters that certain countries were introducing unilateral deterrence measures to deflect asylum applications towards other countries (Ref.??). 



Table 1: Average Number of Asylum Applications per Year, 1985-99

               (per thousand of population)
























Czech Rep.




























United Kingdom








In the light of the collective action problems involved in this area, the German government in 1992 proposed a European wide asylum burden-sharing system.  The German proposal[8] foresaw the distribution of asylum seekers across Europe according to indicative figures that were based on a distribution key composed of three criteria which were given equal weight (population size, size of Member State territory and GDP).[9]  The centrepiece of the German draft was the introduction of a compulsory resettlement mechanism which would have worked as follows: ‘Where the numbers admitted by a Member State exceeds its indicative figure [...], other Member States which have not yet reached their indicative figure [...] will accept persons from the first State’.[10]  This proposal, however, did not find the necessary support among other countries, with the UK in particular being strongly opposed to such a scheme.[11] 


Since then, European states have instead agreed on a number of (limited) steps towards harmonising asylum policy regulations across Europe, including joint visa regimes and a common refugee definition.[12]  At the same time, states have continued to undertake unilateral measures aimed at deterring asylum applications to their country.  Such unilateral deterrence measures have covered the whole range of policy options available to policy makers in this field such as measures on access, status determination and those concerning the integration of asylum seekers.  These measures have been based on a number of apparently widely held assumptions.  First, asylum seekers are seen to be well informed (either through personal networks or their traffickers) about the relative openness and attractiveness of different destination countries' asylum regime.  Second, they are expected to choose to apply to those countries which have the most attractive asylum policy package, in terms of access, determination and integration/welfare measures.  In other words, 'asylum shopping' has been regarded as a widespread phenomenon.  Finally, there appears to be a belief, that countries with relatively more attractive asylum policy will come to be regarded as a 'soft touch' and will consequently have to cope with a disproportionately high number of asylum applications.  Despite the fact that these assumptions are at least questionable, they have formed the basis for the introduction of unilateral deterrence measures which have come to be viewed by both policy-makers (and many academics) as having been highly effective.  For example, the dramatic reduction in the number of applications received by Germany between 1992 and 1994 has widely been attributed to the 1993 restrictions that were introduced to the German Basic Law and the legislation pertaining to foreigners.  The adoption of so called 'safe third country provisions' in particular has be regarded as having been highly effective as it enabled German border guards to refuse certain categories of asylum seekers entry to German territory.[13]  The 71 percent drop in asylum applications in Germany between 1992 and 1994 has often been attributed to the restrictive changes introduced to German asylum policy in 1993.[14]


Some more recent single-country research results (Holzer, Schneider and Widmer 2000; Robinson and Segrott 2002), however, have cautioned against being overly confident about the effectiveness of asylum policy in steering migration flows.  The quantitative analysis of the Swiss case showed that only within limits Switzerland was able to influence the inflow of asylum seekers between 1986 and 1995.  The study showed that the Swiss government was partly successful in manipulating the relative number of refugees it recognized to achieve its deterrence objectives.  However, the study concluded that deterrence measures can be expected to be unsuccessful 'if the push factors in a region nearby to the receiving states reach a critical level' (Holzer, Schneider and Widmer 2000: 1205).  Research conducted in the UK draws even more sceptical conclusions.  The research that was based on 65 interviews with asylum seekers found that most of the respondents knew very little about UK asylum policy before their arrival.  The study found that they certainly did not have sufficient knowledge to make an informed choice based on rational evaluation of reception conditions and welfare benefits on offer by several possible destination countries (Robinson and Segrott 2002: 46; 63). 


These studies, as other research on patterns of asylum seeking in Europe, suffer from their focus on individual countries which makes generalisation difficult.  Moreover, qualitative research based on survey data, whereby asylum seekers are asked about their travel route, their preferences as to particular countries of destinations and why they applied in a particular country can be problematic.  Although, qualitative analyses based on large-scales sample have a strong appeal (who else than the asylum seeker knows why s/he applied in a particular country), research of this kind suffers from two difficulties in particular.  First, survey analyses of the required kind is very costly especially when one is interested in systematic comparative analyses across countries (let alone over time).  Interviews are especially costly because they often have to rely on interpreters.  Second, and potentially even more problematic is the fact that asylum seekers might have a strong incentive to emphasise certain determinants over others, as they know that certain answers might compromise their asylum application or residence status.[15]  It is for these reasons that the systematic quantitative analysis in this paper, despite its own limitations, will make a contribution to the existing literature.[16]



3. Which Host Country?  Migration Theory and Pull Factors


Although still an under-theorised area of study, one can identify a number of prominent theories and models of international migration (Kritz et al. 1992, Massey et al. 1993, Meyers 2000).  One of the most commonly known theoretical migration models is the so-called push-pull model.  In its most abstract from, the push-pull model suggests that there are push factors in countries of origin that cause people to leave their country, and positive or pull factors that attract migrants to a receiving country.  Although this model cannot simply be transferred to the area of forced migration,[17] it still offers a number of insights for research on the direction of asylum flows.  In a similar vein, and of specific interest to this paper, the idea of pull factors is used in the area of forced migration, to explain the patterns of asylum applications across different possible destination countries.  The analysis of such pull factors will help us explain the direction of forced migration flows, an issue that has so far been a widely neglected aspect of migration research.  A review of the theoretical literature on migration,[18] produces five categories of pull-factors—economic, historic, political, geographic and policy related—which will be introduced below. 



Economic Factors

Economic theories of migration have only limited applicability in the area of forced migration, in which displaced persons often will have little or no time for deliberations of 'utility maximisation'.  However, despite the important substantive differences between economic and forced migration, economic considerations can still be expected to play a role in the area of forced migration.  Like other migrants, asylum seekers will often face financial and other constraints which will influence their choice (limited as it might be).  In many cases even forced migrants will have some choice as to their country of destination and can therefore be expected to take economic consideration into account.  Moreover, we do of course know (Kunz 1981; Zolberg et al. 1995) that in a world of high cross-country income differentials and highly restrictive admission policies, persecution is not the only push factor behind the large number of asylum applications of recent years.  Given the above, economic factors must be included when analysing the incidence of asylum applications across the OECD. 


