narrative/story in the findings/emperical data

profileyorker
DiscriminationagainstNativeSwedes.pdf

INTERNATIONAL MIGRATION PAPERS

86E

Discrimination against Native Swedes of Immigrant Origin in Access to Employment

A research study on Stockholm, Malmö and Goth-

enburg, utilizing the ILO discrimination practice

testing approach

________

Karin Attström

This report is the result of a research project commissioned and financed by the Swedish Integration Board, at the behest of the Government of Sweden.

INTERNATIONAL MIGRATION PROGRAMME

INTERNATIONAL LABOUR OFFICE GENEVA

Copyright © International Labour Organization 2007 First published 2007 Publications of the International Labour Office enjoy copyright under Protocol 2 of the Universal Copyright Convention. Nevertheless, short excerpts from them may be reproduced without authorization, on condition that the source is indicated. For rights of reproduction or translation, application should be made to the ILO Publica- tions (Rights and Permissions), International Labour Office, CH-1211 Geneva 22, Switzerland, or by email: [email protected]. The International Labour Office welcomes such applications. Libraries, institutions and other users registered in the United Kingdom with the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP [Fax: (+44) (0)20 7631 5500; email: [email protected]], in the United States with the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 [Fax: (+1) (978) 750 4470; email: [email protected]] or in other countries with associated Reproduction Rights Organiza- tions, may make photocopies in accordance with the licences issued to them for this purpose. Attström, K. Discrimination against native Swedes of immigrant origin in access to employment Geneva, International Labour Office, 2007 ISBN (printed version) 978-92-2-120101-4 ISBN (web pdf)) 978-92-2-120102-1 Also available in Swedish, International Migration Paper 86Sw The designations employed in ILO publications, which are in conformity with United Nations practice, and the presentation of material therein do not imply the expression of any opinion whatsoever on the part of the Inter- national Labour Office concerning the legal status of any country, area or territory or of its authorities, or con- cerning the delimitation of its frontiers. The responsibility for opinions expressed in signed articles, studies and other contributions rests solely with their authors, and publication does not constitute an endorsement by the International Labour Office of the opin- ions expressed in them. Reference to names of firms and commercial products and processes does not imply their endorsement by the International Labour Office, and any failure to mention a particular firm, commercial product or process is not a sign of disapproval. ILO publications can be obtained through major booksellers or ILO local offices in many countries, or direct from ILO Publications, International Labour Office, CH-1211 Geneva 22, Switzerland. Catalogues or lists of new publications are available free of charge from the above address, or by email: [email protected] Visit our website: www.ilo.org/publns Printed in Switzerland

i

Foreword

This Report presents the findings of a national situation testing study measuring discrimina- tion against native Swedes of immigrant background, conducted in Sweden during 2006.

The International Labour Office was commissioned to carry out this study by the national authority responsible for integration policy, the Swedish Integration Board (Integrations- verket) acting on behalf of the Government of Sweden.

The ILO is mandated to elaborate, promote and supervise international standards regarding the world of work; to provide orientation and technical assistance to its tripartite constituents; and to address issues affecting workers, employers and governments world-wide. The ILO has addressed the treatment of migrant workers since its inception in 1919. It has elaborated standards and measures to uphold workers’ rights and dignity and to protect them from all types of discrimination in employment.

Employment is a fundamental means of participation in society, and discrimination implies a waste of valuable human resources. Thus, discrimination in access to employment not only endangers the success of any efforts for migrants’ integration, but also leads to social ten- sions and economic loss for individual enterprises as well as the national economy as a whole.

The ILO has conducted research to determine the occurrence of discrimination in access to employment in Belgium, Germany, Italy, the Netherlands, Spain, and the United States, and most recently in France and Sweden. This research obtained statistically significant data documenting that migrant- or immigrant-origin workers are discriminated against when try- ing to find a job.

Practice Tests are carried out through persons posing as job-seekers whose characteristics are carefully matched except for nationality/national origin. These practice tests document real hiring situations and focus on actual behaviour—rather than on subjective statements—of employers seeking to fill vacancies. Test outcomes cannot be influenced by socially desirable answers, as the employers tested are not aware of the experiment. Practice tests thus clearly have an advantage over attitude testing.

ILO testing has showed discrimination in access to employment to be of considerable impor- tance in all countries covered by the research. Candidates of immigrant origin often had to make three to four times as many applications as candidates of “national extraction” in order to get to the next step of the hiring process.

The ILO carries out this testing to help member governments and social partners understand discriminatory behaviour in the labour market. Such information is a critical resource in or- der to motivate and shape effective remedies.

ILO testing research has had significant impact in countries where it has been conducted. In Belgium, for example, the ILO study was credited with shaping the content of national legis- lation adopted in 2003 to put into effect the EU Directive on racism (Council Directive 2000/43/EC). Campaigns against discrimination were established at regional and federal lev- els by the three national trade union federations and the National Federation of Employers adopted an anti-discrimination code of practice for its constituents.

ii

This report is published as part of the ILO working paper series International Migration Pa- pers with the purpose of disseminating the results of research on discrimination in Sweden, research commissioned by the Government of Sweden. As with previous studies, we hope that presentation of these results with contribute significantly to efforts by the government and social partners to effectively combat discrimination.

The objective of the International Migration Papers is to convey current and innovative studies on global labour migration issues and trends to decision makers, migration policy implementers, and researchers. In doing so, the ILO hopes to stimulate dialogue and policy development on regulating labour migration to contribute to economic growth and employ- ment in both origin and destination countries, while ensuring respect for the rights of migrant workers and the economic and social integration.

The ILO would like to express appreciation to the coordinator of the project, Karin Attsröm, and to Tomas Bjornkilde, Director of Ramboll Management, Sweden, without whom the complex process of research, testing, and preparation of this report would not have taken place. We would like to highlight the constant support and accompaniment provided by col- leagues at the Swedish Integration Board, among others: Abdullahi Aress, Andreas Carlgren, José Alberto Diaz, Stina Lyles, Erika Sallander, and Lena Schröder.

And of course we present a big thanks to all the supervisors and testers in the three cities that made this research happen. We also thank August Gachter, CSI, Vienna and Enrico Allasino, Director of the recent testing in Italy, for their methodological support, guidance and review, and to Åke Sander of Gothenburg University and Lisa Simeone for research support.

Finally, I note with appreciation the efforts of my colleague Patrick Taran, Principal Migra- tion Specialist at the ILO International Migration Programme, responsible for the coordina- tion and management for this research through to its conclusion and publication of the re- sults. I equally note appreciation to Séverine Dalban Moreynas of the ILO for her contribu- tions and her oversight for the project, and to Laura Freeman for her editorial support.

Ibrahim Awad Director International Migration Programme

Table of contents

1. Introduction 1

2. Legislative framework and policies on anti-discrimination 3 2.1 Legal and institutional framework for combating discrimination in Sweden 4 2.1.1 Main principles and definitions in the legislative framework 4 2.1.2 Equality bodies and institutional framework 5 2.1.3 Enforcement and application of the law 5 2.1.4 Other issues and current developments 7

3. General overview of immigrants situation on the labour market in Sweden 8 3.1 Immigrants’ participation in the labour market 9 3.1.1 Employment rates according to region of birth and duration of residence 10 3.1.2 Type of job 11 3.1.3 Wages 12

4. The target group – young people and the labour market 13 4.1.1 The Swedish youth labour market in international comparison 14 4.1.2 Previous research 16 4.1.3 Other previous studies on discrimination 17

5. Methodology of the study 18 5.1 Application of the methodology in Sweden 20 5.1.1 Recruiting and training of testers 21 5.1.2 Selection of advertisements and openings 23 5.2 Internal validity of the testing 24 5.2.1 Statistical test for tester bias 24

6. Results of the testing 29 6.1 Results for Sweden as a whole 29 6.1.1 Results at the first stage “Telephone call” 31 6.1.2 Results at second stage “Invitation to an interview” 37 6.1.3 Results at third stage “Offer of a job” 38 6.2 Results for the three cities 39 6.3 Results by sectors 40 6.4 Results of Sweden in relation to previously conducted testing 43

7. Analysis and Initial Conclusions 45 7.1 What do the results say about inequalities in access to employment? 45 7.2 Differences among sectors 46 7.3 Differences among cities 46 7.4 What does testing say about the labour market for young job-seekers? 46 7.5 Reference to other countries 47

8. Some perspectives on the future 49

9. References 50

1

1. Introduction

From the end of World War II until the 1973 oil crisis, labour migration to Sweden was quite extensive. Sweden never instituted a guest worker system; instead, these migrants and their families were granted permanent residence permits. Coming predominantly from Finland, Southern Europe and Turkey, they worked mainly in the manufacturing sector and their level of education was somewhat low. Since the mid-1970s, labour migration has significantly de- clined. In fact, only about 10% of the almost 40,000 residence permits granted to non-Nordic citizens in 2004 were predicated on labour market needs (Statistics Sweden, 2005).

Following the oil crisis and subsequent industrial downturn throughout Europe, refugees and their families became Sweden’s primary immigration category. In the 1970s they came from countries in Latin America, followed by refugees from countries in the Middle East and Af- rica, and in the 1990s from former Yugoslavia. Currently, almost 72% of the foreign born population has resided in Sweden for ten years or more.

In the Swedish immigration debate, labour market participation of the foreign-born is one of the most important and frequently-used measures of integration into Swedish society. Espe- cially in terms of governmental policy, the reduction of immigrant unemployment has been repeatedly advocated as the most important objective in terms of combating segregation and increasing integration. For example, according to the 2001 report of the Swedish Integration Board (Integrationsverket):

The most important cause of marginalization and exclusion is the lack of work and of the

ability to support oneself … The most important task for the policy of integration is to create

the necessary prerequisites for people to support themselves … Employment is the principal

lever for integration. A place in the Swedish labour market is the key for each individual to

be able to build his or her own life project (p. 17).

The idea behind the so-called arbetslinje (work line) is that increased labour market integra- tion will lead to increased integration in all other areas, just as unemployment will lead to increased segregation in society. This position heeds Frank Bovenkerk’s warning1:

…that immigrant groups will eventually end up being part of an ethnic under-

class. … Rejection in the labour market results in restriction of possibilities of

finding suitable housing and this in turn results in relocation in neighbourhoods

where children encounter fewer chances to follow good education. Poor educa-

tion makes for poor chances in the labour market and so on. … Rejected groups

find themselves suffering from a multitude of social problems (poverty, crime,

substance abuse, etc.) that makes it hard to escape from this vicious circle. … In

combination these processes can be expected to pervert values of equality, breed

social conflict and end up in political turmoil. An unequal society divided by eth-

nicity or “race” produces both resistance by the minorities and provocations by a

strong ultra-right.

Fighting labour market discrimination is the most strategic and effective way to break this circle. Prof. Bovenkerk and the International Labour Organisation (ILO) developed a meth- 1 1992, pp. 1. This assessment is also mentioned by Wrench in several places, 1994, 1997, 1998, 1999.

2

odology which measures the level of unequal treatment on the labour market, i.e. the level of discrimination. A pair of testers, one of native origin and one of immigrant origin, are matched to apply for the same jobs. The testers are chosen and trained to be as similar as pos- sible, both in terms of their job applicant characteristics and in terms of their actual appear- ance, body language, attitude, personality etc. The difference in the outcome of their applica- tions is then calculated and used as a measure of manifest discrimination.

To date, the ILO has conducted research based on this methodology in Belgium, Germany, Italy, the Netherlands, and Spain. Similar research has been conducted by academic institu- tions in Denmark and Switzerland to measure access to employment by immigrant groups. Most recently, testing was conducted simultaneously in Sweden and France. This timing has allowed for a sharing of ideas and mutual problem-solving over the course of the testing process.

In 2005, the Swedish government decided to conduct a study to determine levels of Swedish labour market discrimination. The Swedish authority responsible for integration policy de- velopment, the Integration Board (Integrationsverket), was given the responsibility for carry- ing out the study. It commissioned the International Labour Organisation (ILO) to conduct the study in accordance with the standards of the Bovenkerk/ILO methodology.

This study focused on potential discrimination experienced by native Swedish youth (20-24 years) from families originating in the Middle East. Tests were conducted in Stockholm, Malmö and Gothenburg. In order to minimise the risk of other variables influencing testing outcomes—such as language skills, educational background and cultural aspects—the study chose to target members of a minority group with the same qualifications and experience as their majority counterpart. This allowed testing outcomes to reflect the degree of discrimina- tion based purely on ethnicity/national origin.

The first two sections of the report serve as a general background, with facts on the legisla- tive and institutional framework against discrimination and a general overview of immigrant labour market participation, as well as youth unemployment. This is followed in Section Four by a review of previous studies and research, specifically examining the situation for natives of immigrant origin on the labour market, as well as available data regarding labour market participation for the target groups in the three cities. In Section Five the ILO methodology and its implementation in Sweden is described in detail. The outcome of the testing is pre- sented in Section Six, and lastly, analysis and conclusions in Section Seven.

The study was performed by Rambøll Management, a large private research consultancy company which conducts research, evaluations and policy development within the public sec- tor, on the national and international levels. This research project was led by Karin Attström. Contributions to the report have been made by city supervisors responsible for managing the testing in each city: Anna von Bothmer, Karin Norvell from Rambøll Management and Marianne Blomsterberg from the Gothenburg University Department of Work Science. Background material was provided by Åke Sander from the Gothenburg University Depart- ment of Religious Studies. The study and report have also benefited from the support and supervision of Patrick Taran and Séverine Dalban-Moreynas of the ILO, Geneva. August Gächter, Zentrum für Soziale Innovation, Vienna, provided methodological supervision and valuable input to the interpretation of data, as well as a revised calculation annexed to this report.

3

2. Legislative framework and policies on anti-discrimination

The right to equality of treatment among immigrant workers, their descendents and persons considered “native” is established in international normative standards. Such rights have also been generally incorporated into national legislation throughout Europe, notably Sweden. The notion that discrimination is unacceptable in principle and in law is unambiguous within the international normative documents on which legislation in Sweden is based. Discrimination is expressly proscribed in the Universal Declaration on Human Rights, as well as in the seven international conventions characterized as the fundamental instruments defining basic, uni- versal human rights and ensuring their explicit extension to vulnerable groups world-wide2. Sweden has ratified six of these conventions, notably the Convention for the Elimination of Racism and Racial Discrimination (CERD). The ILO Discrimination (Employment and Oc- cupation) Convention No.111, 1958, also ratified by Sweden, states in Article 1 that:

For the purpose of this Convention the term “discrimination” includes;

(a) any distinction, exclusion or preference made on the basis of race, colour, sex,

religion, political opinion, national extraction or social origin, which has the ef-

fect of nullifying or impairing equality of opportunity or treatment in employment

or occupation;

(b) such other distinction, exclusion or preference which has the effect of nullify-

ing or impairing equality of opportunity or treatment in employment or occupa-

tion as may be determined by the Member concerned after consultation with rep-

resentative employer’ and workers’ organisations, where such exist, and with

other appropriate bodies.

The ILO Migrant Workers (Supplementary Provisions) Convention, No. 143 (1975)—also ratified by Sweden—stipulates in Article 10 that, “Each Member for which this Convention is in force undertakes to declare and pursue a national policy designed to promote and to guar- antee, by methods appropriate to national conditions and practice, equality of opportunity and treatment in respect of employment and occupation...”

ILO standards for migrant workers are primarily concerned with equality of labour market opportunity and treatment in the workplace. Many countries broadly adhere to this objective in the economic and social spheres, ratifying ILO Conventions and doing their best to fulfil the obligations deriving from them. One might expect, therefore, that discrimination would no longer be part of the legislation or practices of these countries. Unfortunately, a great deal of circumstantial evidence suggests that this assumption does not hold in certain respects, within both private and public enterprises. Reports of unequal treatment towards ethnic mi- norities have generated a need for constant monitoring and advocacy within the field of non- discrimination and equal opportunities.

2 International Covenant on Civil and Political Rights, International Covenant on Economic, Social and Cultural Rights, Convention for the Elimination of Racism and Racial Discrimination (CERD), Convention Against Tor- ture (CAT), Convention for the Elimination of Discrimination Against Women (CEDAW), the Convention on the Rights of the Child (CRC), and the 1990 Convention on the Protection of the Rights of All Migrant Workers and Members of Their Families (CMR).

4

2.1 Legal and institutional framework for combating discrimination in Sweden

Protection against unequal treatment can be found at both constitutional and legislative levels in Sweden. The European Convention on Human Rights has also been incorporated into na- tional legislation.

