Introduction

The new indicator KILM 15 provides infor-mation on the extent to which the supply ofskills matches the demand for skills. It is acomplement to KILM 14 that presents statisticson the level and distribution of the knowledgeand skills base of the labour force and theunemployed.

Table 15a presents an index summarizingskills mismatch between labour supply anddemand by educational attainment and table15b provides information on the incidence ofskills mismatch between job requirements andqualifications.

The information in both tables is availableby sex and age group (15 years and over; 15 to24 or 15 to 29; and 30 years and over). Themismatch index in table 15a has been calcu-lated for 48 economies, and information intable 15b is available for 49 economies.

Use of the indicator

The issue of skills mismatch has receivedrenewed attention in advanced economiesfollowing the global economic crisis in 2008-2009, but various forms of mismatch are alwayspresent in labour markets. Addressing skillsmismatch issues is often a complex undertak-ing because of the many factors that influenceskills demand and supply, including for exam-ple the level of economic development of acountry, technological change, demographicsand mobility of workers. At the same time, theextent to which skills supply and demand aresuccessfully being matched is a major factorshaping economic and labour market outcomes,economic growth, productivity and competi-tiveness. Therefore, the formulation and imple-mentation of effective education and trainingpolicies, including responsive education andtraining systems, is a continuous challenge forall countries. Meeting this challenge requireslinking skills development to employment andeconomic development, involving social part-ners and key stakeholders in skills develop-

ment systems, and effective labour marketinformation and analysis systems.1

There is no internationally agreed methodto measure skills mismatch. Skills mismatch isan encompassing term which refers to varioustypes of imbalances between skills offered andskills needed in the world of work, and itapplies equally to the employed and the unem-ployed. Furthermore, in most countries skillsand competencies per se are not measured byregular statistical programmes. That is why skillproxies are used, such as qualifications, yearsof schooling and occupations.

The type of skills mismatch presented intable 15a reflects differences between unem-ployment rates by level of educational attain-ment (see also KILM 14). Such differencesindicate that the level of educational attainmentof workers is an important determinant of theprobability of finding a job besides the level ofunemployment. Differences across groupsof workers with different levels of educationalattainment are summarized in an index, andthe higher the value of the index, the higher thelevel of mismatch according to this measure.For example, the skills mismatch index for theage group 30 and above in Lithuania (26.6 percent) was much higher than in Portugal (8.0 percent) in 2012. This is due to the fact that inLithuania the unemployment rate for workers

 with primary education is much higher (33 percent) than those for workers with secondaryand with tertiary education (15 per cent and4.1 per cent, respectively). In Portugal, thedifferentials are far more limited; unemploy-ment rates for workers with primary, secondaryand tertiary education were 14.5, 13.8 and8.4 per cent, respectively. Put differently, thisindicates that the unemployment rate inPortugal is more a reflection of problems withaggregate demand rather than specific prob-lems related to education, at least whencompared to Lithuania.

The level of the index by itself does not giveinformation concerning the group of unem-

1  See ILO Recommendation No. 195 (2004) on humanresources development; http://www.ilo.org/dyn/normlex/ en/f?p=NORMLEXPUB:12100:0::NO:12100:P12100_INSTRUMENT_ID:312533:NO.

KILM 15.  Skills mismatch

 

96 KILM 15 Skills mismatch

characteristic of the labour market as a wholein developing economies.4 

Measurements of the incidence of skillsmismatch between jobs held by workers and thequalifications they possess vary widely, and aresensitive to methods that have been used,among other factors. Based on the same meth-odology that has been used to produce table 15b(see more details in Definitions and sourcesbelow), the Global Employment Trends forYouth 2013 report estimates that the averageincidence of overqualification in a sample ofEuropean economies in 2010 was 10.1 per cent(and ranged from 3.6 to 16.8 per cent), whileunderqualification averaged 28.1 per cent (andranged from 15.9 to 45.6 per cent). The samereport finds that the risk of overqualification ishigher for migrants, younger workers andpersons with disabilities. Given the strong risein educational attainment and the structuralchange in labour markets occurring in parts ofthe developing world, it is less clear whether

 younger workers are more likely to be overqual-ified outside the advanced economies as well.

Research suggests that mismatch between jobs held by workers and the qualifications theypossess has negative consequences for work-ers, enterprises and the economy.5 For the over-educated, wages are often higher than for the

 well-matched at the same job, but returns tothe years of schooling beyond the requiredlevel are lower. Undereducated workers earnless than the well-matched at the same job, butmore than workers with the same educationallevel and a matching job. Studies show thatoverqualified workers are less satisfied withtheir job then the well-matched, which in turnis likely to affect productivity. Evidence alsosuggests that the overeducated are more likelyto engage in job search and therefore add toturnover of staff. At the macroeconomic level,skills mismatch may raise unemployment rates,reduce labour market flexibility and reduceoutput and productivity growth.

