Abstract
Recent studies have used quantile regression (QR) techniques to estimate the impact of education on the location, scale and shape of the conditional wage distribution. We conduct a similar investigation of the role of work-related training. We utilise both ordinary least squares and QR techniques to estimate associations between work-related training and wages for private sector men in ten European Union countries. For the majority of countries, the association between training and hourly wages varies little across the conditional wage distribution. However, there are considerable differences across countries in mean associations between training and wages.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Although there has been a recent surge in the estimation of wage equations using quantile regression techniques (see Fitzenberger et al. 2001, for some applications), to our knowledge there are no studies investigating the association between work-related training and the conditional wage distribution.
There is an extensive literature on the evaluation of particular labour market programmes, using a variety of techniques. For example, Heckman et al. (1998) estimate the average effects of training on the treated, Heckman et al. (1997) look at the distribution of treatment effects using a non-instrumental variable (IV) framework, and Abadie et al. (2002) examine the training effect on different quantiles of the wage distribution using the IV framework. More recently, Lechner and Melly (2007) develop bounds estimators of quantile training effects. Our interest here is in work-related training and not in a labour market program. Since we do not have a suitable instrument, we do not treat training as endogenous.
The ECHP includes two datasets for Germany: the six-wave dataset (derived from the German Socio-Economic Panel survey), which excludes many shorter training spells (communication from the German Institute for Economic Research, DIW Berlin), and the original three-wave dataset. In the three-wave dataset, interview dates are treated as confidential, so it is not possible to construct job tenure or know whether training was before or after the previous interview.
The modal interview month is October, corresponding to a reference period of 22 months. The British data do not include training dates. However, they are derived from the British Household Panel Survey, where the reference period only slightly exceeds 1 year. Since events are generally very short in Britain, there should be little chance of double counting. For France, we do not use training dates as they are missing for the majority of events. For the Netherlands, the end dates of training are not available so we use start dates only to identify events begun since the previous interview. Notice that we are unable to use the duration data since there are a number of missing observations for some countries, and for that reason we focus on events rather than duration.
Specification 1 includes e i1 to pick up any un-observed training that occurred in and prior to wave 1 and it also includes accumulated training to pick up the history of training subsequently. But e i1 is not included in Specification 2, on the tacit assumption that history does not matter. But if we assume history does not matter, we should only use the latest training receipt rather than using accumulated training in the model. Thus Specifications 1 and 2 are polar: Specification 1 assumes a non-decaying effect of training receipts and Specification assumes a rapid decaying of training effects.
In an interesting paper also using the ECHP, Bassanini and Brunello (2008) utilise data for seven countries from the 1996 wave to explore the degree to which training incidence is correlated with wage compression. Our estimates are not directly comparable with theirs since they pool cross-country data and partition full-time men aged 30–60 years into clusters (by country, education level, occupation and sector) in order to test the degree to which cluster-specific measures of the training wage premium are correlated with general training incidence.
Martins and Pereira (2004), using different non-harmonised data sets and employing years of schooling as their measure of education, found that returns to schooling increased over the wage distribution for their 16 different countries. Our results for tertiary education provide some support for their findings, although our specification includes many more explanatory variables along with a control for pre-sample unobservables. As is typical for studies that estimate the returns to schooling, Martins and Pereira used very few controls (experience and its square). Such a parsimonious specification is inappropriate for our purposes, where the focus is on training received after entering the labour market. Their estimating samples also included men in both public and private sectors aged 15–65 working at least 35 h/week. See also Angrist et al. (2006), who find, using three US census datasets, that over time the returns to college graduates have been increasing across the distribution. Budria and Pereira (2005) use a variety of European data sources and also find that the returns to tertiary education are typically increasing across the distribution.
