Prelims

Half Title Page

SKILL MISMATCH IN LABOR MARKETS

Series Page

Research in Labor Economics

Series Editor: Solomon W. Polachek

IZA Co-Editor: Konstantinos Tatsiramos

Recent Volumes:

Title Page

RESEARCH IN LABOR ECONOMICS VOLUME 45

SKILL MISMATCH IN LABOR MARKETS

EDITED BY

SOLOMON W. POLACHEK

State University of New York at Binghamton and IZA

KONSTANTINOS POULIAKAS

European Centre for the Development of Vocational Training (CEDEFOP), University of Aberdeen and IZA

GIOVANNI RUSSO

European Centre for the Development of Vocational Training (CEDEFOP) and IZA

KONSTANTINOS TATSIRAMOS

University of Nottingham and IZA

United Kingdom – North America – Japan – India – Malaysia – China

Copyright Page

Emerald Publishing Limited

Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2017

Copyright © 2017 Emerald Publishing Limited

Reprints and permissions service

Contact: permissions@emeraldinsight.com

No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. Any opinions expressed in the chapters are those of the authors. Whilst Emerald makes every effort to ensure the quality and accuracy of its content, Emerald makes no representation implied or otherwise, as to the chapters’ suitability and application and disclaims any warranties, express or implied, to their use.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-1-78714-378-4 (Print)

ISBN: 978-1-78714-377-7 (Online)

ISBN: 978-1-78714-921-2 (Epub)

ISSN: 0147-9121 (Series)

List of Contributors

Editorial Advisory Board

• Orley C. Ashenfelter

  Princeton University

• Francine D. Blau

  Cornell University

• Richard Blundell

  University College London

• Alison L. Booth

  Australian National University

• David Card

  University of California

• Ronald G. Ehrenberg

  Cornell University

• Richard B. Freeman

  Harvard University

• Reuben Gronau

  Bank of Israel

• Daniel S. Hamermesh

  University of Texas

• James J. Heckman

  University of Chicago

• Alan B. Krueger

  Princeton University

• Edward P. Lazear

  Stanford University

• Christopher A. Pissarides

  London School of Economics

• Yoram Weiss

  Tel-Aviv University

• Klaus F. Zimmermann

  University of Bonn

Preface

The 2008 global financial and economic crisis led to a significant increase in unemployment rates in most developed economies, yet despite the rising supply of labor a high share of employers claim that they cannot find the right talent and skills. Concerns have arisen that economic restructuring and changing skill needs associated with new technologies and workplace organization practices will not be met by an adequately skilled workforce. These considerations have placed the issue of skill mismatch – the incongruence between skill supply and skill demand – high up in the policy agenda.

Although it is accepted that skill mismatch entails a potential misallocation of economic resources, there is considerable confusion in the policy debate with regards to different types of education/skill mismatch (e.g., over- or undereducation, over- or underskilling, skill shortages, skill gaps, skill obsolescence) and the best policy measures to tackle them. There is also lack of clarity about the economic cost of different forms of skill mismatch, which can distort decision-making and policy priorities. For example, policy discourse has mostly focused on the productivity and wage loss associated with undereducation or with jobseekers and workers who suffer from skill gaps. But in the absence of severe basic skill deficiencies, remedial training and on-the-job learning can support the skill development of workers and effectively close skill gaps. This is a win-win situation in which both organizations benefit from higher productivity and employees experience professional growth. What is less acknowledged in the policy sphere is the fact that graduates in developed economies are increasingly accepting jobs that require lower qualifications than their own (overeducated) for involuntarily reasons and are paid less than their well-matched classmates. Although there is a clear link between qualifications, which are visible to employers when screening job applicants, and levels of pay, this is not necessarily the case for skills. It takes time for individuals’ skills to be revealed in the workplace. Consequently, the wage mechanism may not necessarily be the best proxy for the lost or surplus productivity of workers whose skills are underutilized by their jobs. Yet, skill underutilization tends to be associated with cognitive decline (“use it or lose it”) and skill atrophy and fosters employee demotivation. Such “hidden” costs of skill underutilization often go unnoticed in the policy debate.

