Company training and services - Focus on low skills : Focus on Low Skills 9781845447380, 9780861769452

Citation preview

ijm_cover_(i).qxd

3/19/04

2:45 PM

Page 1

Volume 25 Number 1 2004

ISBN 0-86176-945-7

ISSN 0143-7720

IJM International Journal of Manpower Company training and services: focus on low skills Guest Editors: Rita Asplund and Wiemer Salverda

An Interdisciplinary Journal: Human Resources Management Labour Economics

www.emeraldinsight.com

International Journal of Manpower

ISSN 0143-7720 Volume 25 Number 1 2004

Company training and services: focus on low skills Guest Editors Rita Asplund and Wiemer Salverda

Access this journal online __________________________

3

Editorial advisory board ___________________________

4

Abstracts and keywords ___________________________

5

Introduction: company training and services with a focus on low skills Rita Asplund and Wiemer Salverda ________________________________

8

Company training and low-skill consumer-service jobs in Ireland Gerard Hughes, Philip J. O’Connell and James Williams ________________

17

Why do business service firms employ fewer apprentices? A comparison between Germany and The Netherlands Wendy Smits and Thomas Zwick___________________________________

36

The role of training in changing an economy specialising in tourism Vicente Ramos, Javier Rey-Maquieira and Maria Tugores ______________

Access this journal electronically The current and past volumes of this journal are available at:

www.emeraldinsight.com/0143-7720.htm You can also search over 100 additional Emerald journals in Emerald Fulltext at:

www.emeraldinsight.com/ft See page following contents for full details of what your access includes.

55

CONTENTS

CONTENTS continued

Training, task flexibility and the employability of low-skilled workers Jos Sanders and Andries de Grip___________________________________

73

Training and industrial restructuring: structural change and labour mobility in West Germany and Sweden Tomas Korpi and Antje Mertens __________________________________

90

Labour market effects of apprenticeship training in Austria Helmut Hofer and Christine Lietz __________________________________

104

Human capital spillovers in the workplace: evidence for the service sector in Britain H. Battu, C.R. Belfield and P.J. Sloane_______________________________

123

About the Guest Editors _________________________________ 139 About the authors_________________________________________ 140 Note from the publisher ____________________________ 143

www.emeraldinsight.com/ijm.htm As a subscriber to this journal, you can benefit from instant, electronic access to this title via Emerald Fulltext. Your access includes a variety of features that increase the value of your journal subscription.

How to access this journal electronically To benefit from electronic access to this journal you first need to register via the Internet. Registration is simple and full instructions are available online at www.emeraldinsight.com/ rpsv/librariantoolkit/emeraldadmin Once registration is completed, your institution will have instant access to all articles through the journal’s Table of Contents page at www.emeraldinsight.com/0143-7720.htm More information about the journal is also available at www.emeraldinsight.com/ ijm.htm Our liberal institution-wide licence allows everyone within your institution to access your journal electronically, making your subscription more cost effective. Our Web site has been designed to provide you with a comprehensive, simple system that needs only minimum administration. Access is available via IP authentication or username and password.

Key features of Emerald electronic journals Automatic permission to make up to 25 copies of individual articles This facility can be used for training purposes, course notes, seminars etc. This only applies to articles of which Emerald owns copyright. For further details visit www.emeraldinsight.com/ copyright Online publishing and archiving As well as current volumes of the journal, you can also gain access to past volumes on the internet via Emerald Fulltext. Archives go back to 1994 and abstracts back to 1989. You can browse or search the database for relevant articles. Key readings This feature provides abstracts of related articles chosen by the journal editor, selected to provide readers with current awareness of interesting articles from other publications in the field. Reference linking Direct links from the journal article references to abstracts of the most influential articles cited. Where possible, this link is to the full text of the article. E-mail an article Allows users to e-mail links to relevant and interesting articles to another computer for later use, reference or printing purposes.

Additional complementary services available Your access includes a variety of features that add to the functionality and value of your journal subscription:

E-mail alert services These services allow you to be kept up to date with the latest additions to the journal via e-mail, as soon as new material enters the database. Further information about the services available can be found at www.emeraldinsight.com/usertoolkit/ emailalerts Emerald WIRE (World Independent Reviews) A fully searchable subject specific database, brought to you by Emerald Management Reviews, providing article reviews from the world’s top management journals. Research register A web-based research forum that provides insider information on research activity world-wide located at www.emeraldinsight.com/researchregister You can also register your research activity here. User services Comprehensive librarian and user toolkits have been created to help you get the most from your journal subscription. For further information about what is available visit www.emeraldinsight.com/usagetoolkit

Choice of access Electronic access to this journal is available via a number of channels. Our Web site www.emeraldinsight.com is the recommended means of electronic access, as it provides fully searchable and value added access to the complete content of the journal. However, you can also access and search the article content of this journal through the following journal delivery services: EBSCOHost Electronic Journals Service ejournals.ebsco.com Huber E-Journals e-journals.hanshuber.com/english/index.htm Informatics J-Gate www.j-gate.informindia.co.in Ingenta www.ingenta.com Minerva Electronic Online Services www.minerva.at OCLC FirstSearch www.oclc.org/firstsearch SilverLinker www.ovid.com SwetsWise www.swetswise.com TDnet www.tdnet.com

Emerald Customer Support For customer support and technical help contact: E-mail [email protected] Web www.emeraldinsight.com/customercharter Tel +44 (0) 1274 785278 Fax +44 (0) 1274 785204

IJM 25,1

EDITORIAL ADVISORY BOARD Professor David J. Bartholomew The London School of Economics Professor Derek Bosworth Manchester School of Management, UMIST, UK

4

Professor Martin Carnoy School of Education, Stanford University, USA Professor Peter Dawkins Melbourne Institute for Applied Economic and Social Research, Melbourne University, Australia

Professor John Mangan University of Queensland, Brisbane, Australia Professor Stephen L. Mangum Ohio State University, Ohio, USA

Professor Morley Gunderson University of Toronto, Canada

Professor Barrie Pettman International Management Centres, UK, and Founding Editor of International Journal of Manpower

Professor Thomas J. Hyclak Lehigh University, Bethlehem, USA

Professor David Sapsford Management School, Lancaster University, UK

Professor Susan E. Jackson Rutgers University, New Jersey, USA

Professor P.J. Sloane Department of Economics, University of Wales, Swansea

Professor Harish C. Jain McMaster University, Canada Professor Geraint Johnes Management School, Lancaster University, UK Professor Meni Koslowsky Department of Psychology, Bar-Ilan University, Israel

International Journal of Manpower Vol. 25 No. 1, 2004 p. 4. # Emerald Group Publishing Limited 0143-7720

Professor Lord Richard Layard Centre for Economic Performance, London School of Economics, UK

Professor Klaus F. Zimmerman Department of Economics, University of Bonn, Germany

Introduction: company training and services with a focus on low skills Rita Asplund and Wiemer Salverda Keywords Training, Skills, Employees development This special issue of the International Journal of Manpower aims to make a contribution to broadening our limited understanding of the role and impact of employer-provided training for low-skilled service sector workers. It brings together seven of the papers that were presented at the international conference “Adapting Education and Training for the Enhancement of Low-Skilled Jobs” held at Helsinki in May 2002. The papers are situated at the crossroads where three different strands of research and policymaking meet: the training of the low skilled, the system of vocational training and the role of training for the service sector. The contributions cover an interesting variety of European countries: Ireland, Germany, the Netherlands, Austria, Sweden, Spain and the UK, with diverging levels of low-skilled (un)employment, vocational training and service-sector employment.

Company training and low-skill consumer-service jobs in Ireland Gerard Hughes, Philip J. O’Connell and James Williams Keywords Training, Skills, Service industries, Ireland This paper identifies market forces which induce employers to provide training in Ireland. It investigates if they are present in sufficient strength in the consumer service sectors with a high concentration of low-skill jobs to provide a basis to upgrade such jobs. Data from a survey of firms on training incidence, duration, and cost are used in OLS regressions to investigate the determinants of training at national and sector level. The results show that firm size, the proportion of skilled workers, foreign ownership, perception of changing skill requirements and tightness of the labour market all influence employers’ training decisions. Analysis of sector-specific effects indicates that firms in consumer

service sectors are unlikely to respond to market forces by increasing training to a level which would encompass low-skill jobs. However, policies involving the school system and company-based training could help to enhance low-skill jobs in consumer service sectors.

Abstracts and keywords

5

Why do business service firms employ fewer apprentices? A comparison between Germany and The Netherlands Wendy Smits and Thomas Zwick Keywords Apprenticeships, Skills, Business support services, Germany, The Netherlands This paper analyses why in Germany and The Netherlands the share of apprentices in the business service sector is lower than in other economic sectors. A theoretical introduction surveys the potential reasons that could be responsible for this. The subsequent empirical analysis shows that the level of skill apprentices gain is the main explanation for the relatively low supply of apprenticeships in German business service enterprises. In The Netherlands, the option to hire skilled employees from full-time schools instead of training apprentices seems to be crucial. For these reasons, this paper proposes to offer obligatory extra formal training in areas such as IT skills and foreign languages for the apprentices in business service firms in Germany in order to increase the attractiveness of the dual apprenticeship system for prospective apprentices as well as business service firms.

The role of training in changing an economy specialising in tourism Vicente Ramos, Javier Rey-Maquieira and Maria Tugores Keywords Education, Training, Tourism management, Balearic Islands This paper compares the training requirements of alternative tourism development strategies which are differentiated by the quality of service

International Journal of Manpower Vol. 25 No. 1, 2004 Abstracts and keywords q Emerald Group Publishing Limited 0143-7720

IJM 25,1

6

offered. The paper focuses on the Balearic Islands and uses an original database that consists of a representative sample of Balearic hotels. This database includes data on both employers and employees and enables the identification of differences in job characteristics, as well as differences in human capital, with respect to both education and on-the-job training, depending on the category of the hotel. The article uses a discrete choice model to identify the characteristics of both employer and employee that determine the provision of training. It concludes that educational level is not a strong constraint on the mobility of workers between categories, and we show that on-the-job training has a role to play in the transition to alternative tourism development strategies.

Training, task flexibility and the employability of low-skilled workers Jos Sanders and Andries de Grip Keywords Training, Skills training, Labour mobility, Flexible labour This paper analyses whether low-skilled workers’ training participation and task flexibility contribute to their firm-internal and firm-external mobility, and find that both training participation and task flexibility contribute only to firm-internal employability. However, the workers’ participation in training plays a much more explicit role in their firm-internal career than their task flexibility does, as the former appears to be an important means to increase their opportunities in the firm-internal labour market. Neither the low-skilled workers’ participation in training nor their task flexibility contributes to their external employability. Task-flexible, low-skilled workers are less likely to expect to be externally employable than non-task flexible workers are. The focus of the low-skilled workers on their firm-internal employability can be explained by the fact that such workers usually have more opportunities to improve their position in the firm-internal labour market than in the external labour market.

Training and industrial restructuring: structural change and labour mobility in West Germany and Sweden Tomas Korpi and Antje Mertens Keywords Labour mobility, Skills, Vocational training, Change management, Germany, Sweden While the structural changes that have taken place in the labour markets of the industrialised world over the past decades are well documented, less is known about how individuals respond to this changing environment. This includes the extent of intersectoral mobility during the work career, skill differentials in mobility, the impact of the type of training on mobility and changes in mobility patterns over a long period of time. Against this backdrop, the purpose of this paper is to examine intersectoral labour mobility during the first 15 years of working careers in Sweden and West Germany. The analyses show that individuals in both countries tend to move away from industry into other sectors during their careers, but that this tendency is rather weak. While there are some mobility differences among educational categories, the differences between transition probabilities of German apprentices and Swedish vocational school students are insignificant. In the face of the massive transformation of employment structures, the importance of variation in the curricula is probably minuscule.

Labour market effects of apprenticeship training in Austria Helmut Hofer and Christine Lietz Keywords Apprenticeships, Training, Pay, Labour market, Austria, Unemployment In Austria, the apprenticeship system provides all citizens, including the less able among them, with a training option. Based on social security data, this article examines earnings and the stability of the occupational career of young workers with an apprenticeship diploma. As control groups, workers with a full-time secondary school education and workers who did not receive any further education after completing their

compulsory education were used. One of the main findings is that workers with an apprenticeship diploma are much better, off than those without further education. The article finds the following ranking with respect to education: high-school graduates, ex-apprentices and unskilled workers, with more pronounced differences between ex-apprentices and unskilled workers.

Human capital spillovers in the workplace: evidence for the service sector in Britain H. Battu, C.R. Belfield and P.J. Sloane Keywords Human capital, Workplace, Performance criteria An individual’s human capital has a strong influence in earnings. Yet, individual worker-level estimations of earnings rarely include the characteristics of co-workers or detailed firm-level controls. In particular, co-workers skills are ignored which may be particularly significant where team work is important. This paper utilises a unique matched work-place data set to estimate the

effect on the earnings of co-workers’ education and training in the Hotel and Catering sector, which contains a high proportion of low paying establishments and in the Retail sector which contains a large absolute number of low paying establishments. The data are derived from the 1998 British Workplace Employment Relations Survey. This is a national sample based on interviews with managers in 2,191 establishments with at least ten workers. In addition, a survey of up to 25 randomly selected employees in each establishment was undertaken which included questions on education, training, pay and job satisfaction, as well as a range of other personal and workplace characteristics. We have, therefore, a matched workplace employee sample which is essential for this type of analysis. The results suggest that there are strong co-worker effects in the earnings of individuals when controlling the individual’s own level of education. While there are also high returns to training for individual workers, there are no similar spillover effects from the training of co-workers in these sectors. Nevertheless, this suggests that there could well be a pay-off to the professionalisation of service sector jobs.

Abstracts and keywords

7

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

IJM 25,1

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Introduction Company training and services with a focus on low skills

8

Rita Asplund The Research Institute of the Finnish Economy, Helsinki, Finland, and

Wiemer Salverda Amsterdam Institute for Advanced Labour Studies (AIAS), University of Amsterdam, Amsterdam, The Netherlands Keywords Training, Skills, Employees development Abstract This special issue of the International Journal of Manpower aims to make a contribution to broadening our limited understanding of the role and impact of employer-provided training for low-skilled service sector workers. It brings together seven of the papers that were presented at the international conference “Adapting Education and Training for the Enhancement of Low-Skilled Jobs” held at Helsinki in May 2002. The papers are situated at the crossroads where three different strands of research and policymaking meet: the training of the low skilled, the system of vocational training and the role of training for the service sector. The contributions cover an interesting variety of European countries: Ireland, Germany, the Netherlands, Austria, Sweden, Spain and the UK, with diverging levels of low-skilled (un)employment, vocational training and service-sector employment.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 8-16 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410524965

The past few decades have seen a marked deterioration in the employment situation of low-skilled workers, both in Europe and North America. The relative decrease in the labour market position of low-skilled workers on the European labour markets has mostly taken the form of higher unemployment, which is often blamed on downward rigid wages (Dre`ze, 2002; Krugman, 1994). In the USA, in contrast, the worsened employment situation of low-skilled workers has manifested itself primarily in the form of a decline in relative as well as real wages. But even in countries with flexible wages, such as the USA and the UK, the non-employment of low-skilled labour has remained strikingly high (Glyn and Salverda, 2002). Falkinger and Grossmann (2001) have pointed out some of the mechanisms underlying the fact that flexible labour markets provide no guarantee against the unemployment of low-skilled workers. Several economic shocks have been put forward in the literature to explain why the labour market position of low-skilled workers has deteriorated relative to that of more skilled workers. Among these explanations are: skill-biased technological progress, work organisational changes, increased competition from low-wage economies, increasing levels of education, higher replacement ratios, and accelerating destruction of simple or routine jobs. More recent additions to this list of key explanations are a worsening over-education problem (De Grip and Borghans, 2000), and an increase in skill-intensive

foreign-owned firms. Simultaneously, a growing body of literature has shown that the high and persistent unemployment rates of low-skilled workers are probably not the outcome of one single explanation, but are rather the result of a complex mix of several contributing factors (Gautier, 2002). Another conclusion that can be drawn about the explanations on offer is that they represent the kind of economic shocks that most probably have turned the observed underemployment of low-skilled workers into a lasting trend. Against this background it is hardly surprising that academics as well as politicians are searching for effective tools to combat the situation, that is, to make it more attractive for employers to hire low-skilled workers and also to retain them longer in employment. A commonly emphasised remedy is investment in the human capital of the low skilled in order to adjust their inadequate levels of skill and competence. However, for such policies to be useful low skills need to be measured properly, and the learning processes through which these skill disadvantages can be overcome, need to be identified. Despite intensive research on the labour market situation of the low skilled during the past decade or so, such evidence is only emerging. Who, then, is low skilled in an advanced industrialised country? A multitude of measures have been used in the empirical literature, mostly dictated by the available data. In a recent study, Steedman and McIntosh (2001) approached the problem of finding a definition of a low-skilled worker that is applicable across the European countries. They compared several data sources, particularly the International Adult Literacy Survey, and concluded that the International Standard Classification of Education (ISCED) provides the most suitable means of measuring skills over time and space. More precisely, they recommended those individuals that are categorised into ISCED 0-2 (education at the lower secondary level or below) be labelled as low skilled. This formal education based measure is shown to be a valid and reliable proxy for skills in a situation with little choice, as measures including skills acquired outside the education system do not exist. The validity seems restricted to the European countries; the comparison of ISCED with levels identified in the International Adult Literacy Survey seems too broad to adequately consider the American situation. With the endorsement of the principle of life-long learning, increasing attention has been paid especially to the importance of employer-provided training, including formal as well as informal learning achieved on or off the job. Most of this research has focused on the design, implementation and economic implications of this type of training (OECD, 1999, and the references therein), while few studies have tried to evaluate whether or not it is an effective form of training and which factors determine its effectiveness (Van der Klink and Streumer, 2002). Little is known about the international differences in employer-provided training or their causes and consequences simply because of the rather poor comparability of available data (Brunello and

Company training and services 9

IJM 25,1

10

Medio, 2001; Elias and Davies, 2004; Nestler and Kailis, 2002a, b; OECD, 1991, 1993, 1999). Moreover, the existing studies of employer-provided training mostly focus on the whole workforce, with a distinction made according to a varying set of employee characteristics, including education. Only few studies are explicitly focused on low-skilled workers. One possible reason for this disinterest in research on the training of the low skilled is that formal education and post-school training seem to be the complements rather than substitutes due to the reluctance of private enterprises to invest heavily in the training of low-skilled workers (OECD, 1991, 1999). Fouge`re et al. (2000) concluded, based on evidence for France, that training programmes for unemployed and low-skilled young adults appear neither to improve their human capital investments nor to affect their post-training wages or employment probabilities. Kramarz (2000) added that the participation in such training schemes may become self-reinforcing, leading people from one scheme to the next rather than to stable employment. Indeed, these results repeated the conclusions drawn by Bassi and Ashenfelter (1985) concerning the low-skilled workers for the USA, 15 years earlier in 1985. On the whole, low-skilled workers seem to face a substantial risk of falling into a trap, in which their poor skills are combined with few job opportunities and a low return on training, as recently shown by Burdett and Smith (2002). The growing interest in alternative instruments aimed at improving the employment situation of low-skilled workers is well explained by these rather discouraging assessments concerning the training of low skilled. For instance, Dre`ze (2002) argued for wage subsidies instead of unemployment benefits on behalf of low-skilled workers. This is well in line with the notion by Fouge`re et al. (2000) that payroll tax subsidies have significantly improved the employment probabilities of young minimum wage workers in France, while training programmes have not. However, the efficacy of such measures has also been put into doubt (Kramarz, 2000, Mu¨hlau and Salverda, 2000). Government intervention to provide incentives in order to correct for less than “optimal” levels of training from an individual, company and societal perspective is, however, rather difficult in an area so heavily the domain of the private sector (OECD, 1999). This, however, needs to be weighed against both the commonly observed under-provision of employer-provided training to low-skilled workers and the potential gains obtained from government intervention that deals in an appropriate way with obstacles raised by market failure and systems failure (Keep and Mayhew, 1999), as well as labour market and capital market imperfections (Stevens, 1994, 1999). Likewise, the empirical evidence generally concerns the whole economy with the descriptive or statistical analyses accounting, to a varying degree, for employer and job characteristics, including the branch of activity. A common outcome of these analyses is a remarkable variation across service-sector

branches when it comes to both volume and costs of training (Nestler and Kailis, 2002a, c, d, e; OECD, 1999). Specific analyses of the service sector per se are scarce, though, despite of its alleged employment potential. The scant evidence available is not very encouraging, either. For instance, Van der Klink and Streumer (2002) studied two Dutch service companies and concluded that on-the-job training “is not entirely an effective training method although more research is needed in this area” (p. 196). When, finally, combining the three key elements of low-skilled, employer-provided training and service sector, literature research engines produce few, if any, successful hits. One of the extremely scarce exceptions is a study by Krueger and Rouse (1998), in which they compared the impact of a workplace education programme focused on low-skilled hourly workers in two US companies, one in the manufacturing sector and the other in the service sector. They found that the training programme had no significant impact on earnings in the service company and that trainees were equally likely to exit from the company as non-trainees. In view of our limited understanding of the role and impact of employer-provided training for low-skilled service sector workers, this special issue of the International Journal of Manpower aims to make a contribution to fill this gap in our knowledge. The papers This special issue brings together seven of the papers that were presented at the conference “Adapting Education and Training for the Enhancement of Low-Skilled Jobs” which was organised by the European Low-wage Employment Research network, LoWER, and The Research Institute of the Finnish Economy, ETLA, at Helsinki in May 2002. The papers are situated at the crossroads where three different strands of research and policymaking meet: the training of the low skilled, the system of vocational training and the role of training for the service sector. The focus is the company level but the perspectives vary between the effects on the companies on one hand and on low skilled labour on the other hand. The contributions cover an interesting variety of European countries: Ireland, Germany, the Netherlands, Austria, Sweden, Spain and the UK, with diverging levels of low-skilled (un)employment, vocational training and service-sector employment. The opening paper on “Company training and low-skill consumer-service jobs in Ireland” by Gerard Hughes, Philip J. O’Connell and James Williams investigates what determines the incidence, duration and cost of training in private-sector firms in Ireland. They ask whether market forces could be relied upon to convert the existing low-skill jobs in consumer-service sectors, that is primarily retail trade and hotels and catering where a good part of the low skilled are found, into stepping-stones to recognised qualifications through the provision of more formal training. An answer is sought with the help of the

Company training and services 11

IJM 25,1

12

data from a survey of firms on incidence, duration and costs of training. The analysis of sector-specific effects indicates that the firms in consumer services are unlikely to respond to market forces by increasing training to a level which would encompass the low-skill jobs. The authors conclude that policies that involve both school system and company-based training could be of help to improve the low-skill jobs in these sectors. The second paper, by Wendy Smits and Thomas Zwick, centres on the question, “Why do business service firms employ fewer apprentices? A comparison between Germany and The Netherlands”. It investigates why the share of apprentices is lower in the business service sector than in other sectors of the economy in Germany and the Netherlands. They analyse the data from two telephone surveys among the firms on apprenticeship training. With the additional help of the German IAB Establishment Panel, it is shown that for Germany the limited advantage to business-sector firms of the additional skills that can be obtained by apprentices is the main explanation. For the Netherlands, in contrast, the alternative option that firms – although they have no strong preferences – can hire qualified employees coming from full-time education, seems to play an important role for explaining the low training share in business services. Smits and Zwick direct their recommendations at the German situation, advising to improve the training in foreign languages and IT skills in the apprenticeship system as a means to enhance its attractiveness to the firms. The third paper also treats the firm’s perspective. Vicente Ramos, Javier Rey-Maquieira and Maria Tugores contribute on, “The role of training in changing an economy specialising in tourism”. They dispose of special matched data on the on-the-job training by hotels in the Balearic Islands helping them to analyse the effects of employer (especially the quality level of the hotel) and employee characteristics on the provision of training. They find that a person’s formal educational qualification is not so important a determinant for the provision of training but that the hotel’s quality level is. They conclude that a policy of on-the-job training can play an important role for a strategy that is aimed at moving the economy away from low-quality mass tourism to high quality and that it can mitigate the costs to labour that are associated with such a structural shift. The four other papers analyse the effects on low-skilled labour. Again with a focus on the company level, in “Training, task flexibility and employability of low-skilled workers”, Jos Sanders and Andries de Grip analyse whether the perception of and participation in training and the task flexibility of low-skilled workers contribute to their firm-internal and external mobility. They find an effect only for internal mobility and then mainly for the participation in training. The latter seems to enhance opportunities enabling mobility to other jobs within the firm. They think the surprising absence of a positive effect on external mobility – task flexibility even lowers an individual’s expectation of

inter-firm mobility – can be related to the observation that low-skilled employees usually have more opportunities for internal than for external improvement. Tomas Korpi and Antje Mertens contribute on “Training and industrial restructuring: structural change and mobility labour in West Germany and Sweden”. The paper examined the inter-sectoral labour mobility during the first 15 years of the work career in relation to the individual’s type and level of education and sector of the job. In spite of the massive and parallel transformation of the employment structure at the aggregate level in favour of the service sector, individual tendencies to move away from manufacturing to other sectors are weak. The authors find no significant effect on the observed mobility patterns of the difference between apprenticeship training in Germany on the one hand and vocational-school training in Sweden on the other hand. “Labour market effects of apprenticeship training in Austria” are addressed by Helmut Hofer and Christine Lietz. The Austrian apprenticeship system is an important instrument providing training options for people with otherwise low earnings prospects. Based on employment histories following from social security data Hofer and Lietz examined the effects of apprenticeship qualifications on the level of income and stability of the occupational career in comparison with people with a school education as well as others who have not finished any further education after compulsory school. They find a clear positive effect of apprenticeship vis-a`-vis the lack of any qualification, but not in comparison to a school degree. Strikingly, the benefit for women apprentices seems to reside in career stability, not in the level of income. The authors relate this to the lower quality of apprenticeship training in the service sector where women mostly go. They conclude that the apprenticeship system does help people who pass it with good result, and that it should be maintained. It should be improved, however, for service-sector jobs, particularly in high-skill branches. The final paper, “Human capital spillovers in the workplace. Evidence for the service sector in Britain” by Harminder Battu, Clive Belfield and Peter Sloane, throws up the question what the spill-over effects may be of co-workers’ human capital on an individual’s earnings. Teamwork as well as the passing on of acquired knowledge argue the existence of such effects. Strikingly, little attention is paid to this issue in the customary research on the influence of an individual’s human capital in earnings. Battu et al., utilise matched data from the 1998 British Workplace Employment Relations Survey to estimate the effects in comparison to the rest of the economy in the Hotel and Catering sector, with a high proportion of low-paying establishments, and in the Retail sector, with a large absolute number of low-paying establishments. The authors find substantial co-worker effects, independent from the individual’s own education – in line with the rest of the economy. However, although additional on-the-job training in the two service sectors does appear to impact

Company training and services 13

IJM 25,1

14

individual earnings, more for men than for women, no spill-over effects are found. The former finding can provide support to a policy of professionalisation of service-sector jobs. The authors speculate that the latter outcome may relate to the less complex nature of this particular type of service work, which could act to reduce the advantages of teamwork and knowledge transfer. These contributions inspire several observations. First, we think that low skills and the service industries together with formal vocational training and company training behaviour do make up a very interesting and important area for research. We hope that this journal issue adds significantly to the very few results that are available, but it is clear that a more systematic and comprehensive approach is warranted. As a collection, the papers have touched upon the role of the market and company behaviour, upon on-the-job training and formal vocational education, but often in a different individual context and certainly not in a precisely defined broad international comparison. Second, the interrelationship between individual behaviour – companies, as well as workers – and the context generated (or not) by policy deserves much more attention. At first sight, it does not seem that one country is doing better with regard to the low skilled than the others – but it remains to be seen whether this view will stand up to further scrutiny. One would wish for a methodical review of the role of both vocational training for low-skilled occupations on the one hand and on-the-job training in the industries where these occupations are concentrated on the other hand, and the potential and actual links between the two. This would be complemented by a second wish, for a further reflection on the implications of the fact that we are increasingly dealing with services instead of manufacturing. In many countries, the latter seems better provided by vocational education and company training schemes while services are in the background. However, the training provisions for manufacturing did not come about overnight and without effort and they certainly will not for services. Equally certainly, however, there is no a priori reason that they cannot mature for services. Nevertheless, there may be other or new problems that should be considered. For example, formal education has expanded far beyond the state in which the training schemes for manufacturing developed. Potentially, this could decrease the need for company training but it is essential to realise that the low skills may have been at a loss in this expansion, losing attention as their numbers dwindled and, by extrapolation, would seem to disappear altogether in due course. However, a certain number are still there and will not go away, implying the need for a reconsideration of their position. It is exactly here that a fresh and integrated view on vocational education and company training could be of help. Another example is the gender dimension. In virtually all countries women have been responsible for any net job growth in recent decades and obviously this correlates highly with the service sector. This fact may have

weakened service-sector training efforts. The continuing professionalisation of female employment can help to change this, but, again, such a change will not come about automatically, if only because low-skilled women often seem left behind in the expansion of female employment. The latter is precisely one of the important issues that should be accounted for in reconsidering the position of the low skilled – together with that of low-skilled men, who are faced with the shrinking of manufacturing and the corresponding training trajectories. In spite of these gaps, which we desire to be filled, and the challenges that abound in the field, we are optimistic as we do find support for the conviction that training is useful for both low skilled and companies employing them. If one thing also has become clear, it is the need for an integrated view on vocational education and company training. Individual companies and government policymakers have to act together. A nice and encouraging illustration of this can be found in the American (sic!) hosiery industry in North Carolina, where companies and local government have co-operated in establishing training schemes, provided at the Hosiery Technology Center, that enabled firms to effectively use new technology without requiring higher levels of formal education among workers (Willis et al., 2003). It is not in services, but it is an example worth following. References Bassi, L.J. and Ashenfelter, O. (1985), “The effect of direct job creation and training programs on low-skilled workers”, in Hallock, K.F. (Ed.), The Collected Essays of Orley Ashenfelter. Vol. 2. Education, Training and Discrimination, Economists of the Twentieth Century series, reproduced, Edward Elgar Ltd, Cheltenham, pp. 249-67. Brunello, G. and Medio, A. (2001), “An explanation of international differences in education and workplace training”, European Economic Review, Vol. 45, pp. 307-22. Burdett, K. and Smith, E. (2002), “The low skill trap”, European Economic Review, Vol. 46, pp. 1439-51. De Grip, A. and Borghans, L. (Eds) (2000), Skill Utilization and Bumping down, Edward Elgar Publishing Ltd, Cheltenham. Dre`ze, J.H. (2002), “Economic and social security: the role of the EU”, The Economist, Vol. 150 No. 1, pp. 1-18. Elias, P. and Davies, R. (2004), “Employer-provided training within the European Union: a comparative review”, in Sofer, C. (Ed.), Human Capital over the Life Cycle: A European Perspective, Edward Elgar Publishing Ltd, Cheltenham. Falkinger, J. and Grossmann, V. (2001), “Skill supply, supervision requirements and unemployment of low-skilled labor”, International Journal of Manpower, Vol. 22 No. 1/2, pp. 69-82. Fouge`re, D., Kramarz, F. and Magnac, T. (2000), “Youth employment policies in France”, European Economic Review, Vol. 44, pp. 928-42. Gautier, P.A. (2002), “Unemployment and search externalities in a model with heterogeneous jobs and workers”, Economica, Vol. 69, pp. 21-40. Glyn, A. and Salverda, W. (2002), “Is wage flexibility the answer?”, Challenge, Vol. 41 No. 1, pp. 32-43.

Company training and services 15

IJM 25,1

16

Keep, E. and Mayhew, K. (1999), “The assessment: knowledge, skills and competitiveness”, Oxford Review of Economic Policy, Vol. 15 No. 1, pp. 1-15. Kramarz, F. (2000), “The French experience of youth employment programmes and payroll tax exemptions”, in Gregory, M., Salverda, W. and Bazen, S. (Eds), Labour Market Inequalities, Problems and Policies of Low-wage Employment in International Perspective, Oxford University Press, Oxford, pp. 141-56. Krueger, A. and Rouse, C. (1998), “The effect of workplace education on earnings, turnover, and job performance”, Journal of Labor Economics, Vol. 16 No. 1, pp. 61-94. Krugman, P. (1994), “Past and prospective causes of high unemployment”, Reducing Unemployment. Current Issues and Options, Proceedings of a Symposium in Jackson Hole, Wyoming, Federal Reserve Bank of Kansas, Kansas City, pp. 68-81. Mu¨hlau, P. and Salverda, W. (2000), “Employment effects of low-wage subsidies: the case of ‘SPAK’ in The Netherlands”, in Salverda, W., Lucifora, C. and Nolan, B. (Eds), Policy Measures for Low-wage Employment in Europe, Edward Elgar Publishing Ltd, Cheltenham, pp. 67-92. Nestler, K. and Kailis, E. (2002a), “Continuing vocational training in enterprises in the European Union and Norway (CVTS2)”, Statistics in Focus, Population and Social Conditions, Theme 3-3/2002. Nestler, K. and Kailis, E. (2002b), “First survey of continuing vocational training in enterprises in candidate countries (CVTS2)”, Statistics in Focus, Population and Social Conditions, Theme 3-2/2002. Nestler, K. and Kailis, E. (2002c), “Costs and funding of continuing vocational training in enterprises in Europe”, Statistics in Focus, Population and Social Conditions, Theme 3-8/2002. Nestler, K. and Kailis, E. (2002d), “Providers and fields of continuing vocational training in enterprises in Europe”, Statistics in Focus, Population and Social Conditions, Theme 3-10/2002. Nestler, K. and Kailis, E. (2002e), “Disparities in access to continuing vocational training in enterprises in Europe”, Statistics in Focus, Population and Social Conditions, Theme 3-22/2002. OECD (1991), Employment Outlook, Paris. OECD (1993), Employment Outlook, Paris. OECD (1999), Employment Outlook, Paris. Steedman, H. and McIntosh, S. (2001), “Measuring low skills in Europe: how useful is the ISCED framework?”, Oxford Economic Papers, Vol. 3, pp. 564-81. Stevens, M. (1994), “A theoretical model of on-the-job training with imperfect competition”, Oxford Economic Papers, Vol. 46 No. 4, pp. 537-62. Stevens, M. (1999), “Human capital theory and UK vocational training policy”, Oxford Review of Economic Policy, Vol. 15 No. 1, pp. 16-32. Van der Klink, M.R. and Streumer, J.N. (2002), “Effectiveness of on-the-job training”, Journal of European Industrial Training, Vol. 26 No. 2/3/4, pp. 196-9. Willis, R.A., Connelly, R. and DeGraff, D.S. (2003), “The future of jobs in the hosiery industry”, in Appelbaum, E., Bernhardt, A. and Murnane, R.J. (Eds), Low-Wage America: How Employers Are Reshaping Opportunity in the Workplace, Russell Sage Foundation, New York, NY, pp. 407-45.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Company training and low-skill consumer-service jobs in Ireland

Company training

17

Gerard Hughes, Philip J. O’Connell and James Williams Economic and Social Research Institute, Dublin, Ireland Keywords Training, Skills, Service industries, Ireland Abstract This paper identifies market forces which induce employers to provide training in Ireland. It investigates if they are present in sufficient strength in the consumer service sectors with a high concentration of low-skill jobs to provide a basis to upgrade such jobs. Data from a survey of firms on training incidence, duration, and cost are used in OLS regressions to investigate the determinants of training at national and sector level. The results show that firm size, the proportion of skilled workers, foreign ownership, perception of changing skill requirements and tightness of the labour market all influence employers’ training decisions. Analysis of sector-specific effects indicates that firms in consumer service sectors are unlikely to respond to market forces by increasing training to a level which would encompass low-skill jobs. However, policies involving the school system and company-based training could help to enhance low-skill jobs in consumer service sectors.

Introduction In recent years, there has been renewed interest in the importance of education and training in furthering the goals of economic progress, fuller employment and social integration. This resurgence of interest coincides with a new emphasis on life-long learning, both responding to rapid changes in the organisation and technology of economic activity as well as increased uncertainty and flexibility in labour markets. Investment in training by firms plays an important part in overall investments in human capital and company-sponsored training of employees plays a particularly important role in life-long learning. However, there is mounting evidence to suggest that participation in company-sponsored training is strongly related to initial educational attainment (OECD, 1999; Shields, 1998), so the lower skilled are less likely to receive continuing training than those already in possession of higher skills. This suggests that current patterns of company-sponsored training are likely to exacerbate rather than to mitigate labour market inequalities. In general, the rapid growth in employment in Ireland has entailed a substantial increase in high skilled employment (Duffy et al., 2001; Sexton et al., The authors are grateful for helpful comments on an earlier draft to this journal’s anonymous referees and to Rita Asplund, Steve Bazen, Wiemer Salverda and other participants at the ETLA and LoWER conference in Helsinki on 24-25 May 2002, on “Adapting Education and Training for the Enhancement of Low-Skilled Jobs”.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 17-35 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410524974

IJM 25,1

18

2002), and has been characterised as an upgrading of the labour force in a modernising economy (O’Connell, 2000). Nonetheless, alongside these developments, there has also been substantial growth in low-skill and low wage occupations. The fastest growing occupational groups in Ireland in the 1990s include personal service and sales workers, as well as, of course, managers and professionals. One way to increase added value and incomes in low skilled consumer-related occupations would be to increase investment in training such workers. Upgrading the skills of shop assistants, waiters and waitresses, canteen workers, cleaners, hairdressers, and other personal service workers through education and training could increase the possibility of reasonably paid jobs for such workers and thus render careers in consumer services more attractive. The purposes of this paper are to investigate what determines the incidence, duration and cost of training in private sector firms in Ireland and whether market forces could be relied upon to convert the existing low-skill jobs in consumer service sectors into stepping stones to recognise qualifications through the provision of more formal training. The paper is based on a survey of vacancies in Irish companies in 1998. In the next section we provide an outline of company-sponsored training in Ireland. This is followed by a description of the survey. We then discuss the identification of consumer service sectors in Ireland which employ mainly low-skill workers using data on employment and earnings by sector and level of education by occupation. Summary statistics are provided of the incidence, duration, and cost of training in different sectors. Variables which are likely to influence the provision of company training are presented and the reasons why they are expected to have an influence on the provision of training are outlined. Regression models are estimated to test the influence of these variables on training and the paper finishes with some conclusions based on the implications of the regression results. Company training in Ireland For many years government policy in Ireland has been to support company training in certain sectors through the market since employers and employees benefit most from the returns to investment in training. Job-related training in ´ S the Ireland is, therefore, primarily the responsibility of employers with FA government’s Training and Employment Authority providing support in three ways (Harper and Fox, 1999): (1) a levy/grant system for the textiles, clothing/footwear, food, drink, and tobacco, and chemical and allied product industries which is funded by an annual levy of 0.1 per cent of payroll; (2) an apprenticeship system for the construction, engineering, printing and motor trade industries which is funded by an annual levy of 0.25 per cent of payroll;

(3) a training support scheme which provides grants to small firms in manufacturing, internationally traded services (primarily financial and business services), and construction to assist them in providing training for their employees. As government support for company training is primarily focused on manufacturing rather than service sectors, it is worth exploring the differences in the training effort between manufacturing and services. We are specifically interested in whether the companies in consumer service sectors are likely to provide training for low-skill workers which would be sufficient to result in an upgrading of these jobs in Ireland.

