ArticlePDF Available

Educational Mismatches versus Skill Mismatches: Effects on Wages, Job Satisfaction, and On-the-Job Search.

Authors:
Article

Educational Mismatches versus Skill Mismatches: Effects on Wages, Job Satisfaction, and On-the-Job Search.

Abstract and Figures

Education-job mismatches are reported to have serious effects on wages and other labour market outcomes. Such results are often cited in support of assignment theory, but can also be explained by institutional and human capital models. To test the assignment explanation, we examine the relation between educational mismatches and skill mismatches. In line with earlier research, educational mismatches affect wages strongly. Contrary to the assumptions of assignment theory, this effect is not explained by skill mismatches. Conversely, skill mismatches are much better predictors of job satisfaction and on-the-job search than are educational mismatches. Copyright 2001 by Oxford University Press.
Content may be subject to copyright.
Educational mismatches versus skill
mismatches: effects on wages, job
satisfaction, and on-the-job search
By Jim Allen* and Rolf van der Velden{
* Research Centre for Education and the Labour Market, University of Maastricht,
Postbus 616, 6200 MD Maastricht, The Netherlands;
e-mail: J.Allen@ROA.UNIMAAS.NL.
{ Research Centre for Education and the Labour Market, University of Maastricht
Education-job mismatches are reported to have serious effects on wages and other
labour market outcomes. Such results are often cited in support of assignment
theory, but can also be explained by institutional and human capital models. To test
the assignment explanation, we examine the relation between educational mismatches
and skill mismatches. In line with earlier research, educational mismatches affect wages
strongly. Contrary to the assumptions of assignment theory, this effect is not explained
by skill mismatches. Conversely, skill mismatches are much better predictors of job
satisfaction and on-the-job search than are educational mismatches.
1. Introduction
Education-job mismatches are reported to have serious effects on a number of
labour market outcomes. Overeducation is known to affect labour turnover (Topel,
1986; Hersch, 1991), occupational choice (Viscusi, 1979), and job satisfaction
(Tsang and Levin, 1985). A major line of research has been developed regarding
the effect of education-job mismatches on wages. Empirical results suggest that
both individual human capital and job characteristics are related to wages. Indi-
viduals working in jobs for which a lower level of education than their own is
required (overeducation) are often found to earn less than individuals with the
same level of education working in jobs for which their own level is required
(adequate education), but more than individuals working in an equivalent job
with the level of education actually required (Duncan and Hoffman, 1981;
Hartog and Oosterbeek, 1988; Sicherman, 1991; Hersch, 1991; Cohn and Khan,
1995; Van Smoorenburg and Van der Velden, 2000). Conversely, individuals work-
ing in jobs for which a higher level is required (undereducation) often earn more
than individuals with the same level of education working in jobs for which their
own level is required, but less than individuals with the level of education actually
required in such jobs. The wage effects of overeducation are usually stronger than
the wage effects of undereducation.
#
Oxford University Press 2001 Oxford Economic Papers 3 (2001), 434452 434
All rights reserved
Such results are often cited in support of so-called ‘assignment’ models of the
labour market (Sattinger, 1993), whereby the returns to additional investments in
human capital depend in part on the match between the worker and the job.
The basic idea is that, although higher education raises productivity in general,
the actual level of productivity realised is also determined by the match between
educational level and job level. Working in a job below one’s own level imposes a
limitation to the utilisation of skills. The lower level of the job in effect imposes
a ceiling on the worker’s productivity, resulting in lower wages. Conversely, work-
ing in a job above one’s own level in effect raises this ‘productivity ceiling’, allowing
workers to be more productive than they would be when working at their own
level. However, in this case, the worker’s own abilities are the main factor limiting
productivity. Because workers employed in a job at their own level are already
performing at a level close to their own personal productivity ceiling, the wage
benefits of working above one’s own level are generally modest. This accounts for
the observed asymmetry in the wage effects of over- and undereducation.