Neo-classical economic migration theory (Ranis and Fei, 1961; Harris and Todaro 1970, Todaro 1976) explains the decision to migrate as one of income maximisation in which wealth differentials and differences in employment opportunities constitute important pull factors.  International migration is expected to be determined by geographic differences in the supply and demand of labour.  Ultimately, in this view, it is wage differentials which can explain movements from low-wage countries to high-wage countries.  In it micro-economic extension (Sjaastad, 1962; Borjas 1990), rational actors (be it individuals or larger units such as families or households) decide to migrate in the expectation of a positive, often monetary, net return from migration.  In this framework, the decisive factor is income differentials as well as the probability of employment in the destination country.  In other words migration decisions can be seen as being guided by processes of income maximisation and risk minimisation.



Historical Ties, Networks & Path Dependency

Historical ties between countries of origin and destination countries often lead to transport, trade and communication links which tend to facilitate movements of people from one country to the other.  Material links are often accompanied by ideological or cultural links.  Colonial legacies often explain why administrative and educational systems in third world countries mirror those of a past colonial power and often continue to be reflected in migration flows long after independence (Fassmann and Muenz 1992).  For example, Pakistanis and Bangladeshis learn English, are raised in a British-style education system and keep up links with the UK through the Commonwealth.  Language ties, communication links and cultural networks that are responsible for the diffusion of particular consumption patterns, can be responsible for channeling international migration to particular destination countries (Massey et al., 1993: 446-7). 


Moreover, the fact that migrant or refugee communities have been established in certain destination countries as a result of historic ties, will often lead to the growth of migrant networks that may foster future migration flows.  Such networks are sets of interpersonal ties between earlier migrants already resident in a destination country and potential migrants in countries of origin that are based on family ties, friendships or shared community origin.  Such ties can significantly reduce the costs and risks of migration, thus channelling migration flows in the direction of earlier migration flows.  By passing on information about access to a particular country and its employment opportunities, they constitute a form of social capital (Hugo, 1981; Taylor, 1986).  Once migration connections have been established, the presence of relatives, friends, and/or others from the same community of origin may form a strong incentive to choose a particular destination.  Migration may thus be seen as a self-sustaining diffusion process (Massey et al. 1993), which governments will find difficult to control.


Following the same line of reasoning we could also expect a certain degree of path-dependency from one period of migration to the next.  Such a process is likely to be the result of two dynamics.  First, there will be a reduction of costs and risks for migrants as they can rely on the support of personal networks.  Second, there will be certain persistence of existing migration routes and patterns, as agents and traffickers will have incurred sunk costs by investing in the creation of networks which they will be reluctant to give up (Pierson 2000).



Political Values ('Liberalness')

Concerns about personal security and their acceptance into a new host society will be important considerations for potential migrants, in particular forced migrants who are leaving their country of origin because of concerns about their, or their family's, safety.  The reputation of a country in terms of its 'liberal credentials' as well as its track record on issues such as adherence to human rights standards, international humanitarian crises, community relations, reception and integration of foreigners, naturalisation policies and the like, can be expected to play a role in a migrant's consideration about the relative attractiveness of countries of destination. 



Ease of access, in particular geographic proximity, between a country of origin and a country of destination can also be expected to be a pull factor in migration patterns.  Despite technological developments which have made geographic distance less of an issue than it was in the past, most migrants' resources are limited and smaller distances will often mean lower costs of transport and hence easier access.  In other words, geographic distance can often be regarded as a proxy for the costs of movement.  Although other factors, such as length and relative openness of countries' territorial borders, will also play a role as to how accessible a country of destination is for migrants, geographic distance will often be the key element.  We can also expect an interaction with other pull factors already discussed.  Although geographic proximity does not guarantee the establishment of cross border ties, geographic proximity can clearly facilitate the formation of such ties.



Deterrence Policy

States often regard asylum burdens as a 'zero sum' phenomenon, in which a reduction of one country's burden will result in increasing burdens for other countries.  The assumption is that there is a certain number of migrants each year who intend to claim asylum and that the role of national asylum policy is to restrict the inflow into a particular country to an acceptable proportion.  This means that policy makers will try to use migration policy instruments to make sure that their country will not be seen as a 'soft touch', i.e. an overly attractive destination country that will attract an unacceptable proportion of asylum seekers.  Three sets of such instruments in particular are at their disposal: (1) access control, (2) the determination process and (3) migrant integration policy.  Access control policy refers to the rules and procedures governing the admission of foreign nationals and its instruments include visa policy, regulations for carriers, safe third country provisions, etc.  Rules concerning determination procedures relate to entry into a country's refugee recognition system, appeal rights, and rules concerning protection that is subsidiary to the rather narrowly defined Geneva Convention criteria for full refugee status.[19]  Finally, integration policy is concerned with rights and benefits given to asylum seekers inside a country of destination (e.g. work and housing conditions, rules on freedom of movement, welfare provisions, educational opportunities, etc.).  Policy-makers can introduce changes in the regulations in these three areas in an attempt to raise the deterrence effect of their policy, which in turn is expected to make their country less attractive to asylum seekers in relative terms.


The various above explanations of pull-factors for migrants and in particular asylum seekers, are obviously not mutually exclusive.  On their own, as well as in combination, they can be expected to help explain why asylum seekers apply for asylum more in some OECD countries than in others.  Individuals might engage in cost-benefit calculations that make them choose richer countries with more employment opportunities over poorer ones with fewer work opportunities; they might try to reduce risks and costs by using existing networks; that might prefer a more liberal country of destination over a less liberal one; they might be more likely to end up in a country that is relatively closer to their country of origin; and finally their decision might be affected by the relative asylum policy-mix of different potential countries of destination.  The purpose of this paper is not to examine necessarily competing theories but to test the relative strength of hypotheses that can be drawn from the above discussion of different possible pull factors.



4. Methodology


To do this, the paper uses time-series cross-section of aggregated data for 20 OECD countries for the period 1985 to 1999 collected from UNHCR, OECD and the U.S. Committee for Refugees.[20]  This quantitative analysis allows the testing of hypotheses regarding the existence and relative strength of the different potential pull-factors.  In doing so, the paper will control for particular country and time effects such as differences in reception capacity and fluctuations of the absolute number of asylum applications over time. 