References
Abadie A, Angrist J, Imbens G (2002) Instrumental variables estimates of the effect of subsidized training on the quantiles of trainee earnings. Econometrica 70(1):91–117
Acemoglu D, Pischke J-S (1999) The structure of wages and investment in general training. J Polit Econ 107(3):539–572
Almeida R, Carneiro P (2006) The return to the firm investment in human capital. IZA Discussion Paper No. 1937
Angrist J, Chernozhukov V, Fernandez-Val I (2006) Quantile regression under misspecification, with an application to the US wage structure. Econometrica 74(2):539–563
Arias O, Hallock K, Sosa-Escudero W (2001) Individual heterogeneity in the returns to schooling: instrumental variables quantile regression using twins data. In: Fitzenberger B, Koenker R, Machado JAF (eds) Economic applications of quantile regression. Physica-Verlag, Heidelberg, pp 7–40
Arulampalam W, Booth AL (2001) Learning and earning: do multiple training events pay? A decade of evidence from a cohort of young British men. Economica 68(271):379–400
Arulampalam W, Booth AL, Elias P (1997) Work-related training and earnings growth for young men in Britain. Res Labor Econ 16:119–147
Arulampalam W, Booth AL, Bryan ML (2004) Training in Europe. J Eur Econ Assoc 2(2/3):346–360
Arulampalam W, Booth AL, Bryan ML (2007) Is there a glass ceiling over Europe? An exploration of asymmetries in the gender pay gap across the wages distribution. Ind Labor Relat Rev 60(2):163–186
Ashenfelter O, Lalonde RJ (eds) (1996) The economics of training. Edward Elgar, Cheltenham Northampton
Bassanini A, Brunello G (2008) Is training more frequent when wage compression is higher? Evidence from the European community household panel. Labour Econ 15(2):272–290
Bassanini A, Booth AL, Brunello G, De Paola M, Leuven E (2007) Workplace training in Europe. In: Brunello G, Garibaldi P, Wasmer E (eds) Education and training in Europe. Oxford University Press, Oxford, pp 43–323
Blundell R, Dearden L, Meghir C (1999) The determinants and effects of work-related training in Britain, IFS report R50. Institute for Fiscal Studies, London
Booth AL, Bryan ML (2007) Who pays for general training in private sector Britain? Res Labor Econ 26:85–123
Booth AL, Snower DJ (eds) (1996) Acquiring skills: market failures, their symptoms, and policy responses. Cambridge University Press, Cambridge
Buchinsky M (1998) Recent advances in quantile regression models. J Hum Resour 33(1):88–126
Budria S, Pereira PT (2005) Educational qualifications and wage inequality: evidence for Europe. IZA Discussion Paper No. 1763
Card D (1999) The causal effect of education on earnings. In: Ashenfelter O, Card D (eds) Handbook of labor economics, vol. 3A. Elsevier, Amsterdam, pp 1801–1863
Fitzenberger B (1998) The moving blocks bootstrap and robust inference for linear least squares and quantile regressions. J Econom 82(2):235–287
Fitzenberger B, Kurz C (2003) New insights on earnings trends across skill groups and industries in West Germany. Empir Econ 28:479–514
Fitzenberger B, Koenker R, Machado JAF (2001) (eds) Economic applications of quantile regression. Physica-Verlag, Heidelberg
Gonzales X, Miles D (2001) Wage inequality in a developing country: decrease in minimum wage or increase in education returns. In: Fitzenberger B, Koenker R, Machado JAF (eds) Economic applications of quantile regression. Physica-Verlag, Heidelberg New York, pp 135–148
Heckman J, Smith J, Clements N (1997) Making the most out of social experiments: accounting for heterogeneity in programme impacts. Rev Econ Stud 64(4):487–535
Heckman J, Smith J, Taber C (1998) Accounting for dropouts in evaluations of social experiments. Rev Econ Stat 80(1):1–14
Koenker R (2005) Quantile regression. Cambridge University Press, Cambridge
Koenker R, Bassett G (1978) Regression quantiles. Econometrica 46(1):33–50
Lechner M, Melly B (2007) Earnings effects of training programs. CEPR discussion paper no. 6400. Centre for Economic Policy Research, London
Lillard LA, Tan HW (1992) Private sector training: who gets it and what are its effects? Res Labor Econ 13:1–62
Lynch LM (ed) (1994) Training and the private sector: international comparisons. University of Chicago Press, Chicago
Martins PS, Pereira PT (2004) Does education reduce wage inequality? Quantile regression evidence from 16 countries. Labour Econ 11(3):355–371
OECD (1999) Training of adult workers in OECD countries: measurement and analysis. In: OECD (ed) OECD employment outlook 1999. OECD, Paris, pp 134–175
OECD (2003) Upgrading workers’ skills and competencies. In: OECD (ed) OECD employment outlook 2003. OECD, Paris, pp 237–296
Acknowledgements
For helpful comments, we thank the Editor, two anonymous referees, Marco Francesconi, Amanda Gosling, Kevin Hallock, Jennifer Smith, Ian Walker and seminar participants at the Institute for Social and Economic Research at the University of Essex and the European Association of Labour Economists Conference 2004. The paper is a revised version of our 2004 IZA Discussion Paper No. 984 of the same name. We are grateful for financial support from the Leverhulme Trust under the Award F/00213/H “Training in Europe: Its Causes and Consequences”.
Author information
Authors and Affiliations
Corresponding author
Additional information
Responsible editor: Christian Dustmann
Appendix: Selection of estimating samples
Appendix: Selection of estimating samples
Unless otherwise stated, we applied the initial selection described in “Section 2” of the text. We then dropped observations with missing or invalid data on the variables in the wage equations, that is principally: training, fixed term or casual contract, occupation, industry, region, establishment size, tenure, part-time status, education, health status and marital status. Where the number of missing values was non-trivial (typically where this would have necessitated a drop in sample size of 5% or more as a consequence), we also included a dummy variable for missing value observations in order to preserve the sample sizes. Finally, we kept only continuous sequences of observations from the first wave (ECHP wave 1) to ensure a complete record of training for each individual. The table details the number of observations remaining at each of these selection stages.
Rights and permissions
About this article
Cite this article
Arulampalam, W., Booth, A.L. & Bryan, M.L. Are there asymmetries in the effects of training on the conditional male wage distribution?. J Popul Econ 23, 251–272 (2010). https://doi.org/10.1007/s00148-008-0209-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00148-008-0209-4