According to assignment theory, the allocation is optimal when workers are
allocated top-down according to their skills, whereby the most competent
worker is assigned to the most complex job and the least competent worker is
assigned to the simplest job.
1
The incidence of educational mismatches can thus be
explained by differences in the shares of complex jobs and skilled workers. How-
ever, as Hartog (2000) has pointed out, over- and undereducation can also be
explained in other ways. According to search and matching theory, temporary
mismatches may occur as a result of imperfect information. The fact that over-
education is typically higher in the phase of the transition from school to work is
often taken as evidence for this interpretation. However, others have argued, from
the point of view of human capital theory, that the high incidence of overeducation
among school-leavers reflects these workers’ lack of work experience (Groot and
Maassen van den Brink, 1996).
This human capital argument can be stated in a more general way, to provide an
alternative explanation for the effects normally attributed to assignment theory.
Although the intuitive appeal of assignment theory is considerable, the wage
equations used are not directly derived from assignment theory (Hartog, 2000).
The observed wage differences for jobs below and above one’s own level might just
as easily reflect individual differences in human capital within education levels,
which are roughly sorted according to job level. Those working ‘below their own
level’ are in that case less productive on average than those working ‘at their own
level’, not because the job imposes limitations on their productivity, but because
they have less human capital on average to begin with. Similarly, those working
jim allen and rolf van der velden 435
..........................................................................................................................................................................
1
In addition to the effects of a mismatch between required and actual level of education, many studies
also take into account the effects of working in a job for which a field of education different from one’s
own field is required.
‘above their own level’ have according to this view more human capital on average
than those working at their own level.
Institutional theories offer yet another possible explanation for the same
observed wage effects. They point out that employers often forced to base wages
on easily observable characteristics of employees or jobs, rather than on individual
performance (Thurow, 1975). From such a perspective, the observed wage differ-
ences could be accounted for by the fact that both the formal education of employ-
ees and the required education for the job are frequently incorporated in wage
scales as determined in collective bargaining agreements. The differences may not
reflect individual differences in productivity, but rather the value assigned to edu-
cation and job categories in such agreements.
Wage analyses based on attained and required education cannot in themselves
provide a definitive answer to the question of which explanation is correct. The aim
of this article is to shed some further light on this discussion, by examining the
relation between the education-job match on one hand and the utilisation of indi-
vidual skills on the other. The basic idea is that the above-mentioned theories differ
in the way they postulate a relation between the ‘formal’ education-job mismatch
(hereafter called educational mismatches) and the actual mismatch between
acquired and required skills (hereafter called skill mismatches).
In assignment theory, educational mismatches imply skill mismatches. Skill mis-
matches are believed to account for the observed wage effects of over- and under-
education. If the assignment explanation is valid, we should therefore find that
individuals working below their own level are underutilising their knowledge and
skills to a significant degree, while individuals working above their own level lack
some of the knowledge and skills that are required in order to perform optimally in
their job. We would also expect the effects of education-job match in wage analyses
to be accounted for to a large extent by mismatches in the utilisation of knowledge
and skills. The same applies to effects on other outcomes like job satisfaction and
intention to quit.
This paper puts the assignment theory to the test. If educational mismatches do
not imply skill mismatches, and skill mismatches do not account for the effects of
educational mismatches on wages and other outcomes, assignment theory is
seriously challenged. In the paper we address the following questions:
(i) to what extent do skill mismatches correspond to mismatches between
available and required education in jobs?
(ii) to what extent can wage effects of educational mismatches be accounted for by
skill mismatches?
(iii) to what extent and in what ways do skill mismatches and educational mis-
matches influence job satisfaction?
(iv) to what extent and in what ways do skill mismatches and educational
mismatches influence employees’ decisions to actively seek other employ-
ment?
436 educational vs skill mismatches
2. Data
The data used for the analyses were collected for the project ‘Higher Education and
Graduate Employment in Europe’, an international comparative study of the
labour market situation of graduates from tertiary education in 11 European
countries and Japan.