Dependent Variable

When looking at the issue of asylum from an international 'burden-sharing' perspective, it is more interesting to focus on the number of relative asylum applications (i.e. applications per capita) across time and place, than on the absolute numbers of applications which has tended to dominate the public debate on asylum.  Relative figures are the crucial reference figures if states want to check the success of failure of their attempt to achieve a more equitable distribution of burdens as a result of international co-operation.  Also, it is also hardly surprising (nor objectionable) that in absolute terms, larger countries will tend to attract more inward migration than smaller countries.  This paper therefore seeks to explain the number of asylum-applications in each country and for each year of the data set, relative to the population of each of these countries while controlling for variations in the number of total applications and overall population growth across all the OECD countries included in the dataset.[21]  To calculate the dependent variable, I used the annual UNHCR statistics of asylum applications and OECD data on population developments.  To arrive at the observations for my dependent variable for each country and each year I divided the number of asylum applications by the country’s population size and put this figure is relation to the total number of asylum seekers divided by the total population of all countries under investigation.[22]



Explanatory Variables

The explanatory variables are constructed in such a way as to allow for the examination of the five above theories on key pull-factors for asylum applications. 


First, to test for economic pull factors, the paper analyses OECD data on annual GDP growth (in percent) and the total number of registered unemployed.  The expectation is that a country's relative burden will be positively correlated with its economic performance and negatively with its numbers of unemployed. 


Second, to test the importance of geographic factors, I determine the average distance between the capital of a destination country and the capitals of the top five countries of origin in each year.[23]  The expectation is that countries of destination which are geographically closer to the top five countries of origin in any particular year will attract relatively more asylum-seekers. 


Third, to test the role played by a country's liberal reputation, I used overseas development aid (OECD data measured in million $) as a proxy variable for a country's 'liberalness'.  I expect a more liberal country (i.e. one with relatively high ODA/GDP ration) to attract a relatively high number of asylum seekers. 


Fourth, to test network/historic ties theories, I add (at t-1) the stock of foreign population from the top five asylum countries (at time t).  The expectation is that countries with an already relatively large stock of foreign nationals from the main countries from which asylum seekers are originating in a particular year, will receive a relatively greater number of applications in relation to their size. 


Finally, to analyse the importance and the effectiveness of asylum/deterrence policy measures I use a deterrence index fluctuating between 0 (lowest deterrent effect) to 5 (highest deterrent effect).  To calculate the index, I analysed two sets of annual yearbooks, the OECD's 'Trends in International Migration' (SOPEMI) and the US Committee for Refugees' 'World Refugee Survey' for the years 1984-1999.  Each describes and analyses developments in national asylum policy measures for each country in the paper's data set. 


Five measures in particular stand out that have been widely regarded by policy makers as having the potential to significantly influence an asylum seeker's decision as to which country to apply to (UK Home Office 2002).[24]  First, in the area of access control, arguably the most important measure was the introduction of so called 'safe third country' provisions, which mean that persons seeking asylum will be refused entry into a country, if on their way to this country they travelled though another state which the first country regards as safe and where the asylum seeker could have already applied for asylum.  If an asylum seeker's travel route only transpires in the course of the determination procedure, he or she can be sent back to the 'safe third country'.  The introduction of 'safe third country provisions' across Europe meant that asylum seekers travelling 'overland' to Europe were no longer able to legitimately claim asylum in the country of their destination, as the responsibility for their case was shifted on a neighbouring country through which they had travelled.  To account for introduction safe third country provisions, I created a dummy variable which takes the value 1 for each year that safe third country provisions were applied in a country and the value 0 for all other years.


Second, with respect to a country's determination procedures, the most important potential pull factors that can be influenced by national policy-makers are the rules concerning the granting of subsidiary protection status which allows an asylum seeker to remain in a country of destination even though their application for full refugee status under the Geneva Convention was refused.  Destination countries have complete discretion in defining the requirements that protection seekers have to fulfil to be awarded such subsidiary status which means that within Europe the percentage of asylum seekers allowed to stay in a country on the basis of the award of some protection status varies from single figures to over 70 percent (UNHCR 1999).  Again, I create a dummy variable which takes the value 1 if a country of destination is below the average with regard to the percentage of asylum seekers it allows to stay in its country in a particular year and which takes the value 0 if the percentage of protection seekers allowed to stay is above the OECD average.


Finally, much of the discussion of the past few years has focused on the potential pull-effects entailed in a third category of asylum policy, namely that of integration measures for asylum seekers.  Here three policy aspects are often regarded as being crucial: first, freedom of movement vs. a compulsory dispersal policy; second, cash welfare payments vs. a system of vouchers; and third, the right to work under certain conditions vs. a general prohibition to take up employment as an asylum seeker.  The first of these concerns the right of asylum seekers to move freely within their country of destination until their asylum claim has been determined.  While a federal state such as Germany has always had central reception centres from which asylum seekers are be dispersed to the different Länder according to their relative population size, some unitary states—most notably the UK—have recently introduced similar measures.  Although dispersal measures first and foremost are an attempt to alleviate pressures from particular (particularly metropolitan) areas which are faced with a strong concentration of asylum seekers, such measures are also hoped to deter unfounded asylum claims (Ref.??).  Second, the 'cash' payment of welfare benefits cash rather than a payment 'in kind' or through a voucher system has sometimes been regarded as a pull-factor for asylum seekers.[25]  This has led a number of OECD countries to stop giving asylum seekers cash benefits and to replace cash payments by the direct provision of housing, food and health care.  In 1999, the UK and Ireland introduced a voucher system for asylum seekers, despite the fact that the two governments were advised that such a system would be more costly to administer than a cash-based system.[26]  However, governments have been attracted to vouchers due to the deterrent effect that has sometimes been ascribed to such non-cash schemes.  Finally, allowing asylum seekers to work while their claim to asylum status is being assessed has also sometimes been regarded as a potential pull factor for asylum seekers.  All countries of destination have work restrictions for asylum seekers in place.  However, a number of countries have gone further and now prohibit asylum seekers to undertake any work until their asylum claim has been accepted. 


To assess the potential deterrence effect of the above measures, three dummy variables were created which take the value 1 (for each year and country) for the existence of a dispersal scheme, a non-cash based system of benefits, and a law which prohibits asylum seekers to work until their claim has been accepted.  Adding all the dummy variables for all five of the above potential deterrence measures for each country and each year, results in a country's deterrence index for a particular year.[27]  The expectation is that the higher the index for a particular country in a particular year, the lower that country's relative attractiveness will be and hence its relative burden stemming from asylum applications.  Table 2 summarises the expected relationships between the variables discussed above.