Company training

19

The survey Information on the provision of training by employers in different sectors was collected by the Economic and Social Research Institute (ESRI) in a national survey of vacancies in Ireland in 1998 (Williams and Hughes, 1999). The survey was commissioned by Forfa´s (the National Policy and Advisory Board for Enterprise, Trade, Science, Technology and Innovation) on behalf of the Irish government’s Expert Group on Future Skills Needs. Basic information on the survey is given in Table I. Of the 2,685 firms selected for the sample, completed responses were received from 1,068 to give a response rate of 40 per cent. This is considered as an acceptable response rate for a postal survey with telephone follow-up calls to encourage firms to complete the questionnaire. A grossing factor is used to enable us to make estimates for the entire population of relevant firms in the country according to the size and sector. In preparing the firm-based weights, each firm is assigned a weight which allows us to provide estimates of what the response from the population would have been, if all relevant private sector firms had been successfully interviewed. Within each size-sector category, each firm (in both sample and population) is considered as a single separate entity. Table II shows data on training activities of firms by sector from the 1998 Vacancies Survey, weighted to be the representative of the national population of private sector firms. Finance and business services stands out as the sector

Number Successfully completed 1,068 Returned too late 60 Refused 96 Non response 1,447 Returned but unusable 14 Total 2,685 Source: ESRI National Survey of Vacancies in Ireland (1998)

Per cent 40 2 4 54 0 100

Table I. Response outcomes from the 1998 survey of vacancies

IJM 25,1

20

Table II. Training indicators by economic sector (weighted)

Per cent of employees Training days/total receiving training employment (days) Traditional manufacturing 12.9 0.5 Hi-tech manufacturing 17.2 0.9 Construction 7.5 0.1 Distribution 9.3 0.5 Finance and business services 21.9 1.0 Transport, personal and other services 13.5 0.5 Total 12.4 0.5 Source: ESRI National Survey of Vacancies in Ireland (1998)

Expenditure on training courses as per cent of total labour costs 1.4 1.6 0.3 0.6 1.5 0.5 0.8

with the most training: 22 per cent of employees in the sector received training, with an average intensity of 1 day of training per employee, and at a cost of 1.5 per cent of payroll. The sector with the lowest training indicators was construction, with only 7.5 per cent of employees receiving training, with average intensity of one-tenth of a day, and at a cost of 0.3 per cent of payroll. This is followed closely by the Distribution sector, in which 9 per cent of the employees received training at a cost of 0.6 per cent of payroll. The proportion of employees receiving training in the transport, personal and other services sector was a little higher than the national average, training intensity was average, and training expenditure was below average. Definition of low-skill consumer service sectors The impact on employment of a change in the composition of demand in favour of consumer services is likely to be more in sectors where large number of low-skill workers are employed. Consumer service sectors include retailing, catering, personal care, transport and communications, and finance and business. However, not all of these sectors have high concentrations of low-skill jobs. The cross classification of employment in 1997 by occupation and sector (Table III) shows that just over two-thirds of all sales workers are employed in the distribution sector and nearly half of all personal service workers are employed in transport/personal/other services. Very few sales or personal service workers are employed in the finance/insurance/business services or other sectors apart from public administration. In terms of the number employed, sales and personal workers account for over 40 per cent of employment in the distribution sector and over 34 per cent in the transport and communications sector. Data in Table IV for 1997 show that sales and personal service workers have almost the lowest average hourly earnings of full-time employees in Ireland. Only labourers and other unskilled workers have less hourly earnings (Sexton

Agri.

Agricultural 90.9 Managers and proprietors 1.1 Proprietors in services 0.1 Professional occupations 0.4 Associate professionals 0.1 Clerical occupations 0.6 Skilled workers (maintenance) 0.2 Other skilled workers 0.2 Production operatives 0.6 Transport/communications 0.4 Sales workers 0.0 Security workers 0.1 Personal service workers 0.3 Labourers and others 4.9 Total 100.0 Number employed (000s) 134.2 Source: Labour Force Survey 1997

Occupation and sector 0.3 7.4 0.3 4.2 2.8 9.7 5.4 24.9 32.3 4.0 3.9 0.6 1.1 3.2 100.0 160.4

12.6 11.1 35.1 0.7 1.8 0.2 0.4 2.1 100.0 129.3

11.2 44.0 6.8 3.5 0.4 0.5 0.8 21.3 100.0 96.7

0.5 2.3 0.0 4.6 0.9 3.2 7.5 3.3 2.2 3.6 39.1 0.4 1.4 1.0 100.0 179.8

0.1 11.4 13.3 2.6 1.5 12.6 1.3 1.2 1.3 1.2 6.1 3.7 6.5 1.0 100.0 139.8

0.4 9.9 0.2 26.3 6.5 34.5

Occupational share within sectors (per cent) Trad. Fin./Insur./Bus. manuf. Constr. Distrib. svcs

0.1 7.8 0.0 7.3 10.4 10.3

Hi-tech manuf.

4.8 1.8 1.1 18.7 7.7 1.0 26.8 2.1 100.0 198.7

2.3 8.8 7.8 4.2 1.1 11.9

Trans./Pers./Other svcs

Total 9.8 6.9 3.0 12.1 6.2 13.3 4.7 8.5 8.6 4.4 7.8 2.6 8.2 3.7 100.0 1338.4

Public admin. 0.9 4.8 0.1 30.6 16.8 17.0 0.7 2.0 0.9 0.9 0.4 8.4 13.8 2.6 100.0 299.3

Company training

21

Table III. Employment in Ireland in 1997 by occupation and sector

IJM 25,1

22

Table IV. Average hourly earnings of full-time employees in 1997

Occupation Professional occupations Managers Associate professionals Security workers Skilled workers (maintenance) Clerical occupations Transport/communication Other skilled workers Production operatives Sales workers Personal service workers Labourers and others Total

£/h 13.08 12.43 9.34 7.88 7.79 7.38 6.69 6.36 5.95 5.78 5.40 5.01 8.07

Source: Sexton et al. (1999, Table I)

et al., 1999). Sales workers receive just over 70 per cent of national average earnings while personal service workers receive just over two-thirds of the average. Analysis by Sexton et al. (2002) of 1996 Census of Population occupation data on the highest level of education completed show that the median years of education of sales and personal service workers was 8.9 years compared with the national average of 9.7 years. These measures of time spent in education equate to about lower secondary level for sales and personal service workers and to around upper secondary level for the average worker. The employment, earnings and education data for Ireland, therefore, indicate that sales and personal service workers in the distribution and transport, personal and other services sectors are low skill low paid workers. Having identified the occupations and consumer service sectors with high concentrations of low skill workers we will pay particular attention to the training effort in the distributive services, transport, personal, and other services sectors and to the training provided for the occupational categories sales and personal services. Factors likely to influence training provision The literature on training in Ireland (O’Connell (2002)) and the UK (Green, 1993; Green et al., 1996) suggests a number of factors which could influence training decisions: size of firm, the share of employment accounted for by each occupational group, foreign or domestic ownership, exposure to foreign competition, and staff turnover. In addition, there are suggestions in the literature that men have a higher probability of receiving training than women (Shields, 1998), but recent studies in the UK indicate that the likelihood is about the same for men and women (Clarke, 2002).

However, as the ESRI survey does not provide information on employment by sex we cannot include a variable in our models for the male/female composition of employment. As well as the variables mentioned for which we have information, the strength of the demand for labour by firms and increasing skill requirements could influence training decisions. Firms’ perceptions of changing skill requirements are captured in the ESRI survey by asking firms if, compared with the situation a year back, they feel the average worker needs more or less skill to ensure that the business can keep running effectively or that skill needs are static. The strength of the firm’s demand for labour is indicated in the ESRI survey by the vacancy rate, i.e. the number of vacancies divided by total employment. We now briefly consider how variables representing the characteristics of firms, labour market conditions, economic sector, and other factors are likely to influence the incidence, duration, and cost of training. Size of firm There are a number of reasons why larger firms are expected to provide more training than the smaller firms. First, firms providing training incur a fixed cost so if it can be spread over a large number of employees it will lower the average training cost per employee. Second, larger firms tend to pay higher wages thereby helping to retain individuals they train for longer on their payroll than smaller firms. Consequently, the returns to training investment may be higher for large firms than for small firms. Third, the loss of production due to having trained employees may be higher for small firms than for large, as O’Connell (2002) notes. In our models of training intensity and training expenditure we use five size categories: 1-9 employees, 10-49 employees, 50-99 employees, 100-249 employees and 250 or more employees. Size of firm is represented by a dummy variable equal to 1, if the firm is in a particular category and 0 otherwise. The smallest firm size category is omitted from the regressions and used as the reference category. Positive and increasing coefficients are expected as firm size increases, if firm size is a significant determinant of training intensity and training expenditure. Occupational share of employment Occupational structure is expected to influence training. Firms employing larger proportions of employees in higher occupational categories may invest more in their training since they expect a higher rate of return from such investment than from investment in employees in lower occupational categories. There may be a contrary effect if there are economies of scale in providing training for employees in larger occupational groups. On balance the first influence should dominate. Hence, we expect to find positive coefficients for terms measuring the share of employment accounted for by higher occupational groups, managerial, professional and technical workers.

Company training

23

IJM 25,1

24

Economic sector Economic sector is also expected to influence training. As noted earlier, government support for company training in Ireland is primarily focused on designated sectors within the manufacturing and internationally traded services. If government policy for these sectors is effective we expect higher coefficients for manufacturing sectors and lower or insignificant coefficients for distribution and transport, personal and other services. Foreign or domestic ownership One of the continuing driving forces behind Ireland’s economic development is the support provided through training grants and other financial incentives to encourage foreign companies to locate in Ireland. Many of these foreign firms are multinational companies, which tend to use more efficient production, management and marketing techniques than Irish firms (Go¨rg and Strobl, 2003). They also employ more highly skilled workers than domestic firms and generally appear to have a more positive attitude to training than Irish firms. Foreign owned firms are likely, therefore, to provide more training for their employees than domestic firms. The questionnaire distinguishes between Irish owned private and public companies, subsidiaries of overseas companies, international franchises in Ireland, semi-state companies, co-operative companies, and other companies. These seven types are collapsed down to three categories, Irish, foreign, and other, which are represented in our model by dummy variables. The Irish group is used as the reference category and the coefficient on the foreign variable is expected to be positive. Perceptions of changing skill needs A long-running debate about skills centres on the extent to which modern production methods result in deskilling or upskilling of the labour force (Hirschorn, 1984). At present there is a strong evidence that technological and organisational change has led to an increase in the skills required for employment (Machin and Van Reenen, 1998; OECD, 1994) and to a continuing need for the labour force to undergo training. How firms see their needs for skills in the context of such changes may be an important determinant of training provision. Firms that see an increasing need for skills to keep their business running effectively are more likely to engage in training than firms, which think that the need for skills is declining or stable. In the survey, firms were asked to consider the overall skills needed to keep their business running effectively compared to the situation a year back and to say if they regard the need for skills for the average worker as decreasing, static, or increasing. These responses are represented in our model by three dummy variables. Using the “static” response as the reference category a positive coefficient is expected for the “increasing” category with an insignificant or negative coefficient expected for the “decreasing” category.

Vacancy rate At the time of survey Ireland was experiencing the greatest boom in employment it had ever experienced. Over a quarter of all firms said that they had vacancies, and somewhat less than a quarter said that they had current vacancies, which were proving difficult to fill. These developments were reflected in a high vacancy rate at the time of the survey. Individual firms may meet their labour requirements in such circumstances by attracting labour from other firms but this is not a solution, which can work collectively for firms experiencing labour shortages. One response, which can provide a solution to a generalised labour shortage, is to upgrade the skills of the existing employees by providing training in skills, which are scarce in the labour market. If firms do respond in this fashion one would expect to find that firms with high vacancy rates are more likely to provide training than firms with low rates. The coefficient on the vacancy rate variable is therefore expected to be positive. Exposure to foreign competition The degree of competition which firms face in their markets may have some influence on their training. A rough measure of how much competition firms face is provided by the export share of total sales. Firms, which are highly exposed to foreign competition may train more than firms which sell the bulk of their output on the domestic market. If they do, there should be a positive coefficient on the variable measuring export share of total sales. Staff turnover The influence of staff turnover is ambiguous. If employment expands in response to a strong increase in demand, or if there is a substantial turnover due to a high quit rate, firms may provide more induction training than firms that are not experiencing rapid growth or high turnover. On the other hand, if high turnover is due to loss through poaching of trained staff, firms may provide less training. In different variants of our model we will use three measures of staff turnover, the recruitment rate, the quit rate, and the turnover rate (the sum of quits and recruitment divided by the labour force), to examine the relationship between staff turnover and training provision. Descriptive statistics on variables in the training model We analyse three measures of training as dependent variables in our models to capture the coverage, intensity and cost of training: (1) training participation, measured as the percentage of all employees receiving training; (2) the intensity of training as indicated by the number of days training per employee; and (3) the percentage of labour costs spent on training.

Company training

25

IJM 25,1

26

Initial inspection of these dependent variables indicated non-normal distributions. Transformation of the dependent variables by taking natural logarithms resulted in a more normal distribution. The dependent variables specified in our models are therefore: (1) the natural logarithm of the percentage trained; (2) the natural logarithm of the training days per employee; and (3) the natural logarithm of the percentage of labour costs spent on training. Examination of the dependent variables also revealed that a significant number of firms did not provide any training. This means that training provision is left-censored at zero. Tobit regressions were, therefore, estimated and compared with the results of OLS regressions. The Tobit and OLS regressions gave estimates, which are reasonably in close agreement. Consequently, we present the OLS results, as they are easier to assess than the Tobit regressions. Descriptive statistics for the variables considered above are reported in Table V. Information is provided by over 1,000 firms for all of the variables we are interested in, apart from the training cost variable, for which the number of responses received was around 900. The firms in our sample provided training for about 20 per cent of their employees and the average intensity of training was less than 1 day per employee. Training costs amounted to about 2 per cent of total labour costs on average. Two-thirds of the firms employed less than 50 people and one-third employed 50 or more. The largest occupational groups were managers, clerks, and production operatives while the smallest were personal service and security workers. Over a quarter of the firms were located in the financial, insurance, and business services sector while one-fifth of them were in the distribution sector. The construction and transport, personal and other services sectors each accounted for 10 per cent of the firms in the sample. Around one-fifth of the firms were foreign and nearly 60 per cent of all firms thought that their skills needs were increasing. The average vacancy rate (unweighted) was 9 per cent and one-quarter of total sales were made on export markets. The average recruitment rate was 25 per cent, the average quit rate was 18 per cent and the combined turnover rate was 44 per cent. Factors influencing company training at the national level In estimating the models of participation, intensity and cost of training, we first try to identify the variables which influence different measures of training at the national level and to see if the economic sector in which firms operate and the occupational structure of the firm’s labour force exert an independent influence on training provision. Then we investigate the influence of firm size and other structural variables at sector level to try and establish if employer provided training could provide a basis for upgrading low-skill jobs in the consumer service sector.

Variable Dependent variables Participation: per cent trained Ln of per cent trained Intensity: training days per employee Ln of training days Training costs as per cent of total labour costs Ln of training costs

N

Min.

Max.

Mean

Standard deviation

1,050 0.00 1,050 29.21 1,035 0 1,035 29.21

1.00 0.19 0.00 2 4.50 23.23 0.78 3.15 2 3.91

0.26 3.76 1.70 4.34

897 0.00 897 29.21

20.00 1.62 3.00 2 3.22

2.66 4.76

Independent variables Size 1-5 1,068 0 1.00 Size 10-49 1,068 0 1.00 Size 50-99 1,068 0 1.00 Size 100-249 1,068 0 1.00 Size 250+ 1,068 0 1.00 Occ. share managers 1,068 0 1.00 Occ. share professionals 1,068 0 1.00 Occ share assoc. prof. 1,068 0 0.92 Occ. share clerical 1,068 0 1.00 Occ. share maintenance 1,068 0 1.00 Occ. share production operatives 1,068 0 0.95 Occ. share transport 1,068 0 0.93 Occ. share sales 1,068 0 1.00 Occ. share security 1,068 0 0.94 Occ. share personal service 1,068 0 1.00 Occ. share labourers 1,068 0 0.92 Traditional manufacturing 1,068 0 1.00 Hi-tech manufacturing 1,068 0 1.00 Construction 1,068 0 1.00 Distributive services 1,068 0 1.00 Financial services 1,068 0 1.00 Transport and other services 1,068 0 1.00 Foreign ownership 1,062 0 1.00 Other ownership 1,062 0 1.00 Increasing skill needs 1,048 0 1.00 Decreasing skill needs 1,048 0 1.00 Vacancy rate 1,068 0 53.3 Per cent of sales exported 1,024 0 100.00 Recruitment rate 1,064 0 20.00 Quit rate 1,060 0 20.00 Turnover rate 1,058 0 40.00 Source: ESRI National Survey of Vacancies in Ireland (1998)

0.30 0.37 0.12 0.13 0.09 0.19 0.11 0.07 0.14 0.12 0.14 0.04 0.11 0.00 0.02 0.07 0.19 0.13 0.10 0.20 0.27 0.10 0.18 0.04 0.59 0.03 0.09 25.36 0.25 0.18 0.44

0.46 0.48 0.32 0.34 0.28 0.21 0.19 0.14 0.18 0.22 0.26 0.13 0.21 0.04 0.10 0.16 0.39 0.34 0.30 0.40 0.44 0.30 0.38 0.21 0.49 0.18 0.23 37.47 0.73 0.68 1.34

Our OLS regression results for the determinants of training at the national level are reported in Table VI. Firm size has a highly significant, positive and increasing effect on participation in training in equation 1, on the intensity of training in equation 2, and on the cost of training in equation 3. In the training

Company training

27

Table V. Descriptive statistics for training measures and variables likely to influence training provision

IJM 25,1

Variable

Ln (per cent trained) Coefficient t-value

Ln( training days per employee) Coefficient t-value

Ln (cost of training) Coefficient t-value

(Constant) 2 7.55*** 2 11.21 2 7.56*** 2 9.58 2 8.30*** 2 9.12 Size 10-49 1.32*** 4.56 1.49*** 4.45 2.28*** 5.89 28 Size 50-99 2.47*** 6.15 2.79*** 5.93 4.01*** 7.35 Size 100-249 2.93*** 7.11 3.35*** 7.00 5.05*** 9.28 Size 250+ 3.28*** 6.72 3.70*** 6.49 4.50*** 7.16 Managers 2 1.03 21.29 2 1.19 2 1.28 0.26 0.22 Professionals 1.94* 2.49 2.26* 2.49 2.78** 2.65 Assoc. prof. 2.83** 3.08 3.16** 2.97 2.38* 2.00 Clerical 1.63* 2.13 1.29 1.44 1.72 1.67 Maintenance 2 0.37 20.57 2 0.66 2 0.87 2 1.13 2 1.30 Transport 2 1.46 21.44 2 1.88 2 1.60 2 1.46 2 1.05 Sales 2 0.70 20.91 2 0.94 2 1.05 2 1.68 2 1.59 Security 2 0.73 20.26 0.27 0.08 2 0.12 2 0.03 Personal service 0.53 0.44 2 1.18 2 0.83 2 1.07 2 0.65 Labourers 2 0.82 21.05 2 0.50 2 0.54 2 0.31 2 0.29 Trad. manuf. 0.25 0.56 0.61 1.15 1.18* 1.96 Hi-tech manuf. 0.22 0.48 0.60 1.09 1.36* 2.19 Distribution 0.41 0.84 0.64 1.12 0.61 0.91 Financial services 0.80 1.67 1.26* 2.25 1.83** 2.83 Transport 0.45 0.82 0.96 1.51 0.62 0.83 Foreign ownership 0.80** 2.59 0.56 1.55 0.76 1.87 Other ownership 0.30 0.58 0.57 0.94 2 0.06 2 0.09 Increasing skills 1.30*** 5.88 1.57*** 6.04 1.40*** 4.71 Decreasing skills 0.76 1.30 0.95 1.37 1.20 1.47 Table VI. 1.13* 2.49 1.63** 3.11 2.06*** 3.52 OLS regressions of per Vacancy rate Adjusted R 2 0.26 – 0.24 – 0.28 – cent of employees F 15.02 – 14.37 – 15.55 – trained, average 3.24 – 3.76 – 4.00 – duration of training and S.E.E. N 1026 – 1011 – 879 – expenditure on training Notes: Significance level *P, 0.05, **P , 0.01, ***P , 0.001; and reference categories size: 1-9; on preferred construction; Irish-owned; static skills explanatory variables

participation and intensity equations the coefficients on the firm size variables increase, as expected, as firm size increases. This pattern is also evident in the training cost equation except for the largest firm size where the coefficient is somewhat smaller than the coefficient on the preceding size category. These results indicate that employees are far more likely to receive training in larger firms than in smaller ones. The size effect on training is robust in all three regressions with the coefficients on all size categories being highly significant at the 0.1 per cent level. Firms with a large proportion of their labour force in high- and intermediate-skill professional, associate professional and clerical occupations are likely to train more of their employees and to train them for longer periods than firms with a large proportion of their labour force in

low-skill occupational categories. They are also likely to spend more on training high-skill workers than on low-skill workers and in particular, to spend more on professional and associate professionals than other occupational groups. The sector variables do not appear to have as much influence on training provision as the occupational variables. None of the sectors appear to train more of their employees than the reference category (construction) when the influence of the other variables is considered. The financial services sector is the only one in which firms provide a significantly longer duration of training than the firms in the reference category. This sector together with traditional manufacturing and hi-tech manufacturing also spends a larger percentage of its payroll on training than the other sectors. In terms of the percentage of labour costs spent on training, the financial services sector spends most relative to the reference category and the hi-tech and traditional manufacturing sectors are the second and third highest spenders on training. The sector in which low-skill consumer service workers are concentrated, distribution, does not differ from the reference sector in the number of employees trained, the duration of their training or in the percentage of payroll spent on training. As hypothesised, foreign firms appear to be more focused on training their employees than the Irish firms. The coefficient on the foreign variable in the training participation equation is positive and significant at the 1 per cent level. However, training duration and training expenditure for foreign firms appear to be no different from the length of training provided by the Irish firms or their expenditure on training[1]. Firms’ perceptions of how their skill needs are changing exert a very strong influence on whether they train or not. Where skill needs were thought to have increased compared with the situation 1 year back, firms were far more likely to train their workers. Firms which considered that they had increasing skill needs were also far more likely to train their employees for a longer period and to spend more than firms who saw no change in the skills needed to run their business effectively[2]. The vacancy rate has a significant positive influence on the number of employees which the firms train, and very significant influence on both training intensity and expenditure. None of the competitiveness or turnover variables were added to the explanatory power of the training regression models, so the results for these variables are not reported. Most of the variables which determine training effort at the national level in Ireland are similar to those found to play a role by O’Connell (2002) in research using survey data for 1993 (Fox, 1995) on continuing vocational training in Ireland. The level of employment, occupational shares, sector, and foreign ownership, were all found to have a significant influence on whether the firms were trained or not. Our results show that firms’ perceptions of their skill needs also have an important additional influence on training decisions and that the

Company training

29

IJM 25,1

30

percentage of workers receiving training, the intensity of training and expenditure on training are all likely to increase as the vacancy rate increases. Factors influencing training in consumer service and other sectors Having established that firm size, occupational composition, economic sector, ownership, increasing skill needs, and tightness of the labour market influence training at national level we would like to know which of these market-related variables have significant effects in different sectors. We are particularly interested in whether the effects of such variables in the distributive services and transport, personal, and other services sectors are sufficiently strong to be relied on to result in training which could lead to upgrading of low-skill sales and personal service jobs in these sectors. To simplify presentation of results for individual sectors we focus on the factors, which influence the percentage of employees trained in different sectors. Table VII shows the OLS results for the traditional manufacturing, hi-tech manufacturing, and construction sectors. They show that the independent Traditional manufacturing Coefficient t-value

Hitech manufacturing Coefficient t-value

Construction Coefficient t-value

(Constant) 29.82*** 2 11.34 27.37*** 25.35 2 7.82 2 2.65 Size 10-49 2.83*** 4.03 0.628 0.64 1.00 0.94 Size 50-99 5.78*** 6.68 3.07* 2.42 1.26 0.95 Size 100-249 5.05*** 6.00 3.79** 3.30 3.56* 2.41 Size 250+ 5.49*** 6.05 4.46** 3.57 3.85* 2.05 Managers 4.17* 2.38 21.67 20.58 2 1.54 2 0.40 Professionals 20.45 2 0.22 6.16** 2.80 1.82 0.41 Assoc. prof. 5.36 1.75 0.28 0.10 0.91 0.24 Clerical 2.24 1.32 4.71 1.03 3.86 0.87 Maintenance 0.05 0.05 21.81 21.57 1.01 0.35 Transport 3.66 1.20 10.34* 2.21 2 4.13 2 0.56 Sales 2.26 0.94 0.31 0.10 2 6.30 2 0.55 Security 236.78 2 1.64 22.85 20.08 64.55 1.66 Personal svcs 212.38 2 0.88 0.85 0.04 260.19 2 0.46 Labourers 20.04 2 0.03 23.01 21.46 0.62 0.20 Foreign ownership 0.98 1.81 0.12 0.19 2 0.19 2 0.11 Other ownership 22.22 2 1.49 21.13 20.86 2 3.39 2 0.95 Increasing skills 0.62 1.43 0.62 1.08 1.00 1.27 Decreasing skills 20.48 2 0.37 1.39 0.46 2 0.81 2 0.51 Vacancy rate 4.61** 2.70 6.47** 3.22 1.85 0.64 Table VII. 2 Adjusted R 0.32 – 0.34 – 0.09 – OLS regressions of Ln 5.87 – 4.76 – 1.51 – of per cent of employees F S.E.E. 2.81 – 2.87 – 3.47 – trained in the traditional N 200 – 139 – 103 – and hi-tech Notes: Significance level * p ,0.05, ** p , 0.01, *** p , 0.001; and reference categories size: manufacturing and 1-9; Irish-owned; static skills construction sectors

variables, which influence the percentage of employees, trained at the national level generally also influence the percentage of employees trained in the traditional manufacturing and hi-tech sectors. Employees in the traditional manufacturing sector are more likely to be trained if they work in larger firms, if their firm has a larger proportion of workers in managerial occupations, and if the firm has vacancies. For the hi-tech sector the coefficients on the firm size variables increase as the size increases and all the coefficients are significant except for size category 10-49. The most important influences on training in the hi-tech sector appear to be the shares of professional and transport workers and the vacancy rate. Hence, hi-tech firms with a larger share of professional and transport workers are more likely to train a larger percentage of the labour force than other firms, and training effort rises as the vacancy rate rises. For the construction sector, the F statistic indicates that the overall fit of the equation is poor and the only factor which appears to have any influence on the percentage of the labour force which is trained is firm size. None of the variables representing firm ownership or perceptions of skill requirements are significant in the traditional manufacturing, hi-tech manufacturing or construction sectors. The OLS regression results for the distribution, financial, insurance, and business services, and transport, personal, and other services are presented in Table VIII. The overall fit of the equation for the transport and other services sector is reasonable with the main influences on training participation being firm size and firms’ perceptions of increasing skill requirements. The fit of the model for the financial services sector is poor with only 12 per cent of the variance explained by the influence of the largest firm size, increasing skills and the vacancy rate. The fit for the distribution sector is better with 22 per cent of the variance explained and with significant coefficients on some of the firm size and occupational share variables and on the increasing skills variable. However, the signs of all of the coefficients on the occupational share variables for the distribution sector are negative and those on the managerial and sales worker share variables are significantly different from zero, indicating that firms in this sector with larger proportions of their work force in managerial and low-skill jobs provide less training than firms with lower proportions of their work force in low-skill jobs. Neither foreign ownership nor the vacancy rate appears to play a role in training decisions in the distribution sector. However, firms in the distribution sector, which say that the skills they need to run their business effectively are, increasing, are more likely to train their employees than other firms. For all three-service sectors, firm-size effects for the largest firms are less evident than for the manufacturing sectors. The analysis of group-specific effects shows that even though training provision in services sectors is responsive to firms’ perceptions of changing skill requirements, the high concentration of sales workers in the distribution sector reduces the percentage trained and the high concentration of personal

Company training

31

IJM 25,1 Distribution Coefficient t-value

Financial, insurance and business services Coefficient t-value

Transport, personal and other services Coefficient t-value

(Constant) 23.06 2 1.32 2 5.34** 22.63 25.54* 2 2.12 Size 10-49 2.10*** 3.44 0.49 0.93 1.71 1.77 Size 50-99 1.36 1.34 1.22 1.52 3.14** 2.65 Size 100-249 3.12* 2.56 1.77 1.94 2.91* 2.04 Size 250+ 9.05 1.43 2.45* 2.28 1.53 1.04 Managers 26.69** 2 2.63 2 1.17 20.54 24.40 2 1.41 Professionals 24.13 2 0.98 1.05 0.51 22.57 2 0.72 Assoc. prof. 21.91 2 0.50 2.41 1.10 4.02 1.06 Clerical 21.47 2 0.50 0.79 0.37 21.31 2 0.46 Maintenance 22.41 2 0.91 2 1.61 20.60 2.99 0.67 Transport 24.61 2 1.69 2 10.04 21.86 24.54 2 1.56 Sales 25.17* 2 2.20 2 2.01 20.83 21.05 2 0.35 Security 210.19 2 1.17 0.03 0.01 215.73 2 0.77 Personal svcs 28.98 2 1.37 2 2.84 20.94 0.33 0.12 Labourers 25.30 2 1.83 0.57 0.19 24.29 2 1.34 Foreign ownership 1.54 0.99 0.83 1.54 0.90 0.70 Other ownership 2.03 1.29 0.02 0.03 1.30 1.18 Increasing skills 1.52** 2.93 1.10* 2.24 2.67*** 3.61 Decreasing skills 0.48 0.37 2.89 1.96 2.03 1.31 Vacancy rate 22.07 2 1.28 1.23* 2.03 21.64 2 0.72 Table VIII. Adjusted R 2 0.22 – 0.12 – 0.333 – OLS regressions of Ln F 3.99 – 3.04 – 3.73 – of per cent of employees S.E.E. 3.42 – 3.36 – 3.12 – trained in the N 205 – 274 – 105 – distribution, financial Notes: Significance level * p ,0.05, ** p , 0.01, *** p , 0.001; and reference categories size: and transport services 1-9; Irish-owned; static skills sectors

32

services workers in the transport, personal and other services sector exerts no influence on the percentage trained. These results suggest that companies in consumer service sectors are unlikely to provide additional training for low-skill workers in response to changes in market forces which would be sufficient to result in an upgrading of these jobs in Ireland. Conclusions At national level, the factors analysed in this paper which influence the incidence, duration and cost of company training in Ireland are the number of people employed, the skill composition of the labour force, economic sector, foreign ownership, perception of skill requirements and the tightness of the labour market. Larger firms provide more training for their employees for a longer period and they allocate a higher proportion of labour costs to training. Firms, which have a high proportion of high-skill employees in their labour force, are likely to train more and firms in manufacturing are likely to spend

more on training than firms in services. The pace of technological change, as reflected in firms’ perceptions of their skill requirements, plays an important role in decisions to provide training and how much to spend on it. The tightness of the labour market exerts a strong independent influence on the incidence, duration, and cost of training. If firms cannot find the skills they require for their business due to labour shortages they are far more likely to train their existing labour force than firms, which can recruit workers with the skills they need on the open market. Firms in the distribution and transport, personal, and other services sectors do not differ significantly in their training effort from firms in the construction sector which has the lowest incidence, duration, and expenditure on training of any of the sectors in our sample. Low-skill sales and personal service workers do not differ in the training they receive from production operatives who are themselves low-skilled workers. The analysis of group-specific effects shows that firms in consumer service sectors, in which most low-skill sales and personal service workers are employed, are unlikely to respond to market forces by increasing expenditure on training to a level which would encompass sales and personal service workers and result in an upgrading of low-skill consumer service jobs in Ireland. What could help to enhance low-skill jobs in consumer services is an approach on two fronts involving the school system and company-based training. International research on training by the OECD (1999) indicates that there are complementarities between schooling and training. Hence, policies, which improve general education and literacy levels, can result in a greater supply of skills leading to a greater demand for them. This would encourage workers to look for more training and provide an incentive for employers and workers to invest in continuing vocational training. The important role, which the education system could play in eventually increasing training in the workplace, has also been noted by Shields (1998). This approach will pay dividends only in the long run as students emerge from the school system with the literacy, knowledge, and learning skills which they need to position them to receive continuing training in the workplace. As the effect of this education-based approach would be very gradual it would also be desirable to provide additional training for the existing workforce in low-skill consumer service jobs, which would provide them with more marketable skills and raise their living standards. One possibility would be to extend the Training Support Scheme to provide training for low-skill workers in the consumer service and other service sectors. This programme provides training subsidies, which are concentrated on small firms employing up to 50 employees in manufacturing and internationally traded services. Up to 80 per cent of training costs can be provided for very small firms and up to 20 per cent can be provided for larger firms. It is primarily a demand-driven

Company training

33

IJM 25,1

34

scheme, which allows companies to identify their training needs on the basis of a business development plan and to purchase training services in the open market. The principle on which funding is provided to companies is that, where necessary, the purchasers of training should be subsidised rather than the providers of training. An evaluation of the Training Support Scheme by O’Connell and Lyons (1995) found that the best results were achieved by small firms who would not have been able to pay for training without the subsidy. They argued that a case could be made for extending the scheme to the services sector since that is where most employment growth is occurring[3]. That case is supported by the results reported in this paper. Consideration should, therefore, be given in identifying how the Training Support Scheme could be adapted to cover consumer service sectors and to establish what skills would be required to upgrade low-skill jobs through the provision of grant aid to improve the skills of the existing workers. Notes 1. Given the importance of foreign direct investment in Ireland, and the fact that foreign firms were found to train a higher percentage of their employees, we estimated a series of interaction terms between foreign ownership and firm size. None of the interaction terms achieved statistical significance, and they are not reported in tabular form. 2. We also estimated a series of interaction terms between increasing needs and firm size. Again, however, none of the interaction terms proved significant. 3. The scheme was subsequently revised to increase the targeting on small firms but the sectoral coverage remained unchanged.

References Clarke, A. (2002), “Who trains? Employers’ commitment to workforce development”, Labour Market Trends, Vol. 110 No. 6, pp. 319-24. Duffy, D., FitzGerald, J., Hore, J., Kearney, I. and MacCoile, C. (2001), Medium-Term Review, No. 8, September, Economic and Social Research Institute, Dublin. ´ S – The Training and Employment Authority, Fox, R. (1995), Company Training in Ireland, FA Dublin. Go¨rg, H. and Strobl, E. (2003), “Multinational companies, technology spillovers, and plant survival”, Scandinavian Journal of Economics, Vol. 105 No. 4, pp. 581-95. Green, F. (1993), “The determinants of training of male and female employees in Britain”, Oxford Bulletin of Economics and Statistics, Vol. 55 No. 1. Green, F., Machin, S. and Wilkinson, D. (1996), An Analysis of Workplace Training and Skill Shortages, HMSO, London. Harper, D. and Fox, R. (1999), Vocational Education and Training in Ireland 1998, CEDEFOP monograph. Hirschorn, L. (1984), Beyond Mechanization, MIT Press, Cambridge, MA. Machin, S. and Van Reenen, J. (1998), “Technology and changes in skill structure: evidence from seven OECD countries”, Quarterly Journal of Economics, Vol. CXIII No. 4, pp. 1215-44.

O’Connell, P.J. (2000), “The dynamics of the Irish Labour Market in comparative perspective”, in Nolan, B., O’Connell, P.J. and Whelan, C.T. (Eds), Bust to Boom? The Irish Experience of Growth and Inequality, Institute of Public Administration, Dublin. O’Connell, P.J. (2002), “Does enterprise-sponsored training aggravate or alleviate existing inequalities? Evidence from Ireland”, in Schoemann, K. and O’Connell, P. (Eds), Education, Training and Employment Dynamics: Transitional Labour Markets in the European Union, Edward Elgar, Cheltenham, pp. 285-300. O’Connell, P. and Lyons, M. (1995), Enterprise-related Training and State Policy in Ireland: The Training Support Scheme, Policy Research Series Paper No. 25, The Economic and Social Research Institute, Dublin. OECD (1994), “Medium-term perspectives on labour supply and occupational change”, OECD Employment Outlook 1994, Organisation for Economic Co-operation and Development, Paris, pp. 71-100. OECD (1999), “Training of adult workers in OECD countries: measurement and analysis”, OECD Employment Outlook 1999, Organisation for Economic Co-operation and Development, Paris, pp. 135-75. ´ S/ESRI Sexton, J., Hughes, G. and Finn, C. (2002), Occupational Employment Forecasts 2015, FA Manpower Forecasting Studies. Report No. 10. Sexton, J., Nolan, B. and McCormick, B. (1999), “Review of earnings trends in the Irish economy since 1987”, Quarterly Economic Commentary, Special Article, December, Economic and Social Research Institute, Dublin. Shields, M. (1998), “Changes in the determinants of employer-funded training for full-time employees in Britain, 1984-1994”, Oxford Bulletin of Economics and Statistics, Vol. 60 No. 2, pp. 189-214. Williams, J. and Hughes, G. (1999), National Survey of Vacancies in the Private Non-Agricultural Sector 1998, Economic and Social Research Institute, Dublin. Further reading Hughes, G., McCormick, B. and Sexton, J.J. (2000), Occupational Employment Forecasts 2005, ´ S/ESRI Manpower Forecasting Studies. Report No. 8, Economic and Social Research FA Institute, Dublin.