In labour law there are four laws that ban discrimination on the grounds of ethnicity and re- ligion, as well as other beliefs, gender, disability and sexual orientation. There is also a law from 2001 that prohibits discrimination in higher education on the same grounds. Further- more, there is a criminal law provision that bans unlawful discrimination by merchants on the grounds of ethnicity (and homosexual orientation) in regards to the provision of goods and services.

As a result of the EU directives against discrimination, various amendments to some of the existing laws and a new civil law went into effect on 1 July 2003. The new law, The Prohibi- tion of Discrimination Act (2003:307), prohibits discrimination related to ethnic origin, relig- ion and other beliefs, sexual orientation and disability in regards to labour market pro- grammes; starting or running a business; occupational activity; membership of, participation in and benefits from organisations of workers or employers or professional organisations; and goods, services and housing. The prohibition of discrimination on grounds of ethnic origin or religion or other beliefs also applies to social services; local and national transport; services for disabled people and housing adaptation allowances; social insurance and related transfer systems; unemployment insurance; as well as health and medical services. The powers of the Swedish anti-discrimination ombudsmen were expanded correspondingly.

The adoption of the new civil law, as well as various amendments in other laws–particularly those covering discrimination in working life–means that Sweden has largely complied with the requirements of the EC anti-discrimination directives. A parliamentary committee has recently proposed consolidated legislation covering all or most discrimination grounds and areas of society. Consolidation of the anti-discrimination ombudsmen is also proposed. The Committee also examined how to implement the EC Directive’s provisions on age discrimi- nation.

2.1.1 Main principles and definitions in the legislative framework

As indicated in Section One, Sweden has a number of different laws intended to counteract discrimination. The civil laws, regardless of the discrimination ground at issue, cover both direct and indirect discrimination.3 In addition, harassment, instructions to discriminate and victimisation are defined and prohibited.

In regard to exceptions, it should be noted that the ban against direct discrimination does not apply in regard to employment, promotion or training for promotion if a specific ethnic back- ground, religion or other belief, sexual orientation or specific disability is necessary due to the nature of the work, or the situation in which the work is being carried out.

3 Direct discrimination occurs when rules and practices explicitly exclude or give preference to certain individu- als just because they belong to a particular group. These forms of discrimination are rooted in prejudices and biased perceptions of the abilities or work ethics of such individuals, irrespective of their skills and work experi- ence. (Sheppard, 1989) Indirect discrimination occurs when norms and procedures -whether written or verbal- that appear to be neutral, e.g. do not have any specific requirements in terms of sex, race or age, end up affecting disproportionately members of groups of a particular sex, race or age.

5

In addition, the 2003 law’s provisions banning discrimination on the grounds of ethnicity, religion or other beliefs, sexual orientation and disability in labour market activities (non- employment related training, etc.) specifically state that the prohibition does not apply to the application of “positive action” measures that are intended to promote equal opportunities regardless of ethnicity.

The broad protection related to ethnicity can be traced specifically to the requirements of the EC Race and Ethnicity Directive. It has been an idiosyncrasy of Swedish law that discrimina- tion due to religion has long been associated with ethnicity. This has passed into the 2003 laws, though discrimination on religious grounds is also now clearly specified in its own terms.

As indicated above, Swedish law is essentially in compliance with both of the EC anti- discrimination directives. One major area of importance that had not been covered by the anti-discrimination legislation is primary, secondary and vocational education. However, on the 1st of April 2006, a new Act came into force, the Act Prohibiting Discrimination and Other Degrading Treatment of Children and School Students (2006:67). The intention of the Act is to promote the equal rights of children and pupils in order to combat discrimination on the grounds of sex, ethnic origin, religion or other beliefs, sexual orientation and disability.4

2.1.2 Equality bodies and institutional framework

Four ombudsmen are responsible for supervising the application of labour and civil anti- discrimination laws, as well as monitoring compliance. They are the Equal Opportunities Ombudsman (JämO - gender equality); the Ombudsman against Ethnic Discrimination (DO); the Ombudsman against Discrimination due to Sexual Orientation (HomO); and the Disabil- ity Ombudsman (HO). Their mandate varies somewhat, but is generally centered on work life issues. In addition they have a common mandate in regards to higher education. The 2003 law expanded their respective obligations in a variety of ways. Finally, the Swedish Integra- tion Board is charged with examining and counteracting ethnic discrimination at a struc- tural/institutional level.

The Ombudsman against ethnic discrimination (DO) was set up in 1986. The office investi- gates complaints and, as a final resort, has the right to represent the complainant in the labour court. The DO is also supposed to ensure that employers actively strive to promote equal rights and opportunities in the workplace. As a result of the 2003 law, the office expanded in terms of personnel and caseload. In 2005, 34 people were working at the office.5

The Discrimination Board is a special body that can hear certain appeals involving discrimi- nation cases. Created by the Anti-Discrimination Act of 1999 (1999:130), the Board handles appeals of fines that have been imposed by the ombudsmen. At the request of the Ombuds- men, the Board decides these cases with the aim of ensuring that employers follow their obli- gations to promote equal rights and opportunities regardless of ethnicity or religion or other beliefs. The Board’s decisions are not subject to appeal.

2.1.3 Enforcement and application of the law

In terms of criminal anti-discrimination law, the police and prosecutors handle complaints as part of the ordinary penal process. In civil situations, victims of discrimination can either 4 Fact sheet, Ministry of Education, Research and Culture Sweden, U05.059, November 2005. 5 The Ombudsman against Ethnic Discrimination, Årsredovisning 2005.

6

submit their cases to the Ombudsman against ethnic discrimination (DO) or file independent lawsuits In addition, cases related to working life can be brought on behalf of the victim by a trade union. In fact, unions are supposed to be the primary source of advice and representa- tion in labour law cases.

Other than unions and the ombudsmen, there are not really any non-governmental organisa- tions that today support victims in filing their complaints. However, there are an increasing number of government local anti-discrimination bureaus that provide advice to victims. As- suming the victim has agreed, organisations can support such complaints and act on behalf of the victim. As with much Swedish legislation, there is not much in the way of case law con- cerning ethnic discrimination. However, in 2002, the DO won a case in which the 1999 Anti- Discrimination Act was fully examined by the Labour Court for the first time. This should lend credence to the idea that legal advocacy can be an effective anti-discrimination strategy. The increasing number of voluntary settlements in favour of victims has also indicated this.

On December 4, 2002, the first Labour Court judgement was issued dealing directly with the 1999 measures against ethnic discrimination in working life. The court agreed with the DO claim that the employer violated the law by placing demands that were too high in regard to the need for perfect Swedish language skills in relation to a job as a telephone interviewer. It is significant that the case was decided in favour of the victim and that it involved language skills. It appears that onerous requirements regarding Swedish fluency have frequently been used as a screening mechanism during hiring. Perhaps this case will convince employers to seriously assess the skill levels related to the job in order to apply them on a case-by-case basis to individual job applicants.

Concerning evidentiary standards, the following wording was introduced into the 2003 Act: “If a person who feels that he/she has been discriminated against or exposed to reprisals shows that the circumstances give reason to presume that he or she has been discriminated against or exposed to reprisals, the respondent shall show that discrimination or reprisals have not occurred.” In other words, the employer bears the burden of proof. This conforms to bur- den of proof standard set in the EU “Race Directive” of 2000 referred to above. In Sweden, this rule has applied to the practice of labour law cases for a number of years. Judging from an increase in the number of complaints submitted to the DO, there seems to be a growing awareness of the law against discrimination in working life. This may be partially due to a greater effort among unions towards increasing their members’ awareness of anti- discrimination laws. As for the new law, it is too early to determine the effectiveness of its civil provisions. However, the related penal code on unlawful discrimination has been highly ineffective. Essentially, convictions have been obtained only when the perpetrator confesses.

In regards to sanctions, damages under Swedish law are relatively small in comparison with Canada, the UK and the US. A fine of SEK 80 000 (about € 8,700) would hardly deter a large employer. However, combined with the threat of publicity, they may have some effect. This may partially account for the increasing number of settlements reached by DO on behalf of its plaintiffs. Whereas a court can only award damages, a settlement can lead to the victim getting the job he or she applied for, as well as other work-related benefits. While sanctions are increasingly being applied, other related factors may also come into play in addressing discrimination, such as the introduction of “contract compliance” clauses in public contracts for goods or services that require anti-discrimination policy and practices by suppliers. This can be an important complement to the law, particularly if the public authority has retained the right to cancel the contract if discrimination is found to occur.

7

The workload for the Ombudsman against Ethnic Discrimination (DO) is continuing to in- crease. In 2005, the office received 888 complaints that amounted to a 12% increase over previous year6. Of these cases, 334 were complaints concerning working life, while 554 re- lated to other fields of society. In 2005, 326 work-related cases were closed, 30% without investigation. Reasons cited included prior resolution or inadequate information from the complainant. Often, a trade union had taken over the handling of the case. On the other hand, of the 208 cases that were investigated in 2005, 15 cases were resolved by reaching settle- ments, 27 situations were otherwise resolved, and 3 cases led to a Labour Court judgement. In 163 cases the DO came to the conclusion that there was no ethnic discrimination involved.

2.1.4 Other issues and current developments

In the beginning of 2006, a new national action plan for human rights was presented by the Swedish government7. The action plan involves several areas within the human rights spec- trum, including an agenda to combat discrimination and racism from 2006 to 2009.

In the beginning of 2006, the Parliamentary discrimination committee proposed a new com- prehensive law against discrimination on the grounds of sex, ethnicity, religious belief, sexual orientation and disability. According to the national action plan, a proposal will be prepared which will go into effect no later than 2008. The discrimination committee also proposes the merger of the four ombudsmen into one authority with the duty of monitoring all grounds of discrimination.

The action plan further directs certain state authorities to establish an anti-discrimination strategy. The strategies should be adjusted to the authorities’ different activities and ensure continuous, long-term coverage of all discrimination grounds. The starting point of this pro- ject should be internal, e.g. starting with the discrimination that may be taking place within each respective authority. This work should further include agency clients and discriminatory structures of organization and service delivery.

The national action plan will also require some of the major state authorities to introduce anti-discrimination clauses into their public procurement contracts. The cooperation between the four different ombudsmen and the anti-discrimination bureaus is to be developed further. Furthermore, there are plans to conduct situation tests, such as those utilized in this study, in order to substantiate cases reported to the different ombudsmen. The Ombudsman against Ethnic Discrimination has already looked into this matter but has suggested in a report to the government that further inquiry is needed. The national action plan on human rights has also indicated a need to apply such tests to other areas of social life.

The plan also includes introduction of activities to combat structural discrimination8 on the basis of proposals made by the governmental Inquiry on Power, Integration and Structural Discrimination.

6 Ombudsman against Ethnic Discrimination. Årsredovisning 2005. 7 Skr. 2005. A written government communication 2005/06:95. 8 Structural discrimination refers to discrimination inherent to or institutionalised in social patterns, institutional structures and legal constructs that reflect and reproduce discriminatory practices and outcomes.

8

3. General overview of immigrants situation on the labour market in Sweden

A significant proportion of the population living in Sweden was born abroad; in 2005, more than 12% was foreign born. Among those of working age (16-64), 15% was born in a country other than Sweden. In an international context, this means that the proportion of foreign born is larger in the Swedish labour market than in countries such as the UK and France. Further- more, Sweden has a larger proportion of refugees in its foreign-born population than most other European countries, as well as a comparatively high proportion of foreigners with terti- ary education. The foreign born population in Sweden consists of approximately three major groups:

• the labour migrants of the 1950s and 1960s, predominantly from Finland, Southern Europe and Turkey

• the refugees from the 1970s and onwards, from countries in Latin America, the Middle East, Africa and most recently from the former Yugoslavia.

• people moving within the free labour markets of the Nordic countries and the EU.

Because the foreign-born population stems from different “immigration waves”, average length of residency in Sweden differs considerably according to region of birth.

Table 1 Foreign born population 16-64 years old with different duration of residence, dis-

tributed on region of birth, 2005 (%)

Nordic

Other EU15

Other Europe

Africa Asia Latin America

USA Canada Australia and NZ

Total

Residence:

0–4 years 12.5 10.9 21.2 9.2 38.3 4.7 3.0 100 5–9 years 7.7 9.1 25.4 8.5 41.6 5.0 2.6 100 10–19 years 6.6 4.1 33.3 9.4 37.6 7.7 1.3 100 20 or more 41.3 10.1 17.2 3.1 20.1 6.7 1.4 100 Total number of

persons 187,049 71,533 210,283 59,546 273,458 56,869 15,702 874,440

% 21.4 8.2 24.0 6.8 31.3 6.5 1.8 100 Source: STATIV, database at the Swedish Integration Board.

As can be seen in Table 1, almost one quarter of the foreign-born population originated in other Nordic countries, another quarter in non-EU European countries and almost one third in Asia and the Middle East. Foreign born with very long residence in Sweden predominantly came from other Nordic counties (almost 50%), while those with short duration of residence mainly come from countries in Asia (almost 40%). Of the more than 220,000 persons who have migrated to Sweden in the last five years (2001-2005), almost 40% were born in Asia and the Middle East. More than 25,000 of them were fleeing Iraq.

9

The target group of this research were Swedish-born youth with a Middle Eastern back- ground. This is quite a large group; more than half of the immigrants from Asia and West Asia (almost 150 000) were born in countries in the Middle East. Only 15% of the immi- grants from the Middle East have been in Sweden for 20 years or more, which means that their Swedish-born children are still relatively young. However, the number of second- generation youth of working age will increase substantially in the future.

3.1 Immigrants participation in the labour market

The Swedish labour market has undergone dramatic changes in the last few decades. Until the end of the 1980s, Sweden was a country with very low unemployment rates. At its peak, 83% of the population aged 16 to 64 was employed, and unemployment was down to just 1.5% of the labour force. This changed drastically at the beginning of the 1990s, when Swe- den experienced its most severe labour market crisis since the 1930s. Employment fell to around 70% and unemployment rose to almost 8% by the mid-1990s. The employment situa- tion has recovered since then, but without reaching the peaks recorded at the end of the 1980s.

Figure 1 Employment index, foreign born compared to native born, 16-64 years old.

0

20

40

60

80

100

120

140

1 9

5 0

1 9

6 0

1 9

6 7

1 9

7 5

1 9

7 8

1 9

8 7

1 9

8 8

1 9

8 9

1 9

9 0

1 9

9 1

1 9

9 2

1 9

9 3

1 9

9 4

1 9

9 5

1 9

9 6

1 9

9 7

1 9

9 8

1 9

9 9

2 0

0 0

2 0

0 1

2 0

0 2

2 0

0 3

2 0

0 4

2 0

0 5

Sources: 1987-2005 (represented by the dotted line) Statistics Sweden, Labour Force Surveys, 4th quarter, own

calculations. Single years 1950-2001 (represented by the diamonds) Ekberg and Hammarstedt (2002)

The trend in relative employment rate of foreign born is illustrated by a simple index in Fig- ure 1. By combining the results of a research report (Ekberg & Hammarstedt, 2002) with La- bour Force Surveys, it has been possible to construct an index9 covering the entire period since 1950. All index values lower than 100 mean that the employment rate is lower for for- eign-born than for native-born Swedes.

Until the middle of the 1970s, the employment rates of the foreign-born Swedes were some- what higher than those of the native-born Swedes (the index in Figure 1 has a value higher

9 The index is the employment rate of foreign born divided by the employment rate of native born multiplied by 100.

Employment

10

than 100). This relation, however, changed in the mid-1970s. Since then, the employment rate of the entire foreign-born population has been lower than for the native–born population, in- dicating a continuously widening “employment gap.” This widening gap may have ceased in the mid-1990s. It is, however, too early to ascertain whether this break represents a shift in a deteriorating trend or just reflects fluctuations in the business cycle.

3.1.1 Employment rates according to region of birth and duration of residence

The relative employment rate of people born abroad is a very crude measure of labour market conditions. Employment and unemployment rates differ considerably according to both dura- tion of residence in Sweden and region of birth. Nevertheless, calculating employment in- dexes for people born outside of Sweden who have spent differing periods of time in Sweden reveals some typical traits:

• Native-born Swedes have higher employment rates than all groups of foreign-born peo- ple. Not even those who have spent 20 years or more in Sweden have the same employ- ment rates as natives.

• This employment gap between native- and foreign-born individuals decreases relative to length of residence in Sweden.

• The time trend differs considerably between foreign-born individuals with long and short durations of residence.

• Foreign-born individuals who have been in Sweden for ten years or more have a slightly declining employment rate compared to the native-born Swedes.

• Those with shorter residence in Sweden are more affected by fluctuations in the business cycle than natives.