This volume contains 11 original research papers which deal with the linkages between education and skills and the causes and consequences of different types of skill mismatch. They focus on the various avenues through which workers can develop their skills and on how workplaces utilize (or not) such skills. Many of these papers have benefited from new microdata that were collected by the European Centre for the Development of Vocational Training (Cedefop), containing information on the incidence and determinants of skill mismatch in 28 European countries, the European Skills and Jobs Survey. ¹

The first paper in this volume highlights the pitfalls of literally interpreting claims of skill shortages on behalf of employers as a need for immediate policy concern, without properly acknowledging complex underlying labor market dynamics. Most notably, employers regularly complain about a shortage of qualified scientists, and this has spurred multiple policy initiatives aimed at the further proliferation of STEM (sciences, technology, engineering and mathematics) graduates. Arnaud Chevalier uses a survey focused on the labor market decisions of a cohort of graduates from British universities, observed three years after graduation. His analysis shows that there is considerable “leakage” of science graduates to non-science occupations. Simultaneously, accounting for the selection of individuals into field of study and into occupation, he shows that the higher wage returns of those studying science, in the order of 6–10%, are conditional on being matched to a specific science occupation. This contradicts claims that scientists are poached by higher wages in other occupations (“the pull factor”). By contrast, the analysis finds that science graduates would enjoy higher earnings were they employed in a science occupation. Such graduates also report lower levels of career satisfaction and are more likely to be overeducated (supportive of “a push factor” theory). Overall, the paper is a testament to the significance of graduates finding a job that matches own skills and qualifications. It questions the policy emphasis on educating more science graduates, given that science skills do not command a high return in all occupations.

Despite the high policy attention on skill shortages, since the 1990s, the academic literature has increasingly emphasized that a significant portion of the workforce of both developed and developing economies is affected by skill underutilization. Earlier studies focused much on the issue of overeducation, namely the phenomenon whereby individuals, usually tertiary graduates, are employed in jobs with lower education requirements compared to their own qualification. But the measurement approaches used to identify the incidence of overeducation (subjective, objective or realized matches) are often criticized. Correctly capturing the incidence of qualification mismatch hinges critically on what is classified as a “graduate job.”

In the next paper, Golo Henseke and Francis Green propose a new skills-based indicator of graduate jobs, based on recent data collected as part of the OECD’s Survey of Adult Skills (PIAAC) in 31 countries. Drawing on detailed information about work tasks, which is used to “objectify” individuals’ self-reported qualifications needed for their jobs, they filter the share of graduate jobs in countries by exploring the within-occupation heterogeneity in skill needs. This approach performs better in terms of accounting for wages and job satisfaction of graduates relative to a traditional delineation, which has tended to focus narrowly on professional occupations. Altogether, the authors show that about a third of labor (32%) is deployed in graduate jobs in their sample, yet there is considerable variation in deployment of graduate jobs across countries. Industry and establishment-size composition account for part of this variance, yet cross-country differences are also related to the relative quality of their higher education systems. Correctly identifying the demand for graduate skills is critical for educational policy, since expansionary higher education targets are unlikely to yield the growth dividends expected and will foster rising overeducation, if there are not enough graduate jobs to adequately utilize the higher skill supply.

Overeducated graduates are affected by a non-trivial wage and job satisfaction penalty (typically averaging around 14–15%), relative to well-matched graduates with the same education. Many academic papers have focused on the magnitude and determinants of the overeducation wage penalty, yet the mechanisms underpinning it have not been clearly understood. Multiple theories have been offered for why graduates accept jobs that are not commensurate with their academic credentials. Some of them, such as the argument that the overeducation wage penalty reflects the (unobserved) human capital deficiencies of such workers, or that mismatched jobs are merely stepping stones and part of career mobility, or that compensating job attributes underlie the choice of a mismatched job, imply that overeducation is mostly of voluntary nature. Mismatch should then be of limited concern for policymakers. Other theories, however, place greater emphasis on the information deficiencies of labor market agents, which foster matching frictions. Productivity constraints of jobs, such as limited availability of jobs with high skill needs, and personal/social constraints may also account for the placement of graduates into lower-skilled posts.