Company training

35

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

IJM 25,1

36

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Why do business service firms employ fewer apprentices? A comparison between Germany and The Netherlands Wendy Smits Research Centre for Education and the Labour Market (ROA), University of Maastricht, Maastricht, The Netherlands, and

Thomas Zwick Centre for European Economic Research (ZEW), Mannheim, Germany Keywords Apprenticeships, Skills, Business support services, Germany, The Netherlands Abstract This paper analyses why in Germany and The Netherlands the share of apprentices in the business service sector is lower than in other economic sectors. A theoretical introduction surveys the potential reasons that could be responsible for this. The subsequent empirical analysis shows that the level of skill apprentices gain is the main explanation for the relatively low supply of apprenticeships in German business service enterprises. In The Netherlands, the option to hire skilled employees from full-time schools instead of training apprentices seems to be crucial. For these reasons, this paper proposes to offer obligatory extra formal training in areas such as IT skills and foreign languages for the apprentices in business service firms in Germany in order to increase the attractiveness of the dual apprenticeship system for prospective apprentices as well as business service firms.

Introduction The labour markets in Europe are characterized by a shift of employment to tertiary sectors, mainly to service enterprises, and a decrease in the level of employment in the primary and secondary sectors. Within the service sector, the strongest employment dynamics are experienced in the knowledge-intensive firms such as business services (Zwick and Schro¨der, 2001). Most firms in the dynamic business service sector are relatively young and small. In addition, the qualification demand and the share of highly qualified employees in the business service sector are higher than in other service sectors and the manufacturing sector. Therefore, one would expect that firms in the business service sector invest much in the training of their workforce in order to qualify their employees suitably. The opposite seems to be true, however, at least when we consider apprenticeship training. In Germany and in The Netherlands, two countries where the apprenticeship system is well developed, we see that International Journal of Manpower Vol. 25 No. 1, 2004 pp. 36-54 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410524983

The authors thank Rita Asplund, Andries de Grip, Wiemer Salverda, and a referee for their useful comments. They are also indebted to the Institut fu¨r Arbeitsmarkt- und Berufsforschung (IAB) for the possibility to use the IAB establishment panel.

firms in the business service sector train significantly fewer apprentices Germany and than firms in other sectors. The Netherlands In this paper, we will investigate the reasons for the lower motivation of business service enterprises to supply apprenticeships, compared to firms in other sectors. This issue is relevant both politically and scientifically. In Germany, it is felt that the rather poor performance of the apprenticeship 37 system in the business service sector threatens the attractiveness of the system as a whole, since the most capable young people who wish to find employment in a sector with a promising future currently choose to continue higher education instead of serving an apprenticeship. This reduces the attractiveness of the pool of applicants and deteriorates the firms’ perception of the dual apprenticeship system as a suitable means to get skilled employees. Therefore, it is an important political question how the willingness to train apprentices can be increased in the business service sector. Although this issue is high on the political agenda, only few empirical studies are available that focus on the motivation of firms to provide apprenticeships. The literature on the supply of apprenticeships concentrates largely on the reasons why firms offer general training and on the impact of innovations and new technologies on qualification demand. Both elements are not specific for the business service sector, however. There are no empirical studies that analyse whether (business) service enterprises derive the same competitive advantage from the apprenticeship systems as manufacturing companies do (Culpepper, 1999). The apprenticeship systems are frequently focused on the requirements of the manufacturing industry or established service sectors such as banks or insurance companies. There are, for example, far fewer service profession certificates than there are manufacturing occupations. An additional difference between the sectors may result from the fact that apprenticeships lead to a qualification at an intermediate level, while firms in the business service sector require relatively more skills of a higher level. Finally, qualifications in information-intensive service sectors are often rather abstract and theoretical, and the question therefore may be whether these skills can be taught at the work floor. If, from a pedagogical point of view, the skills needed in the business service sector are best learnt at school, then the apprenticeship system with its current mixture of formal school training and practical training in firms is probably not suited for these skills. To the best of our knowledge, international comparisons regarding the willingness of enterprises to provide apprenticeships are not available yet. German firms rely completely on the dual apprenticeship system for the acquisition of skills on this qualification level, whereas Dutch firms can choose between employees trained in full-time schools and those trained in a dual apprenticeship system. The difference between the German and the Dutch apprenticeship systems provides us with a natural experiment so that we can

IJM 25,1

38

see whether the limited (pedagogical) suitability of the apprenticeship system for teaching complex abstract skills is indeed a reason for the low motivation of firms in the business service sector to participate in this system. By comparing the situation in the business service sector with that in other sectors, we gain more specific insights in the economic and institutional factors that determine the success of the apprenticeship system in a sector of industry. The absence of empirical studies on the determinants of the supply of apprenticeships is mainly a result of the lack of suitable data sets, because there seem to be no data that have been specifically collected for this question. In this paper, we use three establishment level data sets that comprise questions on the salient factors determining the willingness of German and Dutch firms to offer apprenticeships. The supply of apprenticeships Economic reasons The motivation of firms to supply apprenticeships has received much attention in the theoretical economic literature during the last decades. It usually makes a distinction between current production considerations and a future need for qualified labour (Lindley, 1975; Stevens, 1994a). In the first case, the main reason for employing apprentices is their contribution to the production process. The firm then replaces unskilled employees by apprentices. This motive may be important if apprentices’ wages are relatively low compared to the unskilled wage rate and if training costs are low (Soskice, 1994). In the second case, the firm trains apprentices to fill its future need for qualified workers. The question that then arises is: What are the future benefits of training to the firm? The firm can also hire skilled workers, who have been trained elsewhere and thus save on training costs. Furthermore, since apprenticeship training provides skills that are also useful in other firms, the firm risks losing its internally trained workers to competitors. This feature of apprenticeship training raises the question why firms should be willing to supply apprenticeships even when there are considerable costs involved, as is the case for many firms in Germany and The Netherlands. In fact, if apprenticeship training were completely general, as in the definition of Becker (1964), that is the training could be used in many firms, the increase in productivity would be fully reflected in the future market wages of the apprentices. Since apprentices are free to move once they have completed their training, the training firm cannot pay them less than these market wages. The training firm will not profit from the training and as a consequence is not prepared to pay for it. This theoretical argument is difficult to reconcile with the empirical finding that firms supply apprenticeships and bear a substantial part of the training costs (Franz and Soskice, 1995; Harhoff and Kane, 1997; Soskice, 1994). In recent years, economists have put forward many arguments why firms nevertheless have an interest in supplying apprenticeships instead

of hiring skilled labour on the market. Most of these arguments have in Germany and common that, although apprenticeship training generates skills that are useful The Netherlands in more firms than only the training firm, the market for these skills is not perfectly competitive. This creates a wedge between the productivity of apprentices and the market wages, which induces firms to provide apprenticeships. 39 The first argument concerns incomplete information with respect to the contents of the apprenticeship programme. External firms find it difficult to judge the quality of a training programme, and therefore the increase in productivity is not fully reflected in the market wages for skilled employees (Chang and Wang, 1996; Katz and Ziderman, 1990). This argument has little validity in Germany and The Netherlands, because in both countries apprenticeship training is highly certified and its contents are transparent and standardised. A second argument is informational asymmetry with respect to the quality of the apprentice’s skills. It is argued that during the apprenticeship period, the training firm learns about the qualifications of the apprentice, and therefore the training firm is able to select the best workers (Acemoglu and Pischke, 1998; Elbaum and Singh, 1995; Euwals and Winkelmann, 2001; Franz and Soskice, 1995). If firms dismiss apprentices who do not fit their requirements after their apprenticeship and if apprentices do not quit voluntarily on a large scale, the firms that seek skilled workers with a degree from the dual system can only choose from a second-rate pool of applicants. A third argument is that the market for the skills that an apprentice learns may be quite small. Even if the qualifications are useful only in a limited number of firms, they are still transferable (Stevens, 1994b), but not generally as in the definition of Becker. If the number of firms in a trade is limited, the apprentice might not succeed in obtaining a job in the trade once he or she has completed the training. The training firm thus has some monopsony power and can appropriate some of the returns to training. This argument may have some validity for both countries. Blechinger and Pfeiffer (2000) show that in Germany occupational change after an apprenticeship leads to a considerable decline in the applicability of apprenticeship training. Moreover, in both countries the skilled wage rate is often determined at the sector level. Therefore, it is quite likely that due to this monopsony power the skilled wage rate in some sectors is below productivity. There is, however, no empirical evidence that this is in fact the case. At the same time, it is clear that this argument does not apply to the business service sector. There are only few skills at the apprenticeship qualification level that are exclusively needed in the business service sector. Finally, some of the training may be firm-specific (Smits, 1997). It is often inevitable that an apprentice also learns some skills that are firm-specific, since a firm may have its own unique equipment and machinery with which the apprentice learns to work, and the apprentice also learns something about the

IJM 25,1

40

culture of the organisation. Moreover, if there is an interaction between general and firm-specific skills, for example, if general skills can be used more efficiently when the worker has some firm-specific knowledge and skills, it is possible for the training firm to appropriate some of the returns to general training (Acemoglu and Pischke, 1999; Kessler and Lu¨lfesmann, 2000). These theories predict a wedge between productivity and the wages the training firm pays for internally trained workers. These augmentations of Becker’s human capital theory show that firms may indeed have a long-term interest in providing apprenticeship training. The empirical evidence on these theories is, however, scattered and often contradictory (Smits and Stromback, 2001). Many theories will have some validity for both Germany and The Netherlands, but none can fully explain the success of the apprenticeship system in Germany and The Netherlands only. Institutional factors seem to be equally important. Institutional peculiarities in Germany and The Netherlands The dual apprenticeship system is the backbone of professional qualification in Germany. The majority of employees acquired their skills during their period of apprenticeship, and most firms indicate that employees with a degree from the dual apprenticeship system are the main source of qualified labour. Because of that, the dual apprenticeship system is considered a successful institution to induce employers to participate in human capital investments by work-based training. Consequently, the attractiveness of the dual apprenticeship system is high on the political agenda, and the state heavily subsidises the schools that provide the theoretical part of the education (Lindner, 1998). For Germany, several additional institutional determinants of the willingness to provide apprenticeships, which are only weakly related to the direct short- and long-term use of apprentices, have to be mentioned. Several authors stress that in Germany employers’ organisations and chambers of commerce minimise the risk of opportunistic behaviour of firms (Culpepper, 1999). This reduces poaching and provides information, for example, on innovations in training practice for training firms. Harhoff and Kane (1997) mention a norm within the German business community to provide apprenticeship training that applies to all sectors of the economy. Franz and Soskice (1995) argue that social networks for small enterprises and in rural regions or smaller cities also urge firms to provide training. Harhoff and Kane (1997) add evidence to this conjecture by the observation that firms in the countryside are more willing to provide training. Harhoff and Kane (1997) argue that work councils controlled by unions also reduce the ability of non-training firms to attract workers. They can do so by influencing wage agreements and salary scales for skilled workers trained internally or by other firms (Soskice, 1994). In Germany, the wage mark-up for skilled workers attracted from the labour market is legally constrained, and

therefore firms cannot pay considerably more for workers hired from other Germany and firms. The Netherlands Finally, the German government promotes apprenticeships by regular good-will campaigns. Every year, the German government campaigns for an increase in the number of training firms by directly asking firms that do not participate in the dual apprenticeship system to do so. The goal is that every 41 applicant can find an apprenticeship place in the same year. This increases public awareness of the social importance of apprenticeship training (Franz et al., 2000, p. 65). In The Netherlands, the apprenticeship system is well developed but less important than in Germany. Social pressure to train apprentices is less and the importance on the political agenda is also lower than in Germany. Most vocational qualifications at the intermediate level can be obtained both by serving an apprenticeship and by attending full-time education. Although both routes, apprenticeships and full-time education, are formally equivalent, for some occupations an apprenticeship is the most common route while for other occupations it is full-time education. On average, about 25 per cent of qualified school-leavers at the secondary vocational level served an apprenticeship (Borghans and Smits, 1996; Borghans et al., 2000). Apprenticeships are most common in the sectors of industry that are well organised. In many of these sectors, there are sector agreements on training. These agreements dictate that all firms have to pay training contributions irrespective of whether they train, and firms that train apprentices receive subsidies. This is, for example, the case of the building trade. In the building trade, the share of apprentices in the total number of school-leavers qualified for this trade is more than 50 per cent. The same is true for metal industry occupations and motor mechanics. In care occupations, such as nursing, the share of apprentices used to be about 70 per cent, but this figure has fallen in recent years due to the reorganisation of training in hospitals. In administrative occupations, on the other hand, the share of apprentices is very low, between 10 per cent for business administration and 19 per cent for secretaries. An exception are probably IT occupations, which are also classified under administrative occupations. There are no reliable figures available for IT occupations, because a great deal of training for such occupations is provided by private schools and these courses were only recently integrated in the formal education system. Since most training for IT occupations was originally initiated by firms, it is very likely that the share of apprentices is higher in these occupations. The type of apprenticeship also differs considerably between occupations. In most technical occupations, apprentices usually do not only have an apprenticeship agreement with their training firm but also an employment contract. In this case, the apprentice receives wages. In administrative occupations, apprentices often have only an apprenticeship agreement and no employment contract, which means that they do not earn salary, although

IJM 25,1

42

sometimes they receive some compensations. In this case, the cost of employing apprentices is much lower. The data The analysis makes use of two German data sources. These sources contain answers of firms to questions about their training behaviour and their opinion on apprenticeship training. The ZEW-infas questionnaire is a telephone interview of 1,502 firms in selected service sectors, conducted in April 2000. The answers are weighted in order to obtain representative results for the sectors included in the survey. The focus of the questionnaire are the willingness of firms to participate in the dual apprenticeship system, barriers against firm training, and how topical the qualifications acquired during the apprenticeship are. Detailed information can also be found in the work of Zwick and Schro¨der (2001). The IAB Establishment Panel is a yearly panel performed since 1993 asking German establishments about their behaviour concerning training and other personnel measures. Firms in this panel are drawn from all establishments in Germany with at least one employee, who has a social security number. Therefore, only establishments consisting of employees not covered by social insurance (mainly farmers, miners, artists, and journalists) along with public enterprises with only civil servants are excluded. There is a large set of questions that are asked every year about production, investments, sector, employee structure, business strategy, and vocational training. Every year, other questions are added on an irregular basis. In 2000, additional detailed information was collected on the reasons why firms did not offer apprenticeships although they were eligible to have apprentices. A detailed description is given by Ko¨lling (2000). For The Netherlands, we have made use of a telephone survey among Dutch firms that was conducted in August 1999. The survey addressed both their willingness to train apprentices and their preferences with respect to workers, who have completed an apprenticeship or full-time education. All economic sectors were covered. About 1,000 firms were interviewed, of which 87 firms operated in the business service sector. In order to get representative results, all answers were weighted (Borghans et al., 2000). Apprenticeships in the business sector The business service sector is one of the most rapidly growing sectors in Germany and The Netherlands (e.g. Zwick and Schro¨der, 2001). It comprises mainly renting, electronic data processing, research and development, and services for enterprises and covers the enterprises with the NACE codes 71-74 (Hass, 1995). The share of the apprentices among the number of employees with a degree from the dual apprenticeship system is, in Germany, slightly lower in the service sector than in industry (Table I). The Mikrozensus also

Professional education

Industrial sector

No degree Apprenticeship Apprenticeship degree Master/technician Polytechnic degree University degree Ratio apprentices/skilled workers Source: German Mikrozensus, wave

Services (except business services)

Business services

0.14 0.13 0.04 0.04 0.63 0.60 0.10 0.06 0.05 0.05 0.04 0.13 0.07 0.06 1996, own calculations

0.13 0.03 0.48 0.07 0.10 0.19 0.06

Germany and The Netherlands

43 Table I. Professional education shares in Germany

indicates that in business services the qualification demand is greater than in the industrial sector and the share of apprentices is relatively low. The share of employees with an apprenticeship degree as well as the ratio of apprentices/skilled workers are also significantly lower than in other sectors. This leads us to the observation that firms in the German business service sector have a lower willingness to offer apprenticeships and prefer employing higher skilled employees. Higher education is also much more important in the Dutch business service sector than in most other service sectors (Table II). In 1998, 43 per cent of the people working in the business services had attained at least higher vocational education, while only 37 per cent had intermediate education. In the economy as a whole, these figures were 36 and 45 per cent, respectively. Only approximately 5 per cent of the firms in the business services train apprentices, while the average in the Dutch economy is 15 per cent. Reasons for willingness to train apprentices Training costs Apprenticeship costs are one of the main reasons for German firms not to train (Table AI in the Appendix). Therefore, the costs may also be a reason for the differences in the motivation to train between business services and other sectors. There are some differences in apprenticeship costs among professions.

Professional education No degree Lower vocational education Intermediate vocational education (including apprenticeship degree) Higher vocational education University degree Source: ROA/CBS, averages 1997/1998

Business services

Total

0.06 0.14

0.07 0.21

0.37 0.26 0.18

0.45 0.18 0.09

Table II. Professional education shares in The Netherlands

IJM 25,1

44

In technical professions, the net costs are significantly lower than in commercial professions. The apprentice’s productivity is also lower during the apprenticeship in the industrial and technical professions, however, and therefore the total costs differ only about 10 per cent in Germany (von Bardeleben et al., 1997). A multivariate probit regression that explains why German firms that are eligible to train do not offer apprenticeships by the fact that apprentices are too expensive, shows that business service firms mention this reason less frequently than firms in other sectors (Table III). Therefore, we can conclude that training costs do not have a different impact on the willingness of the German business service firms to train than on firms from other sectors. Also in The Netherlands, training costs are one of the main deterrents for firms to employ apprentices (Table AIII in the Appendix). Only 45 per cent of the training firms in the business services report that they recover the expenses of the training during the training period, while for the economy as a whole this figure is 66 per cent. The main reason for this difference in costs is that firms in the business service sector cannot easily integrate apprentices into their production processes. Only 32 per cent of the training firms in the business service sector state that apprentices play an important role in their productions process, while the average across all sectors is 64 per cent. Although 72 per cent of the firms in the Dutch business service sector say that apprentices are given the same tasks as skilled workers, 69 per cent indicate that these concern tasks to learn from. For the economy as a whole, these figures are 56 and 46 per cent, respectively. Apprentices in the business service sector very often have the same tasks as skilled workers, but since they have to learn from these tasks, and hence make more mistakes, they are far less productive. The higher costs due to the low productivity of apprentices during their training period are partly compensated by lower wage costs. In the Dutch Reason

Table III. Probit estimations to explain which firms do not offer apprenticeships, differences between business services and other sectors in Germany

Coefficient

z-value

Apprentices are too expensive 2 0.148* 2 1.70 Apprentices quit although the firm would like to keep them 2 0.136 2 0.90 Preference for hiring experienced skilled employees 0.0201** 2.19 Training contents are outdated due to technological progress 0.346*** 2.35 Existing professions in the dual apprenticeship system are not compatible with the qualifications required 0.224** 1.98 Apprentices spend too much time in school 2 0.430*** 2 3.38 Notes: Significance levels: *p , 0.1; **p , 0.05; ***p , 0.01. The multivariate probit estimations entail the following dummy explanatory variables: business service sector, firm size 1-19, firm size 20-199, firm size 200-499, East Germany and constant. The number of observations is 2,018. Here the coefficient of the dummy business services is given, i.e. the coefficient shows the differences between this sector and all other sectors. The complete probit estimation of the first row is displayed in Table AII in the appendix, as an example Source: IAB Establishment Panel, wave 2000, own calculations

business service sector, the majority of the training firms do not pay salary for Germany and their apprentices but only some allowance for expenses. In this case, the The Netherlands apprentice has an apprenticeship contract with the firm but not an employment contract. In most other economic sectors, it is common that the apprentice has both an employment and apprenticeship contract. On average, 60 per cent of all firms in the economy give their apprentices an employment contract, while in 45 the business service sector this is only 30 per cent of the firms. Table IV shows that training costs are not more important in the decision not to train for firms in the business service sector than for firms in other economic sectors in a multivariate analysis. As a result, we can conclude that in The Netherlands training costs are not a decisive issue in the explanation of the differences in the willingness to train either. Returns to training In Germany, the certification system of apprenticeship degrees assures that all apprentices can be employed in every firm of the sector (Harhoff and Kane, 1997; Lindner, 1998). The tests apprentices must pass to receive their certificates focus upon skills that are generally applicable in the sector and they are monitored by local chambers of commerce and work councils (Franz and Soskice, 1995). This applies to all professions. Therefore, the basic knowledge that is necessary for acquiring a degree can be labelled as general, well-known to everybody, and marketable elsewhere (Euwals and Winkelmann, 2001; Franz et al., 2000; Soskice, 1994). The fact that apprentices leaving the firm immediately after receiving their certificates earn a positive mark-up also points in this direction (Harhoff and Kane, 1997). This system decreases the monopsony power of firms, because apprentices are free to offer their human capital to firms in a different sector. In addition, firms hiring apprentices from other firms are confident that such employees fulfil certain minimum standards

Reason

Coefficient

Wald-test

Apprentices are too expensive 20.070 1.43 Apprentices involve too much administration 20.025 0.15 Apprentices quit although the firm would like to keep them 0.015 0.15 Difficult to fulfil legislative prerequisites to train apprentices 0.093** 4.57 Difficult to find suitable apprentices 20.005 0.22 Notes: Significance levels: **p , 0.05. The logit estimations entail the following dummy variables: business service sector, firm size 1-9, firm size 10-99 and a constant. The number of observations is 592. Only the coefficient of the dummy for business services is reported here. The complete logit estimation of the first row is displayed in Table AIV in the Appendix, as an example Source: ROA Survey, wave 1999, own calculations

Table IV. Logit estimations to explain which firms do not offer apprenticeships, differences between business services and other sectors, in The Netherlands

IJM 25,1

46

that are tested and certified by government-sponsored third parties (the chambers of commerce). As the certification system is the same for the entire economy, these arguments apply to business services and other sectors alike. Other sources of monopsony power may play a role, however, as argued earlier. For example, several authors find evidence that German training firms have superior information on apprentice ability which allows them to employ the most able workers at a wage below productivity. There is no empirical evidence, however, that this theory – if it is correct – has more validity in certain sectors than in others. Like Germany, The Netherlands also has a comprehensive system of certification for the apprenticeship system so that training meets certain minimum quality standards. Nevertheless, there are differences in training quality between firms (Smits, 1999). However, there is no evidence that training quality is more or less variable in the business services than in other sectors of industry. Firm-specificity seems to play a role in some firms in the Dutch business service sector. Training firms in the business service sector report more often (54 per cent) that it is difficult to employ workers trained in other firms (the average over all economic sectors is 24 per cent). At the same time, only 30 per cent of the firms in the business service sector that do not train report that it is difficult to employ workers trained elsewhere. So firm-specificity is a reason for training only a small group of firms in the business service sector. We may conclude that there is little empirical evidence that the potential future returns to training are lower in the business service sector than in other sectors, neither in Germany nor in The Netherlands. There may be differences in future returns, however, if there are differences between the shares of apprentices who stay after completing their training. The share of apprentices taken over in the business service sector is lower than the average in the German economy (Table V). A multivariate probit estimation shows, however, that there are no differences between business service firms and other firms in their perception of apprentices’ quitting as a reason not to offer apprenticeships (Table III). In addition, a consequence of the social pressure not to poach apprentices from other firms is that apprentices leaving although the establishment would like to keep them is not a widespread problem in Germany (Table AI in the Appendix). Economic activity

Table V. Share of apprentices taken over in Germany

Share of apprentices taken over

All sectors except business services Business services Total Source: IAB Establishment Panel, wave 2000, own calculations

0.68 0.57 0.67

The probability that a Dutch firm in the business service sector offers Germany and employment to an apprentice after completing his or her training hardly differs The Netherlands from the probability in other sectors. In the economy as a whole, the average probability is 74 per cent, while in the business service sector it is 77 per cent (Table VI). However, whereas a total of 25 per cent of all apprentices already have an employment contract before the end of their training period, this share 47 is only 5 per cent in the business service sector. Correcting for the low share of apprentices in permanent employment, the actual chance that an apprentice can stay in the training firm is much lower than the average for the economy. The chance that a worker accepts an employment offer once the apprenticeship is served is slightly higher in the business service sector than the average for the economy (76 against 72 per cent, Table VI). So, we can see that there is no large difference between the retention rates of the Dutch business service sector and other sectors either. Moreover, Table IV shows that also in a multivariate analysis business service firms do not refrain from training because of apprentices quitting after the training more often than firms in other economic sectors. Social recognition Social recognition of providing training in the dual apprenticeship system and consciousness of the positive impact of training on society is strong in Germany. Almost all German service firms agree with the statement that firms have a social responsibility to train. More than 94 per cent of the service firms that offer apprenticeship training and a stunning share of more than 90 per cent of the firms that do not offer apprenticeships point out that firms have a social responsibility to offer training, according to the ZEW-infas survey. Social recognition is also disseminated into the young and dynamic business service firms, in which more than 92 per cent accept the social responsibility of firms. In The Netherlands, on average 50 per cent of all training firms say that a very important reason to train is the willingness to contribute to schooling in the sector. At the same time, 30 per cent of the firms that do not train say that this is an important consideration for future training. In the business service sector, only 25 per cent of the training firms say that this is a very important reason to train. It becomes clear that social recognition of training plays a much smaller role in The Netherlands in comparison to Germany, and Dutch

Economic activity

Chance of employment offer by the firm

Business services 0.77 Total 0.74 Source: ROA survey, wave 1999, own calculations

Chance that the apprentice accepts employment offer 0.76 0.72

Table VI. Average expected retention rates in Dutch sectors

IJM 25,1

48

business service firms feel less committed to train by social pressure than firms from other sectors. Eligibility for training Another important determinant for offering training in both countries is that instructors and training programmes must meet several legal requirements before being entitled to offer training. According to the IAB Establishment Panel in 2000, 39 per cent of all German firms indicated that they did not meet the requirements for training apprentices. The share of business service firms is slightly above this average (Table VII). The ZEW-infas study also shows that the requirements for achieving the eligibility for training play a major role in the explanation why firms do not train. Of the non-training firms, 48 per cent (the share is the same for business service firms) say that they do not have the legal eligibility for training, while 34 per cent (38 per cent of the business service firms) state that the bureaucracy involved in obtaining eligibility is a deterrent. The share of firms indicating that they are not eligible to train is especially high, of course, in sectors with low shares of firms participating in apprenticeship training. On average, 60 per cent of those enterprises that are eligible to train also employ apprentices (BMBF, 1999). Like in Germany, nearly half of the business service enterprises in The Netherlands that do not train give as a reason for this that it is too difficult to fulfil the legislative prerequisites for apprenticeships, which is also a problem in other economic sectors. On average, 40 per cent of the firms mention legislative problems. Therefore, in The Netherlands and Germany eligibility criteria are one of the most serious barriers to training. Business service firms in The Netherlands do face some extra problems in comparison to other sectors, however (Table IV). Availability of suitable apprentices The ZEW-infas questionnaire on the service sector detected that one of the main obstacles to more training is the qualification level of applicants for apprenticeship vacancies. Around 40 per cent of the firms explain that they have great problems finding suitable apprentices. A multivariate analysis on the

Table VII. Share of companies that do not fulfil legislative prerequisites for apprenticeship training and share of firms offering training in Germany

Economic activity

Share of companies that do not fulfil legislative prerequisites for apprenticeship

All sectors except business services 0.39 Business services 0.42 Total 0.39 Source: IAB Establishment Panel, wave 2000, own calculations

Share of firms offering apprenticeship training 0.50 0.42 0.49

basis of the IAB Establishment Panel indicates, however, that there is no Germany and difference between business service firms and firms from other sectors (not The Netherlands shown here). While some firms do not offer apprenticeships because they cannot find suitable apprentices (Table AI in the Appendix), this phenomenon applies to all sectors alike. On the other hand, business service firms mention significantly more often that they prefer to cover their qualification needs by hiring 49 experienced and skilled employees (Table III). Dutch firms also have problems finding suitable apprentices. More than half of the firms in the business sector say that it is difficult to find qualified applicants for apprenticeships. This figure does not deviate from the average over all economic sectors, however (Table IV). Qualification of apprentices Another important determinant for the willingness to train apprentices in the business service sector are the qualifications acquired by apprentices during their period of apprenticeship. Although not many firms mention that it is a reason not to train (Table AI), the qualifications needed by skilled employees with an apprenticeship degree in the information-intensive business service sector are probably different from those in other sectors. A first indication that supports the idea that the qualifications of apprentices show gaps, especially when it comes to handling new technologies, is that the share of these employees is negatively related to the investments in IT technology and innovation expenditures in the German service sector (Jacobebbinghaus and Zwick, 2002). A second indication is the subjective evaluation of the applicability of the knowledge acquired during their apprenticeship by the employees with a degree from the dual apprenticeship system. Workers, who are programming or working with computers can make less use of their apprenticeship training when compared to colleagues, who do not use these tools (Blechinger and Pfeiffer, 2000). The last piece of evidence is that significantly more business service firms than firms in other sectors point to the fact that the contents of apprenticeship training has been rendered obsolete by technical progress, and therefore they do not offer apprenticeships (Table III). In the ZEW-infas questionnaire, the firms have also been directly asked, which qualifications are lacking in which professions. The answers of business service firms are centred on commercial professions, but some of the new IT professions were also named. The qualification gaps mentioned most frequently were IT skills and foreign languages (Zwick, 2001; Zwick and Schro¨der, 2001). It seems that the process of fitting professions to the needs of firms does not adequately take account of the new qualification requirements of the young and small information-intensive firms of the business service sector. In Germany, the contents, job titles and requirements of professions are equally determined by employers, organisations, and unions. The German Ministry of Economic Affairs acts as a neutral referee between unions and employer

IJM 25,1

50

associations, and the process is closely tracked by government-sponsored specialists from the Bundesinstitut fu¨r Berufsbildung (BIBB). This procedure is time-consuming and conservative (Zwick and Schro¨der, 2001). A consequence may be that firms in new or quickly changing business areas do not feel that the professions and qualifications taught in the dual apprenticeship system are compatible with the qualifications they require and therefore look for other sources of adequately skilled employees instead. This seems to apply especially to business services (Table III). Business service firms perceive significantly larger gaps between the qualifications obtained during the apprenticeship and those the firm require than other firms. For The Netherlands, we do not have comparable evidence, and therefore we cannot judge if the training contents are also evaluated significantly more negatively by business service sector firms than by other firms. Choice between full-time education and an apprenticeship In The Netherlands, for most occupations there are two possible routes to obtain a qualification: serving an apprenticeship or attending full-time education. Formally, both routes are equivalent but it is very well possible that full-time education is better suited for some occupations and an apprenticeship for other occupations. Occupations that demand a great deal of theoretical knowledge are probably better learned at school while occupations that require practical skills are better learned at the workplace. For most occupations that are important for the business service sector, theoretical knowledge is quite important. This may also explain the low share of apprentices in the Dutch business service sector. In the ROA survey, employers were asked about their opinion on the quality of workers who served an apprenticeship and workers who completed full-time education. In most economic sectors, employers seem to have a strong preference for workers who served an apprenticeship: on average, 53 per cent of all employers have a preference for former apprentices, while a quarter is indifferent between graduates from either route. In the business service sector, however, 40 per cent of the firms judge both routes equally, while only 36 per cent have a preference for former apprentices and 24 per cent for school-leavers from full-time education. We can conclude that in the business service sector one important factor for employing fewer apprentices is the opportunity to hire employees who received training from full-time education programmes. The main difference between the reasons of firms that are eligible to train for not offering apprenticeships in the business service sectors and other sectors in Germany seems to be the applicability of the qualifications taught during the course of studies. An interesting question is therefore, whether full-time schools such as those in The Netherlands may be an appropriate alternative for the dual apprenticeship system. The argument may be that, in particular, skills in IT techniques or foreign languages cannot be taught during the practical part

of an apprenticeship but have to be acquired in the more theoretical setting of Germany and schools. Or alternatively, there are IT skills such as configuring networks or The Netherlands working with large databases that can be taught at the workfloor but require thorough theoretical knowledge before they can be learnt in practice. This hypothesis is backed by a clear difference in the reasons for not having apprentices between the business service sector and the other sectors in 51 Germany: while many firms in other sectors explain that they do not train apprentices because they spend too much time in school, establishments in the business service sector give a significantly lower priority to this reason (Table III). Conclusions This paper assessed empirically the different reasons for firms to offer apprenticeships. We focused on the differences between business service firms and firms in other sectors of the economy. This focus was motivated by the fact that firms in the business service sector have significantly fewer apprentices than other sectors. The business service sector is one of the most thriving sectors of the economy, and therefore the attractiveness of the apprenticeship system itself may suffer from a lack of opportunities for apprentices to find employment in this sector. In Germany, the main reason for the difference in willingness to train apprentices between business service sector firms and other firms is the applicability of the qualifications that apprentices acquire. Business service firms complain more often that the qualifications, especially in IT skills and foreign languages, do not fit their requirements. We know from other studies that information-intensive and innovative enterprises employ significantly fewer employees with an apprenticeship degree, and that employees with an apprenticeship degree can use only a small part of the knowledge acquired during their apprenticeship when they work with computers. In The Netherlands, unlike in most other economic sectors, firms in the business service sector do not have a clear preference for workers who have served an apprenticeship over school-leavers from full-time education. Apparently, firms derive less benefit from training own workers than from recruiting qualified school-leavers in this sector. This fact may partly explain the low training rates in the Dutch business service sector. On the basis of the experiences in the Dutch business service sector with its choice between full-time school leavers and apprentices, no recommendation for a shift in the education system for apprentices to full-time education can be made for Germany. Probably, the best way to increase the motivation of business service firms to train more apprentices is to close the gaps in the qualifications of apprentices by offering additional obligatory formal training courses in foreign languages and IT skills. These contents are more important in this information-intensive sector than in other sectors. This would increase

IJM 25,1

52

the share of time the apprentices in the business service sector spend in school, but it seems as if the firms do not consider this as a major problem. References Acemoglu, D. and Pischke, J. (1998), “Why do firms train? Theory and evidence”, Quarterly Journal of Economics, Vol. 113 No. 1, pp. 79-119. Acemoglu, D. and Pischke, J. (1999), “The structure of wages and investment in general training”, Journal of Political Economy, Vol. 107 No. 3, pp. 539-72. Becker, G. (1964), Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education, The University of Chicago Press, Chicago, IL. Blechinger, D. and Pfeiffer, F. (2000), “Technological change and skill obsolescence: the case of German apprenticeship training”, in Heijke, H. (Ed.), Education, Training and Employment in the Knowledge Based Economy, Macmillan, Basingstoke, pp. 243-78. Borghans, L. and Smits, W. (1996), Ontwikkelingen in het leerlingwezen tot 2000, Discussion Paper ROA-R-1996/6, Maastricht. Borghans, L., Smits, W., Vlasblom, D. and Jacobs, A. (2000), Leren en werken in het Nederlandse beroepsonderwijs, Vraag- en aanbodontwikkelingen voor de BBL 1999-2004, ROA-R-2000/2, Maastricht. Bundesministerium fu¨r Bildung und Forschung (BMBF) (1999), Berufsbildungsbericht 1999, Bonn. Chang, C. and Wang, Y. (1996), “Human capital investment under asymmetric information: the Pigovian conjecture revisited”, Journal of Labour Economics, Vol. 14 No. 3, pp. 505-19. Culpepper, P. (1999), “The future of the high-skill equilibrium in Germany”, Oxford Review of Economic Policy, Vol. 15 No. 1, pp. 43-59. Elbaum, B. and Singh, N. (1995), “The economic rationale of apprenticeship training: some lessons from British and US experience”, Industrial Relations, Vol. 34 No. 4, pp. 593-622. Euwals, R. and Winkelmann, R. (2001), Why Do Firms Train? Empirical Evidence on the First Labour Market Outcomes of Graduated Apprentices, IZA Discussion Paper 319, Bonn. Franz, W. and Soskice, D. (1995), “The German apprenticeship system”, in Buttler, F., Franz, W., Schettkat, R. and Soskice, D. (Eds), Institutional Frameworks and Labor Market Performance, Routledge, London, pp. 208-34. Franz, W., Steiner, V. and Zimmermann, V. (2000), Die betriebliche Ausbildungsbereitschaft im technologischen und demographischen Wandel, Nomos, Baden-Baden. Harhoff, D. and Kane, T. (1997), “Is the German apprenticeship system a panacea of the US labor market?”, Journal of Population Economics, Vol. 10, pp. 171-96. ¨ konomische Bedeutung und politische Hass, H. (1995), Industrienahe Dienstleistungen – O Herausforderung, Beitra¨ge zur Wirtschafts- und Sozialpolitik, Institut der deutschen Wirtschaft, Ko¨ln. Jacobebbinghaus, P. and Zwick, T. (2002), “New technologies and the demand for medium qualified labour in Germany”, Schmollers Jahrbuch, Vol. 122 No. 2, pp. 179-206. Katz, E. and Ziderman, A. (1990), “Investment in general training: the role of information and labour mobility”, The Economic Journal, Vol. 100, pp. 1147-58. Kessler, A. and Lu¨lfesmann, C. (2000), “The theory of human capital revisited: on the interaction of general and specific investments”, iCEPR Discussion Paper 2533, London. Ko¨lling, A. (2000), “The IAB Establishment Panel”, Schmollers Jahrbuch, Vol. 120, pp. 291-300.