This pattern is valid for men as well as for women. However, the differences between natives and newly arrived immigrants are more extreme in the case of women. The general pattern remains when we use statistical methods controlling for differences in age, education and marital status.

Figure 2 Employment rates, foreign born, according to region of birth and duration of residence, 16-64 years old, 2003 (%)

Source: STATIV, database at the Swedish Integration Board.

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

%

Nordic EU Other Europe Africa Asia Latin America

Women Men

11

Figure 2 illustrates that, irrespective of the individual’s duration of residence, the labour mar- ket situation also differs according to region of birth. Closest to the employment rates of na- tives are those born in other Nordic countries, followed by other EU15 countries. The largest “employment gap” is found for people from countries in Africa, Asia and the Middle East, especially those who have spent a short period of time in Sweden. In general, men have higher employment rates than women, but women from Nordic and African countries who have resided in Sweden for an extended period of time are an exception to this gender pat- tern. The pattern remains using statistical methods comparing persons with the same age, education and marital status.

3.1.2 Type of job

The severe employment crisis in the beginning of the 1990s hit the manufacturing sector most severely. At the end of the 1980s, the manufacturing sector employed almost 30% of all na- tive-born men and more than 40% of foreign-born men. This share has since been continu- ously declining, especially for people born abroad. In 2004, less than a quarter of employed men (native or foreign born) worked in the manufacturing sector. The same decline applies to women, although on a smaller scale. This decline has been partially compensated for by a growth in the private and public service sectors.

Thus, foreign-born men and women are no longer over-represented in manufacturing. On the contrary, the 2004 Labour Market Survey found the largest numbers of foreign-born employ- ees working in the “Hotel & Restaurant”, “Other private services”, and “Health Care” sectors.

Table 2 Distribution of employed across industrial sectors, foreign-

and native-born, 16-64 years old, 2004 (%)

Women Men

Sector Native born Foreign born Native born Foreign born Manufacturing 8.5 9.9 23.7 24.8 Construction 1.0 0.2 11.1 4.4 Trade 11.5 9.7 13.8 13.5 Hotel & Res- taurant

3.0 5.7 1.7 8.7

Bank & Insur- ance

4.8 3.2 7.3 4.8

Other private service

8.6 11.8 9.0 11.7

Government, education, re-

search

25.6 21.6 11.8 10.2

Health & Care 27.8 30.9 4.6 7.6 Other 9.2 7.1 17.0 14.2 Total 100 100 100 100

Source: Statistics Sweden, Labour Force Surveys, 4th quarter 2004 in Swedish Integra- tion Board (2005). Figures in bold indicate statistically significant differences between foreign born and native born (5% level).

Recent surveys further suggest that foreign-born workers are more likely to have “atypical jobs”, i.e. temporary and part-time jobs. There is, however, one notable exception to this pat-

12

tern: employed foreign-born women are more likely to work full-time than native born women.

3.1.3 Wages

There are several studies on income and wage differentials between foreign- born workers and natives. The overall picture is the same as has been shown through data on employment. Up to the middle of the 1970s, migrants and natives had comparable incomes. Later, hourly wages and annual incomes of foreign-born workers decreased compared to those of natives. This wage gap decreases relative to length of residence, but still exists even for those who have been in Sweden for 20 years or more. Irrespective of personal history and education level, the income gaps are largest for people born in countries outside Europe. These differen- tials in annual income are largely explained by the more tentative “labour-market attachment” of foreign-born workers, i.e. lower employment levels and, when employed, higher probabil- ity of atypical jobs (Edin & Åslund 2001, Le Grand & Szulkin 2002, Nekby 2003, Rashid 2004).

13

4. The target group – young people and the labour market

At the target age for this study, most young people have recently finished their upper secon- dary education, and are either entering the labour market or continuing their education at the university level. Many of them choose to support their education by interrupting their studies with periods of time in the labour market. This means that there is no evident measure of la- bour market involvement. That the employment rate is low does not necessarily imply failure, as the explanation could be a high proportion studying at university. The unemployment rate gives a better picture, as it shows the proportion of young people in the labour force that are unsuccessfully looking for a job10.

Figure 3 Unemployed percentage of the labour force, 20-24 and 25-54 years old.

Source: Statistics Sweden, Labour Force Surveys, calculations by the Swedish Integration Board. Swedish national definition of unemployment. Statistics Sweden has changed the definitions of unemployment in the Labour Force Surveys, which means that 2005 is not comparable with previous years.

Unemployment has always hit young people harder than adults. The severity of the effects of unemployment on young people has however been increasing with time. Figure 3 also shows that the youth unemployment rate is more sensitive to fluctuations in the business cycle than the unemployment rate of adults. When the economy is slowing down, the youth unemploy- ment rate increases faster than the unemployment rate of adults. When the testing was per- formed in the beginning of 2006, the unemployment rate of young people had been sharply

10 The unemployment rate is calculated by dividing the number of unemployed by the number of people in the labour force (labour force = unemployed + employed). The Swedish national definition of unemployed differs from the standard ILO definition used by the OECD and the EU; the ILO definition includes full-time students looking for a job, whereas the Swedish definition does not.

0

2

4

6

8

10

12

14

16

18

20

19 63

19 65

19 67

19 69

19 71

19 73

19 75

19 77

19 79

19 81

19 83

19 85

19 87

19 89

19 91

19 93

19 95

19 97

19 99

20 01

20 03

20 05

20 06

q 2

20-24 years old 25-64 years old

14

increasing for five years and was almost three times as high as the rate for adults. This means that the young men and women in the project were trying to finds jobs on a very constrained labour market.

4.1.1 The Swedish youth labour market in international comparison

Comparing the situation of youth in the labour market in different countries is difficult. There are several statistical pitfalls because countries differ in the way they define employment and unemployment. Using the international ILO definition of unemployment means that partici- pants in most Swedish youth labour market programmes are included among the unem- ployed, while in other countries young people in similar programmes could be included among the employed. This problem is general in international labour market statistics, but of particular importance in the case of youth, where many countries have extensive labour mar- ket programmes targeted at young people. The international figures in this section thus have to be interpreted with great care.

To compare the situation of youth in different countries, the OECD uses an ‘inactivity’ meas- ure, i.e. the proportion of young people that are neither employed nor studying. The latest available figures are from 2003, when the inactivity rate in Sweden was almost 12% for the group aged 20-24. Although this is a high proportion of young people who are neither study- ing nor working, Sweden seems to have a relatively favourable position compared to other OECD countries (Figure 4).

Comparing the unemployment rates, however, gives an idea of the vulnerable position of Swedish youth. With an unemployment rate of about 17.5% in 2005, Sweden is above the EU15 average, which lies around 15% (Figure 5). This impression is reinforced by compar- ing youth and adult unemployment rates. Unemployment seems to hit young people more severely than adults in Sweden than in most other OECD countries.

15

Figure 4 Proportion of young people in the age group 20-24 who were neither studying nor employed, 2003 (%):

Source: OECD (2006)

Figure 5 Unemployment rates for the age group 20-24, EU15 countries,

period 2003-2005

0.00

5.00

10.00

15.00

20.00

25.00

30.00

2003

2004

2005

Source: OECD Statistical database, compilations by Ramböll Management. ILO- definition of unemployment.

0,0

5,0

10,0

15,0

20,0

25,0

N et

he rla

nd s

D en

m ar

k

N or

w ay

Ir el

an d

S w ed

en

A us

tr ia

P or

tu ga

l

S w itz

er la nd

C an

ad a

A us

tr al ia

F ra

nc e

S pa

in

U ni te

d K in gd

om

G er

m an

y

F in la nd

U ni te

d S ta

te s

B el gi um

G re

ec e

Ita ly

16

4.1.2 Previous research on the labour market situation 11

4.1.2.1 The labour market situation of native youth with a foreign background

Several studies have been conducted on the labour market status of the children of immi- grants. Some of them refer to the entire group of native-born people with parents born abroad. Swedish-born “natives” with parents born in non-European countries are relatively young, which means that the results from these studies are relevant for young natives with a non-European background. Research on young people with a foreign background often in- cludes native youth as well as those who immigrated into Sweden before the age of seven (which is the age at which children start school in Sweden). In most cases it is, however, pos- sible to distinguish young natives with a foreign background from those who have themselves immigrated.

The earliest of these studies (Ekberg 1997) found only minor employment and unemployment differentials between young natives with a Swedish and a foreign background, but there were indications that youth with a non-European background were beginning to face problems in the labour market. These indications have been confirmed in several later studies, which all find that native youth with a non-European background have lower employment, higher un- employment and/or lower wages than other young people (Österberg 2000, Lund et al. 2002, Arai et al 2002, Ekberg & Rooth 2003, Vilhemsson 2002, Hammarstedt 2002, Hammarstedt & Palme 2006, Behtoui 2004, Nilsson 2004). All studies except Hammarstedt (2002) and Hammarstedt & Palme (2006) use data that makes it possible to compare young native people with the same level of education. Most studies include information on the parents’ education level and/or socio-economic status. In two of the studies (Arai et al. 2002 and Vilhemsson 2002), it is moreover possible to compare young people with the same grades in school in the Swedish language.

Two recent studies (Behtoui 2006 and Nekby & Özcan 2006) use longitudinal data to follow young people for 10 years from the year they left upper secondary school12. The differences in employment and wages tend to decrease with time, but still exist after ten years, especially for young people with a non-European background. It also seems that these labour market differences are greater for young men than for young women.

Some conclusions can be drawn from these studies. One is that the greatest obstacle on the labour market is to get a job. Wage differences exist among those who have managed to get a job, but are of lesser importance than employment differentials.

Most of these studies conclude that discrimination is a very probable explanation for the la- bour market differences observed between native young people with and without a foreign background. These results point to the importance of learning more about what is happening in the recruitment process. There is reason to suspect that discrimination may be preventing young people with a foreign background from realizing their ambitions and from contributing their qualities and abilities on the Swedish labour market.

11 For a more extensive presentation of Swedish studies on ethnic wage and employment differences and the role of discrimination, see Schröder (2007). 12 Also these two studies have data on educational level and grades from school.

17

Nekby & Özcan (2006) found that these ethnic labour market differences exist only among young people with a relatively low (less than tertiary) education. Similarly, Rooth and Åslund (2006) have found that education pays in terms of higher employment and higher incomes. Their study, however, only considers immigrants in comparison with natives.

4.1.3 Other previous studies on discrimination

Studies on the labour market situation of native youth with and without a foreign background have played an important role in convincing the majority population that discrimination is a serious obstacle. There are a few other studies that have addressed discrimination from other angles and with different type of data. One such study is Rooth (2002) who uses data on peo- ple who were born abroad and adopted as infants by Swedish parents. This study shows that adopted children who don’t “look European” are at a greater risk of unemployment than those who do.

Arai & Skogman Thoursie (2006) have compared earning promotion for immigrants who changed their foreign-sounding surnames to Swedish-sounding or neutral surnames with im- migrants that kept their original names. In the period before the change of name, name- keepers and the future name-changers had the same yearly earnings. For immigrants with surnames that sounded African, Asian or Slavic, the change of surname was followed by a steeper increase in earnings than for those who kept their foreign-sounding surnames. The earnings gain corresponds to an average of about 15%, and is highest for people with a rela- tively low annual income.

Two field experiments were initiated in Sweden in 2006 using correspondence testing as a method to study discrimination in the labour market. Results were available from only one of these studies (Carlsson & Rooth 2006)13 when this report was prepared. Pairs of equally pro- ductive men, 25 to 30 years old, sent written job applications for the same vacancy. One of the applicants had an Arabic-sounding name and the other a Swedish-sounding name. In to- tal, more than 3,000 applications were sent to more than 1,500 employers. Depending on how mutual no-response data is treated, the net discrimination rate against job seekers with Ara- bic-sounding names varied between 10 and 29% This study also found that the net discrimi- nation rate is largest in relatively unskilled jobs.

13 The other experiment was headed by le Grand et al. Department of Sociology, Stockholm University.

18

5. Methodology of the study

The ILO has sponsored research since the mid-1990s to measure occurrence of discrimination in foreign-born and minority access to employment in Belgium, Germany, the Netherlands, Spain, and Italy. Each of these studies utilized the methodology elaborated by Prof. Frank Bovenkerk to document labour market discrimination. This methodology details how to col- lect statistically significant data and reduce confounding variables in order to isolate dis- crimination from other labour market conditions. During 2006, testing was conducted simul- taneously in France and Sweden, thereby creating a unique basis for exchange of knowledge during the testing process and for interpretation of results.

Practice Tests or Situation Testing generally involves coordinating pairs of job candidates, who are matched save for a single variable, usually ethnicity. One majority and one minority tester applies for the same position, and their respective experiences and results are carefully monitored. Employment applicant profiles and CVs for each member of the pair are matched at equal levels of qualifications, educational attainment and job experience.

The European studies have targeted legally resident migrants or their citizen children who have some experience in the receiving country’s labour market and who would be applying to the same job openings as national workers. In Sweden it was decided that the testing variable would be young native Swedes with a Middle-Eastern origin. By choosing a target group which has been born in the country, undergone the same schooling, and acquired the same language level, researchers minimized the influence of other factors which might influence hiring decisions. All testers and project supervisors are thoroughly trained and supported throughout this process, and confidentiality is rigorously maintained.

The methodology establishes that a minimum of 175 “valid and usable” tests of employment vacancies must be conducted are in each labour market area and for each variable in order to arrive at statistically valid conclusions with respect to the occurrence of discrimination. A considerably higher number of vacancies are usually tested in order to obtain this minimum.

The standard testing procedure is comprised of the following steps:

(1) Voice inquiry: Testers call to inquire about the continuing availability of the position. Testing “pairs” are trained to match their personal styles, such as level of assertive- ness. The critical event in Step 1 is the invitation to apply for the job.

(2) Written application: Testers submit formal applications for the job. Testers prepare carefully to ensure consistent responses when companies call back for an interview. The critical event is the invitation to Step 3.

(3) Job interview: Testers engage in face-to-face interaction with the prospective em- ployer. It is essential that the pair of testers make similar first impressions, since re- search indicates that first impressions matter a great deal. Physical appearances and dress are closely matched to ensure that ethnic background is the only marker of dis- tinction. The critical event is the offer of employment.

There are four outcome options:

19

• Neither is offered a job

• The minority tester alone is offered a job

• The majority tester alone is offered a job

• Both are offered jobs. They may not necessarily be offered equal jobs, and such differ- ences in treatment are also recorded.

The individual case is discontinued as soon as the firm declares a preference for one tester over the other. This could take place at any one of the three steps, but has most frequently occurred at Step 1. In each step there is ample opportunity for the employer to discourage one member of the testing pair to advance over the other to the next step or to employment. Such differences in treatment are recorded by the researchers.

The actual identity of the tester, as well as the employer being tested, is kept confidential at all times during testing. The aim of the study is not to single out discriminatory practices among employers or individuals, but to statistically measure the direct discrimination taking place, consciously or unconsciously, in the employment situation. Therefore, cases and re- sults are recorded anonymously, so as not to accidentally reveal the identity of either the test- ers or the employers. Consequently, results and qualitative data in this report are presented in such a way that single cases cannot be identified.

When appropriate job opportunities are identified, the two testers prepare the test by reading the advertisements and, if needed, looked up more information about the workplace on the Internet.

While one tester makes a phone call, the other tester usually waits in another room, unable to hear the conversation. The first phone call was always made by the minority tester. Every phone call was recorded and supervised by the city supervisor. The general interval between the two phone calls was ten minutes.

If contact fails for both of the testers, the advertisement may be reused on another testing oc- casion. In cases where one but not the other tester gets in contact with the potential employer, the tests are considered invalid.

If both testers are told that the vacancy was already filled, the test is considered valid but not usable. If, on the other hand, both or one of the testers are invited to proceed, by either send- ing in a written application or come for an interview, the test is considered valid and usable.

Thus the following terminology is applied to the cases;

A. Invalid – One or both testers failed to make contact with the employer

B. Valid but unusable – Both testers rejected at the first stage, or test not possible to pur- sue due to different reasons

C. Valid and usable – One or both testers were asked to send in their CV, i.e. invited to continue the employment process

In the second stage there are four possible outcomes: neither of the testers is invited to an interview; only the majority tester or only the minority tester is invited; or both are invited to proceed. In the first three cases, the process is then discontinued. If both testers are invited, the testing proceeds to the third stage.

20

The third stage, invitation to an interview, is usually either the result of sending in a written application, or a direct invitation at the first phone call with the employer. When a tester is invited to an interview he or she always agrees upon a time for the interview with the poten- tial employer. If it turned out that only one of the testers had been invited, the interview is cancelled after it becomes evident that both would not be invited.