4  See, for example, Sparreboom, T. and Nübler, I.: Productive transformation, employment and educationin Tanzania, paper presented at the 2013 UNU-WIDERDevelopment Conference on Learning to Compete: Indus-trial Development and Policy in Africa, 24-25 June, Helsinki;http://www1.wider.unu.edu/L2Cconf/sites/default/files/ L2CPapers/Sparreboom.pdf .

5  For an overview of the effects of skills mismatch, seeQuintini, G.: Over-qualified or under-skilled: A review ofexisting literature, OECD Social, Employment and Migra-tion Working Papers, No. 121 (Paris, 2011); http://www.oecd-ilibrary.org/social-issues-migration-health/over-quali-fied-or-under-skilled_5kg58j9d7b6d-en; much of the litera-ture is focusing on developed economies.

ployed which is in a relatively favourable posi-tion. However, the development of the levelsover time shows whether skills mismatch isincreasing or decreasing. If increases ordecreases occur consistently over longer peri-ods, this may point to the need for moresupport of particular labour market groups. Forexample, the index may show an increase overtime due to a (relative) deterioration of thelabour market position of unemployed with alow level of educational attainment. In Spainfor example, the index increased from 11.8 percent in 2002 to 20.2 per cent in 2012 for the agegroup 30 and over. Even though unemploy-ment rates increased for workers of all levels ofeducation in Spain during this period, theincrease was more severe for workers withprimary education, which underlines the needto promote investment in education particu-larly for this group.

The second type of skills mismatch,presented in table 15b, consists of mismatchbetween the qualification requirements of jobsheld by workers and the qualifications these

 workers possess, an indication ofover- orunderqualification. Concerns about this type ofmismatch, in particular overqualification, havebeen growing due to the increasing supply oftertiary-educated workers in advanced econo-mies and their employment in jobs previouslyheld by workers with lower educational attain-ment. Increasing supply may result in competi-tion for jobs, which pushes better educated

 workers into jobs or occupations usually takenby those with lower levels of education.2 Thistype of mismatch is likely to increase in timesof economic crises, when employment oppor-tunities are scarce and unemployment rates arehigh.3 In developing economies, where educa-tional attainment levels are often lower, under-qualification is more likely to be an importantlabour market issue. Widespread underqualifi-cation points at the need for more education,even if it occurs alongside relatively high unem-ployment rates for tertiary-educated workers.The latter often reflect the supply and demandfor skills and secure jobs in the formal sector,

 while underqualification is more likely to be

2  Karakaya, G., Plasman, R., and Rycx, F.: “Overeduca-tion on the Belgian labour market: Evaluation and analysisof the explanatory factors through two types of approaches”,in Compare: A Journal of Comparative and International Education, Vol. 37, No. 4, pp. 513–532 (2007).

3  See ILO: Global Employment Trends for Youth 2013: A generation at risk (Geneva, 2013); http://www.ilo.org/ global/research/global-reports/global-employment-trends/ 

 youth/2013/lang--en/index.htm.

 

  97KILM 15Skills mismatch

ISCO major groups and skill levels used fortable 15b

ISCO major group Broadoccupationgroup

Skilllevel 

1: Legislators, seniorofficials, managers

High-skillednon-manual

Tertiary(ISCED 5-6)

2: Professionals

3: Technicians andassociate professionals

4: Clerks Low-skillednon-manual

Secondary(ISCED 3-4)

5: Service workers, shop,market sales workers

6: Skilled agriculturaland fishery workers

Skilledmanual

7: Craft and relatedtrades workers

8: Plant and machineoperators and assemblers

9: Elementary occupations Unskilled Primary(ISCED 1-2)

Note: Excluding armed forces occupations.

Skills mismatch in the sense of overeduca-tion or undereducation means that workershave either more education or less educationthan is required. Measurement of this type ofskills mismatch in table 15b is based on theInternational Standard Classification ofOccupations (ISCO, see KILM 5). This measureof mismatch starts from a division of majoroccupational groups (first-digit ISCO levels)into four broad groups (see table below) andassigns a level of education to each occupa-tional group in accordance with theInternational Standard Classification ofEducation (ISCED). Workers in a particulargroup who have the assigned level of educationare considered well matched. Those who havea higher (lower) level of education are consid-ered overeducated (undereducated). Forinstance, a university graduate working as aclerk (a low-skilled non-manual occupation) isovereducated, while a secondary school gradu-ate working as an engineer (a high-skilled non-manual occupation) is undereducated.