Lindley, R. (1975), “The demand for apprentice recruits by the engineering industry 1951-1971”, Scottish Journal of Political Economy, Vol. 22 No. 1, pp. 1-24. Lindner, A. (1998), “Modelling the German system of vocational education”, Labour Economics, Vol. 5, pp. 411-23. Smits, W. (1997), “Apprenticeship training under imperfect information”, paper presented at the 1997 EALE Conference, Aarhus. Smits, W. (1999), “Schoolverlaters van de BBL”, Discussion Paper ROA-R-1999/6, Maastricht. Smits, W. and Stromback, T. (2001), The Economics of the Apprenticeship System, Edward Elgar, Cheltenham. Soskice, D. (1994), “Reconciling markets and institutions: the German apprenticeship system”, in Lisa, L. (Ed.), Training and the Private Sector: International Comparisons, University of Chicago Press, Chicago, IL. Stevens, M. (1994a), “An investment model for the supply of training by employers”, The Economic Journal, Vol. 104, pp. 556-70. Stevens, M. (1994b), “A theoretical model of on-the-job training with imperfect competition”, Oxford Economic Papers, Vol. 46, pp. 537-62. von Bardeleben, R., Beicht, U. and Fehe´r, K. (1997), Was kostet die betriebliche Ausbildung? Fortschreibung der Ergebnisse 1991 auf den Stand 1995, Bertelsmann, Bielefeld. Zwick, T. (2001), “Neue Technologien gefa¨hrden die Attraktivita¨t der dualen Ausbildung. Bescha¨ftigungsmo¨ glichkeiten von Fachkra¨ften in informationsintensiven Dienstleistungsunternehmen”, Mitteilungen aus der Arbeitsmarkt- und Berufsforschung, Vol. 34, pp. 74-81. Zwick, T. and Schro¨der, H. (2001), Wie aktuell ist die Berufsausbildung im Dienstleistungssektor? Sektorale Besonderheiten und deren Auswirkungen auf den Qualifikationsbedarf, Nomos, Baden-Baden.

Germany and The Netherlands

53

Appendix

Reason

Business service establishments

Apprentices are too expensive 0.39 The establishment cannot take over apprentices after the apprenticeship 0.37 Apprentices quit although the firm would like to keep them 0.02 The apprentices spend too much time in school 0.15 Training contents are outdated due to technological progress 0.03 We do not find suitable applicants 0.08 Existing professions are not compatible with qualifications required 0.09 We can cover our qualification demand by hiring skilled employees from other firms 0.06 Preference to hire experienced skilled employees 0.14 Source: IAB establishment panel, wave 2000, own calculations

All establishments 0.38 0.33 0.05 0.20 0.02 0.10 0.06 0.04 0.15

Table AI. Reasons not to hire apprentices although the establishment is eligible to train, shares

IJM 25,1

54 Table AII. Probit estimation to explain which firms do not offer apprenticeships, because apprentices are too expensive, Germany

Exogenous variables

Table AIV. Logit estimation to explain which firms do not offer apprenticeships, because apprentices are too expensive, The Netherlands

z-value

Business services 2 0.148 Firm size 1-19 0.744 Firm size 20-199 0.276 Firm size 200-499 2 0.111 East Germany 0.172 Constant 2 1.051 Number of observations: 2018 Wald X 2 (6): 78.67 Source: IAB establishment panel, wave 2000, own calculations

Reason

Table AIII. Importance of reasons not to hire apprentices, share of firms, The Netherlands

Coefficient

Apprentices are too expensive Apprentices involve too much administration Apprentices quit although the firm would like to keep them Difficult to fulfil legislative prerequisites to train apprentices Difficult to find suitable apprentices Source: ROA survey, wave 1999, own calculations

Exogenous variables

2 1.70 2.16 0.80 0.29 2.78 2 3.07 –

Very important

Slightly important

Not important

0.54 0.73

0.27 0.13

0.12 0.12

0.49

0.24

0.23

0.55 0.31

0.24 0.21

0.16 0.36

Coefficient

Business services 2 0.148 Firm size 1-9 2.187 Firm size 10-99 1.1487 Constant 2 3.863 Number of observations: 592 X 2 (3): 13.95 Source: ROA survey, wave 1999, own calculations

Wald-test 1.43 4.57 2.05 14.63 –

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

The role of training in changing an economy specialising in tourism

Role of training in changing an economy 55

Vicente Ramos, Javier Rey-Maquieira and Maria Tugores University of the Balearic Islands, Baleares, Spain Keywords Education, Training, Tourism management, Balearic Islands Abstract This paper compares the training requirements of alternative tourism development strategies which are differentiated by the quality of service offered. The paper focuses on the Balearic Islands and use an original database that consists of a representative sample of Balearic hotels. This database includes data on both employers and employees and allows us to identify differences in job characteristics, as well as differences in human capital, with respect to both education and on-the-job training, depending on the category of the hotel. The article uses a discrete choice model to identify the characteristics of both employer and employee that determine the provision of training. It concludes that educational level is not a strong constraint on the mobility of workers between categories, and we show that on-the-job training has a role to play in the transition to alternative tourism development strategies.

Introduction Motivation: the economy of the Balearic Islands Patterns of specialisation are not irrelevant to the capacity of an economy to generate income and welfare. This is indeed true for an economy specialising in tourism, where the kind of market segments in which the economy specialises determines its competitiveness. On the one hand, a mass-tourism development strategy usually means price competition and therefore, the need to maintain low prices and costs, with consequent dependence on large numbers of visitors who generate congestion problems and externalities at an elevated environmental, cultural and social cost. These externalities not only reduce residents’ utility but also affect the tourism services production function, and hence threaten growth sustainability. On the other hand, a high-quality tourism development strategy implies competition based on differentiation and innovation, a lower number of visitors and therefore a greater potential for sustainable growth. In view of this, there is growing concern about the need to change the pattern of specialisation and to shift resources from low-quality to high-quality tourism services in several mature tourism destinations that depend heavily on mass tourism. This structural change has implications for the labour market, The authors are grateful to the financial support provided by the Conselleria de Treball of the local government of the Balearic Islands. This institution gave them the opportunity to construct the database used in this paper. The authors are also grateful to an anonymous referee of this journal.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 55-72 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410524992

IJM 25,1

56

since there may be substantial differences in the wages and employment skills required in these two kinds of tourism supply. Our paper analyses the role of training differentials among low-quality and high-quality tourism services as well as the training determinants in the tourism accommodation sector, focusing on the Balearic Islands. The Balearic Islands are a very important tourism destination in Spain and are one of the most well-known tourism centres in the Mediterranean. Fewer than 800,000 people reside in the Balearic Islands, while in 2001 the number of tourists totalled nearly 12 million. Most of these tourists (66.23 per cent) stayed in hotels. The Balearic Islands have a supply of about 400,000 beds and the number of overnight stays totalled about 63 million in 2001. Over 55,000 residents work in this sector, representing more than 15 per cent of the working population. Tourism accounts for about 40 per cent of total aggregate demand in the Balearic economy. Within the context of tourism congestion, a social debate has arisen in the Balearic Islands about possible strategies to reduce the environmental and social costs of tourism development and to guarantee its sustainability. In this debate, special attention has been drawn to the need for a structural change in favour of high-quality tourism, given that higher quality tourism services would presumably allow for the same or a higher income from a smaller number of tourists. This paper does not try to evaluate the convenience of moving from low-quality to high-quality tourism services but instead considers its potential labour market effects. We are specifically interested in training workers need in order to shift from mass-tourism to the high-quality tourism sector. This is a crucial aspect in assessing the possibilities of a tourism-based economy to accomplish successfully the structural change commented on above and in evaluating the associated transition costs. We use hotel category as the element that differentiates between high- and low-quality tourism services. The reason for this is that, legislation regulating the different categories of hotels is associated with the quality of the services they offer (e.g. sport facilities, varieties of culinary services, size of rooms, frequency of room cleaning service, availability of types of technological equipment). We define three groups of hotels according to category as follows. (1) One- and two-star hotels: these hotels provide low-quality tourism services. (2) Three-star hotels: these hotels represent middle-quality services. (3) Four- and five-star hotels: these hotels represent the upper end of service quality distribution. We also assume no demand restrictions in the case of a structural change. This is a common hypothesis despite the fact that the Balearic Islands represent a large part of the tourism supply in Spain. Compared to other destinations (e.g.

Greece, Italy, France, Croatia), the supply offered in the Balearic Islands is insignificant, and so the non-existence of demand restrictions is not a very strong assumption. In the analysis of the training differentials between the two tourism sectors, we follow different steps. After analysing the survey design and the database used, we begin the empirical analysis. First, from a descriptive point of view, we compare the differences in the workers’ job characteristics, the differences in human capital in both education and on-the-job training, and the training requirements in hotels of different categories. Secondly, we use a discrete choice model in order to identify the characteristics of both employers and employees that determine the provision of worker training. Finally, in the last section we present our main conclusions. Related literature Investment in on-the-job training is a key element of human capital accumulation. It has received considerable attention in economic policy circles, where it is seen as an instrument to raise productivity and standards of living, especially of low skilled workers. Several contributions show the importance of investing in human capital, and of investing in on-the-job training in particular, as a competitive strategy to generate sustainable growth and wealth. Theoretical papers mainly focus on the division of the costs and benefits of this investment, as well as on the conditions under which the level of training is sub-optimal, starting with Becker (1962), and followed by the work of Acemoglu and Pischke (1999), Stevens (1994) and Tugores (1998), for example. From the empirical point of view, most of the studies have focused on the extent and impact of investing in human capital. Most analyse the quantity of training provided and the determinants of investing in this type of human capital (e.g. job characteristics, previous education, individual characteristics) (Altonji and Spletzer, 1991; Francis et al., 1998; Greenhalgh and Stewart, 1987; Kennedy et al., 1994). Regarding the estimation of returns from the different types of human capital, we can distinguish different types of paper. On the one hand, there are papers that analyse the effects of formal education (e.g. by specifying a wage offer equation, as proposed by Mincer (1974)). On the other hand, there are papers that focus on returns from on-the-job training (Bartel, 1995; Bishop, 1994; Booth and Bryan, 2002; Lynch, 1992). This branch of literature has shown the positive effect of investing in human capital on wages and job stability (Booth and Satchell, 1994; Lillard and Tan, 1992; Mincer, 1983). However, in the academic literature only a few studies focus on training needs within a context of structural change. There is a branch of related literature that estimates occupational projections (Fina et al., 2000, in Spain). We have not found any evidence of this particular type of research within the

Role of training in changing an economy 57

IJM 25,1

58

context of the tourism sector. Our descriptive analysis of differences in educational and training needs according to hotel category thus represents an important step in this direction. However, other empirical studies have been undertaken within the context of human capital accumulation in the tourism sector. A number of papers analyse training needs in the tourism sector (Amoah and Baum, 1997; Baum, 1994; Brotherton et al., 1994; Woodbury, 1992). As far as we know, the issue of training requirements for a shift from mass to high-quality tourism has not yet been analysed in the economic literature.

Survey design and the database The empirical research presented in the next section is based on the analysis of an original database obtained from a survey carried out among a representative sample of Balearic hotels. The field study involved an initial survey conducted in the spring of 2001, while the main body of work was conducted between August and September of that year. The aim of the survey, which was designed by the authors of this paper, was to gather information about labour conditions in the hotel sector of the Balearic Islands, as these data were lacking. The awareness of the potential importance of training decisions, both to the employee’s status and for the hotel’s strategies, was among a wider range of issues considered when designing the survey. In order to obtain a more complete database, we designed separate questionnaires for the two populations to be analysed, as follows. (1) Hotels in the Balearic Islands. The sample size was computed in order to be representative of the three different groups of hotel categories we planned to study. Once representativeness of the three-star hotels was guaranteed, we over-sampled the other two types of hotels in order to capture worker heterogeneity. The sample consists of 130 hotels. (2) Hotel employees in the Balearic Islands. The sample size is 1,900 workers, which ensures representativeness, both for males and females. The stratification of the field study was based on the initial survey and on population characteristics. We control for the geographical distribution, hotel category and job distribution among workers. Once the sample conditions were established, interviews were based on a random sample of the defined subpopulations. The empirical research presented in the next section uses a final database that merges both samples in order to have information on individual and hotel characteristics. For the descriptive analysis and the determinants of training, we drop 15 observations for which there was no information on explanatory variables. The distribution of the 1,885 workers among the three category

groups is 351 in one- or two-star hotels, 763 in three-star hotels, and 771 in fouror five-star hotels. Descriptive statistics In this section, we present some descriptive statistics. The information is divided into four parts. The first reflects differences in personal characteristics of the individuals working in tourism hotels of various qualities. The second is devoted to the differences in educational level by hotel category. The third analyses the role of training in each category. These three parts use the sample of 1,885 individuals described in the previous section. Finally, in the last part, we present information about the differences among the hotels in each of the category groups we have defined. We use information from the hotel sample in this part and hence we have information on 130 hotels.

Role of training in changing an economy 59

Differences in labour characteristics In Figure 1, we present the distribution of workers by contractual relation with the hotel, in each of the categories. Given the highly seasonal nature of the tourism production process, there is a singular and very common contractual relation in the sector, known as “permanent-intermittent”. These workers have a contract for an undefined period, but only work in the hotel for a maximum of 9 months a year. Clearly, the rationale for this legal instrument is supplied by the length of the tourism season. In the lower quality hotel category, the distribution of contractual linkage is clearly biased towards less steady relations: nearly 50 per cent of the workers in this group have a temporary seasonal contract. As we move up the categories, stability in contractual relations increases. In fact, the proportion of individuals with a permanent contract is 65 per cent higher in higher quality hotels than in lower quality ones. Hence, one immediate conclusion is that the movement from low-quality

Figure 1. Contractual relation by category

IJM 25,1

60

to high-quality tourism services supply will contribute to a decline in the degree of labour market instability. Table I compares the differences of wages between the three categories. Three different measures of wages were given. In the first column, we present the net monthly amount of money (in euros) that individuals earn by working in the hotel. In the second column, we control for the possible differences in working hours. Hence we divide the first column by the number of hours that individuals have in their contracts. In the third column, we use the number of hours, the individuals actually work. It is clear from the table that, by any of the different measures, the mean wage increases as we move to higher category hotels. Differences in educational level In this part of the descriptive statistics we analyse the distribution of educational level, both by hotel category and by the hierarchy in the workplace. The distribution of education endowment in percentage terms by category is given in Table II. The information derived from this table shows precisely the lack of a clear relation between category and education endowment. For instance, there is a higher share of lower-than-primary-school educated workers in lower-quality establishments, but there is also a higher share of university workers. This result is very illuminating for the motivation of the paper because it indicates that formal educational endowment is not a strong binding condition for worker mobility between categories. The structure of the hotel is divided into departments with clearly defined hierarchies. In Table III, we analyse the distribution of educational level in the different levels of the hierarchy. The variable “first” comprises those who have the main responsibility in his or her department. The variable “second” Net monthly wage

Contractual W/H

Actual W/H

882.1 924.7 961.9

5.5 5.8 6.0

5.3 5.4 5.6

Table I. 1* or 2* Wage levels (in euro) by 3* hotel category 4* or 5*

Table II. Percentage distribution of educational endowment by hotel category

1* or 2* 3* 4* or 5*

Lower than primary (per cent)

Primary (per cent)

Secondary school (per cent)

Higher education (per cent)

Tourism studies (per cent)

13.7 8.7 9.0

45.9 46.8 47.6

22.5 31.7 28.0

7.7 5.9 6.9

10.3 6.9 8.6

comprises individuals who have the second level of responsibility. The variable “others” comprises those individuals with a lower level of responsibility. The proportion of individuals with a university education or who have studied tourism increases and the proportion of individuals with primary or lower education decreases within the hierarchy at the workplace. Differences in training In this part of the descriptive statistics, we analyse the relation between training and category, hierarchy at the workplace, educational level and stability of the labour contract. We define our variable of interest (“training”) as whether or not the individual has undertaken any formal training course financed by his or her present employer during the last 12 months. In the initial survey we realised that one common training method used in the sector comes from the cooperation between the hotels and the local authorities. In these cases, courses are financed by the local authorities but take place during the working day, and thus the hotels assume part of the cost. Hence, our variable of interest also includes individuals whose training courses are not financed by hotels but take place during their normal working day. The percentage of individuals who receive training in each of the hotel categories is presented in Table IV. The first column displays these percentages for the total sample, while the second and third columns distinguish them by gender. In all three cases, the percentage of trained individuals increases with an increase in category level. It is interesting to note that the growth path is steeper for women. In the higher quality categories, there are small differences in the percentage by sex, while in the lower categories the differences are substantial.

First (per cent)

Second (per cent)

Others (per cent)

7.2 48.9 23.9 8.3 11.5

6.3 46.6 31.3 7.8 7.8

11.1 46.1 28.9 5.8 7.3

Lower than primary Primary Secondary school Higher education Tourism studies

1* or 2* 3* 4* or 5*

All (per cent)

Men (per cent)

Women (per cent)

10.8 21.8 26.7

16.5 25.6 27.2

6.5 17.5 26.0

Role of training in changing an economy 61

Table III. Percentage distribution of educational endowment by job hierarchy

Table IV. Percentage of trained workers by hotel category

IJM 25,1

62

Figure 2. Percentage share of trained workers by job hierarchy

Figure 3. Percentage of trained workers by education

Figure 2 shows the percentage of individuals who receive training by each of the hierarchy variables defined above. It is clear from the figure that the individuals receiving more training are those in the upper levels of the hierarchy. Hence, regardless of the direction of causality there is a higher share of training for those individuals with greater job responsibility. Figure 3 shows the percentage of individuals who receive training financed by the hotel by each educational level. This figure shows that there is a positive relation between training and education. The percentage of individuals who receive training increases as the maximum level of completed studies increases. Note also that the group with a higher incidence of training is the group of individuals who have completed specific tourism studies. Finally, Figure 4 shows the percentage of individuals who receive training by type of labour contract. Notice that permanent and intermittent-permanent workers receive more training. It is also remarkable that there are only three percentage points of difference between conventional and permanent-intermittent contracts.

Hotel characteristics We begin this part of the statistical analysis by providing some general information about the differences between hotels in each of the categories. We then present more interesting results from the hotel survey. Figure 5 shows the seasonal differences by hotel quality. There is a clear tendency towards extending the operating season as the quality of the hotel increases. The rationale is that the higher the quality of services provided, the greater the investment required. Hence the hotel needs a longer operating period to make a profit. In addition, higher quality hotels are able to diversify the kind of services they provide and hence attract a wider range of tourism demand. One implication of this is that the change from low-quality to high-quality services would help palliate the seasonal nature of the Balearic economy. Table V exhibits the relation between the size of establishments, measured in number of workers, and hotel category. Note that higher-quality hotels are bigger. The rational is the same as before: given the elevated amount of investment involved, the bigger the hotel, the greater the possibilities of achieving a satisfactory return due to its scale effects. The size of the

Role of training in changing an economy 63

Figure 4. Percentage of trained workers by type of labour contract

Figure 5. Seasonal pattern by category

IJM 25,1

64

establishment increases rapidly when we move from one or two-star establishments to three-star establishments. Most three-star establishments are medium-sized. Tables VI-IX present information obtained from hotel questionnaires on training performance in hotels. Table VI presents the percentage of establishments that give a positive answer when asked if they financed training for each of the departments in which we divided the structure of the hotel establishment. We consider seven separate departments that are easy to identify for all respondents. It is clear from this table that higher-quality hotels provide much more training than lower-quality ones. As shown in the data, the most remarkable difference is between the lowest-quality group and the other categories. The departments receiving the most training in the one and two-star establishments are kitchen, reception and management. Establishments in highest category hotels make bigger training efforts in the restaurant and kitchen departments. Table VII presents the percentage of hotel managers who answered negatively when asked whether their employee training and education were sufficient. This percentage increases as we move to higher quality hotels. All categories chose kitchen and restaurant as the most problematic departments. In Table VIII, we present a perception question addressed to managers. They were asked to what degree they considered higher category hotels require

Table V. 1-30 (per cent) 31-60 (per cent) 61-90 (per cent) 90-120 (per cent) . 120 (per cent) Percentage distribution of establishment size (in 1* or 2 * 70.0 22.5 5.0 2.5 0.0 number of workers) by 3* 12.0 38.0 36.0 4.0 10.0 hotel category 4* or 5* 7.5 22.5 42.5 17.5 10.0

Table VI. Percentage of establishments financing training by each department

Table VII. Percentage of establishments considering employee training as unsatisfactory

Management Reception Cleaning Kitchen Restaurant Bar Maintenance (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) 1* or 2* 3* 4* or 5*

10.0 62.0 65.0

10.0 60.0 75.0

7.5 56.0 70.0

10.0 52.0 77.5

7.5 50.0 77.5

2.5 54.0 70.0

7.5 54.0 72.5

Management Reception Cleaning Kitchen Restaurant Bar Maintenance (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) 1* or 2* 3* 4* or 5*

2.5 10.0 12.5

7.5 6.0 12.5

12.5 22.0 25.0

15.0 26.0 30.0

12.5 26.0 25.0

15.0 24.0 22.0

15.0 16.0 15.0

more training. Notice that the percentage of managers who consider that more training is needed in high-category hotels increases as we move from low- to high-quality hotels. Finally, in Table IX we illustrate the perceived relevance of training among managers. They were asked to give a score of between 0 and 10 for the importance of education and training when hiring new workers; in fact, the questionnaire asked them to rank the relevance of different individual characteristics when hiring new employees for each of the previously defined hierarchy groups. Each cell in the table displays the mean value for that group, with the minimum and maximum values given in parenthesis. In all three categories the relevance of education and training increases as we move up the hierarchy. Note also that the mean value increases with category. It is also worth mentioning the differences in the lower assessment of the relevance of training when hiring new workers. This minimum value increases from 0 to 5 and 6 for “first” and “second” as we move up the categories. This change is also reflected at a lower scale for individuals with less responsibility.

Role of training in changing an economy 65

Training determinants In this section, we analyse the determinants of formal training financed by the hotel. We use “training” as the dependent variable. Earlier research has shown that the combination of information from hotels and individuals can lead to improvements in the analysis of training determinants (Francis et al., 1998). Thus, we use the characteristics of both the individuals receiving formal training and the hotels in which they work. Hence, we are able to tell whether a given demographic group has a higher or a lower probability of receiving training in any particular kind of hotel. In order to improve our understanding of the effects of including a richer set of explanatory variables, we present four different specifications. It is Management Reception Cleaning Kitchen Restaurant Bar Maintenance (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) (per cent) 1* or 2* 3* 4* or 5*

45.0 62.0 70.0

52.5 68.0 72.5

37.5 48.0 72.5

First 1* or 2 * 3* 4* or 5* Total

7.78 8.26 8.79 8.30

(0-10) (5-10) (6-10) (0-10)

50.0 56.0 75.0

45.0 58.0 75.0

Second 7.61 8.09 8.59 8.11

(0-10) (5-10) (6-10) (0-10)

45.0 52.0 72.5

45.0 46.0 70.0

Table VIII. Percentage of establishment’s managers who consider that the higher category establishments require more training

Other 7.08 7.25 7.76 7.36

Table IX. Mean value of the (0-10) (3-10) importance of education and training when (5-10) hiring new workers (0-10)

IJM 25,1

66

interesting to compare the coefficient and its significance for a given variable once we incorporate more explanatory covariates into the specification. The significance of the model as a whole is also affected by this enrichment. Apart from the baseline model, in the enrichment of the specification we present only covariates that are significant in the final specification. We proceed this way in order to minimise the possible problems of multi-colinearity among the variables of interest and in order to present a precise R 2 measure. In the following paragraphs, we present four specifications. (1) Baseline specification. The starting specification is a very rough one. We include only the dummy variable for young individuals (“age , 24 yrs”), education variables (“secondary school” and “higher education”), a gender dummy and a civil status dummy. (2) Specification B. This second specification incorporates some individual personal characteristics that influence training decisions. We include the number of years that the individual has been working for the same hotel (“Establishment tenure”); a dummy that takes value 1 if the individual has a steady contractual relation, namely a permanent contract; a dummy that controls for cases of individuals with more than one job (we expect individuals with more than one job to receive less training); a dummy for the “hours flexibility” to identify those individuals from whom the hotel is more likely to recover its investment; and a covariate (“relation”) that signals individuals who have an educational level related to their actual job in the hotel. We think it is reasonable to assume that these individuals will be more willing and more prepared to receive and use specific training. Finally, the “overeducated” variable controls for individuals working in jobs requiring less education than they have, who possibly would be less interested in receiving more training. (3) Specification C. In this step, we incorporate three variables that are crucial to our research. On the one hand, we control for the importance of position in the hierarchical structure when trying to explain who will receive training. We define the variable “first” as value 1 if the individual has the top responsibility in his or her department. We define the variable “second” for individuals in the second-ranked level of responsibility. On the other hand, we incorporate the variable “category” to capture the effect of hotel category on the probability of receiving training. Hence, the sign and the coefficient of the “category” variable tell us the effect of the tourism development strategy on the probability of receiving training, once we control for all the other individual and hotel characteristics. (4) Specification D. This final model incorporates hotel characteristics that may influence the probability of receiving training. The enrichment of the model allows us to control for the influence of all the

above-mentioned variables, regardless of the characteristics of the hotel in which the individual works. We control for the size of the hotel (“less than 30” (workers)) because earlier literature presents evidence of the positive effect of the size of the hotel on the probability of training. We divided the Balearic Islands into eight different areas using the type of tourism service supply as criterion and we include two dummies that control for the hotel’s geographical location in the Balearic Islands (“high tourism density”, “city”). The variable “high tourism density” refers to the different locations in the Bay of Palma de Mallorca characterised by a high density of tourism services. The variable “city” includes hotels located in the city of Palma de Mallorca. It is reasonable to assume that these locations have easier access to training facilities. The dummy “hotel chain” takes a value of 1 if the hotel happens to be part of a chain. We also include the covariate “open all year” for hotels that are open for 12 months a year. Finally, we incorporate two variables that capture the importance of management structure for the probability of financing training (“rent hotel” and “property management”). A third possibility included in the reference group is “external management”, which refers to hotels run by an external firm. In Table X, we present the results of the four specifications we propose. The estimation method is based on ML techniques. Standard errors are computed using the Huber/White/Sandwich estimator of variance to guarantee robustness, and hence significance is free from heterokedastic problems. For each of the models, we present the estimated coefficient (in the “coefficient” column) and the transformation that allows us to interpret it. In the case of continuous variables, the column “dF/dx” indicates the change in the probability of receiving training when the variable increases by one unit, taking the mean value of that variable as the initial point. In the case of the dummy variables, the column “dF/dX” indicates the change in the probability of receiving training when individuals fulfil the condition defined in the variable. First of all, it is important to keep in mind that we cannot derive any conclusions from the results of the simplest specifications, given that those specifications suffer from omitted variable bias, as we will show. Hence, as we have mentioned, the interest stems from the change in each variable coefficient and its significance, and from the change in the explanatory power of the model as a whole, once we enrich the model. Results from the baseline model. In columns 1 and 2, we present the results of the baseline model. All the variables included in this specification appear as significant, at least at the 10 per cent level. The signs are the expected ones. The very young and those with a lower than primary school education have a lower probability of receiving formal training financed by the hotel, and the more educated workers have a greater probability of receiving such training.

Role of training in changing an economy 67

Table X. Probit model estimates 0.004 0.081 20.084 0.05 0.141 20.10214

0.0143*** 0.3036*** 2 0.3533* 0.1795* 0.5079*** 2 0.4510**

0.0992

20.071 20.055 0.059 20.033 20.011

M2 dF/dX

2 1.2759*** 2 0.2872** 2 0.2182* 0.2024* 2 0.1236* 2 0.0403

M2 coef.

0.0958 0.0646 0.0586

0.3268*** 0.2242** 0.2174***

0.1146

0.0034 0.0642 2 0.0772 0.0369 0.1225 2 0.0839

2 0.0624 2 0.049 0.0488 2 0.0099 2 0.0183

M3 dF/dX

0.0126** 0.2427** 20.3268** 0.1339** 0.4460*** 20.3599**

21.8088*** 20.2514** 20.1957 0.171 20.0368 20.0676

M3 coef.

Notes:***significant at 1 per cent; **significant at 5 per cent; *significant at 10 per cent

20.126 20.076 0.052 20.062 0.031

M1 dF/dX

0.106 0.058 0.038 2 0.104 0.0883 0.3187 2 0.0981 2 0.0826 0.1406 0.0945 20.4874** 0.3266*** 0.9419*** 20.3805*** 20.3380*** 0.4668** 0.4245** 0.1539

0.003 0.045 2 0.07 0.041 0.128 2 0.062

2 0.065 2 0.039 0.0252 0 2 0.011

M4 dF/dX

0.3702*** 0.2109** 0.1453***

0.0133** 0.1764** 20.3108** 0.1554** 0.4861*** 20.2675*

21.6163** 20.279 20.1607 0.0946 0.0012 20.042

M4 coef.

68

Baseline specification Reference 2 0.6846*** Age less 24 2 0.5169*** Secondary school 2 0.2932** Higher education 0.1729* Woman 2 0.2178* Married 0.1099* Specification B Establishment tenure Stable contract More than one job Hours flexibility Relation Overeducated Specification C First Second Category group Specification D Less than 30 High tourism density City Hotel chain Open all year Rent hotel Property management Pseudo R2 0.0278 Number of observations: 1885

M1 coef.

IJM 25,1

Regarding the sociological variables, married workers tend to receive more training and female workers are less likely to receive financed training. Results from specification B. In this first step, we include some individual characteristics that may reasonably influence the probability of receiving formal training financed by the hotel. There is a first group of individual variables that contributes to increasing the probability of receiving training. These variables are a longer relation with the hotel, a steady contract, and the worker/timetable flexibility in line with hotel requirements. These variables identify those individuals from whom the hotel is more likely to achieve a return on its training investment. Those individuals signalled by the dummy “more than one job” are in the opposite position and the sign for its coefficient is negative. The variable that indicates a positive relation between the human capital characteristic that the individual has and his or her present job (“relation”) also has a positive effect on training probabilities. Finally, individuals who are overeducated for their current job also have a lower probability of being trained. It is important to note that the inclusion of these individual characteristics reduces the value of the coefficient of the variables considered in the baseline model. Moreover, the “married” variable is no longer significant once we have more precise labour stability measures. Results from specification C. Here we incorporate job hierarchy and the hotel category in which the individual works. The probability of receiving training increases with hotel category. Hence, those individuals who move from lower-quality hotels to higher-quality hotels will need to receive more training. Job hierarchy is revealed as a very significant variable in explaining the probability of receiving training. Although this does not seem to be a very surprising result, there is scant empirical work that controls the degree of responsibility that an individual possesses. As the results show, the higher the responsibility, the higher the probability of being trained by the firm. Another important result is that, once we incorporate these variables into our specification, educational variables and the gender variable lose their significance. Our interpretation is that, if the researcher does not incorporate job hierarchy into the analysis, then the correlation between this hierarchy and educational and gender variables yields a significant coefficient. Academic literature on wage equation helps us to understand this process. The signaling theory proposed by Spence (1973) states that formal human capital investment is not based on productivity, but is a signal of unobserved ability. Hence, Farber and Gibbons (1991), propose that the inclusion of educational variables in wage equations implies a control for not only human capital, but also for the unobserved ability of the individual. Our proposal here is that job hierarchy variables may be relevant because training is job-related or because they control the degree of responsibility an individual assumes in his job, but may also be a better proxy than education for some desirable unobservable characteristics. Hence, the coefficient of such variables captures a mixture of

Role of training in changing an economy 69

IJM 25,1

70

the effect of job levels and some unobserved abilities on the probability of receiving training. However, the estimation of the rest of the coefficients improves, given that we have an instrument that allows us to control for individuals’ unobserved characteristics. In the case of gender variable, the negative effect associated with women is merely due to the presence of segregation in the workplace (Ramos et al., 2002). Women tend to be segregated within lower ranks of the job hierarchy and so if the researcher does not control for the job hierarchy, it is easy to find a significant effect of gender. Results from specification D. We complete the enrichment of our specification by incorporating the effect of characteristics of the hotel in which the individual works. As stated in the literature, smaller hotels provide less training. The conventional explanation for this is the presence of economies of scale. Results also suggest that the geographical location of a hotel plays a role in the decision to finance training. In fact, hotels located in more crowded areas, with a higher incidence of tourism activity, are more prone to finance training. The rationale for this result is in line with the above argument on the importance of economies of scale. Not only is the size of the hotel important, but so too is the availability of cooperating with other hotels in order to create training programmes. Likewise, it is easier for local authorities in areas with a higher tourism density to develop training programmes. For hotels that are part of a chain, this fact has a negative effect on the probability that the hotel finances the training of their workers. Obviously, the most likely reason is that the hotel chain will assume part of the training duties of its establishments, which are the unit of analysis of our dependent variable. Those hotels in which the management activities have a longer planning horizon (as it is the case of property management and firms that rent hotel) have a higher incidence of training. Notice that the inclusion of these firm characteristics does not modify the significance of the variables included in specification C, but does have a slight impact on their coefficients. Finally, the explanatory power of the model as a whole increases as we enrich the model. It is worthwhile mentioning that we enrich the model only by adding variables that are in fact significant, and we reject other irrelevant variables that may only capture some statistical variation. Conclusions Concern is growing in many mature destinations that rely heavily on mass tourism about the need to change the pattern of specialisation and to shift resources from low-quality to high-quality tourism services. The main argument supporting this change is based on structural problems related to the mass-tourism strategy, namely the existence of congestion problems and externalities at an elevated environmental, cultural and social cost; and the mass-tourism strategy usually means price competition and therefore the need to maintain low costs.

The aim of this paper is not to evaluate the convenience of moving from low-quality to high-quality tourism services, but to study the potential labour market effects this proposed change may have. To do so, we analysed the possible role of human capital in determining transition costs by comparing different educational and training requirements between various possible tourism development strategies differentiated by the quality of services offered. We used an original database, created during the summer of 2001, of hotels and hotel workers in the Balearic Islands. These data allowed us to use characteristics not only of workers but also of hotels in order to make our analysis more complete. Our results show the existence of differences both in job characteristics and in human capital. In particular, we found that the crucial type of human capital investment that differentiates between hotel categories is training requirements. That is to say, employers offer and have higher training requirements in top-level categories and for the heads of each department. This result can be confirmed by not only using the descriptive analysis of the data but also the discrete choice model of the determinants of the on-the-job training provision. In order to respond to the main question of the paper – namely the role of training in a productive structural change in the tourism sector – we wish to emphasise two conclusions. (1) Educational level is not binding for the possibilities of moving workers from the mass-tourism strategy to the high-quality strategy. (2) Training plays a crucial role in high-quality tourism hotels. A considerable training effort will be required to adapt workers’ knowledge to fit the requirements of high-quality hotels. Hence, any instrument of political economy designed to stimulate a change from a mass-tourism strategy to a high-quality tourism strategy must be complemented with a training policy in order to minimise associated labour costs.

References Acemoglu, D. and Pischke, J.S. (1999), “The structure of wages and investment in general training”, Journal of Political Economy, Vol. 107 No. 3, pp. 539-72. Altonji, J.G. and Spletzer, J.R. (1991), “Worker characteristics, job characteristics, and the receipt of on-the-job training”, Industrial and Labor Relations Review, Vol. 45 No. 1, pp. 58-79. Amoah, V.A. and Baum, T. (1997), “Tourism education: policy versus practice”, International Journal of Contemporary Hospitality Management, Vol. 9 No. 1, pp. 5-12. Bartel, A. (1995), “Training, wage growth, and job performance: evidence from a company database”, Journal of Labor Economics, Vol. 13 No. 3, pp. 401-25. Baum, T. (1994), “National tourism policies: implementing the human resource dimension”, Tourism Management, Vol. 15 No. 4, pp. 259-66. Becker, G.S. (1962), Human Capital, 2nd ed., Columbia University Press, New York, NY.

Role of training in changing an economy 71

IJM 25,1

72

Bishop, J.H. (1994), “The impact of previous training on productivity and wages”, in Lynch, L.M. (Ed.), Training and the Private Sector. International Comparisons, Chapter 6, Chicago University Press, Chicago, IL, pp. 161-99. Booth, A. and Bryan, M. (2002), “Who pays for general training? New evidence for British men and women”, Discussion Paper No. 486, IZA. Booth, A. and Satchell, S. (1994), “Apprenticeship and job tenure”, Oxford Economic Papers, Vol. 46, pp. 676-95. Brotherton et al. (1994), “Developing human resources for Turkey’s tourism industry in the 1990s”, Tourism Management, Vol. 15 No. 2, pp. 109-16. Farber, H.S. and Gibbons, R. (1991), “Learning and wage dynamics”, NBER Working Paper No. 3764. Fina, L. et al. (2000), “Cambio ocupacional y necesidades educativas de la economı´a espan˜ola”, in Sez, F. (Ed.), Formacion y Empleo, Visor, Madrid. Francis, H., Gittleman, M. and Joyce, M. (1998), “Determinants of training: an analysis using both employer and employee characteristics”, Working Paper, Bureau of Labor Statistics. Greenhalgh, C. and Stewart, M. (1987), “The effects and determinants of training”, Oxford Bulletin of Economics and Statistics, Vol. 49 No. 2, pp. 171-90. Kennedy, S., Drago, R., Sloan, J. and Wooden, M. (1994), “The effect of trade unions on the provision of training: Australian evidence”, British Journal of Industrial Relations, Vol. 32 No. 4, pp. 565-80. Lillard, L.A. and Tan, H.W. (1992), “Private sector training: who gets it and what are its effects?”, Research in Labor Economics, Vol. 13, pp. 1-62. Lynch, L.M. (1992), “Private sector training and the earnings of young workers”, American Economic Review, Vol. 82, pp. 299-312. Mincer, J. (1974), Schooling, Experience and Earnings, Vol. 48, No. 4, Columbia University Press, New York, NY, pp. 817-38. Mincer, J. (1983), “Union effects, wages, turnover and job training”, Research in Labor Economics, Vol. 5 No. 5, pp. 217-52. Spence, M. (1973), “Job market signaling”, Quarterly Journal of Economics, Vol. 87 No. 3, pp. 355-74. Stevens, M. (1994), “A theoretical model of on-the-job training with imperfect competition”, Oxford Economic Papers, Vol. 46, pp. 563-78. Ramos, V., Rey, J. and Tugores, M. (2002), “Ana´lisis empı´rico de discriminacio´n por razo´n de ge´nero en una economı´a especializada en turismo”, Annals of Tourism Research, Vol. 4 No. 1, pp. 239-58 (in Spanish). Tugores, M. (1998), “An on-the-job training model with spillovers and product market competition”, Working Paper No. 12, University of the Balearic Islands. Woodbury, S. (1992), “Employer training needs in Hawaii”, Staff Working Paper No. 92-15, Upjohn Institute. Further reading Huber, P.J. (1964), “Robust estimation of a location parameter”, Annals of Mathematical Statistics, Vol. 35, pp. 73-101. White, H. (1980), “A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity”, Econometrica, Vol. 48 No. 4, pp. 817-38.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Training, task flexibility and the employability of low-skilled workers

Training, task flexibility and employability 73

Jos Sanders and Andries de Grip Research Centre for Education and the Labour Market, Maastricht University, The Netherlands Keywords Training, Skills training, Labour mobility, Flexible labour Abstract This paper analyses whether low-skilled workers’ training participation and task flexibility contribute to their firm-internal and firm-external mobility, and find that both training participation and task flexibility contribute only to firm-internal employability. However, the workers’ participation in training plays a much more explicit role in their firm-internal career than their task flexibility does, as the former appears to be an important means to increase their opportunities in the firm-internal labour market. Neither the low-skilled workers’ participation in training nor their task flexibility contributes to their external employability. Task-flexible, low-skilled workers are less likely to expect to be externally employable than non-task flexible workers are. The focus of the low-skilled workers on their firm-internal employability can be explained by the fact that such workers usually have more opportunities to improve their position in the firm-internal labour market than in the external labour market.