2
In the Netherlands two main types of graduates from tertiary
education are distinguished: those who graduate from university, and those who
graduate from a college for higher vocational education. Two graduate cohorts
were approached at the end of 1998. The first cohort consisted of a representative
sample of those graduating from tertiary education in the academic year 1990–
1991. The second cohort comprised a representative sample of those who
graduated in the academic year 1994–95. For this article the Dutch data from
the former cohort are used, which means that our subjects are individuals who
graduated from tertiary education some seven years prior to the survey. Around
6000 graduates were approached, of which 2723 responded with a completed
questionnaire. The analyses are restricted to the 2460 individuals who at the
time of the survey were in paid employment for at least 12 hours per week.
From these 2460 individuals 901 graduated from university and 1559 from
higher vocational education in the academic year 1991. (For further information
we refer to Allen and Van der Velden (forthcoming).
3. Educational mismatch and skill mismatch
Various measures have in the past been proposed to indicate the match between
education and job. In this article we use an employee self-rating of the level of
education most appropriate for the current job, with response categories: university
plus postgraduate study (1); university only (2); higher vocational education plus
postgraduate study (3); higher vocational education only (4); secondary vocational
education or equivalent (5); or lower (6). By comparing this to the highest attained
level of education of the workers,
3
we can determine whether, and to what extent,
respondents are working above or below their own level. For the Netherlands, Van
der Velden and Van Smoorenburg (1997) have shown that workers’ self-ratings are
far more valid than a commonly used alternative, namely the expert-rating of job
titles. From the point of view of matching theory, ‘appropriate level’ is preferable to
the often-used alternative of ‘required level’. The latter measure may partly measure
formal selection requirements, whereas the former is more likely to refer to actual
job content.
jim allen and rolf van der velden 437
..........................................................................................................................................................................
2
This project was partially funded by the European Commission under the Targeted Socio-Economic
Research (TSER) program (TSER EGS-SOE2-CT97–2023), with additional funding by the Dutch Min-
istry of Education. The project is coordinated by Prof. U. Teichler from the University of Kassel.
3
Many of those who graduated from a university or higher vocational education institute in 1990/1991
subsequently earned a higher degree within tertiary education.
Next we asked the respondents to indicate which field of education is most
appropriate for their job, with response categories: ‘only my own field of education’
(1); ‘my own or a related field’ (2); ‘a completely different field of education’ (3);
‘for this job no specific field is required’ (4); and ‘for this job no specific field (yet)
exists’ (5). We collapsed these five categories into two by distinguishing categories
(1) and (2) from (3), (4), and (5).
Examination of the data reveals that educational mismatches are a common
phenomenon. A considerable percentage of higher vocational education graduates
(14%) and university graduates (8%) were working in jobs for which they
considered a (somewhat) higher level of education was more appropriate.
4
About a third of the higher vocational education and university graduates indicated
that they were working in a job for which they considered a lower level of education
would have been more appropriate. Finally, around one in five graduates were
working in jobs for which their own or a related field of education was not
considered most appropriate. In all, some 50% of university graduates and
56% of higher vocational education graduates worked in jobs for which they
considered a level and/or field of education other than their own to be more
appropriate.
At first sight, it seems rather alarming that such a large proportion of the most
highly trained category of participants in the Dutch labour market have jobs which
don’t match their education. How serious is this? Specifically: to what extent do
such mismatches between own and required education correspond to mismatches
between the knowledge and skills possessed by subjects and the knowledge and
skills required in their work? To investigate this, subjects were asked to indicate the
extent to which they agreed with the following statements:
Statement 1 My current job offers me sufficient scope to use my knowledge
and skills.
Statement 2 I would perform better in my current job if I possessed additional
knowledge and skills.
The responses (on a five point scale) to statement 1 indicate the degree to which
available skills are being utilised in the current job, and the responses to statement 2
the extent to which the respondent possesses the skills required in the job. Under-
utilisation, the skills counterpart of overeducation, is thus indicated by the extent
to which one disagrees with statement 1. A Skill deficit, the skills counterpart of
undereducation, is indicated by the extent to which one agrees with statement 2.