When both testers are invited to an interview, the city supervisor usually accompanies the testers to the place of the interview and waits outside. The testers write a brief report of the interview afterwards, containing information about how the tester was received by the em- ployer and what had been said and done during the interview. If the testers are unable to go to the interview if, for example, the meeting point or worksite is excessively distant, the case may be aborted and considered not valid. When the testers are invited to trial work in place of or in combination with an interview, the tests are considered valid and usable, and are com- pleted.

When both testers actually go for the interview, there are usually three possible outcomes. Either the job is offered to neither of the testers, or the job was offered to the majority or the minority tester only. However, occasionally, both can receive job offers.

All information for each test is kept in a file. This normally contains a copy of each adver- tisement, a copy of both testers’ respective applications for the job, information about the date and time of call, who comprised the pair of testers, the outcome of each phase, and a series of remarks for use in reconstructing and analyzing what happened. At the end of the test, the positive or negative outcome for each test is noted. These files always remain confi- dential.

5.1 Application of the methodology in Sweden

As this methodology was applied in Sweden, two female and two male pairs of testers worked in each of the three cities, Stockholm, Malmö and Gothenburg. A minimum of 175 tests per gender were obtained in each city, amounting to 350 tests per city.

When approximately half of the tests had been conducted, the pairs rotated, meaning that both minority testers had worked together with both majority testers by the end of the testing period. The reason for this was to eliminate possible variables that had more with to do with the tester’s characteristics than with their national background.

In all three cities, low-skilled and semi-skilled occupations were targeted, sectors where per- sons of both immigrant and Swedish ancestry usually apply for employment. The testing was further targeted to gender-specific sectors to avoid extra variables, as presented in the follow- ing table.

Table 3 Targeted sectors for testing in Sweden

Women Men

• Hotel and Restaurant • Retail and Trade • Office and Clerical • Healthcare services

• Hotel and Restaurant • Retail and Trade • Transportation and warehousing • Manufacturing

A focus on the two main sectors of employment for each gender was based on an assessment of current demand in each city for low- and semi-skilled workers. This was the mechanism for acquiring sufficient valid test cases within each sector (175). For females, it was decided

21

to additionally focus on the sectors office-clerical work and/or health-care service, depending on each city profile. For males, additional sectors included transportation-warehousing and/or manufacturing. These sectors were also selected to correspond with the project being simultaneously conducted in France, thereby enabling comparisons across sectors in the countries14.

In identifying job advertisements in Sweden, there were instances where no phone number was given, or where it said explicitly that the employer wanted only written applications. In such cases, the phone call was skipped in the first stage and the testers sent in their CV di- rectly instead.

At the phone call stage as well, most of the potential employers asked the testers to send in their CV. In quite a few cases though, especially within the hotel-restaurant and retail com- mercial sectors, the testers were asked to come for an interview without sending in a written application first. If this applied to both testers, these cases were considered valid. In some cases one of the testers was invited to an interview, while the other tester was asked to send in a CV. These tests were considered valid, usable and completed, since the testers were treated differently.

If both testers were asked to send in a written application, this was done the same day as the telephone call took place. Before sending the CVs, these were adjusted to meet the demands of the potential employer, and approved by the city supervisor. Sometimes the testers were asked to fill in an application form via the Internet. When the applicant’s social security number was required, which often was the case with these forms, the application was com- pleted and the test considered valid but not usable, as following up would risk the anonymity of the testing.

In cases where testers were invited to an interview or meeting, the employer responded by sending an e-mail or by calling the tester’s mobile phone. When testers were not invited to continue, they either received a rejection via e-mail or letter. Most of the time however, the testers did not get a response if they were rejected. Some of the rejections by mail came months after the testing period was completed.

In some cases in Sweden, the interview did not take place because documents were asked for or other requirements were imposed, or the testers were not able to attend because of lack of time or the interview site was excessively distant. In these cases the tests were aborted and considered not valid.

Regarding job offers, it also happened that both testers were offered the job in the retail commercial sector. Sometimes one tester was offered the job after the other one had already turned down the job, which was then considered a non-usable test.

5.1.1 Recruiting and training of testers

For each of the three cities, the respective city supervisor was responsible for identifying and selecting the testers and managing their work. Because the tester recruiting and timeframe was somewhat different in the pilot city of Gothenburg, this process is presented first, fol- lowed by those for Malmö and Stockholm.

14 Due to the parallel nature of the studies in France and Sweden, these comparisons were not available at the time of writing this report, wherefore such analysis will need to be undertaken at a later stage.

22

Once Gothenburg was selected as the pilot city for the test, the recruiting process began in November 2005. Applicants were solicited by e-mail to former and actual students of Gothenburg University. A focus was made on departments expected to have significant num- bers of students with Middle Eastern backgrounds. Some 25 students then attended a meeting to explain the project. During the meeting, interviews took place with the applicants. The city supervisor and the method supervisor then evaluated the interviews and decided to offer four of the women jobs as testers. The selected women began their training in late November. Fol- lowing another call for testers in early December, four men were chosen to begin testing by the middle of that month. All of the testers were fluent Swedish speakers.

As part of their training, the testers were given a thorough presentation of the aims and meth- ods of the project, as well as one day of role playing. Once they had practiced their skills by responding to real job advertisements, they designed their “tester identity” and were taught to prepare job-relevant CVs. Working in pairs, they presented themselves with equivalent edu- cation and labour market experience, etc. The only relevant difference was their ethnic back- ground.

Each tester was supplied with a mobile phone with a cash card, an e-mail address, and a “care of” (c/o) mailing address, so that they could receive written responses to job applications. Should an explanation for this become necessary, they would claim to be temporarily living with friends. Their CVs reflected this alibi as well.

The female testers started their work with real applications in mid- November. After a few weeks as Christmas holidays approached, this testing was suspended due to the scarcity of suitable job offers, to start up again after the holidays.

The male testing started the second half of January 2006. Later, departure of a female tester and two male testers for better employment required new recruitment of testers or finalizing the female testing with one minority tester. As a result and also of need to start considerably more tests than anticipated to obtain the required number of valid and usable cases, the re- searchers were only able to finalize the Gothenburg testing in early June 2006.

The selection and recruiting of testers in Malmö and Stockholm began in January 2006. In Stockholm the testers were recruited from Södertörn University College (Södertörns hög- skola), located in a suburb of southern Stockholm. Information meetings were held separately for male and female students on four different occasions. Brief interviews were carried out with each and every one of them during the meetings. On the basis of the interviews, four male and four female testers were chosen by the city supervisor.

In Malmö the testers were recruited both from Malmö University and Lund University. The city supervisor arranged one information meeting for women and two for men due to difficul- ties in finding matching male testers, both majority and minority. Interviews were carried out with the participants, and four female and four male testers with similar characteristics were finally found.

All the testers, in Malmö and Stockholm, were university students. The minority testers were native Swedes with an immigrant background, with one or both parents born in a country in the Middle East. The minority testers all spoke Swedish at a level equivalent to their Swedish counterparts.

23

City supervisors then convened training meetings with the testers. Two female testers from Gothenburg participated at the meetings to share their experiences about the testing proce- dures.

Though the process went relatively more smoothly in Malmö and Stockholm, a few obstacles did arise. For example, after the introduction course in Malmö, one of the female testers was replaced due to a concern that she did not sufficiently “match” the others. In Stockholm, the two female majority testers decided not to continue to the end when it became clear that the period for the testing would be longer than originally planed. Thus, two new testers had to be recruiting when nearly half of the tests were already completed. Nevertheless, both male and female testing in both Malmö and Stockholm was completed within a three-month timeframe.

5.1.2 Selection of advertisements and openings

As substantial numbers of hotel-restaurant and retail commercial job listings are common to all three cities, these sectors were retained as priorities. For women, it was agreed to seek a similar number of valid and usable cases among either office-clerical work or health-care and similar services, depending on the city profiles.

In Gothenburg, female testers began by applying for health-care jobs and especially as “per- sonal assistants”. These jobs were very common, and many of them were in the private sec- tor. However, the frequency of reference requirements and the severity of patient needs led researchers to discontinue applications in this field. Retail jobs were also commonly adver- tised, particularly those in “call centers”. This work is characterized by high turnover due to poor working conditions, low salary and long hours. The other most common sector in Goth- enburg was hotel and restaurant staffing.

Similar considerations ruled out health-care applications in Stockholm. Instead, the female testers focused only on hotel and restaurant, retail commercial and office-clerical jobs. In Malmö however, health care references were not required to the same extent which made it possible to continue applying for those jobs. Additionally the female testers in Malmö applied for some office-clerical jobs. For men, apart from hotel-restaurant and retail commercial jobs, priority was placed on transportation-warehousing in both Malmö and Stockholm.

Generally, job openings were selected by city supervisors before the testers started working. Occasionally though, the testers also located advertisements, under the oversight of the city supervisors. Advertisements among sectors were chosen based on the period of employment (not less than six months), and on the minimum salary offered in retail commercial jobs, which is often based on commission. It is common in Sweden for companies to contract pro- fessional staffing consultancies. Therefore, eligible advertisements were limited to those that referred to a specific job and a specific company or organization. Every available advertise- ment and opening fitting these criteria was applied for.

The Internet was used as the main channel for seeking job advertisements. The majority of all vacancies in Sweden are advertised via the Public Employment Service (Platsbanken) Inter- net site. Private employment services and daily newspapers were also used, mostly on the Internet.

In Stockholm, the female testers contacted 283 advertisements and the male testers 259, whereas the female testers in Malmö contacted 231 advertisements and the male testers 208. In Gothenburg, not all phone calls were documented due to an initial local methodological

24

misunderstanding. In general, only calls that resulted in some sort of contact were entered into the log. Paradoxically, many tester calls were never answered by employers, particularly at restaurants but even by larger companies that presumably have switchboards. When the advertised jobs seemed very suitable, repeated calls were made, though this delayed the proc- ess.

In all three cities it was sometimes difficult to find sufficient appropriate openings that could be applied for during the testers’ work hours. There were also variations in the frequency of advertisements, with more at the beginning of a week. The hardest times for finding work were in December and during the last few weeks of spring, when seasonal jobs predominated.

5.2 Internal validity of the testing

An important task in the testing analysis is to check results for any possible tester bias. The elimination of extraneous variables is important for all stages of testing, from the telephone call to an actual interview. A rotation of tester pairs is also built into the methodology in or- der to avoid the development of behavioral patterns between testers that might influence the overall results.

During the Swedish research, there were inconsistencies in the pairing of testers. Changes took place in two of the cities, Stockholm and Gothenburg. In Stockholm two new majority female testers were brought on board, and in total, 10 testers have been involved. In Gothen- burg, both male minority testers left before finalizing the testing and one replacement was recruited. A considerable number of tests for both men and women had to be concluded with only one minority tester, with a total of 9 testers involved in the Gothenburg project.

These changes in tester pairings have made the analysis of internal validity all the more im- portant, and also difficult. Ideally there should be no differences in outcomes, and when there are differences, they should be random. When members of some pairs do not conduct many tests together, the chances of random imbalances increase. The overall correlations between tester pair outcomes indicate that factors not accounted for may have been influencing the outcome of testing. As can be seen in the table below, there are clearly outcome differences between pairs in the different cities. However, to assess whether the differences are random or not, we need to look at the statistical significance of these potential biases.

5.2.1 Statistical test for tester bias

The internal validity test must be conducted as a one-sample chi-square test with as many degrees of freedom as there were tester pairs minus one (de Vaus 2002:231, 232). The table to be tested consists of only one column containing the number of net discrimination cases for each pair of testers. This is tested against a column containing the expected number of net discrimination cases for each pair.

The expected number e for pair k of sex j in city I, is computed as

e i j k = c i j k * (a i j -b i j ) /c i j

where c is the number of completed cases, a is the number of cases with majority preference, and b is the number of cases with minority preference. The term (a-b)/c represents the net discrimination rate expressed as a fraction rather than as a percentage.

25

The formula for e contains the net discrimination rate for a given city and sex. This is a weighted net discrimination rate for which, in effect, each pair has entered as many times as it has completed a case. For our purposes, the use of an unweighted average of the pairs’ net discrimination rates might be preferable. This would simply mean computing the net dis- crimination rate for each pair and averaging over the total. The drawback of doing so is that the sum of the actual cases of net discrimination across the pairs would then differ from the sum of the expected cases. The chi-square test, in principle, assumes that the two sums are equal.

The test itself consists simply in computing (a-b-e)2/e for each pair and adding up the results for all pairs. The level of statistical significance can be found in a table of critical values for chi-square. The degrees of freedom are the number of pairs included in the test minus one. If four pairs were included in the test, any chi-square greater than 7.815 would indicate 95% likelihood that the expected values are really different from the actual values, i.e. that more than mere chance was involved in bringing about the differences between the pairs’ actual results. With five pairs included in the test a chi-square of 9.488, and with six pairs a chi- square of 11.070 signify the same. 99 percent likelihood is obtained above chi-square values of 11.341, 13.277, and 15.086, respectively.

A chi-square test should not be performed for a given city and sex if the expected values are too small. According to Erickson and Nosanchuk, “For 2 x 2 tables, the expected values in each cell should be 10 or more. For tables larger than 2 x 2, the mean of the expected values should be six or more for tests at the 5 per cent level” (1992:251). 2 x 2 tables in chi-square testing have one degree of freedom. This is where the minimum ten expected cases applies. The tables we face have only one column. In this situation, the one-sample test, the degrees of freedom are the number of rows minus one. Our tables have between 4 and 6 rows, with a minimum of three degrees of freedom. Therefore the mean six expected cases rule applies. In Sweden this could be met everywhere except among male tester pairs in Gothenburg. Most other literature suggests that expected cell values should be 5 or more. The Gothenburg male results fail this criterion, too.

The restriction on the required number of expected cases cannot be circumvented. There is no alternative but to perform the test on the difference between the completed and the net dis- crimination cases, rather than on the number of net discrimination cases. The result will not be the same.

Rather than testing the cases of net discrimination, it might be preferable to test the cases of majority preference and the cases of minority preference separately. However, there were not sufficient minority preference cases in Sweden to take this route. Finally, the statistical test has to be performed on the actual tester pairs. Personnel changes represent the constitution of new pairs.

The table below presents the relevant information for the chi-square test. The sum for each city and sex can be found in the subsequent table. Columns are as follows: Pair: The first let- ter is the majority tester, the second the minority tester; Completed: All the cases completed by a tester pair; Majority pref: Number of cases with majority preference; Minority pref: Number of cases with minority preference; Observed: Net discrimination, i.e. majority minus minority preference cases; Expected: Expected difference between majority and minority preference cases; Difference: Difference between observed and expected cases; Contribution to chi-square: Each pairs contribution to the total chi-square statistic for a city and sex.

26

Table 4 Chi-square test per pair, sex and city

Pair results for each city and sex, traditional net discrimination rate

Pair Com-

ple- ted

cases

Majo

-rity pref

Mino

-rity pref

Ob-

ser- ved

cases

Ex-

pec- ted

cases

Dif-

fer- ence

Contri-

bution to chi-

square

Com-

ments

c A b n e n-e Gb F BD 81 10 1 9 15 -5.5 2.111 BH 33 8 1 7 6 1.0 0.163 FD 56 14 2 12 10 1.7 0.289 FH 43 13 3 10 7 2.8 1.114 Mm F AB 62 13 8 5 6 -1.1 0.186 AD 59 11 1 10 6 4.2 2.970 CB 57 8 0 8 6 2.1 0.707 CD 46 3 4 -1 4 -5.2 6.394 Stm F BA 59 12 5 7 6 0.7 0.078 BD 45 4 4 0 5 -5.4 5.434 CA 9 0 0 excluded

from test CD 56 17 4 13 6 6.8 7.542 ID 17 5 2 3 2 0.9 0.386 JA 87 16 8 8 10 -1.9 0.358 Gb M CA 65 20 13 7 10 -2.7 0.732 CE 21 5 2 3 3 -0.1 0.004 CI 22 6 4 2 3 -1.4 0.594 GA 55 19 5 14 8 6.1 4.615 GE 28 7 1 6 4 2.1 1.138 GI 19 1 2 -1 3 -4.0 5.298 Mm M CA 52 11 3 8 8 0.4 0.019 CB 61 15 4 11 8 2.9 1.070 DA 41 7 3 4 6 -1.9 0.592 DB 45 6 1 5 6 -1.5 0.326 Stm M FE 48 16 4 12 7 4.8 3.246 FG 63 18 4 14 9 4.8 2.502 HE 79 14 6 8 12 -3.9 1.253 HG 42 2 2 0 6 -5.8 5.771 Sweden 1351 281 97 184 184

In the following table, a calculation of the significance of tester differences is presented. It can be concluded that significant deviations of the traditional NDR from the unweighted av- erage for the city and the sex occur in four of the six cases: Malmö female testers, Stockholm female testers, Gothenburg male testers, and Stockholm male testers.