Introduction In the 1990s, workers’ employability received much attention from both policy makers and human resource specialists in the business community. The paradigm of lifetime employment seems to have been replaced by a new paradigm of lifetime employability marked by a high degree of flexibility. Among others, Arthur (1994), Bridges (1994) and Hyatt (1995) have characterised modern careers as “boundaryless”. We consider employability to be an individual characteristic in terms of a worker’s capacity and willingness to remain attractive in the labour market, that is, a worker’s labour market value. This raises the question, to what extent workers can maintain or increase their employability in the labour market. In the literature on employability, two instruments are frequently mentioned: training participation and task flexibility (i.e. doing tasks that belong to other jobs). These “employability instruments” may contribute to workers’ ability to remain attractive in the labour market and also signal a worker’s willingness to be employable (de Grip et al., 2004). In this paper, we test the hypothesis that training participation and task flexibility contribute to the employability of low-skilled workers. In this analysis, we distinguish between three forms of employability, namely The authors thank Bart Golsteyn, Patrick van Eijs, an anonymous referee and the editors of this volume for their helpful comments on an earlier version of this paper.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 73-89 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410525009

IJM 25,1

74

job-match employability, which refers to a worker’s chance to remain employed in his/her current job within the current firm (same job, same employer); firm-internal employability, which refers to a worker’s chance to switch to another job within the current firm; and firm-external employability, which refers to a worker’s chance to switch to a job in another firm. Although the term job-match employability indicates that participation in training and task flexibility might also be important for low-skilled workers to keep up with the developments related to their current job, we expect that both training and task flexibility are more important for their firm-internal and external employability[1]. Our focus on low-skilled workers is interesting from two viewpoints. First, the labour market position of low-skilled workers is vulnerable currently as their employability is threatened because they are frequently crowded out of their traditional job domains by higher skilled workers (Borghans and de Grip, 2000). Second, low-skilled workers generally participate less often in training than skilled workers do (e.g. Shields, 1998), so one might wonder whether or not low-skilled workers deliberately invest in their employability. We first analyse whether workers’ participation in training and task flexibility affect low-skilled workers’ expectations regarding their employability in terms of the three forms of employability we distinguish. Second, we analyse whether these expectations had been realised two years later. Finally, we analyse whether low-skilled workers’ employability expectations induce a larger participation in training and task flexibility. This enables us to analyse whether low-skilled workers participate in training and demonstrate their task flexibility to increase the chances of realising their expectations. For our analyses we use data from the linked 1998 and 2000 waves of the Dutch OSA Labour Force Survey. We find that both workers’ training participation and task flexibility contribute only to workers’ firm-internal employability. However, workers’ participation in training plays a much more explicit role in workers’ firm-internal careers than their task flexibility does, as the participation in training appears to be an important tool to increase their chances in the firm-internal labour market. Neither participation in training nor task flexibility contributes to the external employability of low-skilled workers. Task-flexible, low-skilled workers are less likely to expect to be externally employable than non-task-flexible workers are. Training, task flexibility and the concept of employability The concept of workers’ employability is not new[2]. It was developed in the 1950s. However, there have been some changes in the focus of the concept in the course of time. In the 1950s and 1960s, employability was seen as an individual’s potential to become employed. The attention focused on a worker’s attitude to employment in general and on the self-perception the workers

develop during their career. Influencing and adjusting attitudes and the perception that the people have of their abilities contributed to the successful labour market re-entry of people who had lost their self-confidence (Soloff and Bolton, 1969). From 1970 onwards, attention became increasingly focused on occupational knowledge and skills instead of on a worker’s attitudes. Not only basic occupational skills but also knowledge about one’s possibilities (Tseng, 1972), about one’s own position in the labour market (Mangum, 1976) and about the employment situation in general play crucial roles here. At the end of the 1970s, partly related to the economic recession in industrialised countries, it was realised that merely having some occupational skills is often not sufficient to remain attractive in the labour market. Hoyt (1978) acknowledged the importance of a worker’s “transferable” skills, which retain their value in many different work situations. Examples of such transferable skills are the social and relational skills that are important not only to get a job, but also to keep it and to move on to another job, if necessary. Moreover, from an employee’s point of view, employability became more important, since the economic recession made it harder both to find and to keep a job. After 1980, the employability concept more and more became a meta-characteristic of workers’ labour market value. This meta-characteristic combines attitudes, knowledge and skills and determines the labour market potential of workers. In this sense, employability has an important influence on a worker’s career whether it is in the beginning, building or final stage (Charner, 1988). In the 1990s the differences between the various views on employability and how it affects people increased. For some authors, only a worker’s labour market potential and skills play a role. Others focus on the possibilities to use a worker’s employability in organisations (Levy et al., 1992), knowledge of the labour market and policies of firms and the government (Outin, 1990), or they emphasise workers’ capacity to influence their career (Bloch and Bates, 1995) and to deal with changes (Hyatt, 1995). In order to structure the employability concept, Thijssen (1997) developed a taxonomy of the existing employability definitions. He distinguished between three types of employability definitions: a core definition, a broader definition and an all-embracing definition. According to the core definition, employability encompasses all individual possibilities to be successful in a diversity of jobs in a given labour market situation. In its core definition, employability only concerns someone’s capacities. The broader definition of employability incorporates both the capacity and the willingness to be successful in a diversity of jobs. In addition, the ability to learn is included as an asset of a worker’s employability. Therefore, in the broader definition, employability encompasses all the individual

Training, task flexibility and employability 75

IJM 25,1

76

characteristics that determine the current and future position in the labour market. In the all-embracing definition, contextual factors and effectuation conditions are added. Effectuation conditions are context-bound factors that facilitate or hamper a worker’s employability, such as the training provided by the firm. In the all-embracing definition, employability encompasses all individual and contextual conditions that determine a worker’s current and future position in the labour market. However, the emphasis is still on workers’ capacities and willingness to be proactive, which gives them a strong position in the labour market. In this paper, we therefore define a worker’s employability as: The capacity and the willingness to be and to remain attractive in the labour market, by anticipating changes in tasks and work environment and reacting to these changes in a proactive way.

It should be emphasised that employability is not a static concept, as a worker’s employability can change over time. Whether or not workers are employable, in the sense that they are able and willing to remain employed, depends on a number of factors, some of which workers can and some of which they cannot easily influence directly. In this paper, we focus on two important factors that can be influenced by the workers themselves, namely their training participation and task flexibility[3]. As mentioned in the introduction, we analyse whether the training participation and task flexibility of low-skilled workers contribute to their firm-internal and external employability. In these analyses it is important, however, to distinguish between the different ways in which workers can remain attractive in the labour market. For instance, Groot and Maassen van den Brink (2000) distinguish between workers’ internal and external employability. External employability refers to the ability and willingness to switch to a similar or another job in another firm, and therefore reflects the value of workers’ human capital in the external labour market. Internal employability refers to a worker’s ability and willingness to remain employed with the current employer, that is, the value of a worker’s human capital in the internal labour market. Here we further distinguish between two groups of internally employable workers: those who remain employed in the same job and those who change jobs within the current firm. We therefore, distinguish between three forms of employability, namely job-match employability, which refers to workers who remain employed in their current job within their current firm; firm-internal employability, which refers to workers who switch to another job within their current firm; and external employability, which refers to workers who switch to a job in another firm. Data The data we used for our empirical analysis were taken from the linked 1998 and the 2000 waves of the Dutch OSA Labour Supply Survey. For 1998 the

total sample size was 4,780 observations. For 2000 the total sample size was 4,185. For the empirical analysis in this paper, we selected the lower educated workers (ISCED 0-2) aged between 16 and 50 who were in paid employment at the time of interview in 1998, who had a permanent contract in 1998 and who participated in both the 1998 and 2000 surveys[4]. This reduced the total sample size to 474. Of these, 92 work in manufacturing and 158 in the services sector. The remaining 224 persons work in other sectors, such as agriculture and fisheries, education or health. Expected labour market position in five years We indicate the perceived employability of low-skilled workers in The Netherlands by the labour market position they expected to have in five years. For this indicator we used the following question in the 1998 survey: If you look five years ahead, what will your position in the labour market be?

For this question, eight replies were possible: . working in a similar job in this firm (job-match employability), . working in a different job in this firm (firm-internal employability), . working in a similar job in another firm (external employability), . working in another job in another firm (external employability), . unemployed, . resigned, . disabled/unable to work, . retired, and . don’t know. This indicator clearly indicates workers’ expectations regarding their future labour market position. Table I shows that practically all the low-skilled workers thought that they would remain active in the labour market for the next five years. Only about 2 per cent expected to be without a job in five years. In this sense it can be concluded that the low-skilled workers in The Netherlands who have permanent contracts are rather optimistic about their overall (i.e. internal and external) employability. Table I also shows that the majority of the low-skilled workers expected to be working in their current or a similar job within the current firm five years on (73 per cent). About 15 per cent of the workers expected to change jobs within the current firm and about 10 per cent expected to leave their current firm to start working elsewhere. This indicates that the great majority of the low-skilled workers rely on their firm-internal labour market. Table I also shows that workers aged between 16 and 34 years were more likely to expect to leave their current job within the next five years than older workers. More than 80 per cent of the 45- to 50-year-old workers expected not to

Training, task flexibility and employability 77

Table I. Expected labour market position of lower educated workers five years from 1998 16 14 15 21 16 10 21 13 11 10 19 16 20 4 10 11 18 13 16 13 11 17 15

73 74 60 61 76 82 68 79 67 80 72 70 73 75 79 30 72 75 73 73 80 72 73

Firm internally employable (per cent)

19 9 7 12 6 17 6 43 8 10 8 12 7 9 9

9 9 25 15 8 4 9 6

Externally employable (per cent)

3 1 3 2 2 4 1 5 2 2 2 2 2 2 2

3 1 4 2 2

1 4

Without a job (per cent)

78

Male Female 16-24 years 25-34 years 35-44 years 45-50 years Elementary job Lower level job Middle or higher level job Manufacturing Services Other Task flexible Not task flexible Satisfied Unsatisfied Overtime No overtime Training No training Part time Full time Total Source: OSA

Job match employable (per cent)

39 21 36 52 67 33 92 8 45 55 71 29 27 73 100

66 34 5 23 47 25 14 47

Total (per cent)

IJM 25,1

leave their current job within five years, compared to 60 per cent of the youngest categories of workers. This probably reflects the job search process at the beginning of workers’ careers (Neil, 1999; Topel and Ward, 1992). It also reflects the idea that older workers are less mobile than their younger colleagues, either because of a higher rate of firm-specific human capital or a stronger aversion to change among older workers (Becker, 1964; Salthouse, 1991), or because employers who recruit new workers would rather invest in a younger worker’s human capital than in that of an older worker (Gallup, 1990). The table also shows that low-skilled workers who are employed in services are more likely to expect to be firm-internally employable than are the low-skilled workers in manufacturing. Low-skilled workers who are doing tasks that are not part of their job are considerably more likely to expect to change jobs within their current firm than are workers who are not task flexible. On the other hand, non-task flexible workers are far more likely to expect to leave their current firm than task flexible workers are. This suggests that low-skilled workers signal their task flexibility to increase their firm-internal employability. The same applies to, though to a far lesser degree, low-skilled workers who participate in training. Of these workers, 16 per cent thought that they would switch to another job within their current firm within the next five years, compared to 13 per cent of the low-skilled workers who did not participate in training. More striking, however, is that whether or not low-skilled workers participated in training had no effect on their expectations that they would stay in or leave their current job. Moreover, it shows that more than one-fifth of the low-skilled workers did not have the idea that it is important for them to participate in training to reduce the risk of losing their job due to skill obsolescence (de Grip and van Loo, 2002). Furthermore, Table I shows that workers who are dissatisfied with their job are far more likely to expect to leave their job than are workers who are satisfied. It is surprising, however, that no less than 30 per cent of the low-skilled workers who were not satisfied still expected to remain employed in their current job with their current employer for another five years. This suggests that these workers might feel insecure about their labour market value or that they feel stuck in their current position, which might contribute to their dissatisfaction. The data also show that 43 per cent of the workers who were not satisfied with their current job expected to leave the firm where they work, whereas only 11 per cent expected “just” to change to another job in the firm-internal labour market. This suggests that workers’ dissatisfaction is more often related to the firm as a whole than to their specific job. Finally, Table I shows that low-skilled workers who have full-time contracts are less likely to expect to remain employed in their current job than those who work part-time. This is quite surprising, as one might expect workers who have a part-time contract to feel less secure about their job than those who have a full-time contract. However, since our data only refer to workers with

Training, task flexibility and employability 79

IJM 25,1

80

permanent contracts, one might also interpret this finding as a confirmation of the idea that part-timers who prefer to work part time to combine work and care will opt for job security and are probably less mobile in the labour market. Estimation results In this section, we analyse whether the training participation and task flexibility of low-skilled workers affect their firm-internal or firm-external employability. First, we present the estimation results of a multinomial logit analysis on the effects of the training participation and task flexibility of low-skilled workers on their employability expectations, in terms of the three types of employability we distinguish (job-match employability, firm-internal employability, external employability). Next, we analyse whether the expectations the workers had in 1998 had been realised in 2000, by performing a multinomial logit analysis on the actual changes in the labour market position of low-skilled workers during the period 1998-2000. Finally, we analyse whether workers’ employability expectations in 1998 affected their training participation and task flexibility during the period 1998-2000. These two binomial logit analyses enable us to show whether or not low-skilled workers invest in training and demonstrate their task flexibility to increase their chances of realising their employability expectations. The determinants of low-skilled workers’ employability expectations We first analyse whether the employability expectations the low-skilled workers had in 1998 were affected by their training participation and task flexibility in the two years before they expressed their expectations. For this analysis we use the following two independent variables: (1) training – participation in courses (1994-1998)[5], and (2) task flexibility – tasks performed outside one’s job (1996-1998). Moreover, we include three dummy variables that indicate the quality of the job the low-skilled workers have: . whether a worker is satisfied in his/her current job; . whether a worker works overtime; and . whether a worker has a part-time contract. We also include a number of covariates, namely gender, age, sector of industry, professional level and tenure (see Appendix Table AI for a list of variables). The estimation results in Table II show that whether or not a low-skilled worker participated in training between 1994 and 1998 has no effect on their perceptions of their employability. Neither the likelihood of workers expecting to change jobs within their current firm nor the likelihood of workers expecting to leave their current firm entirely is affected by previous training participation.

External employability Intercept Training Task flexibility Satisfaction Overtime Part time Male Age Elementary level job Lower level job Middle or higher level job Manufacturing sector Services sector Tenure Firm-internal employability Intercept Training Task flexibility Satisfaction Overtime Part time Male Age Elementary job Lower level job Middle or higher level job Manufacturing sector Services sector Tenure 2 2 Log-likelihood = x2 = N= Df = Notes: *Significant at the 10 per cent level; **significant at the 5 per cent level

b

Standard error

1.66 0.16 2 1.24** 2 2.50** 0.51 0.31 1.20* 2 0.48* REF 2 1.47** 2 0.64 2 0.25 2 0.20 2 0.07*

1.05 0.51 0.44 0.57 0.45 0.67 0.71 0.26 REF 0.64 0.62 0.53 0.53 0.04

2 1.09 0.33 1.78** 2 1.37** 0.21 2 0.67 2 0.23 2 0.65** REF 0.33 0.80 2 0.71 0.68* 2 0.02 449.26 92.84 379 24

1.03 0.38 0.55 0.57 0.32 0.51 0.46 0.22 REF 0.57 0.58 0.45 0.36 0.03 – – – –

Task flexibility, however, rather strongly affects workers’ employability. Task flexible workers expected that they would change jobs within their current firm rather than to remain employed in their current job, but expected to have a smaller chance to leave their current firm than to remain employed in their current job. This suggests that workers who perform tasks that are not part of their job have the idea that this contributes to their chances to move to another job in the firm-internal labour market. Table II also shows that lower skilled workers who have a low-level job (i.e. above the elementary level) are less likely to expect to leave their current job

Training, task flexibility and employability 81

Table II. Multinomial logit estimation of low-skilled workers’ self-assessed employability (reference category: same job, same firm in five years)

IJM 25,1

82

and firm within five years than workers who have an elementary job. This indicates that low-skilled workers with a low-level job have a higher job-match employability than the low-skilled workers who are employed in the elementary jobs. In addition, Table II shows that there is a strong negative relationship between workers’ job satisfaction and their firm-internal employability expectations. However, the negative relationship between job satisfaction and the chance of a worker expecting to leave the firm entirely is even stronger. We conclude, therefore, that dissatisfied workers are more likely to expect to leave their current firm and not just their current job. Low-skilled workers who are employed in the services sector are more likely to expect to change jobs within their current firm. This indicates that the low-skilled workers who are employed in the service sectors have a relatively high firm-internal employability. Table II also shows that the older the workers are, the less likely it is that they expect to leave their current job or their current firm within the next five years. Apparently, older workers gradually lose faith in their external and firm-internal employability. They rely mainly on remaining employed in their current job. Tenure adds to these findings. The longer the workers have worked in their current job, the less likely they are to expect to be externally employable. However, tenure has no significant effect on whether or not a worker expects to change jobs within the firm. Finally, male workers are more likely to expect to leave their current firm within the next five years than female workers. Do workers realise their employability expectations? In order to establish whether or not workers’ medium-term employability expectations in 1998 had already been realised in 2000[6], we analysed the determinants of the actual changes in the workers’ labour market position during the period 1998-2000 by means of a multinomial logit analysis. The dependant variable in this analysis is the effectuation of a worker’s employability in the first two years, in which we again distinguish between a worker’s firm-internal employability, external employability and – as a reference – job-match employability. In order to analyse whether the changes in a worker’s labour market position are in accordance with their employability expectations, we used the following independent variables. (1) Employability expectation: expected labour market situation in five years. . job-match employability, . firm-internal employability, and . external employability (1998). (2) Training 1994-1998: participation in courses (1994-1998).

(3) Training 1998-2000: participation in courses (1998-2000). (4) Task flexibility 1996-1998: tasks performed outside own job (1996-1998). (5) Task flexibility 1998-2000: tasks performed outside own job (1998-2000). Moreover, as in our first analysis, we include the variables on workers’ job satisfaction, overtime work and part-time work, and the covariates gender, age, sector of industry, job level and tenure (all variables refer to the situation in 1998). Table III shows that low-skilled workers who in 1998 expected to move to another job within the firm where they were employed were indeed significantly more likely to have another job within the firm by the year 2000. This indicates that workers’ expectations regarding their firm-internal employability were realistic. However, low-skilled workers who considered themselves externally employable in 1998 were also significantly more likely to change job within the firm than were workers who expected to remain in the same job. Surprisingly, the workers who in 1998 expected to leave the firm within five years were no more likely to have left the firm in the first two years than were the workers who in 1998 expected to remain employed in their job for another five years. One possible explanation for this is that finding a new job outside the firm is more difficult and at least takes more time than finding one inside the firm, and that workers who consider themselves externally employable use inside options as an alternative to outside options that might be more difficult to realise. Table III also shows that low-skilled workers who participated in training in the period 1998-2000 were significantly more likely to move to another job within the firm. Since participation in training during the period 1994-1998 had no such effect, low-skilled workers probably mainly participate in training just before they change jobs internally, or even after the job change[7]. Workers’ task flexibility, however, does not have an additional effect on the extent to which low-skilled workers effectuate their employability. Furthermore, Table III also shows a significant negative effect of job tenure on the external employability of low-skilled workers and a significant positive effect of job tenure on workers’ firm-internal employability. This indicates the firm-specific skills that the workers acquire in the course of their careers or the seniority rules in the internal labour market. Finally, Table III shows that male workers were more likely to leave the firm where they work than female workers. Also, workers in the services sector were less likely to leave their firm than workers in other sectors of the economy. However, the low-skilled workers who are employed in services did not realise their higher firm-internal employability expectations.

Training, task flexibility and employability 83

IJM 25,1 External employability Intercept Expected firm-internal employability Expected firm-external employability 84 Expected job-match employability Training (1994-1998) Training (1998-2000) Task flexibility (1996-1998) Task flexibility (1998-2000) Satisfaction Overtime Part time Male Age Elementary job Lower level job Middle or higher level job Manufacturing sector Services sector Tenure Firm-internal employability Intercept Expected firm-internal employability Expected firm-external employability Expected job-match employability Training (1994-1998) Training (1998-2000) Task flexibility (1996-1998) Task flexibility (1998-2000) Satisfaction Overtime Part time Male Age Elementary job Lower level job Middle or higher level job Manufacturing sector Services sector Table III. Tenure Multinomial logit 2 2 Log-likelihood = estimation of low-skilled x2 = workers’ employability N= effectuation (reference Df = category: job-match Notes: *Significant at the 10 per cent level; employability, i.e. no **significant at the 5 per cent level changes)

b

Standard Error

1.32 0.45 0.85 REF 0.01 2 0.07 2 0.17 2 0.30 2 0.79 0.07 0.17 2 0.73* 2 0.11 REF 2 0.03 2 0.52 2 0.49 2 0.73** 2 0.12**

0.93 0.45 0.52 REF 0.37 0.32 0.35 0.33 0.56 0.32 0.45 0.44 0.20 REF 0.47 0.50 0.42 0.37 0.04

2 3.20** 1.698** 1.252** REF 2 0.67 0.84** 2 0.30 0.08 0.21 0.21 0.42 2 0.77 0.07 REF 0.37 0.77 0.23 0.22 0.06** 511.74 82.37 365 32

1.26 0.43 0.63 REF 0.43 0.37 0.44 0.39 0.70 0.38 0.53 0.51 0.24 REF 0.66 0.66 0.49 0.42 0.02 – – – –

The effect of workers’ employability expectations on training participation and task flexibility One may wonder whether low-skilled workers deliberately participate in training and demonstrate their task flexibility to increase their chances of realising their firm-internal or external employability expectations[8]. Therefore, we analysed by means of two binomial logit analyses whether the employability expectations of low-skilled workers have any effect on their training participation or task flexibility. Table IV shows that low-skilled workers who expected to have a high firm-internal employability were more likely to participate in training courses than workers who expected to remain employed in the same job and the workers who expected to have a high external employability. Since the low-skilled workers who participate in training are more likely to move to another job in the firm-internal labour market than are workers who do not participate in training (see Table III), we conclude that low-skilled workers’ training participation is indeed a vehicle by which they manage to increase their chances of realising their firm-internal employability expectations. Table IV also shows that the employability expectations of the low-skilled workers have no effect on their task flexibility. This indicates that low-skilled workers do not consider their task flexibility as a tool to increase their chances of realising their employability expectations, although it might also be possible that they cannot signal their task flexibility in the jobs they have. Conclusions and discussion In this paper, we analysed whether the training participation and task flexibility of low-skilled workers contribute to their firm-internal and firm-external mobility. We found that both workers’ training participation and task flexibility merely contribute to workers’ firm-internal employability. However, the workers’ participation in training plays a much more explicit role in their firm-internal careers than their task flexibility. Workers who demonstrate a large degree of task flexibility indeed expect to have a large degree of firm-internal employability. The latter, however, does not induce them to demonstrate their task flexibility more often, and their task flexibility does not increase their chances of realising their firm-internal employability expectations. On the other hand, workers’ participation in training does not increase their firm-internal employability expectations. In practice, however, participation in training increases a worker’s chances of moving to another job in the firm-internal labour market, whereas we also found that workers who think they are firm-internal employable are more likely to participate in the training courses. Neither participation in training nor task flexibility contributes to the external employability of low-skilled workers. Task-flexible low-skilled workers are less likely to expect to be externally employable than non-task

Training, task flexibility and employability 85

IJM 25,1

b

Standard Error

Training participation (1998-2000) Intercept 2 1.46** Expected firm-internal employability 0.82** Expected firm-external employability 0.03 86 Expected job-match employability REF Training (1994-1998) 0.91** Satisfaction 0.23 Overtime 2 0.04 Part-time 0.36 Male 0.72** Age 2 0.22 Elementary job REF Lower level job 0.17 Middle or higher level job 0.12 Manufacturing sector 0.21 Services sector 0.18 Tenure 2 0.01 2 2 Log-likelihood = 494.68 x2 = 32.98** N= 365 Df = 13 Task flexibility (1998-2000) Intercept 0.95 Expected firm-internal employability 0.14 Expected firm-external employability 0.37 Expected job-match employability REF Task flexibility (1996-1998) 0.45 Satisfaction 0.28 Overtime 2 0.32 Part-time 2 0.31 Male 2 0.42 Age 2 0.26 Elementary job REF Lower level job 0.63 Middle or higher level job 0.58 Manufacturing sector 0.26 Services sector 2 0.34 Table IV. Binomial logit Tenure 2 0.02 estimation of low-skilled 2 2 Log-likelihood = 428.12 workers’ training x2 = 16.96 participation and task N= 365 flexibility (reference Df = 13 categories: no training and no task flexibility) Notes: *Significant at the 10 per cent level; **significant at the 5 per cent level

0.68 0.31 0.41 REF 0.27 0.45 0.23 0.36 0.35 0.14 REF 0.37 0.38 0.29 0.27 0.02 – – – – 0.76 0.35 0.47 REF 0.27 0.48 0.25 0.38 0.38 0.16 REF 0.39 0.39 0.33 0.28 0.02 – – – –

flexible workers. This shows that low-skilled workers’ task flexibility is merely a firm-internal employability enhancing practice that might reduce their scope of opportunities in the external labour market. Participation in training does not seem to play any role in workers’ perceptions of their external employability or with respect to their actual external employability. These results can probably be explained by the conclusions drawn by de Grip and Wolbers (2002), who found that low-skilled workers usually have more opportunities to improve their position in the firm-internal labour market than they do in the external labour market. This is also shown by our finding that the low-skilled workers who found themselves externally employable were more likely to move to another job in the internal labour market instead of realising their external employability expectations. Notes 1. The extent to which workers can determine their training participation and task flexibility of course also depends on the “effectuation conditions” offered to them. 2. See de Grip et al. (2004) for a more comprehensive overview of the literature on employability. 3. See note 1. 4. We excluded the workers who were older than 50, since in The Netherlands these workers have a fair chance that within five years they will leave the labour market. Moreover, we excluded workers with a temporary contract because these workers will almost automatically leave their current job in the next few years, whereas this does not reflect a strong labour market position as is shown in van Loo et al. (2001). 5. Here we had to use the data for the period 1994-1998 as we did not have the data for the period 1996-1998 only. 6. We do not have the data to analyse this relation for a longer period. 7. Our results confirm the results of de Grip et al. (1998), who found a direct relation (i.e. without considering workers’ employability expectations) between workers’ participation in training and their firm-internal mobility, whereas they did not find a correlation between workers’ training participation and their external mobility. 8. See note 1. References Arthur, M.B. (1994), “The boundaryless career: a new perspective for organizational inquiry”, Journal of Organizational Behaviour, Vol. 15, pp. 295-306. Becker, G. (1964), Human Capital: A Theoretical and Empirical Analysis, with Special Reference to Education, National Bureau of Economic Research, New York, NY. Bloch, S. and Bates, T. (1995), Employability, Your Way to Career Success, Kogan Page, London. Borghans, L. and de Grip, A. (2000), “Skills and low pay: upgrading or over-education?”, in Gregory, M., Salverda, W. and Bazen, S. (Eds), Labour Market Inequalities, Problems and Policies of Low Wage Employment in International Perspective, Oxford University Press, Oxford, pp. 198-223. Bridges, W. (1994), Jobshift, How to Prosper in a Workplace without Jobs, Addison-Wesley, Reading, MA.

Training, task flexibility and employability 87

IJM 25,1

88

Charner, I. (1988), “Employability credentials: a key to successful youth transition to work”, Journal of Career Development, Vol. 15 No. 1, pp. 30-40. de Grip, A. and van Loo, J. (2002), “The economics of skills obsolescence: a review”, in de Grip, A., van Loo, J. and Mayhew, K. (Eds), Research in Labor Economics, Understanding Skills Obsolescence, Vol. 1, JAI Press, Amsterdam/Boston, MA, pp. 1-26. de Grip, A. and Wolbers, M. (2002), Are Low-skilled Workers Better off in Countries where Internal Labour Markets Dominate?, ROA, Maastricht, mimeo. de Grip, A., Heijke, H. and Willems, E. (1998), “Training and mobility”, The Netherlands’ Journal of Social Sciences, Vol. 34 No. 1, pp. 78-98. de Grip, A., van Loo, J. and Sanders, J. (2004), “The industry employability index: taking account of supply and demand characteristics”, International Labour Review, Vol. 143 No. 1. Gallup, J.L. (1990), Ageism: The Problem of the 1990s, Brookstreet Employment Bureau, London. Groot, W. and Maassen van den Brink, H. (2000), “Education, training and employability”, Applied Economics, Vol. 32, pp. 573-81. Hoyt, K.B. (1978), “Employability: are the schools responsible?”, New Directions for Education and Work: Reassessing the Link between Work and Education, pp. 29-33. Hyatt, C. (1995), Lifetime Employability: How to Become Indispensable, Mastermedia Limited, New York, NY. Levy, J.M., Jessop, D.J., Rimmerman, A. and Levy, P.H. (1992), “Attitudes of Fortune 500 corporate executives toward the employability of persons with severe disabilities: a national survey”, Mental Retardation, Vol. 30 No. 2, pp. 67-75. Mangum, G.L. (1976), Employability, Employment and Income, Olympus, Salt Lake City, UT. Neil, D. (1999), “The complexity of job mobility among young men”, Journal of Labor Economics, Vol. 17, pp. 237-61. Outin, J.L. (1990), “Trajectoires professionnelles et mobilite´ de la main-d’œuvre: la construction sociale de l’employabilite´”, Sociologie du Travail, Vol. 32 No. 4, pp. 469-89. Salthouse, T.A. (1991), Theoretical Perspectives on Cognitive Ageing, Erlbaum, Hillsdale, NJ. Shields, M. (1998), “Changes in the determinants of employer-funded training for full-time employees in Britain”, Oxford Bulletin of Economics and Statistics, Vol. 60, pp. 189-214. Soloff, A. and Bolton, B.F. (1969), “The validity of the CJVS scale of employability for older clients in a vocational adjustment workshop”, Educational and Psychological Measurement, Vol. 29, pp. 993-8. Thijssen, J.G.L. (1997), “Employability en employment: terminologie, modelvorming en opleidingspraktijk”, Opleiding en ontwikkeling, Vol. 10 No. 10, pp. 9-14. Topel, R.H. and Ward, M.P. (1992), “Job mobility and the careers of young men”, Quarterly Journal of Economics, Vol. 107, pp. 439-79. Tseng, M.S. (1972), “Self-perception and employability: a vocational rehabilitation problem”, Journal of Counselling Psychology, Vol. 19 No. 4, pp. 314-17. van Loo, J., de Grip, A. and de Steur, M. (2001), “Skills obsolescence: causes and cures”, International Journal of Manpower, Vol. 22 No. 1/2, pp. 121-37. Further reading Dekker, R., de Grip, A. and Heijke, H. (2002), “The effects of training and over-education on career mobility in a segmented labour market”, International Journal of Manpower, Vol. 23 No. 2, pp. 106-25. Doeringer, P. and Piore, M. (1971), Internal Labour Markets and Manpower Analysis, Heath, Lexington, MA.

Appendix

Variable External employability Firm-internal employability Job-match employability Expected firm-external employability Expected firm-internal employability Expected job-match employability Training (1994-1998) No training (1994-1998) Training (1998-2000) No training (1998-2000) Task flexibility (1996-1998) No task flexibility (1996-1998) Task flexibility (1998-2000) No task flexibility (1998-2000) Satisfaction No satisfaction Overtime No overtime Part-time Not part-time Male No male Age Elementary job Lower level job Middle or higher level job Manufacturing sector Services sector Other Tenure

.

Definition

N

Changed firms (1998-2000) 90 Changed jobs within the firm (1998-2000) 57 No changes in labour market position (1998-2000) 316 Expects to change firms within five years (1998) 42 Expects to change jobs within current firm within five years (1998) 70 Expects to remain employed in the same job for another five years (1998) 332 Participated in training between 1994 and 1998 335 Did not participate in training between 1994 and 1998 139 Participated in training between 1998 and 2000 203 Did not participate in training between 1998 and 2000 271 Performed tasks outside own job between 1996 and 1998 325 Did not perform tasks outside own job between 1996 and 1998 142 Performed tasks outside own job between 1998 and 2000 318 Did not perform tasks outside own job between 1998 and 2000 132 Satisfied/highly satisfied with job in 1998 435 Not that satisfied/not at all satisfied with job in 1998 39 Worked extra hours, either paid or unpaid in 1996, 1997 or 1998 215 Did not work extra hours in 1996, 1997 or 1998 259 Worked 32 or fewer hours per week in 1998 116 Worked more than 32 hours per week in 1998 319 Male 311 Female 163 Age in 1998 (474) Had an elementary job in 1998 65 Had a lower-level job in 1998 219 Had a middle, higher or academic job in 1998 180 Worked in the manufacturing sector in 1998 92 Worked in the services sector in 1998 158 Worked in another sector in 1998 135 Number of years a worker had worked in the same job with the same employer in 1998 (473)

Training, task flexibility and employability 89

Table AI.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

IJM 25,1

90

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Training and industrial restructuring Structural change and labour mobility in West Germany and Sweden Tomas Korpi Swedish Institute for Social Research, Stockholm University, Stockholm, Sweden, and

Antje Mertens Max Planck Institute for Human Development and Humboldt University, Berlin, Germany Keywords Labour mobility, Skills, Vocational training, Change management, Germany, Sweden Abstract While the structural changes that have taken place in the labour markets of the industrialised world over the past decades are well documented, less is known about how individuals respond to this changing environment. This includes the extent of intersectoral mobility during the work career, skill differentials in mobility, the impact of the type of training on mobility and changes in mobility patterns over a long period of time. Against this backdrop, the purpose of this paper is to examine intersectoral labour mobility during the first 15 years of working careers in Sweden and West Germany. The analyses show that individuals in both countries tend to move away from industry into other sectors during their careers, but that this tendency is rather weak. While there are some mobility differences among educational categories, the differences between transition probabilities of German apprentices and Swedish vocational school students are insignificant. In the face of the massive transformation of employment structures, the importance of variation in the curricula is probably minuscule.

Introduction Since the early 1980s, the transformation of the world economy has attracted an increasing amount of attention (an early example of this debate is Bluestone and Harrison, 1982, a later example is Reich, 1992). Carried out under headings such as “de-industrialisation”, “the rise of the service economy” and “skill-biased technological change”, the common topic of the debate has been and still is the massive restructuring of the industrialised world’s economies. The decline of agriculture, associated with the rise of industrialised mass production, is now compared to the proclaimed demise of industry and the ascendance of the service sector. This development has been argued to have a number of apocalyptic International Journal of Manpower Vol. 25 No. 1, 2004 pp. 90-103 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410525018

The authors thank Karola Rockmann for research assistance and invaluable help in creating the German data file. They also thank the participants in the LoWER/ETLA Conference in Helsinki, May 2002, for interesting discussions.

consequences, such as mass unemployment and contingent employment among the less skilled, up- and down-skilling in different sections of the labour force, and enhanced inequalities in general. While the structural changes that have taken place over the past decades are well documented, less is known about how individuals respond to this changing environment. One way structural changes can be accommodated is through moves into and out of the employment system, including the early retirement of older workers and the entry into working life of the recently trained young workers. However, this process will most likely not be sufficient to ensure structural change, so that switches between jobs, occupations and industries will become necessary. Moreover, even young workers might enter the labour market with skills that soon become obsolete, and the extent to which individuals leave for more promising pastures is therefore, a question of major interest. Furthermore, the link between industrial restructuring and skills indicates that educational differentials in intersectoral mobility are also of interest, as are changes in these differentials over time. Moreover, we should like to know which educational system or training type allows workers to adapt to varying skill requirements. The training provided within different educational systems may thus encourage job shifts across sectoral boundaries. Alternatively, the training received may impede such an individual adjustment to structural change. In the light of the vulnerability of the less- and medium-skilled, the training obtained by these groups is of particular significance. One crucial distinction seems to be between general and specific skills, that is, between skills that are transferable between different contexts and those that are not. An interesting comparison in this respect is provided by two countries, West Germany and Sweden, where the vocational training obtained at the secondary level differs in specificity. Against this backdrop, the purpose of this paper is to examine intersectoral labour mobility during the working career. Using retrospective life history data, the focus is on young workers in the first 15 years of their working career. First, it is questioned whether mobility among the broad sectors of industry, private services and trade, and those of public services and non-profit organisations, is related to the level of education and the sector of employment. Particular attention is paid to any mobility differences among workers with a vocational education in West Germany and Sweden. These analyses focus on workers who entered the labour market in approximately the first half of the 1970s and who were continuously observed up to the end of the 1980s. Finally, intertemporal changes in sectoral mobility are examined within West Germany. Here, mobility in the period 1960-1990 is examined using different birth cohorts between 1940 and 1971.