The relation between these measures of skill mismatches and educational mis-
matches are shown in Table 1.
438 educational vs skill mismatches
..........................................................................................................................................................................
4
These figures include those who have completed some form of postgraduate study within higher
vocational education and university respectively.
About 15% of all graduates experience a high or very high degree of
underutilisation of skills as indicated by their response to statement 1. The pattern
of answers is clearly related to educational mismatches. The relation is, however, far
from perfect. As might be expected, a large majority of tertiary graduates working
in jobs for which their own or a higher level and their own or a related field of
education was considered appropriate appear to be quite satisfied with the scope
which their job provides for using their knowledge and skills. Somewhat
surprisingly however, a relatively large proportion of graduates working in jobs
below their own level and/or outside their own field also report little or no
underutilisation. This is particularly the case for university graduates. Furthermore,
even graduates whose education matches their jobs sometimes report rather severe
underutilisation. All in all, these results suggest that while a good match in terms of
formal education improves the chances that one will be given the opportunity to
use one’s knowledge and skills, it is neither a necessary nor a sufficient condition
for skill utilisation.
In effect, statement 1 takes the skills of the subject as given, and seeks to deter-
mine the extent to which the subject is in a position to utilise these in the current
work setting. Statement 2 switches the perspective around, taking job requirements
as given, and seeking to establish to what extent the subject is equipped to fulfil
these requirements. Table 2 provides an overview of skill deficits as measured by
the responses to statement 2.
Somewhat unexpectedly, a large proportion of all categories of graduates show
rather high levels of skill deficits according to this measure. This could indicate that
statement 2 has a low threshold value, in the sense that subjects agree with the
jim allen and rolf van der velden 439
Table 1 The relation between educational mismatches and skill underutilisation
None Strong
................................ .................
Skills underutilisation 1 2 3 4 5
Education-job match
University
Higher level of education appropriate 42 42 13 1 1
Own level and field of education appropriate 34 48 11 6 2
Own level and different field of education appropriate 30 48 14 6 2
Lower level of education appropriate 14 37 20 21 8
Total university 28 44 15 10 4
Higher vocational education
Higher level of education appropriate 25 51 16 6 2
Own level and field of education appropriate 30 42 15 10 3
Own level and different field of education appropriate 18 49 20 6 8
Lower level of education appropriate 20 38 21 15 6
Total higher vocational education 25 42 18 11 5
Total university þ HVE 26 43 16 11 4
statement even in the case of relatively minor skill deficits.
5
This measure of
skill deficit is only weakly related to education-job mismatches. Those working
outside their own field appear to be somewhat more inclined to report skill
deficits than those working within their own field. However, there seems to be
little or no relation between job level and skill deficits. One might expect
graduates who work above their own educational level to show higher levels of
skill deficits than graduates working at or (especially) below their own level. This
is not the case.
Taken together, the results shown in Tables 1 and 2 appear to be in conflict with
one of the key assumptions of assignment theory, namely that mismatches between
education and job are accompanied by serious mismatches between available skills
and required skills. The results indicate only a relatively weak relation between
educational mismatches and skill mismatches.
6
From the point of view of the match between skills and skill requirements,
skill utilisation and skill deficits as indicated by the responses to statement 1 and
statement 2 are not simply two sides of the same coin. It is possible, even likely,
440 educational vs skill mismatches
Table 2 The relation between educational mismatches and skill deficits
None Strong
............................... ...................
Skills deficits 1 2 3 4 5
Education-job match
University
Higher level of education appropriate 13 19 21 26 21
Own level and field of education appropriate 7 19 23 37 15
Own level and different field of education appropriate 9 6 22 43 20
Lower level of education appropriate 7 18 22 38 15
Total university 8 18 22 37 16
Higher vocational education
Higher level of education appropriate 5 19 28 34 15
Own level and field of education appropriate 5 18 32 32 13
Own level and different field of education appropriate 10 14 18 33 26
Lower level of education appropriate 10 20 28 30 13
Total higher vocational education 7 19 29 32 14
Total university þ HVE 718263415
..........................................................................................................................................................................