Table 2: Expected Relationship between variables



Independent Variables

Dependent variable: Relative number of asylum applications

Economic pull factors:

            * number of registered unemployed (in t-1)

            * annual real GDP growth (in t-1)




Historical pull factors:

            *stock of foreign nationals from top five             countries of origin (in t-1)



Political pull factors

            * annual ODA payments as % of GDP (in t-1)



Geographic pull factors:

            * average geographic distance between capital of a destination country and capitals of         the top five countries of origin



Policy related pull factors:

            * deterrence index ( in t-1)









Model Estimation

ordinary least squares regression analysis with panel corrected standard errors (Beck and Katz 1995)


To estimate the relationships between these variables and relative burdens for individual countries, the paper uses pooled time-series cross-section (TSCS) ordinary least square regressions (Stimson 1985) with panel corrected standard errors (PCSEs) (Beck and Katz 1995).[28]  Prais-Winston transformations are used to eliminate serial correlation of the errors and to take account of cross-section and panel specific auto-correlation.  In running the regression of the above independent variables on the number of relative asylum seekers, I lagged GDP growth, unemployment, foreign population and the deterrence index by one year as one might reasonably expect that it was the performance of these indicators in t-1, and not current performance, that constituted a pull factor for persons applying in the period t. 



5. Discussion of Statistical Results


Figure 2 shows the distribution of relative burdens as a result of asylum applications per 1000 of population over time in the twenty countries under investigation.  In the majority of countries we can observe significant variations in relative burdens over time.  We also observe that the relative burden of Germany over time has been comparable, and at times was considerably smaller, to that of some smaller countries such as Sweden and Switzerland.  We also observe that many of the other bigger countries, in particular the UK, France, and Japan have attracted far fewer than average applications relative to their population size.  Finally, one can observe that in a number of countries, most notably in Ireland Belgium, Hungary and the UK, relative applications have increased significantly over the past few years. 


The results of the regression analysis are presented in Table 3.  We find that some historic, economic and political pull-factors have a strong impact on the distribution of asylum applications.  There is also evidence that the relative leniency or restrictiveness of a country’s asylum policy (expressed in the deterrence index) has a highly significant effect on the number of applications received.  However, it will be shown below that the impact of the different policy measures that make up that index is highly varied.


Table 3: Determinants for the Relative Number of Asylum Applications



Expected Sign



Number of Unemployed (lagged)




GDP growth (lagged)




Stock of Foreign Nationals (lagged)




 Relative ODA Payments




Average Distance




Deterrence Index




N=227; R-squared = 0.41; ***p<0.01




Economic pull factors

As expected high unemployment figures are negatively related to relative numbers of asylum applications and this relationship is highly significant.  While some displaced persons will have little or no choice in where they end up applying for asylum, as travel options might be limited or predetermined by existing trafficking routes and forced migrants might be under great time pressure to leave their country which does not give them sufficient time to weigh their options, other asylum seekers will have more time and the ability to choose where to apply for asylum.  The data analysed here suggests that economic considerations do play a role when it comes to decisions about where to apply for asylum.  When controlling for the other factors included in the model, one observes that asylum seekers apply in higher numbers in countries which offer greater employment opportunities. This will of course be true for economic migrants who use the asylum route in an attempt to circumvent the restrictive immigration regimes of developed countries.  However, given the strength of the above correlation and the fact that almost 40 percent of all asylum seekers across OECD countries are awarded some protection status, we can reasonably interpret the above results in support of our expectation that labour market considerations also play a role in the considerations of forced migrants.  Seeking physical security from persecution as well as economic opportunities in a country of destination can hardly be regarded as incompatible objectives for people forced to leave their country of origin.  In contrast to the strong effect labour market factors, general economic growth appears to have no significant effect on the distribution of asylum seekers.  Part of the reason for this might be that the OECD countries analysed here are all likely to be perceived as rich, economically thriving, industrialised countries.  For new arrivals to benefit from the economic situation of a host-country, however, employment opportunities are regarded as more important as a country’s performance with regard to short-term economic growth.


Historic factors/Networks

The existence of historical ties and established networks also comes out as highly significant.  The number of people from the main countries of origin already resident in a particular host-country is strongly and positively correlated with the relative number of asylum applicants this country receives.  This appears to support the suggestion that the existence of interpersonal ties, be it with relatives, friends or people from 'back home' already resident in a particular country of destination can act as a strong magnet for asylum seekers.  This is not surprising as it is through such networks, information about the country of origin will be passed on to potential asylum seekers who are still in the country of origin.  Despite the fact that qualitative studies suggest that the amount of information passed from relatives and friends from the destination country to potential asylum seekers before they leave is quite limited (Robinson and Segrott 2002: 41),[29] it appears that any kind of contact, no matter how fleeting will constitute a pull factor for potential asylum seekers that can tip the balance towards the decision to claim asylum in a particular country. 



As expected, favourable perceptions as to the ‘liberalness’ of a potential host-country, show a strong, and positive relationship with the relative numbers of applications that a country receives.  Countries which show a high concern for people beyond their own border and engage disproportionately in efforts to alleviate underdevelopment in the third world (through ODA payments) attract relatively more asylum seekers (Thielemann 2003b).[30] 

[second indicator, extreme right wing votes, from Budge et al]



Given the difficulties and costs involved in long-distance travel, we expected a negative relationship between the average distance of a host country from the principal countries of origin and that host-country’s asylum burden in any particular year.  However, in our dataset of 20 OECD countries, proximity as a pull factor does not produce any significant effect in the analysis.  One might expect that the selection of host-countries analysed here will have influenced this result and that had we included host countries in the developing world, the results might have been different.  However, the result might also be explained by the fact that an asylum seeker’s sense of security might increase, the further away they settle from the country in which they suffered persecution.  Moreover, although the likelihood of established migration networks to exist in neighbouring countries, there might be a lack of other pull factors, in particular economic ones in the countries which will constitute a disincentive for asylum seekers that might outweigh network factors. 