Training and industrial restructuring 91

IJM 25,1

The paper is set up as follows. After briefly reviewing some basic facts on structural change and labour mobility in Section 2, an empirical analysis is made in Sections 3 and 4. Section 5 presents the conclusion.

92

Structural preconditions for individual mobility The proclaimed de-industrialisation is already apparent in the German and the Swedish employment structure. Figure 1 shows the evolution of employment in

Figure 1. Employment shares in West Germany and Sweden

these two countries during the past two decades. As the figure shows, the basis for the restructuring process differs in some notable respects: industry played a slightly greater role in West Germany, whereas public sector employment stood out in Sweden. However, the development from the somewhat distinct starting points is relatively comparable. Over the two decades shown here, employment shares in industry fell by approximately 10 percentage points in both West Germany and Sweden but, at the same time, the service sectors grew. While in West Germany private services increased by 15 per cent, public services in Sweden expanded by a similar amount. The figures also show that despite the large numbers, structural change is a slow process. While there are instances of accelerating change – in Sweden notably in the 1970s and during the recession of the early 1990s – the overall picture is one of slow but steady transformation. Among the factors discussed in connection with the responsiveness of labour markets are the vocational education and the training system. The educational systems in West Germany and Sweden display some interesting variations regarding these factors. Education in West Germany is frequently described as being very structured, highly stratified with early tracking, and having a limited number of routes to specific degrees. Sweden is often seen as the converse: being more open and comprehensive than its German counterpart (Mu¨ller and Shavit, 1998). Of particular interest here is the way vocational training at the upper secondary level is provided in West Germany and in Sweden. The German apprenticeship system and the Swedish vocational schools are thus the standard tracks towards mid-level vocational degrees. The distinctions lie in the general structure of the programmes and the relative weight of workplace training in the two systems. The German system is often referred to as the “dual system of vocational training,” since trainees receive school education at public vocational schools (Berufsschulen) one or two days per week and on-the-job training within firms three or four days per week. However, firms often administer training in workshops rather than at the workplace. Basically, all sectors of the economy offer training and approximately 360 different nationally recognised apprenticeship programmes exist today, which usually last 24-42 months depending on the occupation. The Swedish system of vocational training was reformed in the early 1970s when it was integrated into upper secondary education. The focus here will be on the situation in the 1970s and 1980s, since this is the period covered by our data; subsequent reforms in the early 1990s will be ignored. During the 1970s and 1980s, the Swedish system of vocational training was characterised by having approximately 25 nationally recognised programmes, with subdivisions giving a total of approximately 60 certificates. The duration of

Training and industrial restructuring 93

IJM 25,1

94

training was generally two years and mostly obtained in school. The time spent at a workplace varied substantially, but a rough estimate based on survey results is that an average of approximately one afternoon a week was allocated to firm-based training (SOU, 1986). This, of course, suggests that the specificity of the skills obtained within the two systems varies. It has been argued that German apprenticeships deliver large amounts of general skills, especially certified and supposedly transferable occupational qualifications (Franz and Soskice, 1995). However, it seems likely that the German system, with its extended period of firm-based training, will still generate skills with a relatively higher degree of specificity (Acemoglu and Pischke, 1999, p. F124). The degree of specificity in the two countries will consequently differ, even if there are considerable amounts of general skills accumulated during an apprenticeship training. Given the international fascination in the dual system, the long-term consequences of firm-based training for flexibility in the face of structural change would seem to be of interest. Data and method The Swedish data used here comes from the Swedish level of living survey (LLS) conducted in 1991, which is a survey among representative samples of the Swedish population. Included are those born between 1955 and 1965, entering the labour market no earlier than 1975, since educational reforms prior to that date could severely influence the results. The Swedish sample thus consisted of approximately 1,000 men and women mostly born between 1955 and 1965. The German data have been taken from the German Life History Study (GLHS), which contains information on representative samples of different German birth cohorts. For the intercountry analysis, the birth cohorts used were 1954-1956 and 1959-1961, who were interviewed in 1989 and who entered the labour market approximately between 1968 and 1975. The sample size here was approximately 2,000 men and women. The retrospective character of the data, that is, the likelihood of recall error, prevents the evolution of intersectoral job mobility from being examined in the Swedish-German comparison. An analysis of changes over time is, however, possible for West Germany, where interviews with the various birth cohorts had taken place at different points in time. Thus, analyses have been carried out including the German birth cohorts 1939-1941, 1949-1951, 1964, and 1971 in addition to the cohorts born approximately between 1955 and 1960 that have been used in the comparative analyses. The two surveys used to share many features. Of primary relevance are the retrospective work histories contained in the two surveys. Both work histories include information on self-defined career episodes (jobs, unemployment, etc.) with a duration of at least one month. Following the restrictions in the LLS, the analysed work histories commence with the first job duration of at least six

months that began after the completion of the respondents’ highest educational degree. All employees with valid observations on the variables of interest were included, with some exceptions. Thus, self-employed were excluded, since they could be expected to display very distinct mobility patterns and are of minor relevance for the training debate. German vocational school graduates were also excluded, since full-time vocational schooling in West Germany is relatively rare and very occupation specific, raising issues of self-selection into different types of training. The overall structure of the work history information is thus very similar. Drawing on previous work that observes mobility among 16 industries and other types of mobility (Korpi and Mertens, 2003)), the de-industrialisation thesis was scrutinised more closely and analyses of switches were carried out among the three broad sectors of industry, private services and trade, and public services and non-profit organisations. Owing to the relative insignificance of the agricultural sector and the relatively small number of observations in the data sets, this was not included in the analysis. Although focus was on the discussion of mobility differences related to sectoral affiliation and educational attainment, controls were also included for employment status, sex, employment experience, number of previous switches, parents’ education, firm size, and time varying measures of the national unemployment rate and industry employment growth. In the basic regressions, the standard semi-parametric proportional Cox model was estimated. To test for robustness of the results, piecewise constant models were also estimated with and without controlling for unobserved heterogeneity. While the advantage of the Cox model is that the baseline hazard does not need to be specified, the advantage of the piecewise constant model is that no proportionality between the baseline hazard and the covariates has to be assumed (Kiefer, 1988). Since many individuals enter the estimations more than once, robust standard errors were always used (the Huber/White/Sandwich estimator of variance) that takes this non-independence of observations into account. However, not only could the standard errors be affected by the multiple observations per individual, but also unobserved individual heterogeneity could bias the parameter estimates. To examine the impact of such heterogeneity, shared frailty models were estimated, an equivalent to random-effects models that can be estimated with survival data[1]. In these models, frailties or hazards are not observation specific, but shared across groups of observations, causing observations within groups to be correlated. In our model, the groups correspond to individuals. By estimating piecewise constant models with shared frailties that are assumed to follow a gamma distribution, a test was attempted to observe whether individual heterogeneity indeed influences the results (Cleves et al., 2002). The Cox models are primarily reported, since the outcomes of all the three models were very similar. The

Training and industrial restructuring 95

IJM 25,1

96

Figure 2. All intersectoral and intra-sectoral industry moves within West Germany and Sweden (in per cent)

piecewise constant model with shared gamma frailty is reported only if significant heterogeneity is detected. Empirical analysis A first impression of the process of de-industrialisation is provided by Figure 2, which shows the relative frequency of moves between the three broad industry aggregates: industry, private services and trade, and public services and non-profit organisations. In West Germany, 61 per cent of all moves are intersectoral and in Sweden 63 per cent of all moves – the remainder takes place within the broad aggregates. Another interesting feature is that moves are by no means directed solely away from industry towards the two service sectors. In both countries, shifts take place between all broad aggregates and move in both directions. The flows between the sectors are often of similar magnitude. Closer examination of the pattern of intersectoral mobility reveals only small intercountry differences. In West Germany, 18 per cent of all moves are out of industry and into private services and 14 per cent are moves in the other direction. In Sweden, the corresponding figures are 15 and 11 per cent. The net difference is, in other words, the same. However, in West Germany, a positive

net difference can also be found moving towards the public and non-profit sector, while in Sweden the flows are practically identical in both directions. Overall, it seems that the influence of labour mobility on industrial restructuring obviously plays similar roles in both countries. Disaggregating the moves by educational groups show that these are roughly proportional to the size of each educational group. In West Germany, 75.2 per cent of the moves are, for example, carried out by 73.7 per cent of those with apprenticeship training. Only university graduates appear to deviate from this pattern: in West Germany 8.8 per cent of those with tertiary education account for 4.9 per cent of the moves. The figures in Sweden are similar to those with tertiary education, again showing signs of being the least mobile group. Based on this preliminary analysis, the links were studied between training type and the employment sector in greater detail. In Table I, examination commenced to find whether differences between West Germany and Sweden can be discerned in a multivariate context. Models are estimated with full interactions between all covariates and a dummy variable for Sweden, models in which the main effects correspond to the German effects and the interaction terms identify the difference between West Germany and Sweden. The effects for West Germany and Sweden are shown in Table I, where the Swedish effects

Stratified baseline Germany

I No Sweden

D

Germany

II Yes Sweden

Training and industrial restructuring 97

D

Secondary vocational

20.201* 0.141 ** 2 0.104 0.095 (0.115) (0.132) (0.126) (0.128) Secondary general 20.579 0.029 2 0.501 20.034 (0.359) (0.174) (0.362) (0.170) Tertiary 20.881*** 20.314 * 2 0.711*** 20.371* (0.244) (0.199) (0.252) (0.193) Industry 20.148 0.076 2 0.092 0.002 (0.108) (0.132) (0.113) (0.131) Public service 0.229* 20.274* *** 0.248* 20.272** *** (0.131) (0.146) (0.134) (0.141) No. of persons 2,460 2,460 No. of job spells 3,491 3,491 No. of events 949 949 Log likelihood 2 7,006.3 2 6,372.6 Notes: Reference categories are basic education and private service. Swedish estimates calculated from education, main effects and country-education interaction terms. ***, **, and * are significances at the 1, 5 and 10 per cent levels, respectively, and the D columns indicate significance of country differences. In addition to the variables shown, both models have included employment status, gender, employment experience, number of previous switches, two parents’ education dummies, two firm size dummies, yearly national unemployment rate, and yearly industry employment growth Source: Own calculations using the Swedish LLS and the GLHS

Table I. Intersectoral labour mobility in West Germany and Sweden. Cox models. Standard errors in parentheses

IJM 25,1

98

have been calculated from the estimated main and interaction effects. Also, results are shown from two different models, the difference being that Model II includes separate baseline hazards for each country, that is, a fixed effect for Sweden. The within-country effects are, however, largely independent of the estimated model. With regard to educational differences, both models show that mobility in West Germany decreases with increasing education. The least mobile group are thus workers with tertiary education. A similar pattern is also visible in Sweden, although the mobility reducing effect of education is less pronounced than it is in West Germany. Noteworthy are also the intercountry differences in the educational effects. In Model I there are, thus, indications that mobility is greater in Sweden among all educational categories, in particular among Swedish employees with vocational training and those with university education, who are obviously more mobile than their German counterparts. However, with the introduction of separate baseline hazards for each country, Model II shows that these intercountry differences disappear. The country differences in mobility rates are, therefore, not specifically linked to differences among the educational categories, although inter-industrial mobility rates are generally higher in Sweden. Interestingly, in the two countries no great difference appears in the rate of mobility out of industry. Neither of the estimates for industry differ from the reference category, nor do they differ from each other. There is, in other words, no indication that the rate of de-industrialisation differs much between the two countries, controlling for other mobility related factors. Differences are, instead, apparent in the public sector. Public sector workers and workers in non-profit organisations in West Germany are much more likely to transfer to other sectors than workers in the same sector in Sweden are. Similar rates of de-industrialisation may, however, conceal important intergroup differences. Given the overall shifts in employment, of special interest are the differences among the educational groups in the likelihood of leaving industry and, in particular, in the employment opportunities of vocationally trained workers. In Table II, this is further pursued by introducing interactions between the educational variables and the three sectors. The mobility decreasing effect of education in West Germany is shown here to be evident only in the public sector; the pattern in the other sectors is less obvious. The association between mobility and education in Sweden is rather mixed. Most remarkable, perhaps, is that a difference in the exit rate from industry is manifested only in the case of workers with university degrees. With regard to workers with vocational training, on the other hand, there is no difference in the likelihood of leaving industry. The German data also allow the examination of temporal changes in intersectoral mobility by looking at job starters from different birth cohorts. In addition to the cohorts, born in approximately 1955-60 who were studied in the

Stratified baseline Germany Industry Public service Manufact. Secondary vocational Secondary general Tertiary Private service Secondary vocational Secondary general Tertiary

I No Sweden

20.036 (0.211) 0.504* (0.301)

0.110 (0.193) 2 0.071 (0.219)

20.245 (0.168) 20.928 (0.646) 20.275 (0.320)

0.109 (0.190) 2 0.092 (0.257) 0.746*** (0.304)

20.122 (0.171) 0.030 (0.492) 20.464 (0.406)

0.263 (0.206) 2 0.145 (0.307) 0.123 (0.294)

D

**

Germany

II Yes Sweden

0.156 (0.244) 0.712** (0.330)

2 0.044 (0.188) 2 0.151 (0.207)

20.205 (0.170) 20.869 (0.609) 20.173 (0.325)

0.104 (0.189) 2 0.087 (0.253) 0.674** (0.300)

0.083 (0.211) 0.232 (0.505) 20.188 (0.430)

0.131 (0.194) 2 0.293 (0.298) 2 0.015 (0.278)

D

Training and industrial restructuring 99

*

Public service Secondary vocational

20.282 0.045 20.295 0.051 (0.288) (0.230) (0.298) (0.226) Secondary general 21.093* 0.590** *** 21.143* 0.536* *** (0.584) (0.286) (0.600) (0.281) Tertiary 21.612*** 2 0.965*** 21.548*** 2 0.976*** (0.426) (0.262) (0.434) (0.260) No. of persons 2,460 2,460 No. of job spells 3,491 3,491 No. of events 949 949 Log likelihood 26,983.9 2 6,350.8 Notes: Reference categories are basic education and private services. ***, **, and * are significances at the 1, 5 and 10 per cent levels, respectively, and the D columns indicate significance of country differences. See Table I for further notes

intercountry analysis, those born in approximately 1940-50 and in approximately 1964-71 were included. Model Ia in Table III shows the oldest cohort being less likely and the youngest cohort more likely to switch sectors than the middle cohort studied earlier. However, in contrast to the previous analysis, evidence was found of significant individual heterogeneity within the piecewise exponential models[2]. As can be seen in Model Ib, introducing a shared frailty gamma term for observations that belong to the same individual reduces the increase in the mobility rate over time, leaving only the dummy for the oldest cohort significant. The educational effects found in these models roughly correspond to the results for West Germany, as reported above. Mobility is thus the lowest for

Table II. Intersectoral labour mobility in West Germany and Sweden, educational effects by sector. Cox models. Standard errors in parentheses

IJM 25,1

Model type Chi-square test for heterogeneity Cohort 1940-50

100

Cohort 1964-71 Secondary vocational Secondary general Tertiary Industry Public service Private service 1940-50 1964-71 Industry 1940-50 1955-60 1964-71

Ia Cox n/a

Ib PE (frailty) 14.00***

IIa Cox n/a

IIb PE (frailty) 9.10***

20.480*** (0.109) 0.131* (0.069) 0.036 (0.080) 20.470** (0.232) 20.200 (0.148) 0.067 (0.059) 0.371*** (0.081)

2 0.367*** (0.101) 0.097 (0.072) 2 0.011 (0.081) 2 0.484** (0.218) 2 0.273* (0.144) 0.073 (0.063) 0.356*** (0.083)

20.015 (0.078) 20.528** (0.234) 20.258* (0.146)

2 0.041 (0.080) 2 0.0532** (0.217) 2 0.304** (0.143)

20.369*** (0.139) 0.226** (0.102)

2 0.289** (0.135) 0.199* (0.105)

20.319** (0.133) 20.036 (0.098) 0.034 (0.106)

2 0.199 (0.132) 2 0.028 (0.106) 2 0.018 (0.109)

Public service 1940-50

0.041 0.139 (0.177) (0.172) 1955-60 0.233* 0.224* (0.130) (0.134) 1964-71 0.375*** 0.321* (0.134) (0.139) Table III. No. of persons 4,476 4,476 4,476 4,476 Intersectoral labour No. of job spells 5,807 5,807 5,807 5,807 mobility in West 1,520 1,520 1,520 1,520 Germany. Changes over No. of events 2 12,474.8 25,265.1 2 12,394.8 25,258.7 time, and changes over Log likelihood time by sector. Cox and Notes: Cox proportional hazard and piecewise exponential models with shared gamma frailty. Reference categories are Cohort 1955-60, basic education, private services, and private piecewise exponential models. Standard errors services £ Cohort 1955-60, respectively. ***, **, and * are significances at the 1, 5 and 10 per cent levels. See Table I for further notes in parentheses

those with advanced general or tertiary education. Again, movements from the industry sector are not more likely than movements from the private services, but there is a higher transition rate in the public services than in the other two sectors. In Models IIa and IIb, changes in these sector differences were analysed over time. In the Cox model (IIa), there is a tendency for mobility to increase in all

sectors. In particular the youngest cohort, which entered the labour market in the 1980s and early 1990s, stands out. However, the significances of these differences were reduced when the control for unobserved heterogeneity was introduced. Table IV shows model interactions between sectors, cohorts and educational groups. In Models Ia and Ib examinations were carried out to establish whether educational groups in different sectors have different hazards to terminate their jobs. Among those with only basic schooling, mobility was lowest in the

Model type Chi-square test for heterogeneity Cohort 40-50 Cohort 64-71 Industry Public service Secondary vocational manufact. private service public service Secondary general manufact. private service public service Tertiary manufact.

Ia Cox n/a

Ib PE (frailty) 7.30***

2 0.332*** 20.222** (0.106) (0.099) 0.143*** 0.097 (0.069) (0.072) 0.015 0.014 (0.144) (0.149) 0.636*** 0.685*** (0.189) (0.178)

IIa Cox n/a

Training and industrial restructuring 101

IIb PE (frailty) 12.25***

2 0.552*** 2 0.418** (0.166) (0.166) 2 0.166 2 0.223 (0.196) (0.204) 0.068 0.079 (0.060) (0.063) 0.383*** 0.370*** (0.082) (0.083)

0.013 (0.126) 0.118 (0.124) 0.360** (0.142)

20.001 (0.124) 0.090 (0.122) 0.318** (0.143)

1940-50 2 0.551*** 2 0.473*** (0.146) (0.151) 1955-60 2 0.056 2 0.078 (0.121) (0.133) 1964-71 0.082 0.012 (0.126) (0.138)

2 0.188 (0.344) 0.032 (0.315) 2 0.798** (0.429)

20.264 (0.437) 0.007 (0.305) 20.815** (0.404)

1940-50 2 1.549*** 2 1.500** (0.593) (0.741) 1955-60 2 0.340 2 0.375 (0.353) (0.311) 1964-71 2 0.497 2 0.575 (0.345) (0.342)

0.084 0.043 1940-50 2 0.521* 2 0.500* (0.228) (0.244) (0.272) (0.272) private service 2 0.018 20.073 1955-60 2 0.664*** 2 0.710*** (0.222) (0.219) (0.238) (0.245) public service 2 0.199 20.264 1964-71 0.022 2 0.054 Table IV. (0.224) (0.227) (0.213) (0.217) Intersectoral labour No. of persons 4,476 4,476 4,476 4,476 mobility in West No. of job spells 5,807 5,807 5,807 5,807 Germany, educational No. of events 1,520 1,520 1,520 1,520 effects by sector and Log likelihood 2 12,443.7 2 5,237.1 212,469.8 25,260.1 over time. Cox and Notes: Cox proportional hazard and piecewise exponential models with shared gamma frailty. piecewise exponential Reference categories are Cohort 1955-60, basic education, and private services. ***, **, and * are models. Standard errors significances at the 1, 5 and 10 per cent levels. See Table I for further notes in parentheses

IJM 25,1

102

private services sector (the reference category). In contrast, it was found that vocationally trained workers are more mobile in the public services. There is, in other words, a clear difference in the mobility among public sector workers with only basic education and other public sector employees. In the industry sector, however, there are no apparent marked differences in the exit probability of the educational groups. The mobility of vocationally trained workers in both industry and private services is also basically identical to that of workers with higher educational attainment. This differs from the public sector (including non-profit organisations), where the other two groups clearly show lower mobility than the apprentices. In Model II, the focus was on changes in the educational effects over time. A close inspection of the results indicates that there is evidence for such changes for basically all educational groups. First, a reduction in mobility is shown between the 1940-50 and the 1955-60 birth cohorts, a reduction followed by an increase between the 1955-60 and 1964-71 birth cohorts, which is, however, not significant apart from those with university education. Conclusions The topic of this paper is the fate of the individuals exposed to industrial restructuring. The initial analysis using aggregate data showed that the transformation of the German and Swedish economies mainly run along parallel paths. In both countries, the new structure is one with employment concentrated in two sectors, namely (1) private service and trade; and (2) public service and non-profit organisations, with an industry sector substantially smaller than a few decades ago. This transformation is also visible in individual careers, but only to a rather limited extent. During their careers, individuals in both countries tend to move away from industry into other sectors; however, this tendency is weak and the differential between entries into and exits out of industry appear to be rather small. The entry-exit gap in industry is also small in relation to the overall number of job shifts, many of which are within sectors or between the other sectors. In a multivariate context, this implies that in neither of the countries do employees tend to leave industry more frequently than, for example, private or public services. The slow but steady transformation process also suggests that the intertemporal differences are rather minor. There are some indications that the youngest German cohort is more mobile than its predecessors, yet the evidence is fairly weak. One potential implication of this is that the career consequences of structural change may be greater within than between sectors. Both qualification changes within sectors (Berman et al., 1994) and intra-sectoral job switches might be more important.

There is, therefore, no general flight from the factories, nor does the level of education tend to affect the exit probability. The only exceptions here are those with a university education in Sweden, who leave industry more readily than other Swedish employees, and those with a university education in West Germany. The relatively limited importance of education is also evident as far as there is no difference in the exit probabilities of German apprentices and Swedish vocational school students. The distinction between school- and firm-based vocational training – between more general and more specific occupational training – is therefore, of little relevance to intersectoral mobility patterns. In the face of the massive transformation of employment structure, the importance of the variation in curricula is probably minuscule. .

Notes 1. STATA 7 was used for the analyses. 2. The results from the piecewise exponential models without shared frailty/heterogeneity appear very similar to those from the Cox model and therefore are not reported. References Acemoglu, D. and Pischke, J-S. (1999), “Beyond Becker: training in imperfect labour markets”, Economic Journal, Vol. 109 No. 453, pp. F112-42. Berman, E., Bound, J. and Griliches, Z. (1994), “Changes in the demand for skilled labor within US manufacturing: evidence from the annual survey of manufacturers”, Quarterly Journal of Economics, Vol. 109 No. 2, pp. 367-97. Bluestone, B. and Harrison, B. (1982), The Deindustrialization of America: Plant Closings, Community Abandonment, and the Dismantling of Basic Industry, Basic Books, New York, NY. Cleves, M.A., Gould, W.W. and Gutierrez, R.G. (2002), An Introduction to Survival Analysis using STATA, Stata Press, College Station, TX. Franz, W. and Soskice, D. (1995), “The German apprenticeship system”, in Buttler, F. et al. (Eds), Institutional Frameworks and Labor Market Performance, Routledge, London. Kiefer, N.M. (1988), “Economic duration data and hazard functions”, Journal of Economic Literature, Vol. 26 No. 2, pp. 646-79. Korpi, T. and Mertens, A. (2003), “Training systems and labor mobility”, Scandinavian Journal of Economics, Vol. 105 No. 4, pp. 597-617. Mu¨ller, W. and Shavit, Y. (1998), “The institutional embeddedness of the stratification process: a comparative study of qualifications and occupations in 13 countries”, in Shavit, Y. and Mu¨ller, M. (Eds), From School to Work: A Comparative Study of Educational Qualifications and Occupational Destinations, Clarendon, Oxford, pp. 1-48. Reich, R. (1992), The Work of Nations: Preparing Ourselves for 21st Century Capitalism, Random House, New York, NY. SOU (1986), En trea˚rig yrkesutbildning. Del 1, Riktlinjer fo¨r fortsatt arbete. SOU 1986:2, Utbildningsdepartementet, Stockholm.

Training and industrial restructuring 103

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

IJM 25,1

104

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Labour market effects of apprenticeship training in Austria Helmut Hofer Institute for Advanced Studies, Vienna, Austria, and

Christine Lietz Department of Applied Economics – Microsimulation Unit, University of Cambridge, Cambridge, UK Keywords Apprenticeships, Training, Pay, Labour market, Austria, Unemployment Abstract In Austria, the apprenticeship system provides all citizens, including the less able among them, with a training option. Based on social security data, this article examines earnings and the stability of the occupational career of young workers with an apprenticeship diploma. As control groups, workers with a full-time secondary school education and workers who did not receive any further education after completing their compulsory education were used. One of the main findings is that workers with an apprenticeship diploma are much better off than those without further education. The article finds the following ranking with respect to education: high-school graduates, ex-apprentices and unskilled workers, with more pronounced differences between ex-apprentices and unskilled workers.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 104-122 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410525027

1. Introduction In Austria, low-skilled jobs are much less common than they are in countries such as the United States and Great Britain. Moreover, unemployment – and especially youth unemployment – is low when compared to international standards; in addition, income differences are less pronounced. Without doubt, the Austrian apprenticeship system plays a decisive role in this context, not at least by providing a training option for those who failed at school. The apprenticeship system has recently attracted much international attention. Although it is acknowledged that it functions well as a pathway from school to work and for training even the less gifted, there are fears that the apprenticeship system could be an antiquated model. Empirical and theoretical economic studies focus on the incentives of the various participants (training firms, apprentices, policy makers) and on the labour market effects of the apprenticeship system. Acemoglu and Pischke (1999) picked up Becker’s (1964) theory of human capital formation in which he states that, with competitive labour markets, firms never pay for investments in general training. Acemoglu and Pischke present a model of imperfect labour markets where firm-sponsored general training arises as an equilibrium phenomenon and discuss a variety of evidence, which support the predictions of their non-competitive theory.

Euwals and Winkelmann (2001) have contributed to the ongoing debate about why firms are willing to invest in apprenticeship training, even though many apprentices leave the training firm soon after completing their apprenticeship. They found that apprentices who stay with their training firm after receiving their diploma have higher wages and stay in their first job longer than apprentices who leave the training firm. These results led them to conclude that theories according to which firms use the apprenticeship system to select and retain the more able apprentices, thereby recouping the costs of investing in skills that are in principle portable, are supported. Franz et al. (2000) analysed youth’s chances of labour market entry in Germany, with a main focus on failures during this early stage of the working life. They found some evidence for a permanent income reduction caused by early failures in the working history, for example dropping out of an apprenticeship or failing to pass examinations. Another finding was that to some extent youth unemployment in Germany affects the age group 20-24, because teenagers are absorbed by the apprenticeship system. Harhoff and Kane (1997) described characteristics of the German labour market, which may lead firms to accept part of the cost of general training, even in the face of worker turnover. They compared labour market outcomes of apprentices in Germany with high-school graduates in the United States, and found that the two groups occupy a similar position within the wage structure of their respective country. Ryan (2001) discussed research findings concerning the school-to-work transition in France, Germany, Japan, The Netherlands, Sweden, the United Kingdom and the United States. Germany was chosen mainly because its apprenticeship system is of special interest in this context. Ryan concluded that although school-to-work transition problems run wide and deep, the problems are not as acute as might be thought. For example, the youth labour market has held up quite well in Germany, despite concerns about apprenticeship training. Like many other authors, Ryan believes that the success of the apprenticeship system is underpinned by nation-specific institutions, and he hypothesised that such institutions flourish only in societies in which concern for the integration of youth into socio-economic life is widely shared, and that therein lies the fundamental source of the resilience and effectiveness of the system. Steedman (2001) reviewed the main characteristics of apprenticeship in Austria, Denmark, France, Germany and The Netherlands in order to compare them with current practice in British modern apprenticeship. The main areas examined were the statutory framework and standards, employment prospects, achieving the employer-apprentice match, prior school qualifications of apprentices and motivation to serve an apprenticeship, the management, and financing apprenticeship. Steedman concluded that apprenticeship in Britain, judged as a programme, falls short of the standards elsewhere in Europe on every important measure of good practice.

Labour market effects of apprenticeship 105

IJM 25,1

106

The international literature on apprenticeship refers mainly to Germany. Although the Austrian apprenticeship system also has a long tradition, it is mentioned less frequently (Brandel et al., 1994; Hofer and Pichelmann, 1994; Sto¨ger and Winter-Ebmer, 2001). Therefore, the goal of this study is to present an overview of the Austrian apprenticeship system, to show similarities with and differences from the well-known German system, and to investigate its labour market effects. Based on social security data, we examine earnings and the stability of the occupational career of workers with an apprenticeship diploma relative to workers with a secondary school education and workers who did not finish any further education after completing their compulsory education. Section 2 presents the Austrian apprenticeship system and describes the incentives and various positions of the parties involved, namely apprentices, training firms and policy makers. Section 3 details the data used for the empirical analysis and presents the results. Section 4 comprises the summary and conclusion. 2. The Austrian apprenticeship system 2.1 Institutional framework About 40 per cent of all young Austrians who have completed their compulsory education serve an apprenticeship. Correspondingly, one-third of the Austrian population holds an apprenticeship diploma. These figures illustrate the importance of the apprenticeship system for the pathway from school to work, and for the Austrian economy in providing skilled labour. The latter – the provision of skilled labour – is also supported by a broad spectrum of vocational schooling at the middle- and upper-secondary school level. This type of vocational education has gained ground in the last decades, and in recent years competition between the two tracks of vocational education has developed. A very important feature of the Austrian apprenticeship system is its strong legal and institutional framework. The apprenticeship curricula are largely determined by legally defined occupational profiles. A large number of bodies are involved in the decision-making process concerning the apprenticeship system: the Ministry for Economic Affairs has jurisdiction at the enterprise level, the Ministry of Education and Research has jurisdiction regarding vocational schools, and local governments at the district level have functions covering both areas. The social partners are strongly involved, by for example, fulfilling advisory functions or through collective wage bargaining. In principle, anyone who has completed compulsory education may serve an apprenticeship; there are no formal entry requirements regarding school performance. Enterprises have to fulfil certain legal requirements in order to have the right to train apprentices. In particular, the training firm must employ persons who are qualified as educators (i.e. have passed the relevant exams).

The apprentice or his/her legal guardian and the training firm sign an apprenticeship contract, which establishes a labour relationship between the two parties. The main parameters of this contract (duration, working hours, protection against dismissal, etc.) are determined by negotiations between the social partners, namely the trade unions and representatives of the entrepreneurs. The contract also specifies the amount of compensation the apprentice will receive (i.e. a wage for performing productive work). A minimum amount is usually fixed in centralised collective bargaining agreements. The length of training varies from two to four years, but in most cases lasts three or three and a half years. About 80 per cent takes place at the firm and the remainder at a vocational school; the latter provides the theoretical foundation and supplements the training in the firm (75 per cent of the school time), and enables general education to be completed (25 per cent of the school time). The vocational schools are financed by public means, whereas workplace training is financed by the training enterprise. At the end of their training, apprentices can obtain a legally recognised apprenticeship diploma by passing a final exam. The contents and the execution of this exam are strongly regulated. In its basic institutional structure, the Austrian apprenticeship system is similar to the German one. However, there are remarkable differences as the Austrian system is more traditional. In Germany, the majority of employees acquire their skills while serving an apprenticeship and the apprenticeship system is common in all sectors of the economy. The Austrian system covers a narrower range of more traditional occupations, and there is a strong connection with the manufacturing and construction sectors. The connection with the service sector is comparatively weak and mostly involves traditional service sectors with a comparatively low skill profile. Approximately, 33 per cent of all training firms operate in the services sector, with a strong concentration in the trade and in the hotel and restaurant sectors. Although there are approximately 280 apprenticeship occupations, the distribution of apprentices across these occupations is highly concentrated: about 90 per cent of apprentices are trained in less than 50 occupations. The ten most popular apprenticeship trades are retail trade service, office assistant, hairdresser and wigmaker, joinery, cook, motor vehicle mechanic, electrician, motor vehicle engineer, retail trade service-food products, and bricklayer. Some 70 per cent of female and 42 per cent of male apprentices are trained in one of these ten trades (BMWA, 2003). Apprenticeships are also highly segregated by gender. While many male apprentices opt for a job in the productive sector (e.g. car mechanics, plumbing, carpentry, masonry, metalworking), female apprentices are mainly found in sales, clerical work, hairdressing, and restaurant services. Moreover, the initial age of starting an apprenticeship and the education level are lower in Austria than they are in Germany. In Austria, rising demands

Labour market effects of apprenticeship 107

IJM 25,1

for qualifications have been met by the competing full-time vocational colleges rather than by upgrading the nation’s apprenticeship system (Lassnigg, 2001). Furthermore, the gross costs of providing apprenticeships (e.g. instructor costs, costs of capital goods and material) are higher in Germany, pointing to a higher quality of apprenticeship training (Lassnigg and Steiner, 1997).

108 2.2 Problems of the Austrian system In the 1990s in Austria, there was a continuous drop in the number of apprentices as well as in the number of training firms, hitting bottom in 1996. The decline in the number of apprentices can only to some extent be explained by the demographic development: not only the number of apprentices dropped, but also the share of young people who become an apprentice after completing their compulsory education fell. A consideration of the incentives of the various parties involved may be helpful for a better understanding of the developments concerning apprenticeship. Besides the apprenticeship system, vocational education in Austria is provided by a broad spectrum of vocational schools at both middle- and upper-secondary school levels. In recent years, there has been an ongoing trend of young people opting for full-time school education rather than for an apprenticeship. A major reason for this trend is the general opinion that school education provides young people with better chances for a future working life than an apprenticeship does. Another reason may be that youths who cannot find a training place after completing their compulsory education switch to full-time school education. Concerning the incentives of entrepreneurs, human capital theory may provide some insight. In his Human Capital, Becker (1964) was the first to distinguish between general and firm-specific skills. General skills can be used in every firm, while firm-specific skills are only of value in the firm that provided the training/skills. Becker concluded that firms have no incentives to bear the costs of general training, because the workers once skilled can be poached by free-riding other firms. But skills acquired in the Austrian apprenticeship system are to a remarkable extent general skills, as a result of prescribed occupational profiles. Moreover, Lassnigg and Steiner (1997) found that roughly 66 per cent of the Austrian apprenticeship firms face net costs (benefits minus costs) for training. The question is, then, why does the apprenticeship system work? Becker’s predictions are based on the assumption of perfectly competitive labour markets. By dropping this assumption, Acemoglu and Pischke (1999) developed a model in which firms may sponsor general training in equilibrium. With imperfect labour markets, a “compressed wage structure” may arise. This means that training firms are able to pay their skilled workers less than their marginal product, without the risk that other firms will poach them. If this is the case, firms are compensated for the training costs if the skilled workers stay

with them long enough. Acemoglu and Pischke found several explanations for the existence of a compressed wage structure. One of them is asymmetric information between the training firm and other employers: because potential employers do not know the exact level of skills the worker has achieved, they may not be willing to pay his full marginal product. Another explanation is search costs, as it may not be easy for the skilled worker to find an appropriate new employer[1]. But even if trained workers stay with the training firm and increase profits because of their enhanced productivity, they have to stay long enough to compensate for the firms’ training costs. But evidence for Austria shows that they do not stay long enough (Hofer and Pichelmann, 1994): one year after completing their apprenticeship, the probability of still being employed by the training firm is less than 50 per cent (Lietz, 2001). This fact – and in this context, the high training costs – may contribute to the decline in the number of firms providing apprenticeships. Another reason may be the above-mentioned trend for schools to provide vocational training. If the more able opt for school education, only the less able will serve an apprenticeship. On the other hand, the qualification requirements for the apprentices are subjected to constant change due to technical and economic developments[2]. Policy makers are highly interested in the functioning of the apprenticeship system, because it is widely accepted that the apprenticeship system plays a very important role in keeping youth unemployment figures low and in providing a well-founded education even for the less able. For this reason, the growth of the Lehrstellenlu¨cke (the gap between the number of young people looking for an apprenticeship and the number of available apprenticeships) in the 1990s alarmed the responsible decision makers. Since 1997, a range of measures have been established which are helping to ease the tension, at least in the short run. One of these measures was the creation of a safety net, that is, offering publicly funded apprenticeships to youths who were unable to find one in a firm. Other measures were aimed at reducing the costs of training firms to increase their incentives to train. Over 100 apprenticeship occupations were newly created or reoriented. Although the importance of a properly functioning apprenticeship system is generally recognised, the various decision makers are at odds concerning the manner in which the system can be maintained. For example, the employers’ representatives are demanding the further reduction of costs for training firms, the loosening of regulations and the better qualification of new apprentices through the implementation of preparation courses. They argue for higher costs for youths (i.e. their parents) for school education to redress what in their view is an imbalance in the incentive structure between the two tracks of education.