5
A comparison with the graduates’ responses based on a list of 34 separate skills indicates that the
threshold is indeed somewhat low. Even those who indicated very few or no discrepancies between
available and required skill items often responded positively to statement 2. Nonetheless, there was a
clear positive relation between the number of item discrepancies and the response to statement 2,
suggesting that the statement has at least a certain degree of validity.
6
We also used other indicators of skill utilisation and skill deficit by asking respondents to indicate for a
list of skills whether they were required in the job and whether they possessed these skills. These analyses
show roughly similar results, although the effects on labour market outcomes are less strong.
that some people will simultaneously experience shortages and surpluses of
(different) skills. To illustrate this, Table 3 shows different combinations of the
two measures.
Table 3 reveals that the responses to the two statements are only quite weakly
related: those who disagree with statement 1 are slightly less likely to agree with
statement 2 than those who do not disagree with statement 1 and vice versa. A
small group (6%) of graduates simultaneously report that their job does not offer
them sufficient scope for using their knowledge and skills and that they could do
their work better if they possessed additional knowledge and skills. This group of
graduates does not so much suffer from having too little or too few skills for their
jobs, as from having the wrong skills.
4. The model
In each of the analyses on the effects of educational mismatches and skill mis-
matches on wages, job satisfaction and on-the-job search, we will use similar
models. We will begin with a model containing indicators of acquired level of
education, with as control variables labour market experience before the start of
current job,
7
tenure in current job, and dummies for self-employment and tem-
porary employment. The model specification is:
Y ¼ a
0
þ a
1
X þ a
2
EDUC þ e ð1Þ
With Y ¼ dependent variable under consideration (i.e. log wages, job satisfaction,
looking for another job); X ¼ a vector of control variables, and EDUC ¼ a set of
dummies indicating the acquired educational level.
The dummies representing the educational level are as follows: university plus
postgraduate study (UEþ), university education only (UE), higher vocational edu-
cation plus postgraduate study (HVEþ), and higher vocational education only
(HVE) as the reference category.
jim allen and rolf van der velden 441
Table 3 Combinations of skill underutilisation and skill deficits
Skill deficits
none/weak strong total
Skill underutilisation
None/weak skill match skill shortage 84.8%
41.8% 43.0%
Strong skill surplus wrong skills 15.2%
9.3% 6.0%
Total 51.0% 49.0% 100.0%
..........................................................................................................................................................................
7
We used age at the start of the current job as a proxy for prior labour market experience.
In model (2) we add the mismatches according to the formal educational
requirements for the job
Y ¼ model 1 þ a
3
UNDEREDUC þ a
4
OVEREDUC
þ a
5
OWNFIELD þ e ð2Þ
With UNDEREDUC ¼ degree to which job level is higher than own schooling level,
OVEREDUC ¼ degree to which job level is lower than own schooling level, and
OWNFIELD ¼ dummy indicating if one’s own or a related field of education is
considered appropriate (1), zero otherwise.
The variables OVEREDUC and UNDEREDUC are measured in terms of the
number of years normally required for the subject’s own level of education and
for the level of education considered most appropriate for the current job. Each
level of education is assigned a score depending both on the nominal length and the
difficulty of the educational track. The following scores are used: UEþ (21), UE
(19), HVEþ (18), HVE (17), secondary education and lower (14) (cf. Van der
Velden and Van Smoorenburg, 1997). In line with previous research we distinguish
between the effects of overeducation and undereducation. Years of overeducation
have been calculated by subtracting the actual level of education from the appro-
priate level of education, with all negative scores set to zero. For years of under-
education the reverse has been applied, again setting all negative scores to zero. The
reference category for both variables is having a matching job. Because the model
specification includes the respondent’s actual level of education instead of required
level, the assignment theory predicts in model 2 a negative effect of overeducation,
since in this specification overeducation implies a job at a lower level. Similarly the
theory predicts a (somewhat smaller) positive effect of undereducation.