These findings mean that one has to refrain from generalising the results of single country case studies (like that by Holzer, Schneider and Widmer, 2000) which strongly emphasise the importance of geographic proximity as a pull factor in the case of Switzerland.  Clearly there are instances when geographic proximity does matter, especially when geographic pull factors interact with other pull factors such as existing historical ties (as was the case with refugees from former Yugoslavia fleeing to Germany or Switzerland in the 1990s).  However, the broader analysis across time and space reveals that geographic factors are more limited in their effect than other pull factors. 



The combined effects of deterrence measures, as shown in the deterrence index, comes out in the expected negative direction and does so at significant levels.  The effect of policy-related factors, however, is not as significant as that of historic and economic factors.  Moreover, we find that the measures analysed appear to be quite short-term in their effect.  When we lag the deterrence index by more than one year, it ceases to have any significance.  Finally, if one disaggregates the measures included in the deterrence index (see Table 4), one finds that the index’s significance is due to the strong effect of only two of the five deterrence measures analysed here: (1) not allowing asylum seekers to work until their application has been successful or until they have been allowed to stay in the host-country more permanently on the basis of a subsidiary protection status (2) granting protection status to a smaller percentage of asylum seekers (in relation to total number of applications) than other host-states.  Each of these two measures on its own is significant at the 0.05 level.  Their combined significance in the deterrence index is even stronger. 



Table 4: Impact of Individual Deterrence Measures


Deterrence Measure

Expected Sign



   Prohibition to Work




   Below Average Recognition Rate




   Safe Third Country Provisions




   No Freedom of Movement




   Non-Cash Benefit Payments




N= 227; R-squared = 0.41; **p<0.05



No significant effect in reducing asylum applications, however, could be found for the other three measures which have dominated the public policy debate on asylum in recent years.  There are: (1) measures that allow states to turn asylum seekers back at their borders and return them to so-called ‘safe third countries’; (2) measures that deny asylum seekers freedom of movement within a host country, i.e. the introduction of dispersal schemes and (3) measures which have meant the end to cash benefit payments to asylum seekers, e.g. through the introduction of a voucher scheme.  The following will provide some initial suggestions as to how to explain the  variation in effectiveness of these different measures. 



At the most general level, it seems clear that information about recognition rates and employment opportunities do reach asylum seekers either directly or indirectly through their agents and traffickers, whereas knowledge about more detailed policy measures is either not available to asylum seekers or not considered important enough to determine decisions regarding the choice of country of destination. 


The likelihood of asylum seekers receiving some kind of status that allows them to remain in their host country should they wish to do so, unsurprisingly is of the utmost importance.  We have known for some time that that host countries interpret their international obligations under the Geneva Convention in sometimes very different ways (ECRE 2000) and that recognition rates can vary greatly between host countries at any particular time, even for asylum seekers from the same country of origin (Holzer and Schneider 2002: 43).  Moreover, host countries have also dealt very differently with their discretionary granting of subsidiary protection status to those asylum-seekers who do not qualify for refugee status but who host states feel cannot or should not be sent back to their country of origin (Thielemann 2003b).  Even though asylum seekers will of course not have access to comparative league tables on which to base their decisions, information on whether or not other asylum seekers were allowed to remain in a host country can be expected to be carried back to agents, traffickers and other potential asylum seekers in the countries of origin.


Given the high significance of employment opportunities as a pull factor for asylum seekers (see above), it is not surprising that a policy of not allowing asylum seekers to work until their application has been decided upon (a process that in some countries can take several years) will act as an effective deterrence to those being in a position to choose in which country to lodge their application.  Again this will not only be true for those applicants whose motives are primarily economic, but also for those who are fleeing persecution. 


One common problem behind the other three measures which might in part explain their limited effectiveness, is clearly the issue of how much knowledge about specific policy measures asylum seekers can be expected to possess.  In addition, there are a number of more specific problems with such measures that will also contribute to their limited effectiveness.  In the case of safe third country provision, problems arise as a state who wants to send a potential actual asylum seeker back to another safe third country can only do so if it can establish at least part of the migrant’s transit route.  Often this proves difficult as asylum seekers are either unable or unwilling (having been instructed by their agents) to provide such information.  In particular with persons who apply for asylum only once they are already inside a country, that country can only hope for the cooperation of transit countries who might have already registered a person.  Judging on past experiences, this type of cooperation is often not forthcoming.  Despite efforts by the European Union to institutionalise such cooperation with the Schengen and Dublin Agreements and the joint EURODAC database, progress in this area has so far been limited (Noll 2000).  Cooperation with countries outside the EU is even more difficult as here the application of safe third country provisions requires special bi-lateral or multi-lateral re-admission agreements.  Therefore, the limited effect that safe third country provisions have had so far should not come as a surprise.  Even in the case of the often quoted 71 percent drop in asylum applications in Germany from 1992 to 1994, which have generally been ascribed to the introduction of safe third country provisions with the 1993 changes to the German Basic Law, few observers appear to be aware of the fact that this drop happened against a 53 percent drop in overall applications across OECD countries and a 76 percent drop of applications from former Yugoslavia for which Germany had been the preferred country of destination in 1992, implying again the limits of policy based explanations.


Regarding the remaining two measures—the denial of free movement and cash benefit payments—it is perhaps not surprising that these measures have not deterred asylum seekers in any significant numbers.  The prospect of personal safety from persecution and a green card at the end of a successful determination procedure might make even the prospect a few months on non-cash benefits in North Dakota just about bearable.  The effectiveness of dispersal regimes is further reduced by the fact that, short of a general policy of detainment, there appears to be little a host-state can do to prevent the movement of those determined to join relatives or friends in other parts of the country.  Even the policy of withdrawing housing and welfare assistance in such cases, as practised by some sates, has not always had the wanted deterrence effect.[31]  Finally, the role of cash benefits payments as a pull factor for asylum seekers has without doubt been greatly exaggerated in the media and by policy makers.  Payments of benefits at a level that is often much less than the social assistance minimum will clearly be of limited attractiveness, in particular when many OECD countries have failed to effectively curtail illegal employment opportunities which promise vastly higher rewards.



5. Conclusion


This article has shown that some of the most prominent public policy measures aimed at regulating unwanted migration and at addressing the very unequal distribution of asylum applications among OECD countries are less effective than has often been assumed.  There are four main reasons for this:


First, policy making in this area sometimes appears to exaggerate the degree of choice and the level of information that asylum seekers and their agents are assumed to have.  The concern about widespread and systematic ‘asylum shopping’ must therefore be called into question.  Still, it is clear that those asylum seekers who are in a position to choose between a number of alternative host countries do so in a rational manner, basing their decision on a certain knowledge about the differences between these states. 