Labour market effects of apprenticeship 109

IJM 25,1

110

For their part, the employees’ representatives are demanding the continuation of the safety net with the argument that the government should ensure that there are training possibilities for everyone if the industry is unable to do so. One of their main goals is the maintenance or even improvement of the quality of the training provided by an apprenticeship. Deregulation at the expense of quality is strictly rejected. 3. Data and results The aim of this section is to compare the occupational career and the earnings of workers with an apprenticeship diploma with: (1) same-aged workers with a secondary school education but without a university degree, and (2) same-aged workers who did not undertake any further education after completing their compulsory education. In the following, these three groups are referred to as ex-apprentices, high-school graduates and unskilled workers, respectively. 3.1 Data The study is based on administrative data from the Main Association of Social Insurance Institutions in Austria (Hauptverband der Sozialversicherungstra¨ger). By law, all employers are required to report information on their workforce to the Austrian Social Security Administration. The sample drawn contains employment histories from 1984 to 1998 of 100,000 workers with at least one recorded social-insurance-related episode[3] in 1992 and/or 1996. As the data come from official registers, an important advantage is its accuracy with respect to the duration of the labour market episodes and to earnings. The insurance records allow the differentiation of insurance coverage resulting from employment, unemployment and serving an apprenticeship. A serious limitation is that the data contain no information about certain interesting aspects, such as working time (part or full time), family circumstances, qualification and occupation. Therefore, we have to use indicators to define the three groups. We use the duration of the apprenticeship episodes as an indicator for finishing the training. The other groups are defined according to the date of the first employment episode. Note that working during the summer break is not defined as a first employment episode, as we exclude episodes which occur only in February or in the period of July-September, respectively. The information available in the database does not allow us to control for work disabilities or to investigate the question of endogeneity of the choice between different training options. These limitations need to be considered in the interpretation of our analysis. The duration of an apprenticeship varies, depending on occupation, between two and four years. We assume that apprentices whose training episodes

amount to two years or more have finished training and have been awarded an apprenticeship diploma[4]. We define workers with a first employment episode between the age of 17 and 21 as high-school graduates. The consideration behind this definition is that 17 is the earliest age at which one can graduate from a vocational school, which takes three years. On the other hand, the earliest age at which one can receive a university degree is 22[5]. The group of unskilled workers in our definition consists of persons who have not received any further education. This group includes people who had their first employment episode at an age younger than 17 and those whose apprenticeship episodes add up to less than two years. Our analysis concerns the birth cohort of 1970. Because according to our definition of high-school graduates the observed persons started their working life at the latest in 1992, there are six years (1993-1998) during which their occupational career is observed. Table I shows the distribution of the workers, which is generally in line with the data from other sources. The share of women with an apprenticeship diploma corresponds to official figures (i.e. roughly 33 per cent). In line with official statistics the female share of high-school graduates amounts to slightly more than 50 per cent. The very high share of unskilled female workers reflects partly the lower training participation of women in Austria and partly a slight underestimation of unskilled male workers in our data set.

Labour market effects of apprenticeship 111

3.2 Occupational career Compared with other OECD countries, youth unemployment in Austria is low, and there is virtually no doubt that the apprenticeship system is one of the major reasons for this. It is remarkable, however, that unemployment among those aged 20-24 is considerably higher than among those aged 15-19. Therefore, the question arises whether apprenticeship simply shifts the youth unemployment problem to an older age group rather than, as is generally assumed, helping young people to enter a stable career[6]. With the data in hand, the occupational career (i.e. the employment and unemployment episodes as well as those spent out of the labour force) of the persons in the sample can be observed. This enables us to shed some light on the question. Here, we analyse the occupational careers of the birth cohort of 1970 between 1993 and 1995 and between 1996 and 1998, respectively. These two three-year

Ex-apprentices High-school graduates Unskilled workers Total

Both sexes

Female (per cent)

813 580 231 1,624

37.0 56.7 78.4 49.9

Table I. Analysed persons (birth cohort of 1970)

IJM 25,1

112

periods were chosen because, with our definition of high-school graduates, these people would have commenced their working life in 1992 at the latest and because the use of two time periods allows us to observe changes over time. Table II shows some indicators for employment stability in the two three-year periods. People with a secondary school education were less often unemployed than people with an apprenticeship diploma and unskilled workers in both periods and for both genders. Ex-apprentices were less often unemployed than unskilled workers only in the second period. However, by differentiating between men and women, only female ex-apprentices were worse off in this context. Moreover, to be unemployed (i.e. for at least one day) in a three-year period is not necessarily a bad sign, as it may indicate a voluntary change of employer. To be unemployed long-term can hardly be interpreted positively. Long-term unemployment is defined as being unemployed continuously for six months or longer. This indicator clearly shows more favourable values for ex-apprentices than for unskilled workers, and the bad performance gets worse over time, at least for women. Again, high-school graduates were better off than the other two groups, although this seems to apply mainly to women. In the first period, male high-school graduates were more often long-term unemployed than male ex-apprentices were, while in the second period there was only a small difference in favour of high-school graduates. To be employed throughout the whole three-year period can be interpreted as a positive indicator of a stable occupational career. High-school graduates and ex-apprentices were more often continuously employed than unskilled workers were. To interpret the quite low value for male high-school graduates in the first period, it should be taken into consideration that the probability that they had just finished their education, and therefore had completed their military service, is higher for them than for ex-apprentices. Correspondingly, in the second period, male ex-apprentices and male high-school graduates were continuously employed nearly equally often. Female high-school graduates in both periods had a better employment performance than female ex-apprentices. Being out of the labour force cannot unambiguously be interpreted as negative: the reasons can lie in labour market conditions as well as in personal life planning. The most outstanding feature in this context is that women were more often out of the labour force than men were, which is related to career interruptions for raising children. Altogether the following ranking appears: high-school graduates, ex-apprentices, and unskilled workers. However, differences between ex-apprentices and unskilled workers are more pronounced than those between ex-apprentices and high-school graduates. Moreover, women seem to benefit more from secondary school education, whereas men seem to profit more from an apprenticeship. As mentioned, we are not able to control for the endogeneity of educational choice. Therefore, our results may be biased due to

Total Suffer unemployment Suffer long-term unemployment Continuously employed Out of labour force Women Suffer unemployment Suffer long-term unemployment Continuously employed Out of labour force Men Suffer unemployment Suffer long-term unemployment Continuously employed Out of labour force 32.4 4.5 31.9 6.0 22.8 2.4 34.7 7.0 45.0 7.2 28.3 4.8

35.9 6.0 27.2 6.6 46.7 5.1 35.6 2.0

1993-1995 High-school (per cent)

42.7 5.4 32.5 3.7

Ex-apprentices (per cent)

62.0 12.0 26.0 6.0

33.1 7.2 27.6 11.1

39.4 8.2 27.3 10.0

Unskilled workers (per cent)

39.6 5.3 42.6 2.7

33.9 9.0 26.6 8.6

37.5 6.6 36.7 4.9

Ex-apprentices (per cent)

29.5 5.2 42.2 6.0

21.0 6.1 35.6 11.3

24.7 5.7 38.5 9.0

1996-1998 High-school (per cent)

46.0 12.0 34.0 6.0

42.5 18.2 22.7 16.0

43.3 16.9 25.1 13.9

Unskilled workers (per cent)

Labour market effects of apprenticeship 113

Table II. Indicators for stability of occupational career

IJM 25,1

114

unobserved heterogeneity. In the following, we control for observed heterogeneity (work experience, industry, etc.). However, the selectivity issue must be considered in the interpretation of our results. The advantage of comparing the three groups for fixed periods is that there is no need to take cyclical fluctuations into account. However, the very different durations of participating in the labour market of the three groups, and therefore in working experience, may influence the comparison. The unfavourable outcomes for male high-school graduates compared with male ex-apprentices, especially in the first three-year period, may be explained at least in part by this drawback. To highlight the influence of the various factors on the occupational career, we estimate two regressions for the periods 1993-1995 and 1996-1998, respectively. The two variables we want to explain are “personal unemployment rate” and the “incidence of long-term unemployment”. The first variable is defined as the number of days a person is unemployed over the number of days the person is in the labour force (days employed and unemployed) throughout the three-year period. The second variable takes on the value of 1 if the person is unemployed continuously for si9x months or longer within the three-year period, and 0 if this is not the case. The explanatory variables are gender, education, past employment and unemployment experience, and the industry in which the person is employed longest during the three-year period. To explain the personal unemployment rate we estimate the following specification of a Tobit model[7]: pur ¼ b0 þ b1 female þ b2 hs-graduates þ b3 unskilled þ b4 past employment experience þ b5 past unemployment experience þ b6 primary sector þ b7 construction þ b8 trade þ b9 hotels and restaurants þ b10 other services þ e The coefficient of the dummy variable “female” shows the differential in the personal unemployment rate of women relative to men. b2 and b3 show the effect of secondary school education and compulsory education, respectively, relative to serving an apprenticeship. The variables “past employment” and “unemployment experience” are defined as the number of days a person was employed and unemployed, respectively, before the three-year period (i.e. in 1984-1992 and 1984-1995, respectively) and show the influence of past labour market experiences on the personal unemployment rate. The industry dummies control the industry effects relative to working in the manufacturing sector.

Past employment and unemployment experiences show the expected signs (Tables III and IV). Even if we control these two factors, education had a significant influence on the personal unemployment rate. Persons with secondary school education faced a lower personal unemployment rate than persons with an apprenticeship diploma did, while the personal unemployment rate for unskilled workers was higher than that for ex-apprentices. Remarkably, the positive effect of school education compared to an apprenticeship was almost the same in both periods, whereas the negative effect of no further education got stronger over the years[8]. In the first period, no significant gender difference can be established. The second period shows Variable Constant Female High-school graduates Unskilled Past employment experience Past unemployment experience Primary sector Construction Trade Hotels and restaurants Other services Number of observations Log likelihood Pseudo R2

Coefficient 1993-1995

SE

Coefficient 1996-1998

SE

10.66* 2 2.76 2 10.01* 10.12* 2 2.07* 6.28* 49.31* 5.06 2 5.70 13.34* 2 7.21* 1,536 2 3,372 0.077

4.03 2.26 2.52 3.53 0.26 0.57 4.79 3.68 3.30 3.38 2.77

23.28 7.69* 210.02* 17.62* 21.38* 5.10* 65.23* 17.45* 24.51 14.19* 29.27** 1500 23033 0.084

6.34 2.92 3.26 4.23 0.23 0.51 5.78 4.89 4.48 5.39 3.87

Coefficient 1993-1995

SE

Coefficient 1996-1998

SE

2 1.33* 2 0.16 2 0.12 0.20 2 0.03* 0.13* 1,624 2 310.25 0.10

0.17 0.12 0.14 0.18 0.01 0.03

21.63* 0.28* 20.08 0.35* 0.01** 0.10* 1,624 2 387.52 0.13

0.17 0.10 0.12 0.14 0.01 0.02

Labour market effects of apprenticeship 115

Table III. Personal unemployment Notes: Tobit estimates, past employment and unemployment experience are multiplied by 100 rate, 1993-1995 and * significant at 1 per cent level, **significant at 5 per cent level 1996-1998

Variable Constant Female High-school graduates Unskilled Past employment experience Past unemployment experience Number of observations Log likelihood Pseudo R 2

Notes: Probit estimates, past employment and unemployment experience are multiplied by 100; * significant at 1 per cent level, **significant at 5 per cent level

Table IV. Incidence of long-term unemployment, 1993-1995 and 1996-1998

IJM 25,1

116

that women had a significant disadvantage. Compared to workers in manufacturing, those employed in the hotel and restaurant sector had a higher personal unemployment rate and those in the other services sector had a lower personal unemployment rate. To show the influence of the various factors on the incidence of long-term unemployment, we estimate the following probit model[9]: ltu ¼ b0 þ b1 female þ b2 hs-graduates þ b3 unskilled þ b4 past employment experience þ b5 past unemployment experience þ e The explanatory variables have the meaning described above. The industry dummies are insignificant both separately and commonly and are therefore omitted. Table IV shows the results of the estimates for the two three-year periods. Again, past employment and unemployment experiences have the expected signs. In the first three-year period, all other variables are insignificant (except the constant), while in the second period the missing education of unskilled workers seems to have increased the risk of being long-term unemployed. In both periods, there is again no significant difference in the probability of being long-term unemployed between ex-apprentices and high-school graduates, while being a woman seems to have had a negative influence in the latter period. Summing up, our results suggest that an apprenticeship diploma helps young people to enter a more stable career. Controlling the labour market experiences (and other possible influencing factors) seems to make no difference in the ranking given above (high-school graduates, ex-apprentices, unskilled workers). The more pronounced difference between ex-apprentices and unskilled workers than between ex-apprentices and high-school graduates is confirmed by the regression results. With respect to the personal unemployment rate, we found a statistical significant difference between high-school graduates and ex-apprentices in both periods. Moreover, the significant negative coefficient for the unskilled worker increased over time. Regarding long-term unemployment, the results are somewhat different. In the first period, the coefficients for high-school graduates (unskilled workers) are positive (negative) but not statistically significant. Only in the second period, we do find a significant higher risk of long-term unemployment for the unskilled worker. Therefore, there seems to be a clear evidence that the relative unemployment risk of unskilled workers increases over time. 3.3 Earnings So far our results suggest that school education and serving an apprenticeship improve employment stability. The next question is whether having served an

apprenticeship enhances earnings, as suggested by human capital theory. Here, we analyse the earnings data of the persons for 1992 (i.e. at the age of 22) and 1998 (i.e. at the age of 28). Monthly median earnings[10] serve as a first descriptive statistic. Table V shows the monthly median earnings (in 1998 prices) of the three groups for the years 1992 and 1998[11]. A striking feature is that women earned considerably less than men. However, it has to be mentioned in this context that the data contain no information about working time. Therefore, the eye-catching earnings differences can be explained at least partly by the fact that women, especially women aged 20-30 (who often have to take care of children), work part-time more frequently than men do. In 1998, when the subjects were 28 years old, those with a secondary school education earned slightly more than those with an apprenticeship diploma, whereas unskilled workers earned about 20 per cent less than the ex-apprentices. In 1992, at the age of 22, the difference in earnings between unskilled workers and the other groups was less pronounced, and high-school graduates earned even less than ex-apprentices. However, this may possibly be explained by different labour market experiences of the three groups. In the descriptive analyses, we focus on differences differentiated by gender because of the gender wage differential, which could influence the results due to the different sex composition of our skill groups. Women with a secondary school education earned noticeably more than their contemporaries, and this effect became stronger over time. In 1998, the median earnings of female high-school graduates was 23 per cent higher than that of female ex-apprentices and female unskilled workers. In this view, serving an apprenticeship does not help women to earn a higher income. In 1992, unskilled workers earned only slightly less, and in 1998, the median earnings of female ex-apprentices and female unskilled workers were almost equal. 1992 Earnings in e* Total Ex-apprentices High-school Unskilled Women Ex-apprentices High-school Unskilled Men Ex-apprentices High-school Unskilled Notes: *In 1998 prices;

**

Per cent**

1998 Earnings in e*

117

Per cent**

1,384 1,351 1,212

100.0 97.6 87.6

1,788 1,824 1,432

100.0 102.0 80.1

1,146 1,351 1,122

100.0 117.9 97.9

1,337 1,650 1,337

100.0 123.4 100.0

1,474 1,392 1,433

100.0 94.4 97.2

1,969 2,006 1,824

100.0 101.8 92.6

share of the respective earnings of ex-apprentices

Labour market effects of apprenticeship

Table V. Monthly median earnings, 1992 and 1998

IJM 25,1

118

Contrary to women, men seem to benefit from an apprenticeship diploma. The median earnings of men with an apprenticeship diploma were higher in both years than those of unskilled workers. Men with secondary school education did not earn more than the men with an apprenticeship diploma. In 1992, they even earned less, but this again may be due to the missing consideration of working experiences. To get a better image of the importance of education and other possible influencing factors on earnings, we estimate an earnings equation of the following form for each of the two years: lninc ¼ b0 þ b1 female þ b2 high-school þ b3 unskilled þ b4 tenure þ b5 tenure squared þ b6 experience þ b7 experience squared þ b8 blue-collar worker þ b9 regional unemployment rate þ b10 firm size 50-299 þ b11 firm size . 299 þ b12 firm size unknown þ b13 primary sector þ b14 construction þ b15 trade þ b16 hotels and restaurants þ b17 other services þ b18 sector unknown þ e Note that in 1992, there were no workers with unknown sector. Furthermore, we follow our strategy from the unemployment equations and do not consider gender differences with respect to the effect of skills. The dependent variable lninc is the logarithmic monthly earnings used for calculating the median earnings of the three groups. In addition to the variables used in the unemployment equation, we control for tenure and experience. Tenure measures the duration (in years) of the employment episode to which the analysed earnings refer. A positive sign of b4 and a negative sign of b5 would confirm the assumption of a concave earnings curve with regard to the duration of the employment episode. The influence of working experience is quantified with the variables “experience” and “squared experience”. The variable is defined as the number of days the person was employed before the actual employment episode. To control the occupation and differences in the wage bargaining process, we differentiate between blue-collar and white-collar workers. We use the regional, gender-specific average unemployment rate of the year in question to take the regional labour market situation into consideration. We include dummy variables to control the firm size effects and industry dummies. Reference categories are small firms (, 50 employees) and the manufacturing industry (Table VI).

Variable

Coeff.

t-value

Coeff.

1992 Constant Female High-school graduates Unskilled Tenure Tenure squared Experience Experience squared Blue-collar worker Regional unemployment rate Firm size 50-299 Firm size .300 Firm size unknown Primary sector Construction Trade Hotels and restaurants Other services Industry unknown Number of observations R-squared Notes: OLS regression, *significant

2.781* 20.236* 20.005 20.121* 20.023 0.003 0.064* 20.003* 20.096* 20.019* 0.096* 0.185* 0.035 20.042 0.086* 20.116* 20.083 20.079*

t-value 1998

45.5 10.0 0.26 4.09 1.17 0.80 5.49 3.83 4.49 2.86 4.04 8.09 1.45 0.67 3.67 5.29 1.95 3.55

2.874* 2 0.367* 0.116* 2 0.100** 0.035* 2 0.003** 0.062 2 0.000 2 0.194* 2 0.004 0.080* 0.150* 2 0.116 0.104 0.051 2 0.191* 2 0.067 2 0.125* 2 0.156 1,160

19.31 13.5 4.68 2.36 2.87 2.20 1.70 0.04 7.30 0.63 3.09 5.56 1.67 1.44 1.54 5.94 1.20 4.62 0.29

1,192 0.29 0.40 at 1 per cent level, **significant at 5 per cent level

According to our estimates, unskilled workers in both years earned significantly less than workers with an apprenticeship diploma. The variable “high-school” is significant in the regression for 1998; that is, if we control for other possible influences, secondary school education seems to have led to higher earnings, at least for some years after joining the workforce. In general, the various control variables show the expected signs. Women earned significantly less than men did, as mentioned, this may to a great extent be explained by the fact that women work part-time more frequently than men do. In 1992, the variables for working experience and in 1998, the variables for job duration are significant. This seem to imply that the importance of past work experience diminishes as the length of the current relationship increases. As expected, blue-collar workers earned less than white-collar workers. The increasing differential between blue-collar and white-collar workers over time reflects the better promotion chances of white-collar workers. The regional unemployment rate is only significant in 1992, but shows the expected sign. Workers in medium-sized firms earned more than workers in small firms, and workers in large firms earned even more. In the trade and other services sectors, earnings were significantly lower, which can be explained partly by the high share of part-time workers in these sectors.

Labour market effects of apprenticeship 119

Table VI. Monthly earnings, 1992 and 1998

IJM 25,1

120

Summing up, the analyses of earnings show results quite similar to the analyses of the occupational career (i.e. the ranking high-school graduates, ex-apprentices, unskilled workers), but with a more pronounced difference between unskilled workers and ex-apprentices. At least the descriptive evidence indicates that women seem to benefit a lot from secondary school education, but hardly from an apprenticeship diploma. This is an interesting point for further research. 4. Summary and conclusion The Austrian apprenticeship system is intended to help even people who failed at school to enter a well-founded vocational training and thereby to stable occupational career. One of the main findings of this paper is that workers who have served an apprenticeship are much better off than workers who have completed only their compulsory education. We established the following ranking: high-school graduates, ex-apprentices and unskilled workers, with more pronounced differences between ex-apprentices and unskilled workers. At least descriptive evidence seems to indicate that women benefit more from secondary school education, while having served an apprenticeship seems to help them to enter a more stable occupational career but not to earn a higher income compared to those with no further training. Although we are not able to control for selectivity of educational choice, our results thus seem to confirm the positive effects on the occupational career of serving an apprenticeship. This again demonstrates that the apprenticeship system is not an antiquated model and that all efforts to maintain the system – and particularly the quality of the system – seem to be justified. However, the relatively weaker effects of apprenticeship training for women can be interpreted as a result of the comparatively weaker quality of the apprenticeship training due to the strong concentration in a few traditional trades in the service sector. Therefore, the campaigns to encourage particularly female apprenticeship applicants to take up trades which are not so well known or are non-traditional for women are useful. In general, further attempts to create new apprenticeship occupations in the expanding high-skill service sectors are recommended as they could provide appropriate vocational training opportunities for young people. Notes 1. Peraita (2001) tested the Acemoglu-Pischke model for Spain and found a poor effect of highly compressed wage structures on firm-sponsored general training. 2. Sto¨ger and Winter-Ebmer (2001) found that structural changes are not able to fully explain the decrease in willingness to offer apprenticeship positions. 3. For example, an episode of gainful employment, unemployment, apprenticeship or self-employment. 4. This indicator was first used and discussed in detail by Brandel et al. (1994).

5. The best group to be compared with people with an apprenticeship diploma would be those who have completed vocational school, but because we have no information on qualification we cannot exactly identify this group. Our definition therefore includes people who have completed the more general education oriented gymnasium. 6. If this is true only in part, an additional concern arises: do some firms misuse apprentices as low-paid workers instead of training them? 7. The Tobit estimate is used because the dependent variable possesses an upper and a lower limit, that is, it can take on only values between 0 and 100. For detailed information, see for example Johnston and DiNardo (1997). 8. The terms “positive” and “negative” are used here in a logical manner. The sign of “high-school” is negative and that of “unskilled” is positive, but a lower unemployment rate (negative sign) is a positive sign for a stable occupational career. 9. The probit estimation is used because the dependent variable can take on only the values 0 and 1. For detailed information, see for example Johnston and DiNardo (1997). 10. The earnings data refer to the monthly average base for paying social security contributions. As the data contain no working time information, it is impossible to compute wage rates. 11. Only workers with positive earnings in the particular year are considered, as we concentrate on earning effects. References Acemoglu, D. and Pischke, J.S. (1999), “Beyond Becker: training in imperfect labour markets”, The Economic Journal, Vol. 109 No. 453, pp. F112-42. Becker, G.S. (1964), Human Capital, The University of Chicago Press, Chicago, IL. BMWA (2003), Apprenticeship: Vocational Education and Training in Austria, Federal Ministry for Economic Affairs and Labour, Vienna. Brandel, F., Hofer, H., Lassnigg, L. and Pichelmann, K. (1994), Aspekte der Arbeitsmarktintegration von Lehranfa¨ngern, Institut fu¨r Ho¨here Studien, Vienna. Euwals, R. and Winkelmann, R. (2001), Why Do Firms Train? Empirical Evidence on the First Labour Market Outcomes of Graduated Apprentices, IZA Discussion Paper No. 319, Bonn. Franz, W., Inkmann, J., Pohlmeier, W. and Zimmermann, V. (2000), “Young and out in Germany: on the youths’ chances of labour market entrance in Germany”, in Blanchflower, D. and Freemann, R. (Eds), Youth Unemployment and Joblessness in Advanced Countries, The University of Chicago Press, Chicago, IL, pp. 381-426. Harhoff, D. and Kane, T.J. (1997), “Is the German apprenticeship system a panacea for the US labour market?”, Journal of Population Economics, Vol. 10 No. 2, pp. 171-96. Hofer, H. and Pichelmann, K. (1994), “Zur Arbeitsmarktintegration von Lehrlingen in O¨sterreich”, in Steiner, V. and Bellmann, L. (Eds), Mikroo¨konomik des Arbeitsmarktes, BeitrAB 192, Institut fu¨r Arbeitsmarkt- und Bildungsforschung der Bundesanstalt fu¨r Arbeit, Nu¨rnberg, pp. 327-52. Johnston, J. and DiNardo, J. (1997), Econometric Methods, McGraw-Hill, New York, NY. Lassnigg, L. (2001), “The learning-oriented company and policy perspectives for VET and HRD”, in Nieuwenhuis, L.F.M. and Nijhof, W.J. (Eds), The Dynamics of VET and HRD Systems, Twente University Press, Enschede, pp. 35-57. Lassnigg, L. and Steiner, P. (1997), “Financial aspects of apprenticeship in Austria: results of an empirical study”, paper presented at the European Conference on Education Research (ECER97), Frankfurt/Main.

Labour market effects of apprenticeship 121

IJM 25,1

122

Lietz,

C. (2001) “Auswirkungen der Lehrausbildung auf Berufseinstieg und Arbeitsmarktkarriere”, unpublished diploma thesis, University of Vienna. Peraita, C. (2001), “Testing the Acemoglu-Pischke model in Spain”, Economics Letters, Vol. 72 No. 1, pp. 107-15. Ryan, P. (2001), “The school-to-work transition: a cross-national perspective”, Journal of Economic Literature, Vol. 39 No. 1, pp. 34-92. Steedman, H. (2001), Benchmarking Apprenticeship: UK and Continental Europe Compared, LSE, Centre for Economic Performance Discussion Paper 513, London. ¨ sterreich: Welche Betriebe bilden Sto¨ger, K. and Winter-Ebmer, R. (2001), Lehrlingsausbildung in O Lehrlinge aus?, Institut fu¨r Volkswirtschaftslehre der Johannes Kepler Universita¨t Linz, Arbeitspapier Nr 0110, Linz. Further reading OECD (1999), “Preparing youth for the 21st century: the transition from education to the labour market”, Proceedings of the Washington D.C. Conference, OECD, Paris. OECD (2000), From Initial Education to Working Life: Making Transitions Work, OECD, Paris.

The Emerald Research Register for this journal is available at www.emeraldinsight.com/researchregister

The current issue and full text archive of this journal is available at www.emeraldinsight.com/0143-7720.htm

Human capital spillovers in the workplace: evidence for the service sector in Britain

Human capital spillovers in the workplace 123

H. Battu Department of Economics, University of Aberdeen, Aberdeen, UK

C.R. Belfield National Center for the Study of Privatization in Education, Teachers College, Columbia University, New York, USA, and

P.J. Sloane WELMERC, Department of Economics, University of Wales Swansea, Wales, UK and IZA, Bonn Keywords Human capital, Workplace, Performance criteria Abstract An individual’s human capital has a strong influence in earnings. Yet, individual worker-level estimations of earnings rarely include the characteristics of co-workers or detailed firm-level controls. In particular, co-workers skills are ignored which may be particularly significant where team work is important. This paper utilises a unique matched work-place data set to estimate the effect on the earnings of co-workers’ education and training in the Hotel and Catering sector, which contains a high proportion of low paying establishments and in the Retail sector which contains a large absolute number of low paying establishments. The data are derived from the 1998 British Workplace Employment Relations Survey. This is a national sample based on interviews with managers in 2,191 establishments with at least ten workers. In addition, a survey of up to 25 randomly selected employees in each establishment was undertaken which included questions on education, training, pay and job satisfaction, as well as a range of other personal and workplace characteristics. We have, therefore, a matched workplace employee sample which is essential for this type of analysis. The results suggest that there are strong co-worker effects in the earnings of individuals when controlling the individual’s own level of education. While there are also high returns to training for individual workers, there are no similar spillover effects from the training of co-workers in these sectors. Nevertheless, this suggests that there could well be a pay-off to the professionalisation of service sector jobs.

1. Introduction The highly significant influence of education on earnings is one of the most commonly tested relationships in economics (Lazear, 2000). There is also a sizeable literature on education’s influence on social activities through externalities, i.e. spillover benefits from educated individuals to others (Wolfe The authors acknowledge the Department of Trade and Industry, the Economic and Social Research Council, the Advisory, Conciliation and Arbitration Service and the Policy Studies Institute as the originators of the 1998 Workplace Employee Relations Survey data, and the Data Archive at the University of Essex as the distributor of the data. None of these organisations bears any responsibility for the authors’ analysis and interpretations of the data.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 123-138 q Emerald Group Publishing Limited 0143-7720 DOI 10.1108/01437720410525036

IJM 25,1

124

and Zuvekas, 2000). However, there is surprisingly little evidence on whether these spillover effects occur within the workplace, and what effect these might have on own earnings. In particular, there is no clear indication as to whether educated co-workers raise or lower own earnings. Co-workers’ human capital may positively influence own wages through several routes. Idson and Kahane (2000) refer to a “team dynamic”, but equally strong effects may emerge from information-sharing, from skill complementarity and from training by co-workers (Barron et al., 1997), or from more “rational” behaviour (Behrman and Stacey, 1997). However, co-workers’ human capital need not be complementary to own human capital, for two reasons. First, if co-workers have different amounts of human capital, then there may be a “skills incompatibility” (Kremer, 1993): a firm with a uniform standard of education may have higher productivity. Second, if workers are in competition with each other for high-paying jobs within the firm, they may engage in activities to undermine their co-workers and promote themselves (Winter-Ebmer and Zweimuller, 1997)[1]. Both the levels and distribution of workplace human capital are likely to influence earnings, and the directions of these influences can only be assessed empirically. In an earlier paper by Battu et al. (2001), these propositions were tested using random samples of workers from British workplaces as a whole. This paper is different in two ways. First, it limits the analysis to two service sectors (though results for the complete Workplace Employee Relations Survey (WERS), but excluding the two service sectors, are provided in Appendix Table AI for comparison purposes). Second, it incorporates training into the analysis of workplace effects. In particular, it focuses on the extent to which education and training can lead to better pay for low-skilled workers and perhaps improve the quality of these services as a consequence of the professionalisation of jobs. This paper is structured as follows. In Section 2, the possible externalities from co-worker education and training are discussed, along with pertinent (but mainly indirect) evidence; this discussion allows for the formulation of testable hypotheses about the effects on earnings. In Section 3, the dataset – the 1998 British WERS – is described. In Section 4, the effects of workplace education on own earnings are tested. In Section 5, summary and conclusion is offered. 2. Theory and evidence on human capital spillovers 2.1 Returns to human capital within the workplace The literature on the rate of return from education is vast (see recent reviews by Ashenfelter et al., 2000; Cohn and Addison, 1998). Broadly, this cross-sectional literature indicates that in Western economies the wage premium for each additional year of education is approximately 5-10 per cent (for the UK, see Blundell et al., 2000, and for various European countries Harmon et al., 2001). Moreover, this premium is not substantially altered when the endogenous

decision to become educated is modelled (e.g. from twin studies or natural experiments, Harmon and Walker, 1995; Miller et al., 1997). The impact of own education on own productivity and earnings is strong and robust. In addition, workers’ productivity may also depend on the education and training of co-workers within the firm. Workers do not always work autonomously. Many tasks require group work, with skills diffused through teams and across the workplace; and organisational cultures (e.g. those associated with high-performance management) may depend on the average workforce skill level. Furthermore, firms may deliberately cultivate a “team dynamic”, with information-sharing, co-training, monitoring, and support so as to exploit these spillovers. Idson (1995), for example, looks at how earnings are correlated with production in teams, as well as with behavioural traits such as team size, encouragement and helpfulness. The possibility of human capital spillovers gains support from the substantial literature on the externalities from education (Wolfe and Zuvekas, 2000). Enterprises may even be able to generate more externalities than exist in societies, insofar as they can enforce tighter employment contracts across workers. The possibility of positive spillovers is at least indirectly suggested by evidence such as the positive clustering of high-skilled professionals with high-skilled non-professionals in the US (Bronars and Famulari, 1997), the positive earnings effect from increasing proportions of skilled workers in a firm (Troske, 1999), and from sports performances, where team dynamics clearly occur (Idson and Kahane, 2000)[2]. In addition, there is a positive correlation between all individuals’ earnings and the average education level of a region (Rauch, 1993), and a “brain drain” of educated workers to areas where there are other educated workers (Borjas et al., 1992)[3]. Human capital spillover effects may therefore exist: working with others who have high education levels may independently raise own earnings (and this may explain some of the observed firm-level heterogeneity in earnings, see Abowd et al. (1999) and Bayard and Troske (1999). However, own earnings may be affected by the dispersion of human capital within the workplace. In some respects, this line of argument runs counter to the notion of human capital spillovers. In Kremer’s (1993) o-ring theory, for example, the productivity of high-skilled workers is increasing in the skill levels of co-workers. However, the o-ring theory also predicts that an important determinant of factor payments is the compatibility of standards: where workers are of a “compatible standard”, they will earn more. Perhaps training programmes are easier to implement (Barron et al., 1997; Van Smoorenberg and van der Velden, 2000), or there are fewer co-ordination failures in standardised workplaces. From this theory, an increase in education levels within the workplace may not raise earnings if it also serves to widen the dispersion of education levels, but earnings will be raised where high-skilled workers cluster together.

Human capital spillovers in the workplace 125

IJM 25,1

126

Furthermore, many workers do not have the optimal amount of education for their jobs (see the review in Oosterbeek, 2000). For the UK, over-education is significant and substantial: years of surplus education only weakly affect individual earnings, and they negatively affect job satisfaction (Battu et al., 1999; Dolton and Vignoles, 2000). Thus, if workers have accumulated more education than is optimal, any externalities may be subverted: an over-educated worker, with lower job satisfaction, may distract or demoralise other workers. Monitoring costs for the firm may be pushed up, to avoid workers who adroitly perform the set tasks then distracting co-workers. Thus, in general, co-workers’ years of education may raise earnings, however, where co-workers have excess or these are surplus years of education they may tower own earnings. Tsang (1987), with data from the Bell Company, found that over-education reduces firm output via a negative effect on job satisfaction. Other indirect evidence of the effect of the human capital mix on firm performance includes the negative effect of proportions of unskilled or manual workers (Addison and Belfield, 2001)[4]. A third reason to doubt a positive effect from co-worker human capital is the possibility of intra-firm job tournaments. Where workers with equivalent skills compete for promotions, they may sabotage each other, and so reduce productivity overall. Finally, spillover effects may not be appropriated by workers, but instead by managers in terms of higher profits. This last effect seems unlikely though, since highly educated workers would probably be the most effective at bargaining for higher shares of the workplace surplus. In the case of training, it has been observed that there can be potential negative effects for groups of non-trainees (Pfeiffer, 2000). Further, different types of skill can be substitutes or complements, depending on the nature of labour market institutions and technology. Thus, Machin and van Reenen (1998) found that skilled workers with higher secondary education are substitutes rather than complements for skilled workers with higher education. If, as a simplifying assumption, we regard education as equivalent to general human capital and training offered by the employer as equivalent to specific human capital then in the former case, given that individuals will themselves bear the costs of education, we should observe a close relationship between the level of education and earnings. In the latter case, the relationship between the provision of training and earnings should be less apparent as firms themselves will bear the costs of specific skill acquisition and hope to obtain a return through higher productivity. Where the potential mobility of specifically trained workers is high, there may, however, be some sharing of the returns to such investment in human capital. Thus, there are plausible arguments on both sides, regarding the complementarity or substitutability of workers’ human capital within a

workplace. Ultimately, the net effects can only be decided empirically. But, the empirical investigation has broader implications. If human capital spillover effects are important, and there is clustering of education levels within workplaces, then the earnings premium to education should be reinterpreted. The earnings premium is typically attributed to human capital and not to labour market sorting of workers according to their credentials (Belfield, 2000). Yet, part of the premium may be a consequence of educated workers being hired to workplaces where the average education level is high. These workers then “share” their human capital and so have higher earnings. Education still enhances productivity and so earnings, but part of that enhancement comes from education’s role in securing for a given worker a job which allows for interaction with other skilled workers. 2.2 Model specification The above arguments can be modelled formally An appropriate specification of the relationship between own earnings and co-worker attributes is laid out by Idson and Kahane (2000): ln yij ¼ a1 þ a2 eij þ a3 E j þ a4 eij *E;j þa5 tij þ a6 T j þ a7 tij *T j þ a8 zij þ a9 Z j þ vj þ ui

ð1Þ

In equation (1), own individual earnings yij are determined by: the education eij of individual i at workplace j, the education levels of co-workers Ej, and the interaction between these two education levels. Similar relationships apply to training as indicated by the return to coefficients a5, a6 and a7. A vector of worker and workplace controls zij and Zj are also included (vj,N(0, sj) and ui,N(0, si) are independent and identically distributed (iid) workplace and individual error terms. Under this specification, an additional year of an individual worker’s own education affects their earnings by a2+a4Ej. The coefficient a2 captures the direct effect of years of education, and the coefficient a4 captures the effect of average co-worker education on how own education is valued. An additional cross-workplace increase in education of one year will increase own earnings by a3+a4eij. Co-worker education will impact directly through the coefficient estimated as a3, and indirectly through the interaction coefficient a4. Here, if a3 is non-zero, then its omission (or that of a4) serves to bias upwards a2, the standard measure of the education premium. Based on the above discussion, the expectation is that a3 and a4 will be positive, although as noted above only limited evidence is available on each of these coefficients. Where a3 is positive, own earnings are positively related to co-workers’ education; where a4 is positive, increased years of co-worker education raise wages for those with high education levels. Similar relationships hold for training.