In their paper, Seamus McGuinness and Konstantinos Pouliakas draw on the rich set of determinants of skill mismatch recently made available by the Cedefop European Skills and Jobs Survey. They examine which alternative theoretical frameworks (proxied by blocks of corresponding variables) can account for the overeducation wage penalty as observed in 28 EU countries. Decomposing the wage difference between overeducated and well-matched employees, they fail to find supporting evidence for the job compensation and career mobility theories. For instance, overeducated workers are found to be less likely to be promoted in their jobs or to have accepted them for reasons of career progression. By contrast, in addition to observed human capital differences between overeducated and well-matched classmates, they find that bounded information about the education and skill requirements of jobs and assignment to jobs of lower skill intensity account for a large part of the pay penalty endured by mismatched employees. Nevertheless, the authors highlight the need for customized policy interventions to tackle education mismatch, given that information asymmetries weigh heavier on the wages of overeducated tertiary graduates, whereas job productivity constraints on those of medium-qualified overeducated workers.

Whereas the previous studies concentrate on vertical mismatch, the next paper focuses on another facet of educational mismatch, namely horizontal mismatch. The choice of a degree or major is influenced by professional aspirations and the desire to enter into a given occupation. However, students with a given major usually end up working in a variety of occupations (this holds also for majors that are mostly occupationally-oriented). Michael R. Ransom and Aaron Phipps characterize majors using two indexes: an index of distinctiveness, which captures the degree of segregation of graduates in a given major across the spectrum of occupations; and the degree of variability, which describes the variety of occupations held by graduates with a given degree or major. The paper documents that the degree of variability has increased over time in the United States. Current graduates with a degree tend to be scattered across occupations more widely now than in the past. The increase in the degree of variability is examined in relation to push and pull factors. Push factors are factors linked to educational mismatch, such as when the educational system produces more college graduates with a given diploma than the occupation can absorb. Pull factors are related to technical changes that have pulled college graduates with a given major toward occupations that were previously not associated with that major. The authors conclude that while both push and pull factors are at work, the increased occupational diversity does not appear to have been accompanied by greater skills mismatch.

Mismatch comes about not only in terms of a gap between educational credentials required by employers and the ones possessed by workers but also in terms of skills. It is now well-understood that qualification mismatch is not the same as mismatch in skills, given that there is significant heterogeneity in both the skills of individuals with the same qualification level and in the tasks of jobs within broad occupations. A series of papers in the volume hence expand the definition of mismatch to focus on the discrepancy between workers’ skills and the skills requirements of the jobs they hold.

Before delving deeper into an examination of the determinants and costs of skill mismatches, the paper by Lorenzo Cappellari, Paolo Castelnovo, Daniele Checchi and Marco Leonardi attempts to accurately uncover the wage returns associated with the acquisition of higher skills in relation to formal educational credentials. The authors try to decouple the endogeneity bias between schooling and numerical skills in earnings regressions, as it is acknowledged that educational choices and investment in skills are both influenced by a common underlying process driven by unobserved individual ability. The authors thus rely on an instrumental variables approach and propose two instrumental variables, stemming from educational reforms across birth cohorts and countries. These reforms improved the quality of primary teachers (hiring criteria for primary teachers) and broadened access to the university. The empirical results, based on the OECD Survey of Adult Skills (PIAAC) data, for men and women separately, show that the instrument for the years of education appears to work well. The IV wage effect of education is found to be larger than OLS. By contrast, the instrument for numerical skills appears to perform less well. The authors implement a recursive strategy: the reforms affecting the quality of primary teachers across birth cohort and countries are used to instrument education, which is then used in a skill formation equation. In turn, skills enter into a Mincerian earnings regression. This is equivalent to assuming that education affects earnings only through its impact on skill formation. The estimation of the recursive model shows that education has a sizeable impact on skills (at least for men) and that (numerical) skills have an independent, positive, and sizeable effect on earnings. These findings hence confirm the importance of focusing on skills and skill mismatch as underlying forces affecting individual’s earnings.