Columns for the table below are as follows: Mean e: The mean of the expected difference between majority and minority preference cases across pairs of the given sex in the given city; Chi-sq: The value of a one-sample chi-square test between the observed and the ex- pected difference between majority and minority preference cases; Df: Degrees of freedom; Sig.: The level of significance of the chi-square test. Values of less than 0.05 mean there is at least a 95 percent chance the deviations of pair net discrimination cases in a city and sex from the average are not entirely due to chance; V: Cramer’s V = square root of a term calculated

27

as chi-square divided by the product of the number of cases and the number of columns or rows, whichever is smaller. The latter in our case is one making V identical to Phi.

Table 5 Chi-square on tester pair results per sex and city

One-sample chi-square tests of the pair results on the net number of discrimi-

nation cases

City Mean e

chi-sq df sig. V Commentary

Female Gothenburg 9.5 3.677 3 0.298 0.311 Malmö 5.5 10.256 3 0.017 0.683 not tester bias Stockholm 6.0 13.799 4 0.017 0.667 not tester bias Male Gothenburg 5.2 12.381 5 0.030 0.632 too few expected

cases Malmö 7.0 2.007 3 0.571 0.268 Stockholm 8.5 12.771 3 0.005 0.613

Significant deviations of the traditional NDR of one or more pairs from the average for the city and the sex occur in three of the six cases: Stockholm female testers, Gothenburg male testers, and Stockholm male testers.

If pairs worked in different months, this will also impact their results. Most of the work in Stockholm and Malmö was done between March and May when the labour market changes rapidly. The number of cases is far too small to test pair results on a monthly basis. If pairs worked in different sectors of the labour market, this may also affect their results. Again, the numbers are too small to test.

The Malmö female pair results are significantly different from the unweighted average but on closer scrutiny this does not appear to be due to tester bias. Together with minority tester D, majority tester A was preferred more often than majority tester C was with minority tester A, but with minority tester B, majority tester C was preferred more often than majority tester A was with B. Hence, no clear pattern emerges as to which tester was preferred more or less than the other. The same is true of Stockholm female tester pairs and there is no evident ex- planation for this pattern. It may be due to chance, it may be due to the timing of the testing, the positions being applied to, etc. To make a breakdown of these factors would not clarify the issue, since the groups on which to calculate would be too small. This casts a degree of doubt on the instances where the results seem to indicate tester bias. The pattern in the Stock- holm male results occurs partly at the phone stage, partly at the CV stage. There is no expla- nation for patterns occurring at the CV stage. For the same unknown reason, patterns may also occur at the phone stage. In fact, except for Malmö male testers, all CV stage results are significantly different from the unweighted average for the city and the sex. Statistical tests on the phone stage results are insignificant only for Malmö female testers.

When we have calculated excluding pairs with few cases, such as the Gothenburg male GI pair or the Stockholm female ID pair, the results remain virtually unchanged. Gothenburg male results are risky to perform a chi-square test on. The mean expected number of cases of net discrimination across the six pairs that were at work is only 5. For three of the six pairs the expected number is only 3, for one it is only 4. By all standards, therefore, a test should not be performed.

28

Stockholm male results show a uniquely interesting pattern of their own. The NDR is the dif- ference between the share of cases with majority preference and the share of cases with mi- nority preference. All three measures just mentioned were greater when majority tester F was involved than when majority tester H was, and they were greater when minority tester E was involved rather than G. Consequently, the FE pair has not only the largest NDR and the larg- est share of majority preference but also the largest share of minority preference, and the HG pair is lowest on all three. The difference between the majority testers has a larger impact than that between the minority testers. With a mean of 7 expected cases, Stockholm male re- sults permit chi-square testing of the CV stage by itself. The CV stage is where there is no direct interaction between employer and applicant. The interaction here is in writing, and yet a clear pattern emerges. The chi-square test yielding a value of 11.075 is highly significant (p=0.011). There is no known explanation for such differences occurring at the CV stage. Patterns occurring at the phone stage may be attributable to the same unknown factors. Given the unexplained differences at the CV stage, we really do not know what any differences ob- served at the phone stage may mean and there are insufficient cases to test for biases in net discrimination at the phone stage by itself.

Given the unexplained differences at the CV stage, we cannot draw conclusions as to what the differences at the phone stage may mean, but only conclude that they exist. Therefore, it is difficult to analyse if and to what extent they have influenced the overall results. Since only one similar study has been conducted in Sweden 15, we have to resort to the assumption that the unexplained differences are either due to chance, or to unidentified variables influencing the outcome, and not to tester bias per se.

15 Carlsson’s and Roth’s recently published correspondence testing study produced very similar overall results

29

6. Results of the testing

6.1 Results for Sweden as a whole

In the following section we present the results for Sweden as a whole, divided between men and women.

The study shows that there are notable inequalities in the Swedish labour market regarding access to employment for Swedish employment seekers with an immigrant background, and a foreign sounding name.

The most immediately understandable presentation of the findings is to show the differential between the number of times a ‘majority’ tester needs to apply and the number of times an immigrant origin tester needs to apply to get a positive response in the application process.

The table below summarizing the outcome of the Sweden study shows that there are signifi- cant differences between the majority and minority testers in terms of how many jobs they need to apply to before being preferred or chosen.

Table 6 Number of tries before being preferred, per city and sex

Number of tries necessary before being preferred

Gothenburg Malmö Stockholm Total

Female Male Female Male Female Male Female Male

Tries for the minority

26.4 7.0 11.9 18.4 7.2 10.8 11.1 10.4

Tries for the majority

4.1 3.3 5.0 5.2 3.6 4.1 4.1 4.1

Difference 22.3 3.7 6.9 13.2 3.6 6.7 6.9 6.3 Multiplier 6.4 2.1 2.4 3.5 2.0 2.6 2.7 2.6

On the high end are the females in Gothenburg, where the majority was chosen once every fourth time, whereas the minority applicant had to apply to over 26 job openings before being offered a chance. In Malmö the minority males had to apply to around 18 jobs before being preferred, whereas the majority needed to apply to around 5 jobs. Other results are a bit more moderate, with the lowest difference being for females in Stockholm, where the minority testers “only” had to apply to approximately double the number of openings to be preferred, when compared with majority females.

It is easy to comprehend the difficulties and the frustration of the individual applying for a job when it takes such effort to be chosen or preferred. To be repeatedly rejected takes its toll on self-confidence and motivation, which can be further illustrated by this transcription of paired telephone calls where unequal treatment was clearly manifested.

30

Stockholm – Female – Case 239 – Position as waitress in a restaurant

Call 1 Minority Minority tester: Hello, my name is Zainab XX Employer: Hello Minority tester: I am calling about your advertisement Employer: Wait a minute, I have to get a pen… ok, and your name was? Minority tester: Zainab XX Employer: Reina? Minority tester: No, Zainab! Employer: Can you please spell it? Minority tester: Z A I N A B Employer: Zainab, ok, and how old are you? Minority tester: I am 23 years old. Employer: Have you worked in a restaurant before? Minority tester: Yes, in bars and restaurants. Employer: And what did you do there? Minority tester: I worked as a waitress. I also went to hotel- and restaurant school. Employer: Ok, where do you live? Minority tester: Söder, in Stockholm Employer: Söder, ok, I will take your phone number so that I can get back to you. We have a lot of

applications already, but I make a note here that you have some experience. Minority tester: Yes… Employer: I will call you back when I’ve gone through all the applications. Minority tester: Do you want me to send in a CV or something? Employer: No, not yet, I call you back. Just give me your telephone number. (The tester gives her the telephone number and the phone call is completed) Call 2 Majority Majority tester: Hello, my name is Lovisa XX Employer: Yes hello Majority tester: I am calling about your advertisement Employer: Ok, how old are you Lovisa? Majority tester: I am 23 Employer: Have you worked in the restaurant industry before? Majority tester: Yes I have. In a restaurant in Norrköping and in a café. Employer: Have been serving food? Majority tester: Yes I have. Employer: Okay, where do you live? Majority tester: Södermalm Employer: Söder, okay. (Then asks the tester if she knows how to get from there to the restaurant. The tester tells the employer

that she has a car. Then the employer invites the tester to an interview the day after. An appointment

is made and the phone call is completed).

By considering the measure of cases before being given preference, it is also possible to cal- culate a multiplier, i.e. how many more times a minority applicant has to try before he or she is chosen. As the figure below shows, the most marked result is for Gothenburg. However, this spike does not influence the overall female results, which lie very close to the male re-

31

sults, due to a reverse situation in the other cities (i.e. minority males had to apply to more jobs in order to be preferred than minority females).

Figure 6 How many applications were needed to be given preference?

How many more times than a majority member does a minority member have to apply for a job in

order to be given preference once?

0,0

1,0

2,0

3,0

4,0

5,0

6,0

7,0

Gbg F Gbg M Mm F Mm M Sthm F Sthm M Female Male

6.1.1 Results at the first stage “Telephone call”

In the table below, the results for the telephone stage is presented. As the table shows, a total of 735 female cases and 696 male cases were conducted, , which means that a total of 1431 cases were started in Sweden.

Table 7 Results for Sweden at the phone stage

First stage – Telephone call Female Male

1 Total number of cases 735 696

a) No contact 57 25 b) No applicant accepted 21 51 c) Phone skipped - direct to CV 166 176 1.2 Total valid cases 657 620

1.3 Valid and usable cases phone stage (1.2 – 1c) 491 444 a) Both applicants accepted 411 345 b) Only the majority applicant accepted 41 59 c) Only the minority applicant accepted 25 28 d) dropped/lost cases 14 12 1.4 Net discrimination (1.3b-1.3c) 16 31

Net Discrimination Rate NDR (%) (1.4/1.2 x 100) 2 5

1.5 Continuing to next stage 577 521

32

A considerable number of cases were initiated directly by submission of a CV, i.e. without first calling the employer, as some employment ads specified NOT to call, but to send a CV directly instead. Researchers decided to respond to these announcements as well since they reflect common application procedures in Sweden today. A total of 166 cases for the females and 176 for the males were conducted in this way. Thus the total number of valid and usable cases at the telephone stage of the testing was 491 for the women and 444 for the men (also excluding the cases which were unusable and invalid).

As the table shows, the results at this stage of the testing were fairly low, with a net discrimi- nation rate of 2% for the women and 5% for themen. When looking at the results and exclud- ing the dropped cases in the calculation, the results do not change to any significant extent; the only change is that the NDR for womenincreases one percentage point to 3%

One may question whether the high rate of direct to CV cases may have had an influence on the results, since it is not known how these employers would have reacted to a phone call. If we distinguish between cases that were completed at the phone stage, cases that went through the phone stage but were completed either at the CV or the interview stage, and cases that skipped the phone stage, an interesting pattern emerges. Among completions at the phone stage, majority preference was 1.9 times as likely as minority preference. Among later-stage completions, majority preference was 3.9 times as likely as minority preference. Among cases skipping the phone stage, majority preference was 2.8 times as likely as minority pref- erence, which is very close to the overall result of 2.6 times. In other words, the average across all cases beginning at the phone stage was almost the same as the average across cases skipping the phone. This suggests that, had the cases in which the phone stage was skipped also included the phone stage, the results would have been affected very little, if at all.

Though the results at this stage signal fairly equal treatment of the minority in respect to the majority testers, there is still a difference in outcome. During the calls, there were also a number of situations in which discrimination showed itself in a direct manner Note the fol- lowing transcript from a test conducted in Malmö:

Malmö – Female – Case 85 – Position as a waitress in a restaurant

Call 1 Minority Minority tester: Hello, my name is Jasmina XX and I am calling about the job as a waitress. Employer: Eh, sorry to say, but the vacancy is filled. Minority tester: Okay. Employer: Yes, thank you. Minority tester: Thank you bye. Employer: Bye. Call 2 Majority Majority tester: Hello my name is Carolina. Employer: Hello! Majority tester: I am calling about the job as a waitress that you are advertising for. Employer: Yes, have you experience? Majority tester: Yes I have, one year. Employer: As a waitress in a restaurant? Majority tester: Yes. Employer: Okay, eh… what else do you do?

33

Majority tester: I am from Småland, but moved to Malmö a month ago. I studied one semester but I want to work again, I don’t have a job at the moment.

Employer: Eh… so you are not up to anything else at the moment? Majority tester: No. Employer: Can you come in here tomorrow? Majority tester: Yes, I can do that! Employer: Okay, that’s nice! Majority tester: Okay, bye Employer: Bye

This case clearly demonstrates unequal treatment, as the employer resorted to saying that the vacancy was filled in order to terminate the call from the minority tester. In other cases, the manner of dismissal was more subtle and difficult to discern. However, the following ex- change with the majority tester brought these latent attitudes into bright relief.

Stockholm – Male – Case 179 – Position as a waiter in a restaurant Call 1 Minority Minority tester: Hello, my name is Junus XX Employer: Hello Minority tester: I am calling about your advertisement for a waitress or waiter Employer: It’s waitress in your case, is it? Minority tester: No, its waiter *laugh* Employer: We are looking for a girl Minority tester: You are looking for a girl? Employer: Yes Minority tester: Ok, yes but… Employer: Sorry Minority tester: Well, that’s a pity Employer: Yes, but we actually do! Minority tester: Ok, thank you Employer: Thank you, bye Call 2 Majority Majority tester: Hello, my name is Martin Employer: Hello Majority tester: I am calling about your advertisement for waiters and waitresses Employer: Yes, ok Majority tester: Are you the right person to speak to? Employer: Yes absolutely, what did you have in mind? Majority tester: Well I want to send in an application for the job, but I have some questions about the

salary. Employer: Ok, but you should talk to the restaurant manager about that, it depends on your previous

work experience and so on. Majority tester: Ok, I understand, but I will send in my CV then. Employer: Yes, and what was your name? Majority tester: Martin XX Employer: Yes, Martin. Great Martin, send in your CV. The Manager will read it and get back to you. Majority tester: That’s nice! Thank You. Employer: Yes, thank you and bye, bye! Majority tester: Bye!

34

In this case, the minority tester appears to have been immediately dismissed on the equally illegal grounds of gender. However, the majority tester’s experience reveals this be to have been a false pretext.

However, there were also cases in which unequal treatment was to the disadvantage of the majority caller, as the following case illustrates:

Stockholm – Female – Case 56 – Position as a shop assistant Call 1 Minority Minority tester: Hello, my name is Zainab XXXX, I am calling about your advertisement for a shop

assistant Employer: Yes that’s right Minority tester: I wanted to ask you some questions about the post Employer: But can you please send in a CV like every body else? Minority tester: Yes of course, I can do that! Employer: Do that, that’s easier, because I really don’t have time to talk to you on the phone right

now Minority tester: You don’t? Employer: No, we have clients here at the moment, so it’s much easier if you just send it in. Minority tester: Okay, but can you please tell me when I can expect an answer from you? Employer: No, it says clearly that you are supposed to send in your application by E-mail. Minority tester: Yes, but I just wonder when I can expect an answer. Employer: As soon as I get it, I will read it and get back to you if it something for us Minority tester: Ok, bye Employer: Bye Call 2 Majority Majority tester: Hello, my name is Anna Employer: Is it about the job advertisement you are calling? Majority tester: Yes, exactly. Employer: Ok, did you send in your CV already? Majority tester: No not yet, I just wanted to talk to you about the job first… Employer: Okay, we have already got a terrible amount of applications, so we decided not to accept

any more applications now. We have got more than 100 applications by now. Majority tester: Oh, so there is no need for me to send in one then, is it? Employer: No, it’s not. We have already got too many of them. Majority tester: Yes, I understand, but thank you anyway Employer: Yes, thank you, and bye Majority tester: Bye

In this case, the employer tells the second caller that they have a “terrible” number of applica- tions., It is clear in both calls that the employer is stressed about the fact that several job seekers are calling and sending in their CV.