Human capital spillovers in the workplace 127

IJM 25,1

128

The sign of coefficient a4 offers one test of Kremer’s hypothesis about compatible standards of inputs. If a4 is positive, then own and co-worker human capital are complementary. Further tests are also possible. One simple test is to include the absolute mean dispersion of education levels Ej on the right hand side of equation (1). Greater dispersion should reduce own earnings. To capture non-linear effects, the square of workplace human capital E 2j is inserted; the coefficient for this parameter should be positive. In summary, the following hypotheses are offered. First, workplace education levels raise own earnings. Second, the dispersion of workplace education levels lowers earnings (“skills incompatibility”). Third, co-worker human capital effects are stronger in high-skilled workplaces (“increasing returns to human capital” or complementarity). Similar hypotheses hold for training. It is possible to test each of these hypotheses using matched worker-workplace data. 3. Data and measures To test the hypotheses, the dataset used is the British Workplace Employee Relations (WERS) survey, collected in 1998. The WERS is a national sample of interviews with managers from 2,191 establishments with at least ten workers. The firm-level survey addresses the “management of employees”, with information on workforce composition and workplace performance. In addition, 25 employees from each workplace were randomly selected for individual survey. This survey asked questions about individuals’ education, pay and job satisfaction, as well as a range of personal characteristics. The information set is therefore rich, with detailed information on multiple workers per workplace. For estimation, the sample here is restricted to full-time workers[5] and to workplaces where more than three workers responded to the worker survey. Here we focus on workers in the Hotel and Catering Sector, which contains a high proportion of low-paying establishments and the Retail Sector containing a large absolute number of low-paying establishments. The derivation of the key variables is briefly described here: a full derivation is reported in Appendix Table AI, along with a catalogue of substitute derivations of the key variables. The simplest way to estimate these relationships is to use years of education as the unit of account. First, each worker’s full-time equivalent years of education was calculated, to obtain ei; these calculations were based on reported qualifications. (As qualifications may not translate readily into years of education, full sensitivity analysis is conducted below). Second, workplace education levels Ej were derived using both worker and workplace data. Based on the full worker sample, mean years of education per occupation are calculated. This mean can then be weighted for each workplace, using information on the occupational mix of the entire workforce at each workplace. (Two alternative measures of mean workplace education are available, and these are utilized in the sensitivity testing). Third,

the dispersion of workplace education levels Ej is also calculated: this dispersion measure is the average of absolute differences between own education (across the random sample of workers) and mean workplace education. Fourth, pay levels yij are taken from self-reports across 12 wage bands, and converted into earnings per hour using the reports of hours worked. Median pay across the workplace Yj is also available; this variable is based on the distribution of pay across the workforce, as reported by the manager. As far as training is concerned WERS 1998 asks workers how much training they have had during the last 12 months, either paid for or organized by the employer. Only training away from the normal place of work was to be included, though this could be on or off the premises. A range of answers from none to ten days or more was listed. In this paper, own training (tij) of more than one day in the last year (1) or otherwise (0) was the chosen variable. Workplace training was proxied by a variable (Tj) measuring the percentage of co-workers trained. Basic frequencies for the key variables are reported in Table I (with full frequencies for the other variables detailed in Appendix Table AI). The average years of education per male (female) worker are 12.88 (12.76)[6]. Mean education levels per workplace are 12.82 (s.d. 0.70), so the sample of respondents has slightly more education than the average of their workplace. The dispersion of education across a workplace is 1.88 (s.d. 0.50). For the dependent variables, log pay per hour per individual worker is 1.68; and log median annual earnings per workplace are 7.21. Of the sample 36.49 per cent have received training in excess of 1 day (s.d. 0.48) and the corresponding figure for workplace training is 35.90 per cent (s.d. 0.21).

Education and outcome variables Education variables Years of education per worker: male Years of education per worker: female Years of education per workplace Interaction own – workplace education Dispersion of education per workplace Own training (.1 day in last year) Percent co-workers trained Earnings variables Log pay per hour Log median wage per workplacea Number of workers Number of workplaces

Code

Mean

Standard deviation

eijm eijf Ej eij*Ej Ej tij Tj

12.88 12.76 12.82 164.90 1.88 0.36 0.36

2.51 2.40 0.70 35.85 0.50 0.48 0.21

1.68

0.47

yij Yj Ni Nj

2,461 261

– –

Notes: Unweighted data. See Appendix Table AI for definitions of variables, and for alternative derivations. aN j ¼ 870

Human capital spillovers in the workplace 129

Table I. Service sectors only – frequencies: education and outcome variables

IJM 25,1

130

Such matched worker – workplace data are ideal for testing the hypotheses listed above. There are detailed controls for each worker, workplace information from two sources (the manager and the worker respondents), and full information on education, pay, and job satisfaction. This allows for numerous sensitivity tests and cross-validation of the results. One potential caveat is that this analysis relates to workplaces rather than teams: co-worker, in this sense, refers to those in the same workplace, rather than those doing the same tasks or team-working. (No ability controls are available, either.) Other British datasets do allow for some workplace controls; but, the random sample of workers and the detailed information on both workers and workplaces in the WERS – essential for investigating these arguments – is unique for the British economy. One potential problem is that the cross-section data do not allow for the clear identification of firm effects and effects from working with people with higher or lower educational endowments. In order to identify firm effects and co-worker effects it would be preferable to have longitudinal data[7]. However, we do include a number of firm specific variables and restrict ourselves to two particular sectors where establishments are likely to be rather similar. 4. Estimation and results 4.1 The effects of workplace education on own earnings The main hypothesis is whether earnings are increasing in the education levels of co-workers. Table II reports a series of Mincerian earnings log pay per hour equations, estimated with both own and co-worker levels of education against log pay per hour. As per equation (1), which includes error terms for workplaces and individuals, Random Effects Generalized Least Squares is used[8]. Model (1) includes individual characteristics zi only (detailed in the notes in Table II); it shows that the earnings premium for an additional year of education in our two service sectors is 5.6 per cent (4.6 per cent) for males (females). This is lower than for the WERS as a whole, particularly in the case of females. Model (1) explains 38 per cent (25 per cent) of the variation in earnings; and the fraction of the variance attributable to the workplace error term rj is 20 per cent (25 per cent). The provision of training in the workplace (defined as greater than one day) significantly raises earnings, but the effect is much stronger for men (about 10 per cent compared to about 5 per cent). Model (2) introduces firm-level characteristics Z. The premium to education falls slightly, with an increase in the explained variation to 47 per cent (39 per cent); and the workplace error term variance falls for men but not for women, so rj is now 0.1441 (0.2669). The training effects on earnings increase slightly. There are then relatively few changes between Models (1) and (2) when Z is added. Model (3) includes the average years of education across the workplace Ej as an additional firm-level variable. This variable is statistically significant and

Male coeff (t. stat)

Female coeff (t. stat)

Model (1) a Firm-level characteristics No No Own years of education eij 0.0555*** (12.98) 0.0464***(10.34) Own training tij 0.0956*** (4.39) 0.0421** (2.05) rj 0.2012 0.3351 0.3776 0.2464 R 2 total Model (2) a,b Firm-level characteristics Yes Yes Own years of education eij 0.0511*** (12.01) 0.0412*** (9.27) Own training tij 0.1037*** (4.88) 0.0498*** (2.46) rj 0.1441 0.2669 R 2 total 0.4732 0.3890 Model (3) a,b Firm-level characteristics Yes Yes Own years of education eij 0.0474*** (11.19) 0.0398*** (9.01) Mean workplace years of education Ej 0.1699*** (7.63) 0.1250*** (4.66) Own training tij 0.0951*** (4.25) 0.0442** (2.13) Mean workplace training Tj 0.0294 (0.45) 0.0778 (1.13) rj 0.0939 0.2520 R 2 total 0.5116 0.4257 Model (4) a,b Firm-level characteristics Yes Yes Own years of education eij 2 0.0029 (2 0.04) 0.0223 (0.34) Mean workplace years of education Ej 0.1167 (1.57) 0.1069 (1.47) Interaction eij*Ej 0.0039 (0.75) 0.0013 (0.27) Own training tij 0.0954*** (4.24) 0.0438** (2.10) Mean workplace training Tj 0.0300 (0.47) 0.0793 (1.14) rj 0.0706 0.2524 R 2 total 0.5125 0.5123 Nj (Ni) 261 (1,302) 274 (1,141) Notes: Unweighted data. Significance: ***1 per cent level; **5 per cent level; *10 per cent level. Only full-time workers rj ¼ ðsj Þ2 =½ðsj Þ2 þ ðsi Þ2
aIncluded set of individual characteristics: tenure; tenure squared; age; age squared; ethnicity (1 dummy); disability (1); marital status (3); union member; temporary, fixed term or overtime worker (3); bincluded set of firm-level characteristics are: industry sector (retail distribution (1)), public sector (1), employment size (6), ratio of female workers; ratio of part-time workers; share-ownership scheme (1); profit-related pay (1); workplace older than 20 years (1); teamwork (i.e. whether more than 60 per cent of workers work in teams) labour proportions of operating costs (3); injury rate.

has a substantial effect on own earnings: an across-the-workplace increase in education of one year raises own earnings by 17.0 per cent (12.5 per cent) for males (females). The premium to own education is reduced, although again not substantially. The strength of the coefficient (a3 . 0) suggests that own and co-worker education appear to be strongly complementary. Spillover effects are evident; and the goodness of fit of the wage equation is raised by 3-4 percentage

Human capital spillovers in the workplace 131

Table II. Log pay per hour – individual and mean workplace education and training effects (GLS random effects)

IJM 25,1

132

points. Own training maintains its strong effect on earnings, but again there are no spillover effects arising from the training of co-workers. Model (4) is the full estimation specified as in equation (1), to include the interaction between own and co-worker years of education. However, unlike the case for the whole sample, the interaction term is insignificant and its introduction causes the other education variables to lose significance as a consequence of multi-collinearity. The effect of the dispersion of human capital is reported in Table III. Two tests are applied. First, in Model (5) a direct measure of dispersion Ej is included in place of the interaction term. The upper panel of Table III shows that, adjusting for overall workforce human capital, greater dispersion of education across the workplace has no significant effect on own earnings in contrast to the results for the whole sample. Second, the square of workplace years of education is reported in the Model (6) in the bottom panel of Table III. The coefficients on workplace education are negative, but for its square they are positive: again contrasting with the results for the whole sample. Now for Male Coeff (t. stat)

Table III. Log pay per hour – individual and dispersion of workplace education levels (GLS random effects)

Female Coeff (t. stat)

Model (5) a,b Firm-level characteristics Yes Yes Own years of education eij 0.0470*** (10.99) 0.0399*** (8.98) Mean workplace years of education Ej 0.1627*** (6.82) 0.1269*** (4.46) Dispersion: workplace years of Education Ej 0.0266 (0.83) 2 0.0075 (20.22) 0.0953*** (4.26) 0.0442** (2.13) Own training tij Mean workplace training Tj 0.0266 (0.41) 0.0786 (1.13) rj 0.0939 0.2537 R 2 total 0.5123 0.04259 Model (6) a,b Firm-level characteristics Yes Yes Own years of education eij 0.0476*** (11.26) 0.0397*** (9.01) Mean workplace years of education Ej 2 0.8812*** (2 2.29) 2 1.7905*** (2 4.16) Squared term: E 2j 0.0394*** (2.73) 0.0720*** (4.47) Own training tij 0.0953*** (4.27) 0.0434** (2.09) Workplace training Tj 0.0608 (0.92) 0.1269* (1.88) rj 0.0893 0.2241 R 2 total 0.5156 0.4422 Nj (Ni) 1,320 (261) 1,138 (273) Notes: Unweighted data. Significance: ***1 per cent level; **5 per cent level; *10 per cent level. Only full-time workers rj ¼ ðsj Þ2 =½ðsj Þ2 þ ðsi Þ2
aincluded set of individual characteristics: tenure; tenure squared; age; age squared; ethnicity (1 dummy); disability (1); marital status (3); union member; temporary, fixed term or overtime worker (3); bincluded set of firm-level characteristics are: industry sector (retail distribution¼1); public sector (1); employment size (6); ratio of female workers; share-ownership scheme (1); profit-related pay (1); workplace older than 20 years (1); teamwork (whether more than 60 per cent of workers work in teams); labour proportion of operating costs (3); injury rate

women only there are substantial positive returns from the training of co-workers. 4.2 The effects of workplace education on own earnings: union sample For further exposition, the hypotheses are tested across union and non-union workplaces. This split is interesting, because the results could go either way. Unions may facilitate the sharing of skills across workers, who then collectively bargain over pay. With facilitation, the probability of invidious competition between workers would be lower and so cross-workplace spillovers of human capital should be stronger. Yet, if unions instead demarcate skills and apportion tasks, this would reduce the opportunities for human capital spillovers. Models (3) and (4) as per Table II are re-estimated in Table IV, split by union/non-union workers (the union is identified by having pay bargaining rights). These estimations show that both the returns to own education and training are lower for union workers. Indeed, workplace years of education and training are both insignificant for both men and women in union workplaces. It is possible that educated workers in high demand are likely to join trade unions and since workers in craft occupations are highly unionised there is less requirement for additional training in these occupations. Appendix Table AI provides comparable data to those contained in Table II for all sectors other than our two service sectors. This evidence from outside the service sectors shows substantial spillover effects from co-worker workplace training

Male coeff (t.stat)

Human capital spillovers in the workplace 133

Female coeff (t.stat)

Only union workplaces 0.0421*** (4.86) 0.0362*** (2.94) Own years of education eij Mean workplace years of education Ej 0.0787 (1.57) 0.0492 (0.73) Own training tij 0.0832 (1.62) 2 0.0408 (2 0.72) Workplace training Tj 20.0530 (2 0.34) 2 0.0096 (2 0.05) rj 0.2024 0.2429 R 2 total 0.5106 0.4737 Nj (Ni) 77 (196) 80 (203) Non-union workplaces Own years of education eij 0.0497*** (10.39) 0.0415*** (8.50) Mean workplace years of education Ej 0.1877*** (8.08) 0.1387*** (4.89) Own training tij 0.0948*** (3.81) 0.0532** (2.31) Workplace training Tj 0.0311 (0.45) 0.0833 (1.13) rj 0.0648 0.2525 R 2 total 0.5280 0.0442 Nj (Ni) 241 (1,124) 258 (935) Notes: Unweighted data. Significance: ***1 per cent level; **5 per cent level; *10 per cent level Each estimation includes individual-level and firm-level characteristics, as per Model (3) of Table II. See notes in Table II

Table IV. Log pay per hour – mean workplace education and training levels for union and non-union workplaces (GLS random effects)

IJM 25,1

for both men and women; these spillovers exceed those to own training. Precisely why these effects do not apply in our two service sectors requires further research, but could be related to the labour intensive nature of such activities and the relatively low levels of human capital found that are relative to the rest of the industries and services.

134 4.3 The effects of workplace education on own earnings: sensitivity tests A series of tests for the robustness of the results in Table II were undertaken The sensitivity tests were grouped into three categories: (1) restrictions on the sample for estimation, (2) respecification of the data, and (3) use of alternative derivations of ei, Ej and yij. Table V reports estimations of Models (3) and (4), but with sample restrictions applied. The sample is restricted to those workplaces where it might plausibly be expected that human capital spillovers would be the strongest: these are identified as workplaces where there is high team-working (i.e. firms where at least 60 per cent of workers are reported to work in teams) and where the technology is labour-intensive (i.e. where labour accounts for more than 75 per cent of operating costs). Reducing the sample inflates the standard errors, and this generates some sensitivity across the results. The results obtained earlier are repeated when the sample is restricted to establishments working in teams. It is worth remarking, however, that the teamwork variable was insignificant Male coeff (t.stat)

Table V. Log pay per hour – mean workplace education and training levels for restricted sample (GLS random effects)

Female coeff (t.stat)

Only establishments working in teams 0.0409*** (7.47) 0.0410*** (7.53) Own years of education eij Mean workplace years of education Ej 0.1687*** (5.61) 0.1242*** (3.28) Own training tij 0.0975*** (3.37) 0.0512** (2.01) Workplace training Tj 0.0323 (0.38) 0.1516 (1.64) rj 0.1104 0.2856 R 2 total 0.5406 0.4129 Nj (Ni) 165 (796) 173 (780) Only labour intensive establishments Own years of education eij 0.0459*** (6.58) 0.0510*** (6.90) Mean workplace years of education Ej 0.1948*** (5.43) 0.2157*** (5.71) Own training tij 20.0105 (2 0.27) 2 0.0315 (2 0.81) Workplace training Tj 0.1000 (0.69) 0.1628 (1.17) rj 0 0.0183 R 2 total 0.6168 0.6277 Nj (Ni) 44 (353) 46 (328) Notes: Unweighted data. Significance: ***1 per cent level; **5 per cent level; *10 per cent level Each estimation includes individual-level and firm-level characteristics, as per Model (3) of Table II. See notes in Table II

in the earnings regressions both for the service sector sub-sample and for the whole sample of industries and services. When the sample was restricted to labour intensive establishments (where labour accounts for more than 75 per cent of operating costs) the training variable became insignificant, suggesting that training is more important in technologically advanced workplaces. Respecification of the data and the use of alternative derivations of ei, Ej and Yij made little difference to these results. 5. Conclusion This analysis of workplaces in two service sectors confirms the results of the earlier, more general, study into education spillovers for industry and services as a whole – such spillovers are substantial and these effects are largely independent from the effect of own education. The provision of training at the workplace also has a substantial effect on earnings in our two service sectors, particularly for men. But here, unlike the case for the rest of industry and services, we were unable to detect any training spillover effect. That is, additional training for co-workers had no effect on colleague’s earnings. This might possibly reflect the less complex nature of work in these two sectors compared to other sectors[9]. As in the more general study, there are substantial differences between the union and non-union workplaces with own training, tij, being significant in the latter but insignificant in the former and own education effects being greater in the non-union workplaces and workplace education effects only being significant in non-union establishments. It appears that unions may reduce some of the benefits of education and training through demarcation and other effects. Unlike the case in the earlier study, a greater dispersion of earnings across the workplace is not associated with higher own earnings, in line with the hypothesis of skills compatibility, and with the lower level of human capital required in these sectors. These spillovers suggest that part of the return to human capital comes from the interaction of workers with one another as reflected in team work or the passing on of acquired knowledge, but there are significant differences in the extent to which they impact by gender, which should be the subject of further research. Notes 1. Negative effects may also arise if a sizeable proportion of the workforce has more education than is required for their job: they are over-educated (Oosterbeek, 2000). It is well documented that overeducation generates a wage penalty relative to being fully matched. These co-workers may negatively affect own earnings. 2. Macro-economic models and endogenous growth models draw on the notion of “production externalities”, with economic growth boosted through more efficient social capital investments (McMahon, 2000; Romer, 1994).

Human capital spillovers in the workplace 135

IJM 25,1

136

3. In an unpublished study, Barth (2000) directly tests the effects on co-worker earnings using two types of matched employer– employee data from Norway: an independent effect on own pay from the average level of education within an establishment ranges from 1 to 4 per cent per year of average education. In contrast, Groshen (1991) finds that education levels very weakly reduce establishment wage differentials. 4. Yet, across the workplace, over-education may be minimal: either the employee mix could be adjusted so that over-educated and under-educated workers compensate for each other or physical capital intensities may be varied. However, over-educated workers may guide or assist co-workers who only have the required education and it may be under-educated workers who are less competent, need greater monitoring or require more co-worker support (Tsang et al., 1993). 5. The limitation to full-time workers is stringent given the importance of part-time work in the two service sectors, but it is likely that spill over effects from education will be much stronger for full-time employees, as opposed to workers employed for only a few hours per week. 6. Average years of education are lower for our services sub-sample than for the WERS as a whole by a figure approaching one year. The dispersion of education and the level of earnings are also appreciably lower. 7. We are grateful to an anonymous referee for pointing this out. 8. Random effects GLS is a less biased estimator than OLS, because the data are grouped across workplaces (Moulton, 1987). A Hausman test easily rejects the use of fixed effects GLS. However, diagnostic tests do show multicollinearity across the independent variables. 9. There appears in our services sector sample to be a substantially significant relationship (at the 10 per cent level) between the workplace education levels and relative financial performance of the establishment as judged by management, though no similar relationship emerges for workplace training. References Abowd, J.M., Kramarz, F. and Margolis, D.N. (1999), “High-wage workers and high-wage firms”, Econometrica, Vol. 67, pp. 251-333. Addison, J.T. and Belfield, C.R. (2001), “Updating the determinants of firm performance: estimation using the 1998 workplace employment relations survey”, British Journal of Industrial Relations, Vol. 39, pp. 341-66. Ashenfelter, O., Harmon, C. and Oosterbeek, H. (2000), “A review of estimates of the schooling/earnings relationship, with tests for publication bias”, Labour Economics, Vol. 6, pp. 453-70. Barron, J.M., Berger, M.C. and Black, D.A. (1997), “How well do we measure training?”, Journal of Labor Economics, Vol. 15, pp. 507-28. Barth, E. (2000), External Effects of Education? Evidence from the Wage Structure, Institute for Social Research, Oslo. Battu, H., Belfield, C.R. and Sloane, P. (1999), “Over-education among graduates: a cohort view”, Education Economics, Vol. 7, pp. 21-38. Battu, H., Belfield, C.R. and Sloane, P.J. (2001), “Human capital spillovers in the workplace”, Discussion Paper No. 404, November, IZA, Bonn. Bayard, K. and Troske, K.R. (1999), “Examining the employer-size wage premium in manufacturing, retail trade and service industries using employer-employee matched data”, American Economic Review, Vol. 89, pp. 99-103. Behrman, J.R. and Stacey, N. (1997), The Social Benefits of Education, Michigan UP, MI. Belfield, C.R. (2000), Economic Principles for Education. Theory and Evidence, Edward Elgar, Cheltenham.

Blundell, R., Dearden, L. and Goodman, H. (2000), “The returns to higher education in Britain: evidence for a British cohort”, Economic Journal, Vol. 110, pp. f82-f99. Borjas, G.J., Bronars, S.G. and Trejo, S.J. (1992), “Assimilation and earnings of young internal migrants”, Review of Economics and Statistics, Vol. 74, pp. 170-5. Bronars, S.G. and Famulari, M. (1997), “Wage, tenure, and wage growth variation within and across establishments”, Journal of Labor Economics, Vol. 15 No. 2, pp. 285-317. Cohn, E. and Addison, J.T. (1998), “The economic returns to lifelong learning”, Education Economics, Vol. 6, pp. 309-46. Dolton, P.J. and Vignoles, A. (2000), “The incidence and effects of over-education in the UK graduate labour market”, Economics of Education Review, Vol. 19, pp. 179-98. Groshen, E.L. (1991), “Sources of intra-industry wage dispersion: how much do employers matter?”, Quarterly Journal of Economics, Vol. 105, pp. 869-84. Harmon, C. and Walker, I. (1995), “Estimates of the economic return to schooling for the United Kingdom”, American Economic Review, Vol. 85, pp. 1278-86. Harmon, C., Walker, I. and Westergaard-Nielson, N. (Eds) (2001), Education and Earnings in Europe, Edward Elgar, Cheltenham. Idson, T.L. (1995), “Team production effects on earnings”, Economics Letters, Vol. 49, pp. 197-203. Idson, T.L. and Kahane, L.H. (2000), “Team effects on compensation: an application to salary determination in the National Hockey League”, Economic Inquiry, Vol. 38, pp. 345-57. Kremer, M. (1993), “The o-ring theory of economic development”, Quarterly Journal of Economics, Vol. 107, pp. 551-75. Lazear, E.P. (2000), “Economic imperialism”, Quarterly Journal of Economics, Vol. CXV, pp. 99-146. Machin, A. and van Reenen, J. (1998), “Technology and changes in skill structure: evidence from seven OECD countries”, Quarterly Journal of Economics, Vol. CXIII, pp. 1215-44. McMahon, W.W. (2000), Education and Development, Oxford University Press, Oxford. Miller, P., Mulvey, C. and Martin, N. (1997), “Family characteristics and the returns to schooling: evidence on gender differences from a sample of Australian twins”, Economica, Vol. 64, pp. 119-36. Moulton, B.R. (1987), “Diagnostics for group effects in regression analysis”, Journal of Business Economics and Statistics, Vol. 5, pp. 275-82. Oosterbeek, H. (2000), “Editorial: the economics of over- and under-schooling”, Economics of Education Review, Vol. 19, pp. 129-31. Pfeiffer, F. (2000), “Training and individual performance in Europe: evidence from microeconometric studies”, Discussion Paper No. 00-28, Centre for European Economic Research, Mannheim. Rauch, G.E. (1993), “Productivity gains from geographic concentration of human capital: evidence from the cities”, Journal of Urban Economics, Vol. 34, pp. 380-400. Romer, P.M. (1994), “The origins of endogenous growth”, Journal of Economic Perspectives, Vol. 8, pp. 3-22. Troske, K.R. (1999), “Evidence on the employer size-wage premium from worker-establishment matched data”, Review of Economics and Statistics, Vol. 81, pp. 15-26. Tsang, M.C. (1987), “The impact of under-utilization of education on productivity: a case study of the US Bell companies”, Economics of Education Review, Vol. 6, pp. 239-54. Tsang, M.C., Rumberger, R.W. and Levin, H.M. (1993), “The impact of surplus schooling on workers’ productivity”, Industrial Relations, Vol. 30, pp. 209-28. Van Smoorenberg, M.S.M. and van der Velden, R.K.W. (2000), “The training of school-leavers: complementarity or substitutability?”, Economics of Education Review, Vol. 19, pp. 207-18.

Human capital spillovers in the workplace 137

IJM 25,1

138

Winter-Ebmer, R. and Zweimueller, J. (1997), “Unequal assignment and promotion in job ladders”, Journal of Labor Economics, Vol. 15, pp. 43-71. Wolfe, B.L. and Zuvekas, S. (2000), “Non-worker outcomes of schooling”, International Journal of Educational Research, Vol. 27, pp. 491-502. Further reading Battu, H., Belfield, C.R. and Sloane, P. (2000), “Over-education: how sensitive are the measures?”, National Institute Economic Review, Vol. 171, pp. 82-93. Borghans, L. and de Grip, A. (Eds) (2000), The Over-educated Worker? The Economics of Skill Utilisation, Edward Elgar, Cheltenham. Hartog, J. and Oosterbeek, H. (1998), “Health, wealth and happiness: why pursue a higher education?”, Economics of Education Review, Vol. 17, pp. 245-56. Appendix

Male coeff (t. stat)

Table AI. Log pay per hour – individual and mean workplace education and training effects for the whole sample excluding hotels and catering and retail distribution

Model (1) Firm level characteristics No Own years of education eij 0.0573*** (45.27) Own training tij 0.0883*** (12.60) rj 0.3825 R 2 total 0.3940 Model (2) Firm level characteristics Yes Own years of education eij 0.0572*** (44.90) Own training tij 0.0835*** (11.91) rj 0.3103 R 2 total 0.4540 Model (3) Firm level characteristics Yes Own years of education eij 0.0543*** (42.03) Mean workplace years of education Ej 0.0757*** (11.15) Own training tij 0.0782*** (11.00) Mean workplace training Tj 0.1255*** (3.72) rj 0.2939 R 2 total 0.4880 Model (4) Firm level characteristics Yes Own years of education eij 0.1050*** (7.64) Mean workplace years of education Ej 0.1306*** (8.03) Interaction eij*Ej 2 0.0037*** (2 3.70) Own training tij 0.0779*** (10.95) Mean workplace training Tj 0.1161*** (3.48) rj 0.2835 0.4880 R 2 total Nj (Ni) 1,125 (9,232) Notes: Significance level: *10 per cent level; **5 per cent level; and ***1

Female coeff (t. stat) No 0.0625*** (43.46) 0.0639*** (8.76) 0.3118 0.3723 Yes 0.0629*** (43.89) 0.0668*** (8.64) 0.2354 0.4490 Yes 0.0609*** 0.0486*** 0.0606*** 0.0859*** 0.2201 0.4607

(41.92) (8.00) (7.59) (2.93)

Yes 0.0582*** (3.43) 0.0459*** (2.58) 0.0002*** (0.16) 0.0606*** (7.58) 0.0862*** (2.94) 0.2177 0.4606 1,104 (6,574) per cent level

About the Guest Editors Rita Asplund Rita Asplund is research director in ETLA, The Research Institute of the Finnish Economy, with responsibility for the Institute’s research programme on technology, competence and competitiveness. Her research interests focus on applied labour market economics with special emphasis on economics of education. E-mail: [email protected]

About the Guest Editors

139

Wiemer Salverda Wiemer Salverda is co-ordinator of the European Low-wage Employment Research network (LoWER). He is affiliated to the Amsterdam Institute for Advanced Labour Studies (AIAS) of the University of Amsterdam. His research interests are the functioning of the low-skilled and low-paid labour market including product demand and employer behaviour, wage and income inequality, the Dutch “Polder Model”, labour market policies and pay in the public sector. E-mail: [email protected]

International Journal of Manpower Vol. 25 No. 1, 2004 p. 139 q Emerald Group Publishing Limited 0143-7720

IJM 25,1

About the authors

140

Harminder Battu Harminder Battu is Senior Lecturer in the Department of Economics, University of Aberdeen, Scotland. His main research interests are in the fields of labour and regional economics. In particular, his research has focused on the utilisation of education in the labour market, the modelling of regional economics, regional mobility and commuting behaviour. E-mail: [email protected] Clive Belfield Clive Belfield is Associate Director of the National Centre for the Study of Privatisation in Education, Teachers College, Columbia University, USA. His main research interests are in the economics of education. He is the author of Economic Principles for Education: Theory and Evidence, Edward Elgar, 2000. E-mail: [email protected]. columbia.edu Andries de Grip Andries de Grip is Head of the Division of Employment and Training at the Research Centre for Education and the Labour Market (ROA) and Professor of Economics at the Faculty of Economics and Business Administration, Maastricht University. His research interests include training and mobility, employability, skill mismatches and obsolescence, human resource management, labour market segmentation, atypical employment and upgrading and overeducation. E-mail: [email protected] Helmut Hofer Helmut Hofer works currently as a researcher in the Department of Economics, Institute for Advanced Studies in Vienna, Austria. His main research focus is on labour economics and on business cycle forecasting. He works on various aspects of the Austrian labour market, e.g. wage dynamics, unemployment, and labour market policy evaluations. He earned his PhD (Doktorat) at the University of Economics and Business Administration in Vienna. E-mail: [email protected] Gerard Hughes Gerard Hughes is a Research Professor at the Economic and Social Research Institute, Dublin and a Visiting Professor at the Department of Economics, University College, Cork. He has economics degrees from University College Dublin, the London School of Economics and Trinity College Dublin. His research interests are in the fields of labour economics, social policy and public finance. E-mail: [email protected]

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 140-142 q Emerald Group Publishing Limited 0143-7720

Tomas Korpi Tomas Korpi works as a Researcher at the Swedish Research Council and at the Swedish Institute for Social Research. His research focuses on issues related to unemployment, labour mobility and public policy. E-mail: [email protected]

Christine Lietz Christine Lietz works in the Microsimulation Unit of the Department of Applied Economics, University of Cambridge. Her current main research interests are tax-benefit microsimulation, income distribution and poverty. She earned her Masters (Magister) at the Department of Economics, University of Vienna. In her diploma thesis she explored aspects of the Austrian apprenticeship system. E-mail: christine. [email protected] Antje Mertens Antje Mertens is a Researcher at the Max Planck Institute for Human Development and at the Humboldt University in Berlin. Her research interests center on empirical labour economics, especially labour mobility, training, and wages. E-mail: [email protected] Philip O’Connell Philip J. O’Connell is Research Professor at the ESRI, Dublin. Much of his work focuses on education, training and the labour market. He has written several books on the effects of work-related education and training, and papers on this and other labour market issues have been published in Work, Employment and Society, The Economic and Social Review, The European Sociological Review, The Industrial and Labour Relations Review, and The British Journal of Industrial Relations. His current research interests include the impact of active labour market programmes and workplace change. E-mail: [email protected] Vicente Ramos Mir Vicente Ramos Mir is Associate Professor in the Department of Applied Economics at the University of the Balearic Islands. He holds a Master Degree in economics from Pompeu Fabra University Barcelona (2000). He teaches tourism economics and macroeconomics in the Economics Faculty. He is finishing his PhD dissertation on the labour market characteristics of the tourism sector. His fields of interest are labour economics and tourism economics. E-mail: [email protected] Javier Rey-Maquieira Javier Rey-Maquieira is Vice-Dean in the Economics Faculty of the University of the Balearic Islands. He was the Vice-Chancellor of Economic Planning in this University. He is the Associate Professor of Macroeconomics in the Department of Applied Economics. His topics are in Tourism and Natural Resources. He holds a PhD in Economics at the University of Barcelona. He has directed several research projects on Tourism Economics. He supervised different PhD theses on this subject. He published in different national and international journals. E-mail: [email protected] Jos Sanders Jos Sanders is a Researcher at the Division of Employment and Training at the Research Centre for Education and the Labour Market (ROA) at Maastricht University. He graduated in Human Resources Sciences from Tilburg University in 1998. His research interests include employability, human resources management and training.

About the authors

141

IJM 25,1

142

He is also a part-time Personnel Consultant at the faculty of Health Sciences at Maastricht University. E-mail: [email protected] Wendy Smits Wendy Smits is Research Fellow in the Research Centre for Education and the Labour Market (ROA) at Maastricht University, The Netherlands. Her main research interests are in labour economics, in particular the economics of the apprenticeship system, education and training and labour demand. E-mail: [email protected] Peter Sloane Peter Sloane is Director of the Welsh Economy Labour Market Evaluation and Research Centre (WELMERC), Department of Economics, University of Wales Swansea. He is also a Research Fellow at the Institute for Labor Studies (IZA), Bonn and a founder member of the European Low Wage Employment Network (LoWER). His research interests cover most areas of labour economics and the economics of sport. E-mail: [email protected] Maria Tugores Maria Tugores is Associate Professor and Assistant Director in the Department of Applied Economics at the Universitat de les Illes Balears. She obtained her PhD in Economics at the Universidad Carlos III de Madrid (2000). She teaches industrial economics and labour economics in the Economics Faculty. She has published scientific papers in national and international journals. Her fields of interest are industrial economics, labour economics and tourism economics. E-mail: [email protected] James Williams James Williams is a Research Professor at the ESRI Dublin and is the Head of the Institute’s Survey Unit. He holds a BA, an MA and an MEconSc from University College Dublin and a Diploma in Applied Statistics from Trinity College Dublin. He previously held the position of Senior Economist at Bord Failte Eireann, the Irish National Tourist Organisation. His areas of interest include survey methodology, poverty and income distribution, Geographic Information Systems and regional development. E-mail: [email protected] Thomas Zwick After graduating from Regensburg University, Germany in economics, Thomas Zwick worked as a PhD student at Maastricht University, The Netherlands. Since 1998, he has been a Research Fellow at the Centre for European Economic Research (ZEW) in Mannheim, Germany. His main research interests are microeconomic and microeconometric labour market research, personnel economics and human capital formation. E-mail: [email protected]

Note from the publisher During 2003, Emerald developed its corporate publishing philosophy. We did this through discussion with readers, contributors and editors and we would like to share it with you. We believe that our approach to quality makes us different and unique amongst scholarly publishers. It is based on six core principles, which together form our distinctive philosophy.

Note from the publisher

143

We put quality at the centre of our approach to scholarly publishing All papers published by Emerald go through a quality-assured peer review system; in all but a few practitioner-focused journals this takes the form of double-blind peer review. All papers published by Emerald are expected to make, in some way, an explicit original contribution to the existing body of knowledge. All papers published by Emerald are accessible to a wide range of students, scholars and practitioners in the fields in which we publish. All papers published by Emerald are beneficial in some way to researchers, practitioners, or both. . In 2001, we were audited and certified as “Committed to Excellence” following a European Foundation for Quality Management self-assessment exercise. . We retained our status as an ISO 9000 certified organisation, and our Investors in People (IIP) certification. . More than 30 Emerald journals are listed in the ISI Citation Index.

Continuous improvement of reader, author and customer experience We continue to invest in enabling technology to increase efficiency and effectiveness in content provision, customer service and management. We benchmark against others and against our own standards. We are as clear as possible in our policies, measures, targets and achievements and we do not hide shortfalls, but confront and learn from them. . Emerald papers go through a further post-publication “review” which assesses them on readability, originality, implications for further research and practice. We publish this information, and it can be used as search criterion, on the Emerald database. . In 2002, we were judged as providing the Best Customer Support by the scholarly library publication, The Charleston Advisor. . We provide high levels of dissemination of our authors’ work – nearly 1 million papers per month are downloaded and read by subscribers to the Emerald online portfolio. . In 2004, we will be introducing an online submission and peer review system which will speed up the publication process.

Internationality We operate in a transnational world of scholarly ideas and we believe that this should be reflected within our publications. Working with our authors, we set targets for international representation of authors and editorial team members, and measure against them.

International Journal of Manpower Vol. 25 No. 1, 2004 pp. 143-144 q Emerald Group Publishing Limited 0143-7720

IJM 25,1

.

.

In the past six months, we have published more than 50 themed issues with a specific international focus. In the first half of 2003, papers from 60 different countries were published.

An interdisciplinary approach

144

We set targets, and ask for papers and special issues on interdisciplinary approaches, and new/emergent themes. This gives us better, stronger, and more vibrant journals, and a clear leadership position in our industry. . In the first half of 2003, we published more than 20 themed issues dealing with interdisciplinary approaches to a subject or industry. . We encourage themed issues on leading edge and innovative research topics, and in the past 6 months published 35 such issues.

Supporting scholarly research: The Literati Club We help remove the barriers to publication. We conduct workshops for researchers on publishing issues. We provide help and advice to new researchers. We offer a service for authors whose first language is not English. Our staff regularly present papers at conferences on scholarly publishing themes. . Our scholarly community Web site, the Literati Club, disseminates information about how to write for publication more successfully – we seek to make the process more transparent. . The Emerald Research Register, an online forum for the circulation of pre-publication information is designed to help researchers gain advanced recognition among their peers by publicising their research at the earliest opportunity. . Each year, we distribute grants to researchers working on improving the scholarly publishing dissemination process, and encouraging scholarship in the developing world. . We conducted research workshops at 11 universities and conferences, had papers accepted at nine academic and other conferences, and supported six academic conferences worldwide.

Integration of theory and practice We ask editors and review board members to focus where applicable on application, and beneficial implication for practice. We do so because this gives a clear message to our core supplier and consumer markets – the applied researcher, the reflective practitioner, the students of business and their teachers, the business and management school. . All of our journals will publish a majority of papers that have a direct application to the world of work. . More than 1,000 university libraries worldwide subscribe to the Emerald portfolio, including 97 per cent of the Financial Times Top 100 Business Schools. John Peters Director of Author Relations Emerald – October 2003