The statistics on employment and unem-ployment by educational attainment that wereused to produce table 15a are obtained fromKILM 14a and KILM 14b. The informationpresented in table 15b shows ILO calculationsbased on two sources:

•  The European Social Survey (ESS), rounds 1through 5 (published by Norwegian Social

 As was mentioned before, underqualifica-tion in developing economies is often a mani-festation of low levels of educational attainment.This means that, for example, professional jobs(teachers, engineers, health professionals) aretaken by workers with (at most) secondaryeducation. Low levels of educational attain-ment may also result in underqualification inoccupations which require low levels of skills,such as elementary occupations, which is rarein developed economies.6

Definitions and sources

Table 15a presents skills mismatch betweensupply of labour and demand for labour in theform of an index of dissimilarity based on thedifferences in the shares of educational attain-ment of the employed in comparison with theunemployed. This index captures one dimensionof mismatch, namely mismatch between skillsdemand (defined by the skills of the employed)and skills supply (defined by the skills of theunemployed), both proxied by level of educa-tional attainment, and is defined as follows:

 where: i: an indicator for the level of educa-tion (primary or less; secondary; tertiary); ABS :the operator for the absolute difference;  E 

i /E :

the proportion of the employed with educationlevel i; U 

i /U : the proportion of unemployed

 with education level i; for information regard-ing educational attainment readers may consultthe manuscript of KILM 14.

 Apart from being a measure of mismatchbetween skills supply and demand, the indexcan be interpreted as a summary measure ofthe relative position of labour market groups

 with different levels of education. If primary,secondary and tertiary graduates all have thesame unemployment rate, the index will have a

 value of zero (no dissimilarity between groups), while the index would reach a value of 1(complete dissimilarity) if, for example, allthose with primary and tertiary education areemployed and all those with secondary educa-tion are unemployed.

6  ILO: Global Employment Trends for Youth 2013: A generation at risk (Geneva, 2013), table 10, shows under-qualification and overqualification by ISCO major group inselected developing economies; http://www.ilo.org/global/ research/global-reports/global-employment-trends/ 

 youth/2013/lang--en/index.htm.

 ID Mismatch =  ABS –

i=1

3

12∑    

      E 

i 

 E 

U i 

U 

 

98 KILM 15 Skills mismatch

 An advantage of the ISCO-based measure ofskills mismatch is that the definition ofmismatch does not change over time and theresults are therefore strictly comparable. Adisadvantage of this measure is that, byconstruction, it does not allow for overeduca-tion in major groups 1 to 3. The ISCO-basedmeasure also assumes that job titles alwayshave the same meaning in terms of job contentand have the same educational requirement inall countries, which is not necessarily true.8

For the purpose of measuring mismatch, theupper age bound for young people is extendedto 29 years (including ‘young adults’ aged 25to 29). This is in recognition of the fact thatsome young people remain in education beyondthe age of 24 years, particularly those in tertiaryeducation in developing economies, but thisdifference in age group may hamper a coherentanalysis of youth employment issues (as youthare usually defined as persons aged 15 to 24).

8  For an overview of alternative methods to measureskills mismatch between job requirements and qualifica-tions, see Quintini 2011, op.cit.

Science Data Services), a biennial surveycovering over 30 countries (even thoughcountry coverage differs by round).7

• School-to-work transition surveys, which arehousehold surveys conducted within thescope of the ILO’s Work4Youth partnership with The MasterCard Foundation (see box 15).

Limitations to comparability 

The limitations to comparability withrespect to the employed and the labour forcenoted in KILM 1, 2 and 9 manuscripts applyequally to KILM 15. The same is true for limita-tions on the comparability of indicators oneducation between countries and over timenoted in KILM 14.

7  See http://www.europeansocialsurvey.org/ .

Box 15. Work4Youth: An ILO project in partnershipwith The MasterCard Foundation

The Work4Youth (W4Y) project is a partnership between the ILO Youth Employment Programme andThe MasterCard Foundation. The project has a budget of US$14.6 million and will run for five yearsto mid-2016. Its aim is to “promot[e] decent work opportunities for young men and women throughknowledge and action”. The immediate objective of the partnership is to produce more and betterlabour market information specific to youth in developing countries, focusing in particular on transitionpaths to the labour market. The assumption is that governments and social partners in the project’s28 target countries will be better prepared to design effective policy and programme initiatives oncearmed with detailed information on:

• what young people expect in terms of transition paths and quality of work;

• what employers expect in terms of young applicants;

• what issues prevent the two sides – supply and demand – from matching; and

• what policies and programmes can have a real impact.

Copyright © International Labour Organization 2014