In other cases, the unequal treatment of the majority caller is more difficult to explain or un- derstand:

35

Stockholm – Male – Case 51 – Position as a delivery driver Call 1 Minority Minority tester: Hi, my name is Yashar and I’m calling about the job advertisement. Employer: Okay… Minority: Delivery driver, right? Employer: Yeah, that’s right. Minority: Is the job still available? Employer: Yes, it’s still available. Minority: Okay, I have a few questions. The salary, how much is it, more or less? Employer: Have you worked as a delivery driver before? Minority tester: Yes, I have. Employer: Okay, for what company? Minority: “Bud centre” and “Jet Pak”. Employer: “Jet Pak”? Minority tester: Yes. Employer: Okay, the (not possible to understand)? Minority tester: Pardon me? Employer: Which, Jet Pak? Minority tester: Yes, exactly. Employer: Okay, for how long did you work there? Minority tester: At Jet Pak I worked for… about three or four months. Employer: Okay, and why did you quit? Minority tester: I don’t know, I wanted something new. They didn’t give me many work-hours. You

know, extra hours. Employer: Okay. Minority tester: But I see that your job is full-time. Employer: Okay, you’re looking for full-time. Minority tester: Yes. Employer: Actually, we have full-time jobs. Minority tester: Okay, and how do I do now, do I send my CV, or… Employer: Yeah, sure, but you wanted to know about the salary, right? Minority tester: Yes, exactly. Employer: It’s about 16 to 17 per month. Minority tester: 16 to 17 thousand? Employer: Yes, exactly. Minority tester: Okay, that sounds great. Do I just send my CV or how do we..? Employer: Yes, you can… where do you live? Minority tester: I live in Stockholm. Employer: Yeah, I know you live in Stockholm, where in Stockholm? Minority tester: Stockholm city. Employer: Stockholm city? Minority tester: South side. Employer: South side, okay. And, how old are you? Minority tester: 24. Employer: Okay then. Send us an application and we’ll have a look at it, to see where you’ve worked

and so on… And then we’ll call you. Minority tester: Okay, just one more question. How long does it take before you reply, do you know? Employer: Well, you know, we have a lot of applicants, so… Minority tester: Well, I’ll just hope for the best then. Employer: Yeah, okay. Minority tester: Okay, thanks.

36

Call 2 Majority Majority tester: Hello, my name is Martin. I’m calling about the job as a delivery driver. Employer: Yes? Majority tester: Is the job still available? Employer: No, it isn’t. Majority tester: Okay, thank you. Employer: Thank you, bye.

In this case the second caller is promptly rejected by the employer, who does not even at- tempt to start a conversation, as with the first caller Here, there was a clear difference in the length of the phone call and the employer selection. However, in most cases the length of phone call does not necessarily determine the outcome. During testing, the length of tele- phone calls was routinely recorded and later analysed for possible correlations with ethnicity or employer preference. No consistent correlations were found.

37

6.1.2 Results at second stage “Invitation to an interview”

At the second stage of testing, the outcome could be either that neither was invited to inter- view, that one of the two testers was invited, or both were. As the table shows, the results produced a differential outcome of the testers’ applications, with a net discrimination rate of 12% for women and 10% for men.

Of all the submitted applications: only 13% (39+34) for the women, and 6% (16+20) for the men resulted in both applicants being invited to an interview. The results at this stage demon- strate one of the peculiarities (and difficulties) of the Swedish labour market, i.e. that it is difficult for any young inexperienced person to enter the labour market. Even so, it appears to be even harder for a person with an immigrant background to get a job. A cumulative net dis- crimination rate at 14% for women and 15% for men demonstrates that there was a difference in treatment not readily explained by variables other than discriminatory behaviour on the part of the employer.

Table 8 Sweden as a whole – Results at the second stage

Second stage – Invitation to an interview Female Male

2 Usable cases 577 521

a) Both rejected 77 77 b) No reply to CV 307 283 c) Both testers invited 39 16 d) Only the majority tester invited 88 86 e) Only the minority tester invited 11 24 f) CV skipped - interview direct 34 20 g) dropped/lost 19 15 2.1 Net discrimination (d-e) 77 62

Net discrimination Rate NDR (2.1/1.2 ) % 12 10

Cumulative Net Discrimination Rate % 14 15

Continuing to next stage (c+f) 73 36

Of all the cases that went on to stage two, less than half led to an actual outcome, whether it was that both were rejected, one invited, or both invited. The percentage of cases where no response was given at all to CVs averaged 60 percent. Consequently only a few of these cases went on to an interview.

If we look at the results and exclude the cases which were dropped or lost at some stage in the testing process the NDR for the second stage increases by one percentage point for women and men respectively.

However, if we also exclude all the cases where no response was given whatsoever to the CV, the NDR increases dramatically, to 29 percentage points for women and 21 percentage points for the men. These figures correspond very well to the correspondence testing16 which has been carried out by Carlsson and Rooth during 2006, where the overall NDR was calcu-

16 Correspondance testing is conducted by submitting applications in written form from a majority and a major- ity with similar qualifications, and then calculating the difference in outcome.

38

lated to be 29.6 percentage points (when using an equivalent mode of calculation). In the test- ing conducted concurrently in France, this more rigorous mode was used to arrive at the NDR, by discounting cases where there was no clear demonstration of employer considera- tion. Annex I presents the results of this study calculated by the alternative approach.

The results at stage two present some challenges to the interpretation of the results. In other countries where testing has been performed, the larger share of discriminatory behaviour took place at the first stage, a telephone call, whereas in Sweden the larger share of the discrimina- tion clearly takes place at the second stage, an invitation to interview.

According to information from employers who were given a follow-up call, as many as 300 applicants may submit their CVs for the same job. Employers have little opportunity to per- sonally review all of the potential applicants. Since only a few of cases went onwards to stage three, the offer of a job, the overall basis for calculating the Net Discrimination Rate, NDR, is lower in Sweden than in other countries where testing has been conducted. In the report we therefore also present data in ways which attempt to illustrate the difficulties for entering the labour market (See Chapter 6.4)

6.1.3 Results at third stage “Offer of a job”

For the women, a total of 73 cases continued beyond stage two. For men, however, only 36 cases reached that point. As the table below shows, several of the cases had to be dropped at this stage due to a number of factors, such as employers asking for references or exam papers or the testers’ unavailability at the time of the interview. Since the number of cases is too low to give any sense in terms of net discrimination rate for this stage, only the cumulative net discrimination will be presented.

Table 9 Sweden as a whole – Results at the third stage

Third stage - Offered a job Female Male

3 Usable cases 73 36

a) Interviews did not take place 45 20 3.2 Interviews took place 28 16

a) Job not offered to any of tester 5 6 b) Job offered only to majority tester 5 2 c) Job offered only to minority tester 7 2 d) Job offered to both testers 11 6 3.3 Net discrimination (b-c) -2 0

Cumulative Net Discrimination Rate (%) 14 15

What the results show is a similar pattern as in other countries, namely that the discrimination at the interview stage is often lower. Despite the small number of cases, it appears that in some instances, for example in Malmö, there is actually a negative result for discrimination at the last stage. Logically, this makes sense since an employer who have already made the de- cision to invite an applicant for an interview, presumably does so with the intention to possi- bly employ the individual in question.

When we look at the results and exclude cases that at some point in the testing have been dropped or lost, we arrive at a net discrimination rate of 16 percentage points, for women as

39

well as men. If only cases receiving a response are used, the overall cumulative NDR in- creases to 34 percentage points for women and 32 percentage points for the men. See Annex II for a complete presentation of the results using this calculation.

6.2 Results for the three cities

In the table below, results for the three cities are presented, divided between men and women. The results show that the situation is quite similar in the three cities, with only the female testing in Gothenburg showing a higher degree of discrimination than the others. Regarding the results for men, the cumulative NDR is nearly the same, and it is only among the women that there appears to be some regional variation.

Table 10 Results for the cities

Göteborg Malmö Stockholm

F M F M F M

First stage - CV sent to employer Total number of cases 221 222 231 215 283 259 No contact 12 8 23 6 22 11 No applicant accepted 4 8 7 16 10 27 Phone call skipped 27 58 31 38 108 80 Valid and usable cases phone stage 178 148 170 155 143 141 a) Both applicants accepted 157 107 148 125 106 113 b) Only the majority applicant ac- cepted

8 22 12 21 21 16

c) Only the minority applicant ac- cepted

4 12 5 7 16 9

d) Dropped/lost 9 7 5 2 0 3 Net discrimination (b-c) 4 10 7 14 5 7 Net Discrimination Rate (%) 2 5 4 7 2 3

Cases continuing 184 165 179 163 214 193 Second stage – Invite to interview Useable cases 184 165 179 163 214 193 a) Both testers rejected 24 30 24 19 29 28 b) No reply to CV 104 70 108 115 95 98 b) Both testers invited 16 9 9 1 14 6 c) Only the majority tester invited 36 36 23 16 29 34 d) Only the minority tester invited 3 14 3 4 5 6 e) Dropped/lost 1 2 4 4 16 9 f) CV skipped – interview direct 0 4 8 4 26 12 Net discrimination (c-d) 33 22 20 12 24 28 Cumulative NDR 37 32 27 26 29 35 Cumulative NDR (%) 17 15 11 13 10 15

Cases continuing 16 13 17 5 40 18 Third stage - Offered a job Useable cases 16 13 17 5 40 18 Interviews did not take place 14 9 8 1 23 10

40

a) Job not offered 1 2 1 2 3 2 b) Job offered only to majority tester 1 0 0 2 4 0 c) Job offered only to minority tester 0 1 5 0 2 1 d) Job offered to both testers 0 1 3 0 8 5 Net discrimination (b-c) 1 -1 -5 2 2 -1 Cumulative NDR 38 31 22 28 31 34 Cumulative NDR (%) 19 16 11 15 12 16

The results per city also highlight that the largest amount of discrimination seems to have occurred at the CV stage, where the employers have chosen whom to invite for an interview. As discussed in the previous section this indicates that for employment procedures in Swe- den, the telephone stage is not used as a screening instrument as in other countries. Instead most of the screening is done on the basis of the CV.

6.3 Results by sectors

In the analysis of the results, we have also divided the outcome of the testing according to the primary sectors tested. As has been described in the methodology section, two sectors were chosen as targets for both men and for women: wholesale and retail trade and hotel-restaurant work. In addition, as a third concentration, it was decided to conduct a significant number of tests in the transport-warehouse sector for men and in health care for the women. However, there were not sufficient job openings in these latter specific sectors so the results are pre- sented with an aggregate mix of sectors17 in the category “other services”.

In the table below, the first stage results are shown, disaggregated according to which sectors the cases belonged to. It is clear that the sector wholesale/retail and trade (ISIC 46-53) is dominating the total sample, with over 300 cases for both men and women. Several of the openings within this sector were telemarketing,, with low salaries sometimes purely based on commission Another interesting feature of the testing by sectors is to note the high number of cases within “wholesale, retail and trade” in which the phone call was skipped, i.e. where the applicants were explicitly asked not to call the employer before sending in an application.

17 “Other Services” nonetheless mainly include warehousing/transport for males and for females office/clerical and health/social sectors

41

Table 11 Results by sector, first stage

These sectoral findings are similar to the overall results, apart from the male testing within Hotel and Restaurant (ISIC 55/56). The NDR for males seeking employment within this sec- tor reaches 10%, which is clearly higher than the overall results, and also higher than the male results in other sectors and the female results within Hotel and Restaurant at the first stage. However, these results are not statistically significant, and can only be called upon to demonstrate a tendency.

At the second stage of testing, the invitation to an interview, the same pattern emerged, and the results for the male testing within Hotel and Restaurant reach a cumulative NDR of 22 percent. It is also worth noting that at the second stage, the female testing within “Other Ser- vices” reaches 18 percent NDR, which is considerably higher than other sectors, although also not statistically significant.

First stage – Phone call/CV sent to employer

Whole- sale/Retail/Trade

Ho- tel/Restaurant

Other services (incl health)

Female Male Female Male Female Male

1 Total number of cases 325 346 215 160 191 158

a) No contact 26 9 16 9 15 7 b) No applicant accepted 2 11 18 26 1 14 c) Phone call skipped 79 96 55 30 32 41 1.2 Valid and usable cases

phone stage

218 230 126 95 143 96

a) Both applicants accepted 187 188 91 60 131 83 b) Only the majority appli- cant accepted

20 25 18 22 3 6

c) Only the minority appli- cant accepted

8 12 13 10 4 4

d) Dropped/aborted cases 3 5 4 3 5 3 1.3 Net discrimination (b-c) 12 13 5 12 -1 2

Net Discrimination Rate

NDR %

4 4 3 10 -1 1

1.4 Cases continuing 266 284 146 90 163 124

Both applicants accepted 187 188 91 60 131 83 Phone call skipped 79 96 55 30 32 41

42

Table 12 Results by sectors, second stage

Second stage – Invitation to an

interview

Wholesale / Re-

tail / Trade

Hotel / Restau-

rant

Other services

(incl health)

Female Male Female Male Female Male 2 Valid and usable cases 266 284 146 90 163 124

a) Both testers rejected 31 43 15 9 30 19 b) No reply to CV 140 146 78 51 88 74 b) Both testers invited 28 10 3 3 8 2 c) Only the majority tester invited 33 49 23 16 32 19 d) Only the minority tester in- vited

7 17 3 1 1 5

e) Dropped/aborted 10 10 6 1 3 3 f) CV skipped - interview direct 15 9 18 9 1 2 2.1 Net discrimination (c-d) 26 32 20 15 31 14

Net discrimination at the stage (%)

9 10 11 12 18 10

Cumulative discrimination 38 45 25 27 30 16

Cumulative NDR (%) 13 14 14 22 17 12

2.2 Cases continuing 43 19 21 12 9 4

The second stage of the testing shows quite clearly that there are some differences with re- gard to the outcome of the tester’s applications, depending on which sector they are applying in. The differential treatment appears to be larger for openings within “Other Services” for women, and within “Hotel and Restaurant” for men.

Table 13 Results by sectors, third stage

Third stage – Offered a job Wholesale / Re- tail / Trade

Hotel / Restau- rant

Other services (incl health)

Female Male Female Male Female Male

3 Valid and usable cases 43 19 21 12 9 4

a) Interviews did not take place 27 10 13 6 5 3 b) Job not offered 1 2 2 3 2 1 c) Job offered only to majority tester

3 2 2 0 0 0

d) Job offered only to minority tester

5 1 1 1 1 0

e) Job offered to both testers 7 4 3 2 1 0 3.1 Net discrimination (c-d) -2 1 1 -1 -1 0

Cumulative discrimination 36 46 26 26 29 16

Cumulative discrimination (%) 12 14 14 21 17 12

43

At the third stage no change to this pattern occurs, as can be seen in the table above. The NDR for each sector is similar to the overall percentage, apart from “Hotel and Restaurant” for the men and “Other Services” for the women.

When we look at the testing by the number of tries before being given preference, the differ- ential treatment of women seeking employment within “Other Services” is prominent. Ac- cording to this measure, a minority female would have to apply to five times as many jobs as a majority female, in order to be given preference.

Table 14 Results by sectors, number of tries before being given preference

6.4 Results of Sweden in relation to previously conducted testing

The testing methodology is not designed to compare the level of discrimination in different countries. Such comparisons are highly speculative because of the many factors specific to national labour markets, including legal context, labour market conditions, immigration his- tory and groups tested, and so on.

Nonetheless, it may be instructive to refer to the outcomes of other country studies with the goal of addressing discrimination as a common phenomenon.

As the table below shows, the experience of ‘minority’ testers in Sweden was quite similar to that of other countries in terms of the number of tries necessary for a ‘minority’ tester to be given preference or chosen, relative to a ‘majority’ tester.

Table 15 International references – cases per being given preference 18

Job applications per case of being given preference, young males by city in four countries (including cases of both applicants being rejected at the phone stage) Tries per case of being

preferred

Multiple

Difference

Majority Minority

Gothenburg 3.2 6.7 2.1 3.5 Malmö 4.8 16.9 3.5 12.1 Stockholm 3.6 9.5 2.6 5.9 Brussels 1.6 6.6 4.1 5.0 Antwerp 1.8 7.5 4.1 5.6 Liege 1.3 3.1 2.4 1.8 Barcelona 1.5 13.1 8.6 11.6 Madrid 2.1 9.9 4.6 7.8

18 The testing was conducted during the following time periods Spain: September 1994 to January 1995; Bel- gium: Autumn 1995 to spring 1996 and autumn 1996 to spring 1997; Italy: February to June 2003; Sweden; November 2005 to August 2006

Wholesale /

Retail / Trade

Hotel / Restau-

rant

Other services

(incl health)

Female Male Female Male Female Male

Minority tries 9.5 9.2 8.8 10.1 23.3 15.8 Majority tries 4.1 3.9 3.8 3.5 4.5 5.7 Difference 5.4 5.3 5.0 6.5 18.8 10.1 Multiple 2.3 2.4 2.3 2.9 5.1 2.8

44

Malaga 1.3 21.0 16.0 19.7 Turin 1.5 5.1 3.3 3.5 Rome 2.0 21.8 11.0 19.8 Naples 2.0 6.3 3.1 4.2

What is also notable is that it takes a considerable effort to find employment in Sweden, even for majority jobseekers. The majority testers in Sweden had to apply to nearly double as many jobs as majorities in other countries in order to be given preference. While the study cannot offer an explanation, this may be due to a weak labour market with relatively high unemployment at the time of testing.