Increases in the stock of highly educated workers have significantly boosted labor productivity in recent decades, yet with a slowing rate of growth in the stock of human capital in developed countries the ability of economies to efficiently deploy their existing stock will take on heightened significance. In the next paper, Müge Adalet McGowan and Dan Andrews emphasize that in light of the significant education/skill mismatches revealed by the OECD’s Survey of Adult Skills (PIAAC), there is considerable scope to improve the efficiency of human capital allocation in OECD countries. Whereas most literature has focused on the effect of mismatch on indirect individual measures of productivity (e.g., wages, job satisfaction), few studies have attempted to measure its direct impact on aggregate productivity. Linking mismatch data from the PIAAC micro-dataset with industry-level labor productivity indicators, constructed from firm-level data, the authors decompose the channels of association into within-firm productivity shifts (accounted by differences in managerial quality, intangible assets, etc.) and allocative efficiency shifts between firms in industries over time. It is often claimed that hiring high shares of workers with surplus skills may be beneficial for firms. However, the authors point out another cost of mismatch from the perspective of the economy as a whole. By “trapping” valuable skills in low-productivity firms and starving other more productive firms from access to desired human capital, skill mismatch can entail significant reallocation effects. Indeed, their analysis confirms that there is a statistically significant negative correlation between skill mismatch and productivity. However, this association is driven by the negative impact of overskilling, whereas underskilling fails to exert any significant impact. The main channel through which overskilling bears on productivity is through its effect on allocative efficiency, rather than within-firm factors. This misallocation of skills could potentially account for a non-trivial share of cross-country labor productivity gaps.

In the next paper, Giuseppe Lucio Gaeta, Giuseppe Lubrano Lavadera, and Francesco Pastore investigate the labor market fortunes of Italian PhD students, using a survey carried out in 2009 by the National Statistical Office (ISTAT). Italy is an interesting case to study because the number of PhD graduates has been characterized by an upward trend, whilst the number of R&D workers has remained stable over time, well below the EU average. The authors distinguish between the two facets of mismatch, overeducation and overskilling. In fact, the paper uses two measures of overskilling: one based on the usefulness of skills learned during the PhD for carrying out the job, and another one based on PhD graduates’ reported satisfaction with the use of the skills acquired during the PhD. The paper finds that overeducation is associated with a wage penalty, which is as high as the one associated with being dissatisfied with how the skills learned during the PhD are being utilized. The largest wage penalty is experienced by PhD graduates who are genuinely overskilled (i.e., both overskilled and overeducated).

With most recent studies highlighting the sizeable productivity costs associated with overskilling, Ilias Livanos and Imanol Núñez nevertheless show in their article that a situation of underskilling at the time of recruitment, whereby workers’ skills are below the level needed by the job, is not uncommon in European job markets. Fully 23% of EU workers reported in the Cedefop European Skills and Jobs Survey that they experienced a skill gap at the start of their job. The authors identify three mechanisms that may account for such high levels of apparent “recruitment mismatches.” First, imperfect signaling associated with educational qualifications. Second, labor demand conditions, such as local imbalances in skills demand and supply. These may induce workers to apply for jobs for which they do not possess the full skills set while organizations may be willing to hire such workers because they cannot find suitable candidates in the face of skill shortages. Third, the organization may deliberately hire underskilled candidates (at a wage rent) with a view to offering subsequent skills formation through formal training or on-the-job learning. Using European Skills and Jobs Survey data, the authors find some support for the signaling hypothesis, since the likelihood of underskilling is lower for medium-qualified workers relative to those with higher education degrees. Skill signals associated with past experience (e.g., working in the same occupation in a previous job) also reduce chances of underskilling. The empirical analysis further supports the importance of labor market conditions. Levels of underskilling are found to be inversely related to an indicator of labor market thickness, based on whether individuals turned down job offers at the time of their job search.

Even if skill gaps ensue due to matching frictions or labor market constraints at the time of hiring, Rolf van der Velden and Dieter Verhaest nonetheless contradict the notion that underskilling is inherently costly for workers. Consistent with previous evidence failing to show a strong wage or job satisfaction penalty to underqualification, the authors adopt a learning perspective to skill gaps. They exploit unique information made available by the recent Cedefop European Skills and Jobs Survey, which allows for examination of the relation between varying degrees of initial skill gaps and the extent of skill formation in workers’ jobs, in particular via participation in spontaneous forms of informal learning. They illustrate that challenging work situations, in which required skills are at a higher, but not too high, a level than own skills (a “zone of proximal development”), can result in more skill development. Conversely, learning opportunities are worst when workers are hired in a job for which they have surplus skills (overskilled). This nuanced learning perspective of skill gaps challenges the traditional fixation of policymakers on achieving effective skills matches between workers and jobs. Indeed, the term skill mismatch is usually accompanied by the inherent connotation of a situation that needs to be avoided. By contrast, the paper highlights that although a skill match is good for workers, some skill gaps may even be better in the long -run.