Second, the empirical analysis has shown, however, that the most powerful explanatory factors for an asylum seeker’s choice of host country are clearly not consideration of short-term welfare maximisation by the asylum seeker but legacies of migrant networks, employment opportunities and asylum seekers' perceptions about the relative 'liberalness' of a particular host country, i.e. factors that, at least in the short and medium term, are not fully within the reach of asylum policy makers.  There are a number of plausible reasons why it might not be greatly surprising that individual deterrence measures will be overshadowed by these other pull factors.  Ties with friends or family are likely to prove very strong even in the face of a country's not so welcoming asylum regime.  Moreover, path-dependent processes can be expected to play strong roles because of the sunk costs involved in the creation of forced migration networks.  Restrictive immigration control policies create a profitable niche market for those exploiting the black market of international migration.  Organised trafficking gangs and individual 'entrepreneurs' provide a range of services to migrants for which they are able to charge often extortionate fees.[32]  Such services range from the smuggling across borders, forged documents and visas, arranged employment and lodging, etc.  It is estimated that more than 70 percent of asylum seekers make use of such services (Ref.??).  Networks between these traffickers, agents, potential migrants and legal residents or citizen of destination countries are costly to build and are unlikely to be given up lightly.  In a similar vein, a country's liberal reputation, which will have emerged over decades, is also unlikely to be called into question overnight.[33] 



Third, there are also structural factors that limit the effectiveness of policy measures in this area.  As states tend to copy deterrence measures introduced by other states, the desired impact of such attempts to by one state to make its asylum policy more restrictive relative to other potential host countries, is often limited to very short-term first mover advantages.  It has been shown elsewhere (Thielemann 2003a), using the example of safe third country provisions, that such processes of cross-country policy transfer have been very common in this area. 


Finally, the effectiveness of unilateral policy measures will therefore be further undermined by multilateral efforts of international policy harmonisation which will tend to consolidate existing disparities in the distribution of asylum burdens.  Ongoing efforts by the European Union that concentrate on the harmonisation of asylum rules across Europe should therefore not be seen as a panacea to Europe’s burden-sharing problem.  Unilateral and multilateral policy responses (as indeed academic research) will have to place a higher emphasis on the question of how to tackle migration push-factors, i.e. on addressing those problems which are responsible for people leaving their home country to claim asylum.  International cooperation on restrictive policy responses alone is unlikely to provide a long-term solution for the problem of inequitable distribution of burdens across Europe.  The current almost exclusive focus on minimising pull factors and maximising deterrence appears to be aimed more at aim at least as much at appeasing xenophobic elements among the general public and in parts of the media,[34] than at providing long term solutions to the more fundamental problems underlying the current asylum crisis. 





Barsky, R.F. (1995), ‘Arguing the American Dream a la Canada: former Soviet Citizens’ Justification for their choice of host country, Journal of Refugee Studies, 8(2), 125-41.

Barsky, R.F. (2000), Arguing and Justifying: Assessing the Convention Refugees' Choice of Moment, Motive and Host Country, Aldershot: Ashgate.

Beck, N. (2001), Time-Series-Cross-Section DATA: What have we learned in the past few years?, Annual Review of Political Science, Vol. 4, pp. 271-93.

Beck, N. and J. Katz (1995), What to do (and not to do) with time-series cross-section data, American Political Science Review, Vol. 89, No. 3, pp. 634-47.

Boecker, A and T Havinga (1998), Asylum Migration to the European Union: Patterns of Origin and Destination, Luxembourg: Office for Official Publications of the European Communities.

Boecker, A and T. Havinga (1999), country of asylum by choice or by chance: asylum-seekers in Belgium, the Netherlands and the UK, Journal of Ethnic and Migration Studies, Vol. 25, No. 1, pp. 43-61.

Borjas, George (1990), Friends or Strangers: The Impact of immigrants on the U.S. Economy, New York: Basic Books.

Crisp, Jeff (1999), `Who has counted the refugees?' INHCR and the politics of numbers, New Issues in Refugee Research, Working Paper No. 12, Policy Research Unit, Geneva: UNHCR.

ECRE (European Council on Refugees and Exiles) (2000), Non-State Agents of Persecution and the Inability of the State to Protect, ECRE Working Paper, September.

Fassmann, H. and R. Muenz (1992), Patterns and trends of international migration in Western Europe, Population and Development Review, Vol. 18, No. 3, pp. 457-80.

Fawcett, J.T. (1989), Networks, linkages and migration systems, International Migration Review, Vol. 23, No.3, pp. 671-80.

Freeman, G.P. (1994), Migration Policy and Politics in Receiving States, International Migration Review, Vol. 26, No. 4 (winter), pp. 1144-67.

Freeman, Gary P. (1994), Can Liberal States Control Unwanted Migration?, Annals of the American Academy of Political and Social Science, Vol. 534 (July), pp. 17-30.

Guiraudon, V. and Lahav, G. (2000), A Reappraisal of the State Sovereignty Debate. The Case of Migration Control, Comparative Political Studies, Vol. 33, No. 2, pp. 163-95.

Hailbronner, K. (1993). “The Concept of "Safe Third Country" and Expeditious Asylum Procedures: A Western European Perspective.” International Journal of Refugee Law 5(1): 31-65.

Harris, J.R. and M.P. Todaro (1970), Migration, unemployment, and development: A two-sector analysis, American Economic Review, Vol. 60, pp. 126-42.

Hathaway, J.C. (1992), The Conundrum of Refugee Protection in Canada: From Control  to Compliance to Collective Deterrence, in Gil Loescher (ed.), Refugees and the Asylum Dilemma in the West, University Park: Pennsylvania State University Press.

Hicks, A. (1994), Introduction to Pooling, in: Janooski, T and A. Hicks, The Comparative Political Economy of the Welfare State, New York: Cambridge University Press.

Holzer, T. and Schneider G. (2002), Asylpolitik auf Abwegen, Opladen: Leske & Budrich.

Holzer, T, Schneider G. and Widmer T. (2000), The Impact of Legislative Deterrence Measures on the Number of Asylum Seekers in Switzerland, International Migration Review, Vol. 34, No. 4, pp. 1182-1216.