In terms of the original testing methodology, Sweden shows a significant level of discrimina- tion, yet it appears lower than results obtained in other countries. The overall standard cumu- lative NDR for women was 14%, and 15% for men, as compared with NDR’s of between 25% and 40% in other countries.

This difference is particularly notable at the first stage of application, the telephone call. In other countries, rates ranged from 14% in Denmark to above 30% in the Netherlands and Spain. In Sweden the rate was found to be only 5%.

However, at the second stage, results in Sweden are close to results in other countries. Also, when all cases of “no response” to written applications are excluded, overall Swedish results are similar to those in other countries. As noted above, these calculations revealed an NDR of 32% for women and 34% for men.

In contrast to other countries, Swedish employers do not rely on initial telephone contact as a screening instrument, but encourage candidates to send in CVs. However, employers also reported to have received up to 400 applications for one position. The earlier testing experi- ence in Denmark showed a similar pattern with a relatively low NDR at the first stage of test- ing (14 percent), whereas in other countries the discrimination rate was highest in the first stage.

45

7. Analysis and Initial Conclusions

A few initial conclusions can be drawn from the findings of this study. These conclusions may be helpful in considering policy options to address differential treatment of natives of immigrant origin, as well as immigrants, in access to employment.

7.1 What do the results say about inequalities in access to employment?

The study shows that there are inequalities in the Swedish labour market regarding access to employment for Swedish employment seekers with an immigrant background and a foreign sounding name.

Following the standard ILO methodology comparing the experiences of ‘majority’ and ‘mi- nority’ testers, statistically significant differences were registered.

This testing found a net discrimination rate of 14% for women and 15% for men. If analysis is limited to cases in which a response is given, the overall cumulative NDR increases to 34% for women and 32% for the men.

These results are consistent with a similar study conducted by Carlsson. M and Rooth D.O (2006) using so called correspondence testing19 to test access to employment for male minori- ties with an Arabic-sounding name. Given that correspondence testing is roughly equivalent to stage two in the practice testing, we can conclude that the results of the two tests are simi- lar, with the NDR of 9.6% in Carlsson and Rooth’s study corresponding to a second stage NDR of 10% in this study (using a similar mode of calculation. i.e. using all cases in the de- nominator for calculation of NDR at the written application stage). The similarity in outcome of these two studies reinforces the validity of their respective results in the Swedish context.

In the correspondence testing, the NDR reached a level of 29.4% when using only cases in which at least one applicant was accepted or chosen. As noted above, the ILO study found a NDR of 34% for women and 32 for men when excluding similar cases for which, no response was obtained.

Even more dramatic differences are evident when application experiences between ‘majority’ and ‘minority’ testers are compared. In Gothenburg, the majority applicant was preferred once every fourth time, whereas the minority applicant had to apply to over 26 job openings before being offered a chance. In Malmö, the minority men applied to some 18 jobs before being preferred, while the majority applicants only needed to apply to 5 jobs. The lowest dif- ference, for women in Stockholm, still revealed that minority testers had to apply to approxi- mately double the number of openings of majority women.

19 Correspondence testing is conducted by submitting applications in written form from a majority and a major- ity with similar qualifications, and then calculating the difference in outcome.

46

7.2 Differences among sectors

The results show differences in behaviour across different economic sectors. Comparing re- sults with background data on which sectors predominantly employ persons of immigrant origin or identity, those sectors with a higher proportion of ‘minorities’ show higher rates of discrimination. In other words, a higher level of discrimination was observed in the sectors that presented a statistically significant prevalence of foreign born in the labour force.

On the other hand, foreign-born applicants are occasionally favoured over the native-born in the Hotel and Restaurant sector, as well as in other private services. The third sector tested, Retail and Trade, showed a significant difference favouring the ‘majority’, yet also presented the lowest overall net discrimination rate.

These results are consistent with the research conducted by Carlsson and Rooth (2006), where it was concluded that occupations with a higher percentage of minority workers were inclined to favour majority applicants in the employment processes. As mentioned before, this study does not explore the grounds for differential treatment, so no attempt is made to explain this phenomenon. The Carlsson and Rooth study discusses whether the predominance of minorities in certain sectors is due to a lack of majority candidates willing to undertake the work in question, thereby creating a situation where the majorities are favoured should they apply.

It is often stated that any profession is better than no profession, even though this might mean working below one’s formal competence level. However, as testing has demonstrated, access to employment is by no means easier when applying to the sectors where immigrants consti- tute a significant proportion of the labour force. For an individual seeking employment, it must be extremely discouraging to be repeatedly rejected even for jobs below his or her level of competence.

7.3 Differences among cities

The testing did not show significant differences among cities with regard to the net discrimi- nation rates. However, there are some notable differences concerning overall labour market conditions that affect both majority and minority employment seekers.

This holds true specifically in Malmö, where employment levels are generally lower for both majorities and minorities, and where the testers needed to apply to considerably more vacan- cies to be given preference. This indicates a problem on the demand side of the labour mar- ket, which equals out the net discrimination rate, since it appears to be harder for both a ma- jority and the minority to be given preference.

7.4 What does testing say about the labour market for young job-seekers?

A characteristic of the Swedish testing was been that both testers, the majority as well as the minority, had difficulties in being preferred or chosen by employers. In a large majority of cases, neither of the paired applicants was chosen for an interview.

It took considerable efforts for both the majority and the minority tester to be given an oppor- tunity to continue the application process. This was in a context in which all testers had as- sumed an identity with relevant experience to the position in question, had an address in a

47

well-regarded mixed neighbourhood and had completed high school. Even so, their chances of obtaining jobs that did not require any formal skills or experience were very small.

These results are consistent with the employment level of youths in Sweden, which is one of the European countries with the highest youth unemployment “gap”20 between the ages of 20 to 24.

The testing demonstrates that youth in Sweden without high qualifications have great diffi- culties accessing employment which extend beyond the purview of this study. There seems to be a clear mismatch between labour market supply and demand within low- and semi-skilled occupations.

7.5 Reference to other countries

This testing methodology is not designed to compare levels of discrimination between differ- ent countries. Comparisons are not valid since many factors specific to each national labour market affect outcomes.

Nonetheless, the experience of minority’ testers in Sweden was quite similar to that of other countries in the differential number of tries necessary for a ‘minority’ tester and a ‘majority’ tester to be given preference or chosen.

On the other hand, while the Swedish results show a significant level of discrimination, they appear lower than results obtained in other countries, particularly at the initial application stage.

This study measures the existence of unequal treatment, but cannot determine the grounds for it. A number of factors may explain, individually and in combination, the differences among outcomes in different countries, and within Sweden. These factors include the specific proc- esses and characteristics of national labour markets. The study results coincide with other data showing the Swedish labour market to be particularly tight for youth employment.

A second factor is the time in which studies were done vis-a-vis general economic conditions. Several of the studies were conducted during periods when national economies were experi- encing growth and relatively lower unemployment rates than was the case in Sweden for this study.

Thirdly, the immigration situations and target groups for testing vary among countries. For example, the studies in Italy and Spain addressed male immigrants in countries where immi- gration is a recent phenomenon.

A possible reason for a lower rate in Sweden

A fourth factor is the extent and duration of anti-discrimination legislation, the existence and effect monitoring and enforcement mechanisms, and the extent of efforts and activities to change public and employer attitudes.

Sweden now has a considerable history of legislation, institutions and awareness raising ef- forts by government, social partners and other actors to combat discrimination. For example,

20 OECD Employment Outlook

48

the Swedish Ombudsman Against Ethnic Discrimination is currently celebrating its 20th year of existence.

In other countries, little or none of these elements existed at the time of testing, let alone prior to it. While it is difficult to draw firm conclusions from the outcome in Sweden as related to other countries, recent history of anti-discrimination measures may have had some effect in changing behaviours in the Swedish labour market.

49

8. Some perspectives on the future

The target group for this testing study was relatively small in terms of the number of people within that the age group born in Sweden. However, this group of Swedes is growing as more and more members of the second-generation reach employment age. If account is taken of those who arrived at an early age and are now in school, a steadily growing stream of people of immigrant background will soon be seeking access to employment.

The outcome of this study reinforces concerns articulated in other studies and government commissioned investigations: marginalisation in Sweden remains a serious challenge, par- ticularly as large groups of native youth of immigrant origin and foreign born youth who ar- rived long ago now enter the labour market. Unless young Swedes of immigrant origin are given equal opportunities and access to employment, there is a risk of increasing social ex- clusion and decreasing cohesion in society.

These tendencies have already been observed. For example, the concept “Young outsiders” (SOU 2003:92), has recently been used to designate a considerable number of youth, mainly of foreign origin or background, which are considered to live completely outside of Sweden’s security systems. Powerful measures are required to reverse this exclusion; otherwise a con- siderable proportion of a generation of able and potentially productive citizens are at risk of being lost and marginalised.

While the legal and institutional framework against discrimination in Sweden may be consid- ered quite well developed, its implementation in terms of training, monitoring, promotion of good practices and other aspects may be inadequate and can certainly be improved. Enforce- ment is lacking and sanctions against those found to discriminate are relatively insignificant.

The findings of this study, prepared by the independent research agency Rambøll Manage- ment, will permit the ILO to recommend even stronger measures to promote equal opportu- nity in Sweden. In doing so, the ILO can be expected to draw on its unique and extensive knowledge base and experience in the field of discrimination and equality in employment.

50

9. References

Allasino, Enrico; Reyneri, Emilio; Venturini, Alessandra; Zincone, Giovanna (2004) Labour Market Discrimination Against Migrant Workers in Italy, International Migration Paper 67 (Geneva: ILO)

Arai, Mahmood; Schröder, Lena; Vilhemsson, Roger (2000) En svartvit arbetsmarknad – en ESO-rapport om vägen från skola till arbete. Expertgruppen för studier i offentlig ekonomi (ESO), Ds 2000:47.

Arrijn, Peter; Feld, Serge; Nayer, André et al (1998) Discrimination in access to employment on grounds of foreign origin: the case of Belgium, International Migration Paper 23E (Ge- neva: ILO)

Åslund, O.; Nordström Skans, O. (2005) Ses vi på jobbet? Etnisk segregation mellan arbets- platser 1985 – 2002. IFAU Rapport 2005:17

Will I see you at work? Ethnic workplace segregation in Sweden 1985-2002. IFAU Report 2005:24

Åslund, Olof; Rooth, Dan-Olof (2006) Utbildning och kunskaper i svenska – framgångsfaktorer för invandrade? SNS förlag, Stockholm.

Behtoui, Alireza (2006) Om de hade på rätt plats födda föräldrar. Om ungdomar med utländsk bakgrund i det svenska utbildningssystemet och på den svenska arbetsmarknaden, expertbilaga till Rapport Integration 2005.

Behtoui, A. (2004) “Unequal Opportunities for Young People with Immigrant Backgrounds in the Swedish Labour Market” in Labour Vol. 18, No. 4, 633-660

Bovenkerk, F. (1992) Testing Discrimination in Natural Experiments. A Manual for Interna- tional Comparative Research on Discrimination on the Grounds of “Race” and Ethnic Origin (Geneva: ILO)

Bovenkerk, Frank; Gras, M. J. I.; Ramsoedh, D. et al. (1995) Discrimination Against Migrant Workers and Ethnic Minorities in Access to Employment in the Netherlands, International Mi- gration Papers 4 (Geneva: ILO)

Carlsson, Magnus; Rooth, Dan-Olof (2006) Evidence of Ethnic Discrimination in the Swedish Labor Market Using Experimental Data, Discussion Paper No. 2281. (Bonn: Institute for the Study of Labour IZA)

Cediey, Eric; Foroni, Fabrice (2007) Les Discriminations à raison de « l’origine » dans les embauches en France : Une enquête nationale par tests de discrimination selon la méthode

du BIT, International Migration Paper 85 (Geneva: ILO)

51

Edin, Per-Anders; Åslund, Olof (2001) Invandrare på 1990-talets arbetsmarknad, Forskningsrapport 2001:7. (Uppsala: The Institute for Labour Market Policy Evaluation IFAU)

Ekberg, Jan (1997) “Hur är arbetsmarknaden för den andra generationens invandrare?” in Arbetsmarknad och arbetsliv, Vol. 3, pp. 5-16

Ekberg, Jan; Hammarstedt, Mats (2002) “20 år med allt sämre arbetsmarknadsintegrering för invandrare” in Ekonomisk Debatt, Vol. 30, no. 4.

Ekberg, Jan; Rooth, Dan-Olof (2003) “Unemployment and Earnings for Second-Generation Immigrants in Sweden: Ethnic background and parent composition” in Journal of Population Economics, Vol. 16, No. 4, pp. 787-814.

Hammarstedt, Mats; Palme, Mårten (2006) Intergenerational mobility, human capital trans- mission and the earnings of second-generation immigrants in Sweden, Discussion Paper No. 1943. (Bonn: Institute for the Study of Labour IZA)

Hammarstedt, Mats (2002), “Hur försörjer sig andra generationens invandrare?” in Ekono- misk Debatt Vol. 30, no 5.

Le Grand, Carl; Szulkin, Ryszard (2002) “Permanent Disadvantage or Gradual Integration: Explaining the Immigrant-Native Earnings Gap in Sweden” in Labour, Vol. 16, no. 1, pp 37- 64.

Lundh, Christer; Bennich-Björkman; Li, Ohlsson; Rolf, Pedersen; Peder J.; Rooth, Dan-Olof (2002) Arbete? var god dröj! Invandrare i välfärdssamhället. Välfärdspolitiska rådets rapport 2002, SNS förlag.

McClure I.; Taran P.; Zegers de Beijl R. (2004) Challenging Discrimination in Employment: A summary of research and a typology of measures, International Migration Paper 68 (Ge- neva: ILO)

Nekby, Lena; Özcan, Gülay (2006) Utbildning och arbetsmarknaden – är den svenska ut- bildningen lika för alla? expertbilaga till Rapport Integration 2005.

Nekby, Lena (2003) “How Long Does it Take to Integrate? Employment Convergence of Immigrants and Natives in Sweden” in Empirical Studies on Health Insurance, Employment of Immigrants and the Gender Wage Gap, PhD dissertation, Department of Economics, Stockholm University.

Nilsson, Joakim (2004) “Labour market outcomes of second-generation immigrants in Swe- den”, Masters thesis, Department of Economics, Stockholm University.

Organisation for Economic Co-operation and Development (OECD) (2006) Employment Outlook.

52

Österberg, Torun (2000) “Children with immigrant backgrounds: Ethnic externalities in their labour market performance” and “Intergenerational income mobility among individuals with different ethnic origins” in Economic Perspectives on Immigrants and Intergenerational Transmissions, Ekonomic Studies no. 102, Ph.D. dissertation, Department of Economics, Handelshögskolan, Gothenburg University.

Rashid, Salman (2004) Immigrants’ Income and Family Migration, PhD thesis, Umeå Eco- nomic Studies, no. 625.

Rooth, Dan-Olof (2002) “Adopted Children in the Labour Market – Discrimination or unob- served characteristics?” in International Migration, Vol. 40, no. 1, pp. 71-98.

Schröder, Lena (2007) “From Problematic Objects to Resourceful Subjects: An overview of the ethnic wage and employment gaps in Sweden, and its possible explanations and cures”. Forthcoming in Swedish Economic Policy Review.

Solon, G (1999) “Intergenerational Mobility in the Labour Market”, in Ashenfelter O.; Cards D (editors) Handbook of Labour Economics Vol. 3. Elsevier

SOU 2005:56 Det blågula glashuset – strukturell diskriminering i Sverige. Betänkande av Utredningen om strukturell diskriminering på grund av etnisk eller religiös tillhörighet. P. Lappaleinen (ed.) Stockholm: Frizes

SOU 2003:92 Unga utanför. Linde L. Sojanovic V. Kassman A. (ed) Stockholm: Frizes

SOU 2005:41 Bortom vi och dom: teoretiska reflektioner om makt, integration och strukturell diskriminering. Utredning om makt, integration och strukturell diskriminering. P. de los Reyes & M. Kamali (ed.) Stockholm: Fritzes

SOU 2005:115 Anonymitetsutredningen; Tanja Lindeborg (ed) Stockholm: Fritzes

Statistics Sweden, Labour Force Surveys (various years).