In their paper examining the relation between different types of work-related learning and skill accumulation in Europe, Maria Ferreira, Annemarie Künn-Nelen, and Andries De Grip further confirm that underskilled workers enjoy greater complementarities between formal and informal learning than their well-matched or overskilled counterparts. The authors open the black box of the direct relation between on-the-job human capital investments and workers’ skill development. This is a significant contribution relative to past research that has focused on the training-wages/productivity nexus. The article confirms that participation in formal training or informal learning at work is associated with greater improvement of employees’ skills, compared to those who do not undergo training. Although both forms of on-the-job human capital investment exhibit complementarity in terms of improving employees’ skills on the job, the contribution of informal learning is larger than participation in formal training courses. It is hence confirmed in the paper that the positive impact of work-related training on wages and productivity is mediated via a direct positive contribution to skills formation, especially if such training takes place during work hours and is paid by the employer. Nevertheless, this positive effect on skills growth is moderated by workers’ skill mismatch status, with initially underskilled workers benefiting the most in terms of further skill accumulation, while the overskilled simply manage to counteract any skills depreciation.

The aforementioned conclusions highlight that being employed in non-challenging jobs that underutilize worker’s skills has important implications for the returns to investment in training. This puts some responsibility on employers to retain job tasks and responsibilities at a challenging level for their employees and to invest in “learning workplaces” should they wish to ameliorate the productivity loss associated with overskilling and stimulate greater skills growth of their workforce. In the last paper of the series Giovanni Russo highlights that both the skill requirements of jobs and workers’ human capital can change over time, the former as a result of technological progress and the latter in response to investment in learning. These two change processes are intertwined, as an increase in job complexity may push workers out of their professional comfort zone, inducing additional learning and skills growth. The author hence focuses on the relation between job complexity and the self-reported measure of skill improvement in the job, provided by the Cedefop European Skills and Jobs Survey. The empirical analysis shows a positive and robust relationship between the degree of skills development and the degree of change in job complexity. Overskilled workers tend to experience the slowest degree of skills development, whereas underskilled workers the highest. Overskilled workers exposed to challenging job tasks and learning opportunities nevertheless experience significant skills development as the skills multiplier process (“skills beget skills”) is then accentuated.

The main thrust of the papers in this volume is therefore that the right remedy for tackling skill mismatches and shortages can often be found within workplaces, more so than in education and training institutions, and depends on a multitude of interventions from a wide policy sphere (e.g., guidance and counseling, employment protection legislation, workplace innovation, housing, social insurance). Ensuring continuous (informal) learning opportunities and employers’ commitments to workplace practices that effectively utilize skills (e.g., job rotation, autonomy, task complexity) are necessary conditions for sustainable workplace innovation and productivity. However, mitigating skill underutilization can be an inherently more difficult task and challenge for both public and private actors relative to incentivizing training provisions to mitigate skill gaps.

As with past volumes, we aim to focus on important issues and to maintain the highest levels of scholarship. We encourage readers who have prepared manuscripts that meet these stringent standards to submit them to Research in Labor Economics (RLE) via the IZA website (http://rle.iza.org) for possible inclusion in future volumes. For insightful editorial advice, we thank Simon Adler, Nadia Belghith, Christian Belzil, Camilo Bohorquez, Massimiliano Bratti, Alex Bryson, Daniele Checchi, Daniel Chiquiar, Luca Flabbi, Liana Fox, Dirk van der Gaer, Peter Gottschalk, Elizabeth Handwerker, Enkelejda Havari, Laura Hospido, Stephen Jenkins, Thomas Lemieux, Marco Leonardi, Carl Lin, Jeremy Lise, Alfonso Miranda, Paolo Naticchioni, Andreas Peichl, Nico Pestel, Nicolas Pistolesi, Xavi Ramos, Rebecca Riley, Francois Roubaud, Jacques Silber, Denissa Sologon, Jan Stuhler, Domenico Tabasso, Roger Wilkins, and Yu Zhu.

Solomon W. Polachek

Konstantinos Pouliakas

Giovanni Russo

Konstantinos Tatsiramos

Editors