Hovy, B. (1993), Asylum Migration in Europe: Patterns, determinants and the role of East-West movements', in King, R. (ed.), The New Geography of European Migrations, London/New York: Belhaven Press, p. 207-227.

Hugo, G. J. (1981), Village-community ties, village norms, and ethnic and social networks: A review of evidence from the Third World, in G. F. DeJong and R.W. Gardner (eds.), Migration Decision Making: Multidisciplinary Approaches to Microlevel Studies in Developed and Developing Countries, New York: Pergamon Press, pp. 186-225.

Jacobson, D. (1995), Rights across Borders, Baltimore: John Hopkins University Press. 

Joppke, C. (1997), Asylum and State Sovereignty: A comparison of the United States, Germany, and Britain, Comparative Political Studies, Vol. 30, pp. 259-98.

Joppke, C. (1998), Why Liberal States Accept Unwanted Migration, World Politics, Vol. 50, pp. 266-93.

Kjaergaard, E. (1994). “The concept of "Safe Third Country" in Contemporary European Refugee Law.” International Journal of Refugee Law 6(4).

Kritz, M.M., L.L. Lim and H. Zlotnik (eds.) (1992), International migration systems. A global approach. New York: OUP

Kunz, E.F. (1981), Exile and resettlement: refugee theory, International Migration Review, Vol.15, No.1, pp. 42-51.

Massey, D. S, Arango, J., Hugo, G., Kouaouci, A., Pellegrino, A. and J.E. Taylor (1993), Theories of International Migration: A Review and Appraisal, Population and Development Review, Vol. 19, No. 3, pp. 431-66.

Massey, D., J. Arango, G. Hugo, A Kouaouci, A. Pellegrino, and J.E. Taylor (1993), Theories of international migration: review and appraisal, Population and Development Review, Vol.19, No. 3, pp.432-66.

Meyers, Eytan (2000), Theories of International Immigration Policy—A Comparative Analysis, International Migration Review, Vol. 34, No. 4 (winter): 1245-1282.

Morrison, J. (1998), The Cost of Survival: the trafficking of refugees to the UK, London: The Refugee Council.

Noll, G. (2000). Negotiating Asylum. The Hague, Nijhoff.

OECD, Trends in International Migration: Continuous Reporting System on Migration (SOPEMI), Annual Report, OECD: Geneva, issues 1980-2001.

Pierson, Paul (2000), Increasing Returns, Path Dependence, and the Study of Politics, American Political Science Review, June, pp. 251-66.

Ranis, G. and J.C.H. Fei (1961), A theory of economic development, American Economic Review, Vol. 51, pp. 533-65.

Reader's Digest (2001), Are We a Tolerant Nation?,

Refugee Council (2002), Nailing Press Myths,

Robinson, Vaughan and Jeremy Segrott (2002), Understanding the decision-making of asylum seekers, UK Home Office Research Study No. 243, London: UK Home Office.

Sassen, S. (1996), Losing Control? New York: Columbia University Press.

Schoorl, J., Heering, L., Esveldt, I., Grenewold, G., van der Erf, R., Bosch, A., de Valk, H. and B. de Bruijn (2000), Push and Pull Factor of International Migration: A comparative report, Luxembourg: Office for Official Publications of the European Communities (Eurostat).

Soysal, Y. (1994), Limits of Citizenship: Migrants and Postnational Membership in Europe, Chicago: University of Chicago Press.

Sjaastad, L.A. (1962), The costs and return of human migration, Journal of Political Economy, Vol. 70, No. 5, pp.80-93.

Stimson, J. A. (1985), Regression Space and Time A Statistical Essay, American Journal of Political Science, Vol. 29, pp. 914-47.

Thielemann (2003a, forthcoming), The 'Soft' Europeanisation of Migration Policy: European integration and Domestic Policy Change, Journal of Ethnic and Migration Studies, Vol.28, No. 2.

Thielemann (2003b, forthcoming), Forced Migration and the Politics of Burden-Sharing Beyond the State, Journal of Refugee Studies, Vol.16, No. 3.

Todaro, M. P. (1989), Economic Development in the Third World, New York: Longman.

Todaro, M.P. (1969), A model of labor migration and urban unemployment in less-developed countries, The American Economic Review, Vol. 59, pp. 138-48.

U.S. Committee For Refugees, World Refugee Survey, Washington: US Committee for Refugees, issues 1980-2002.

UNHCR (1999), Refugees and Others of Concern to UNHCR - Statistical Overview, Geneva: UNHCR.

United Kingdom Home Office (2002), Reducing Asylum Applications, internal working paper in preparation for the ministerial meeting of 15 May, on file with author.

Zolberg, A.R., A. Suhrke, and S Aguayo (1989), Escape from violence: Conflict and the refugee crisis in the developing world, New York: OUP.

[1] An earlier version of this paper was presented at the 98th American Political Science Association Annual Meeting, Boston, 29 August – 1 September 2002.

[2] The largest part of the world’s 15 million asylum seekers in 2001 sought refuge in developing countries.  However, since the early 1980s the number of asylum seekers in Europe increased almost tenfold to 970.000 in 2001.  In the period between 1985 and 1999, Switzerland as the largest recipient of asylum seekers on average relative to its population size was faced with 30 percent more asylum applications than Sweden, 40 percent more than Germany, 6 times as many as France and the UK, 30 times more than Italy and 300 times as many as Portugal and Sweden (UNHCR 1999).

[3] Recent examples are the currently strained relations between Denmark and Sweden following the introduction of highly restrictive asylum measures by the new conservative government in Denmark and the controversy about the Sangatte refugee camp which soured relations between France and Britain.

[4] One notable exception is the recent study by Holzer and Schneider (2002).

[5] The words "floodgates", "swamped", "scroungers", "soft touch" and "bogus" are frequently used by newspapers (and at times by politicians) in the context of asylum policy.  Take the following examples from newspaper headlines in the UK: "Our land is being swamped by a flood of fiddlers stretching our resources—and our patience—to breaking point" (The Sun); "Hello Mr Sponger… Need Any Benefits?" (Daily Star, 26/04/2002). "Scandal of how it costs nearly as much to keep an asylum seeker as a room at the Ritz" (The Mail); " …we resent the scroungers, beggars and crooks who are prepared to cross every country in Europe to reach our generous benefits system" (The Sun, 07/03/2001).