Swedish Integration Board (2005) Statistic Report 2004.

Swedish Integration Board (2006) Pocket Facts: Statistics on Integration.

Vilhelmsson, Roger (2002), “The Immigrant-Native Wage Gap in Sweden: Is There a Differ- ence Between the Top and the Bottom of the Wage Distribution?” in Vilhelmsson, R: Wages and Unemployment Of Immigrants and Natives in Sweden, PhD thesis, The Swedish Institute for Social Research, Stockholm University, Dissertation Series 56.

Wrench, J. (1999) “Employers and anti-Discrimination Measures in Europe: Good Practice and Bad Faith” in Wrench (editor) Migrants, Ethnic Minorities and the Labour Market: Inte- gration and exclusion in Europe. (Houndmills: MacMillan Press).

53

International Migration Papers

Cahiers de migrations internationales Estudios sobre Migraciones Internacionales

1. Adjustments to Labour Shortages and Foreign Workers in the Republic of Korea. M.I. Abella; Y.B. Park; W.R. Böhning, 1995

2. Consumption and Investments from Migrants’ Remittances in the South Pacific. Richard P.C. Brown, 1995

3. Training Abroad: German and Japanese schemes for workers from transition economies or developing countries. Christiane Kuptsch; Nana Oishi, 1995

4. Discrimination against Migrant Workers and Ethnic Minorities in Access to Employment in the Netherlands. F. Bovenkerk; M.J.I. Gras; D. Ramsoedh, with the assistance of M. Dankoor and A. Havelaar, 1995

5. Orderly International Migration of Workers and Incentives to Stay: Options for emigration countries. M.I. Abella; K.J. Lönnroth, 1995

6. From Outlawing Discrimination to Promoting Equality: Canada's experience with anti- discrimination legislation. C. Ventura, 1995

7E. Labour Market Discrimination against Foreign Workers in Germany. A. Goldberg; D. Mour- inho; U. Kulke, 1996

7G. Arbeitsmarkt-Diskriminierung gegenüber ausländischen Arbeitnehmern in Deutschland. A. Goldberg; D. Mourinho; U. Kulke, 1995

8E. The Integration of Migrant Workers in the Labour Market: Policies and their impact. W.R. Böhning; R. Zegers de Beijl, 1995

8F. L'intégration des travailleurs migrants sur le marché du travail: Les politiques et leur impact. W.R. Böhning; R. Zegers de Beijl, 1996

9E. Labour Market Discrimination Against Migrant Workers in Spain. Colectivo IOE: M.A. de Prada; W. Actis; C. Pereda, y R. Pérez Molina, 1996

9S. La discriminación laboral a los trabajadores inmigrantes en España. Colectivo IOE: M.A. de Prada; W. Actis; C. Pereda, y R. Pérez Molina, 1995

10. The Jobs and Effects of Migrant Workers in Northern America: Three essays. J. Samuel; P.L. Martin; J.E. Taylor, 1995

11. The Jobs and Effects of Migrant Workers in Italy: Three essays. L. Frey; R. Livraghi; A. Venturini; A. Righi; L. Tronti, 1996

12. Discrimination against Racial/Ethnic Minorities in Access to Employment in the United States: Empirical findings from situation testing. M. Bendick, Jr., 1996

13. Employer des travailleurs étrangers: Manuel sur les politiques et les procédures plus particu- lièrement applicables aux pays à bas ou moyen revenus. W.R. Böhning, 1996

14. Protecting (Im)migrants and Ethnic Minorities from Discrimination in Employment: Finnish and Swedish experiences. K. Vuori, with the assistance of R. Zegers de Beijl, 1996

54

15E. Migration from the Maghreb and Migration Pressures: Current situation and future pros- pects. D. Giubilaro, 1997

15F. Les migrations en provenance du Maghreb et la pression migratoire: Situation actuelle et prévisions. D. Giubilaro, 1997

16. The Documentation and Evaluation of Anti-discrimination Training Activities in the Nether- lands. J.P. Abell; A.E. Havelaar; M.M. Dankoor, 1997

17. Global Nations: The impact of globalization on international migration. P. Stalker, 1997

18. Anti-discrimination Training Activities in Finland. K. Vuori, 1997

19. Emigration Pressures and Structural change: Case study of the Philippines. A. Saith, 1997

20. Emigration Pressures and Structural change: Case study of Indonesia. D. Nayyar, 1997

21. The Evaluation of Anti-discrimination Training Activities in the United Kingdom. P. Taylor; D. Powell; J. Wrench, 1997

22. Pratiques de formations antidiscriminatoires en Belgique. F. Castelain-Kinet; S. Bouquin; H. Delagrange; T. Denutte, 1998

23E. Discrimination in Access to Employment on Grounds of Foreign Origin: The case of Belgium. P. Arrijn; S. Feld; A. Nayer, 1998

23F. La discrimination à l'accès à l'emploi en raison de l'origine étrangère : le cas de la Belgique. P. Arrijn; S. Feld; A. Nayer, 1998

24. Labour Immigration and Integration in Low- and Middle-income Countries: Towards an evaluation of the effectiveness of migration policies. J. Doomernik, 1998

25. Protecting Migrants and Ethnic Minorities from Discrimination in Employment: the Danish experience. N.-E. Hansen, I. McClure, 1998

26. Illegal Migration and Employment in Russia. Eugene Krassinets, 1998

27. The Effectiveness of Integration Policies towards Immigrants and their Descendants in France, Germany and the Netherlands. Jeroen Doomernik, 1998

28. Approche juridique de la discrimination à l’accès à l’emploi en Belgique en raison de l’origine étrangère. B. Smeesters, sous la direction de A. Nayer, 1999

29. The Documentation and Evaluation of Anti-discrimination Training in the United States. M. Bendick, Jr., M.L. Egan, S. Lofhjelm, 1999

30. Illegal Labour Migration and Employment in Hungary. J. Juhász with contributions from M. Cosmeanu; I. Ramond; J. Gmitra, A. Bácskai, 1999

31. Foreign Labour in Lithuania: Immigration, employment and illegal work. A. Sipaviciene, in cooperation with V. Kanopiene, 1999

32. Legal and Illegal Labour Migration in the Czech Republic: Background and current trends. Milada Horákova, 2000

33. Migrant Labour: An annotated bibliography. R. Chen; M. Madamba, 2000

55

34. Settlement and Integration Policies towards Immigrants and their Descendants in Sweden. Charles Westin, 2000

35. United States Policies on Admission of Professional and Technical Workers: Objectives and outcomes. Philip Martin, Richard Chen and Mark Madamba, 2000

36. Employer Sanctions: French, German and US experiences. Philip Martin and Mark Miller, 2000

37. Quotas d’immigration : l’expérience Suisse. Etienne Piguet et Hans Mahnig, 2000

38. The Effectiveness of Employment Equality Policies in Relation to Immigrants and Ethnic Mi- norities in the UK. John Wrench and Tariq Modood, 2001

39. The Ambiguities of Emigration: Bulgaria since 1988. August Gächter, 2002

40. Migration for the Benefit of All: Towards a new paradigm for migrant labour. Eric Weinstein, 2001

41. Migrants in Irregular Employment in the Mediterranean Countries of the European Union. Emilio Reynieri, 2001

42. From Temporary Guests to Permanent Settlers? A review of the German experience. Heinz Werner, 2001

43. From Brain Exchange to Brain Gain: Policy implications for the UK of recent trends in skilled migration from developing countries. Allan Findlay, 2002

44. Migration of Highly Skilled Persons from Developing Countries: Impact and policy re- sponses. B. Lindsay Lowell and Allan Findlay, 2002

44F. L’émigration de personnes hautement qualifiées de pays en développement : impact et répon- ses politiques – Rapport de synthèse. B. Lindsay Lowell et Allan Findlay, 2003

45. Policy Responses to the International Mobility of Skilled Labour. B. Lindsay Lowell, 2002

46. Some Developmental Effects of the International Migration of Highly Skilled Persons. B. Lindsay Lowell, 2002

47. Women Migrant Domestic Workers in Bahrain. Sabika al-Najjar, 2002

48. Women Migrant Domestic Workers in Lebanon. Ray Jureidini, 2002

49. Skilled Labour Migration from Developing Countries: Study on India. Binod Khadria, 2002

50. Skilled Labour Migration from Developing Countries: Study on the Caribbean Region. Elizabeth Thomas-Hope, 2002

51. Skilled Labour Migration from Developing Countries: Study on the Philippines. Florian A. Alburo and Danilo I. Abella, 2002

52. Skilled Labour Migration from Developing Countries: Study on South and Southern Africa. Haroon Bhorat, Jean-Baptiste Meyer and Cecil Mlatsheni, 2002

53. Situación de los trabajadores migrantes en América Central. Abelardo Morales Gamboa, 2002

56

54F. L'immigration irrégulière subsaharienne à travers et vers le Maroc. Lucile Barros, Mehdi Lahlou, Claire Escoffier, Pablo Pumares, Paolo Ruspini, 2002

54S. La inmigración irregular subsahariana a través y hacia Marruecos. Lucile Barros, Mehdi Lahlou, Claire Escoffier, Pablo Pumares, Paolo Ruspini, 2002

55. Skilled Labour Migration from Developing Countries: Annotated bibliography. Allan M. Findlay and Emma Stewart, 2002

56. Skilled Labour Migration from Developing Countries: Annotated bibliography on economic analysis, impact and policy issues. B. Lindsay Lowell, 2002

57. Asian Labour Migration: Issues and challenges in an era of globalization. Piyasiri Wick- ramasekara, 2002

58. Skilled Labour Migration from Developing Countries: Study on Argentina and Uruguay. Adela Pellegrino, 2002

58S. Migración de mano de obra calificada desde Argentina y Uruguay. Adela Pellegrino, 2003

59. Remesas de mexicanos en el exterior y su vinculación con el desarrollo económico, social y cultural de sus comunidades de origen. Mario López Espinosa, 2002

60. Migraciones laborales en América del Sur: La comunidad andina. Ponciano Torales, M. Estela González y Nora Pérez Vichich, 2003

61. Economic Integration in the Caribbean: The development towards a common labour market. Deike Fuchs and Thomas Straubhaar, 2003

62F. Enjeux et défis de la migration de travail ouest-africaine. A.S. Fall, 2003

63. Migraciones laborales en Sudamérica: el Mercosur ampliado. Ezequiel Texidó, Gladys Baer, Nora Pérez Vichich, Ana María Santestevan, Charles P. Gomes, 2003

64. Empowering Filipino Migrant Workers: Policy issues and challenges. Rene E. Ofreneo and Isabelo A. Samonte, 2004

65. Acuerdos bilaterales sobre migración de mano de obra: Modo de empleo. Eduardo Geroni- mi, 2004

66. Acuerdos bilaterales sobre migración de mano de obra: Estudio de casos. Eduardo Geronimi, Lorenzo Cachón y Ezequiel Texidó, 2004

67. Labour Market Discrimination against Migrant Workers in Italy. E. Allasino, E. Reyneri, A. Venturini, G. Zincone, 2004

67I. La discriminazione dei lavoratori immigrati nel mercato del lavoro in Italia. E. Allasino, E. Reyneri, A. Venturini, G. Zincone, 2004

68. Challenging Discrimination in Employment: A summary of research and a typology of meas- ures. P. Taran, R. Zegers de Beijl and I. McClure, 2004

69. Labour Market Effects of Immigration: An Empirical Analysis Based on Italian Data. A. Venturini and C. Villosio, 2004

57

70. Admisión, contratación y protección de trabajadores migrantes: Panorama de la legislación y la práctica nacionales de Argentina, Bolivia, Brasil, Chile, Colombia, Ecuador, España,

Perú, Portugal y Uruguay. E. Geronimi, 2004

72E. Migration Management and Development Policies: Maximising the benefits of International Migration in West Africa. Savina Ammassari, 2006

72F. Gestion des migrations et politiques de développement : optimiser les bénéfices de la migra- tion internationale en Afrique de l’Ouest. Savina Ammassari 2004

73. Migration Prospects after the 2004 Enlargement of the European Union. Gloria Moreno- Fontes Chammartin and Fernando Cantú-Bazaldúa, 2005

74. Identification of Potential for Increasing Employment and Productive Investment in Albania, Moldova and Ukraine Based on Remittances. Gloria Moreno-Fontes Chammartin and Fer- nando Cantú-Bazaldúa, 2005

75. Rights of Migrant Workers in Asia: Any light at the end of the tunnel? Piyasiri Wickramase- kara, 2006

Labour Migration in Africa Series

Série sur les migrations de main d’oeuvre en Afrique

76E. Systems of Statistical Information on Migrant Workers in Central Maghreb. Musette Mo- hamed Saïb, and Belghazi Saad , Boubakri Hassan, Hammouda Nacer Eddine, 2006

76F. Les systèmes d’informations statistiques sur les travailleurs migrants au Maghreb Central. Musette Mohamed Saïb, et Belghazi Saad , Boubakri Hassan, Hammouda Nacer Eddine, 2006

77E. Report on Legislation Concerning International Migration in Central Maghreb. Mohamed Saïb Musette, and Monia Benjemia, Khadija Elmadmad, Azzouz Kerdoun, 2006

77F. Rapport sur les législations relatives à la migration internationale au Maghreb Central. Musette Mohamed Saïb, et Monia Benjemia, Khadija Elmadmad, Azzouz Kerdoun, 2006

78E. Summary Report on Migration and Development in Central Maghreb. Mohamed Saïb Musette, and Youssef Alouane, Mohamed Khachani, Hocine Labdelaoui, 2006

78F. Rapport sur les migrations et le développement au Maghreb Central. Musette Mohamed Saïb, et Youssef Alouane , Mohamed Khachani, Hocine Labdelaoui, 2006

79E. Labour Migration Statistics in West Africa. Hamidou Ba, and Babacar Ndione, 2006

79F. Les statistiques des travailleurs migrants en Afrique de l’Ouest. Hamidou Ba, et Babacar Ndione, 2006

80E. Legislation Relevant to Migrant Workers in West Africa. Hamidou Ba, and Abdoulaye Fall, 2006

80F. Législations relatives aux travailleurs migrants en Afrique de l’Ouest. Hamidou Ba, et Ab- doulaye Fall, 2006

81. A Study of Labour Migration Data and Statistics in East Africa. Joseph M. Shitundu, 2006

82. Migration Legislation in East Africa. Flora Mndeme Musonda, 2006

58

83. The Migration-Development Nexus in East Africa. Humphrey P.B. Moshi, 2006

84E. The Challenge of Labour Migration Flows Between West Africa and the Maghreb. Aderanti Adepoju, 2006.

84F. Les défis liés aux flux migratoire pour le travail entre l’Afrique de l’Ouest et de la Maghreb. Aderanti Adepoju, 2006.

Discrimination testing France & Sweden

85F. Les Discriminations à raison de « l’origine » dans les embauches en France : Une enquête nationale par tests de discrimination selon la méthode du BIT. Eric Cediey et Fabrice Foroni. 2007

86E. Discrimination in Employment against Second-generation Swedes of Immigrant Origin. Att- ström K. 2007.

59

Perspectives on Labour Migration

Perspectives sur les migrations du travail

Perspectivas sobre migraciones laborales

1. Getting at the Roots: Stopping exploitation of migrant workers by organized crime. Patrick Taran and Gloria Moreno-Fontes Chammartin, 2003

2. Aspectos jurídicos del tráfico y la trata de trabajadores migrantes. Eduardo Geronimi, 2002

2 F. Aspects juridiques du trafic et de la traite de travailleurs migrants. Eduardo Geronimi, 2003

3. Globalization, Labour and Migration: Protection is paramount. Patrick Taran and Eduardo Geronimi, 2003

3 F Globalisation et migrations de main-d’oeuvre : Importance de la protection. Patrick Taran et Eduardo Geronimi, 2003

3 S. Globalización y migraciones laborales: importancia de la protección. Patrick Taran y Edu- ardo Geronimi, 2003

5 E Policy Responses to Skilled Migration: Retention, return and circulation. Piyasiri Wickrama- sekara, 2003

5 F Options politiques de réponse à la migration des compétences : Rétention, retour et circula- tion. Piyasiri Wickramasekara, 2003

6. Temporary Foreign Worker Programmes: Policies, adverse consequences and the need to make them work. Martin Ruhs, 2003

7. Protección y asistencia a las víctimas de trata. Eduardo Geronimi, 2003