[6] UNHCR data??.

[7]  There are several possible criteria to measure reception capacity, the most common one being relative population size (per capita).

[8] Council Document 7773/94 ASIM 124.

[9] The form of the suggested redistributive mechanism followed the example of German domestic legislation, which stipulates a similar key for the distribution of asylum seekers among the German Länder.  See section 45 of the German Asylum Procedure Act (Asylverfahrensgesetz).

[10] Council Document 7773/94 ASIM 124.

[11] BMI, Pressemitteilung vom 1.12.1994, FAZ 27.1.1995, p.2; BT-Drs. 13/1070, 55; Integrationsbericht, p.92.

[12] For a comprehensive discussion see Noll (2000).

[13] "Safe third country" provisions mean that asylum seekers are denied access to the refugee status determination procedure on the grounds that they could or should have requested and, if qualified, would actually have been granted asylum in another country.  In practice this means that asylum seekers who have travelled through other countries before reaching their destination will not have their asylum application examined in the country of their choice but will be returned to the other country (Hailbronner 1993; Kjaergaard 1994).

[14] Later in this paper it will be shown, however, that the attributing the drop mainly to changes in German asylum policy is highly questionable.

[15] This problem is certainly confounded when the research in funded by the national body which determines asylum cases, as was the case in the above study by Robinson and Segrott which was financed by the UK Home Office.

[16] Problems with quantitative data analysis stem from the aggregation of data and difficulties stemming from the incongruence of national definitions.  For an extensive discussion see e.g. Crisp (1999) and the chapter on 'Sources and Comparability of Migration Statistics'; in: OECD (SOPEMI) (2002: 269-73).

[17] In the area of forced migration pull factors are not assumed to be the driving forced behind persons leaving their country and push factors are often assumed to be limited to persecution of the kind listed in the Geneva Convention (UN Convention on the Status of Refugees 1951, as amended by the 1967 New York Protocol).  The Convention defines a refugee as a person who "owing to a well–founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group, or political opinion, is outside the country of his nationality, and is unable to or, owing to such fear, is unwilling to avail himself of the protection of that country".  Refugees must therefore be seen as distinct from economically motivated labour migrants, as the former move involuntarily, while for the latter there is an element of choice in their migration decision.  In practice, such classifications are less clear than often assumed, as political and economic causes frequently join forces in producing movement, and freedom of choice is rarely absolute and might be limited in both types.

[18] One objective of this review of the general theoretical literature on migration is to analyse to what extent hypotheses derived from this literature (that deals primarily with economic migration) can be usefully extended to the area of forced migration.

[19] See definition in footnote 16 above.

[20] Due to missing data, the US, Canada and Australia could not yet be included in the data-set.  However, they are expected to be included for the final version of this paper.

[21] Asylum applications per population is the most commonly accepted way of analysing relative burdens in this area.  Controlling for GNP, instead of population size, leads to an almost identical ranking order in terms of relative burdens.  As this analysis here is interested in explaining the distribution of relative asylum burdens over time, it does not seek to assess the role of push-factors responsible for variation in absolute asylum-applications.

[22] Expressed formally: ,

whereby the term B represents the relative number of asylum applications received in country i in year t; a stands for the absolute number of asylum applications received in country i in year t; p for the population of country i in year t; A for the sum total number of asylum applications received across all OECD countries in the dataset in year t and P represents the sum total population figure of all OECD countries in the dataset in year t.

[23] To do this I used a programme developed by John A. Byers which can be accessed at

[24] Other relevant indicators such as a countries' detention and deportation rates, their visa requirements, their readmission agreements with third countries have not yet been included in the index for lack of comparative data.  However, these measure are expected to positively correlate with the other indicators used here and their omission is there not expected to significantly distort the results.

[25] The British government, for example, had long resisted pressures to scrap the UK’s voucher scheme.  Government advisors warned that ‘re-introducing cash benefits would create a “pull factor” for thousands more asylum seekers’ (‘Details of Blunkett’s asylum shake up’, The Guardian, 7 February 2002).

[26] In the light of strong protests by human rights NGOs and rising costs, the UK has recently abandoned its voucher scheme and reintroduced the previous cash-based system.

[27] As a simplifying assumption, I take each of the five policy measure to have the same potential deterrence effects.

[28] 'Pooled', 'panel' or 'TSCS' analysis has become a popular tool for the empirical analysis of issues in Comparative Politics and International Relations.  It involves the analysis of N cross-sections (countries) and T time periods (years).  It increases the number of observations available and allows for the analysis of dynamic factors in cross-national comparative research.  The paper corrects for expected downward bias in standard errors and upward bias in t-statistics (Hicks 1994) by eliminating serial correlation of the errors applying Beck and Katz's standard method of 'panel corrected standard errors' (PCSE).  For a recent review on 'pooling' see Beck (2001).

[29] Robinson and Segrott's survey and interview data suggests a number of reasons for this limited degree of prior contacts—lack of time for those who had to flee at short notice, the dangers involved in risking that others might find out about their emigration plans and loss of contact due to internal flight conditions prior to their decision to leave their country (2002: 41).

[30] Other indicators such as number of racial attacks or number of extreme right wing votes in a country could be added as alternative indicators for the relative 'liberalness' of a country.

[31] Germany for example practises such a policy and has found it wanting in particular as asylum seekers assigned to some parts of Eastern Germany chose to forfeit assistance in the light of a disproportionately incidence of racial violence in areas that until very recently has very little experience with (non-white) foreigners.  This suggests that dispersal schemes sometimes have also failed to achieve their goal of decreasing local residents’ adverse reaction towards asylum seekers which is thought to result from the concentration of such groups in metropolitan areas.

[32] According to IOM figures, these fees range from several hundred to over 30.000 US dollars depending on which country of origin and which host country are involved. [Economist reference]

[33] In interviews, asylum seekers in the UK regularly mention the UK's long-standing democratic tradition as one of the factors that attracted them to Britain (Robinson and Segrott 2002)—a reputation that the introduction of a voucher scheme is unlikely to challenge. 

[34] A recent Mori poll in the UK for example finds that sixty-six per cent of those questioned feel that there are too many immigrants in Britain (Reader's Digest 2001).