In order to ascertain to what extent any effects of educational mismatch are due
to skill underutilisation and/or deficits, models 3 and 4 incorporate measures for
the two types of skill mismatch. In model 3 we replace the terms representing over-
and undereducation by the respondents’ judgements of skill utilisation and deficits,
as measured by the responses to statements 1 and 2. The scores on statement 1 are
recoded such that a high score indicates high underutilisation.
Y ¼ model 1 þ a
6
UNDERUTILISATION þ a
7
DEFICIT þ e ð3Þ
with UNDERUTILISATION = measure based on recoded response to statement 1
and DEFICIT = measure based on response to statement 2.
Models 2 and 3 contain specifications of the effects of the two types of mis-
matches on labour market outcomes. For model 3, assignment theory leads to the
prediction of a negative effect of underutilisation, analogous to the negative effect
predicted for overeducation. The theory also predicts a (somewhat counterintui-
tive) positive effect of skill deficits, which are interpreted here as the skill counter-
part of undereducation. A skill deficit thus indicates a job above one’s own level. A
comparison of these models will provide an indication of which kind of mismatch
has the stronger overall effect on the outcome in question.
442 educational vs skill mismatches
More interesting from our point of view is a model specification in which both
educational mismatches and skill mismatches are included together. This will allow
us to determine the net effect of each kinds of mismatches after controlling for the
effect of the other. This specification is shown in model 4
Y ¼ model 2 þ a
6
UNDERUTILISATION þ a
7
DEFICIT þ e ð4Þ
5. The effects of mismatches on wages
Table 4 presents the results of the analysis on the (natural log) of hourly wages.
Model 1 shows that only 12% of the wage differences can be explained by the
variables in the model. This seems quite low, but but one should bear in mind the
relative homogeneity of the group in terms of basic human capital aspects. There
are in fact quite large differences between the different educational levels acquired.
Having followed university education rather than higher vocational education
yields a wage increase of 23% (exp(0.211)). Having followed university plus
jim allen and rolf van der velden 443
Table 4 Results of regression-analyses with dependent variable ln (hourly wage)
Model 1 Model 2 Model 3 Model 4
Dependent variable:
.......................... .......................... .......................... ..........................
ln(hourly wage) B SE B SE B SE B SE
Human capital
Level of education (reference group HVE)
University þ 0.262* 0.023 0.316* 0.023 0.244* 0.023 0.304* 0.023
postgraduate study
University 0.211* 0.018 0.247* 0.018 0.206* 0.018 0.243* 0.018
HVE þ 0.073* 0.018 0.110* 0.017 0.065* 0.018 0.103* 0.018
postgraduate study
Tenure in current job 0.011* 0.002 0.010* 0.002 0.011* 0.002 0.010* 0.002
Experience before 0.011* 0.001 0.011* 0.001 0.011* 0.001 0.011* 0.001
current job
Job characteristics
temporary job 70.145* 0.022 70.154* 0.021 70.137* 0.021 70.148* 0.021
self-employed 70.033 0.024 70.043 0.023 70.046 0.024 70.050 0.023
Education–job match
overeducation (years) 70.081* 0.005 70.071* 0.006
undereducation (years) 0.036* 0.012 0.035* 0.012
job outside own field 70.036 0.015 0.031 0.015
Skill–job match
skill underutilisation 70.060* 0.005 70.032* 0.006
(statement 1)
skill deficit 0.000 0.005 70.004 0.005
(statement 2)
Constant 2.938* 0.043 2.947* 0.042 3.080* 0.047 3.033 0.047
Adjusted R2 0.12 0.23 0.17 0.24
N 2217.00000 2188.00000 2170.00000 2141.00000
* significant at 1% level
postgraduate study even leads to a wage increase of 30%.
8
Postgraduate study after
higher vocational education increases wages by about 8%. Model 1 shows positive
effects of both experience variables tenure and experience before current job, and a
negative effect for having a temporary job. Being self-employed has no significant
effect on wages.
In model 2 the indicators of educational mismatches are added to the model.
This improves the model fit markedly, resulting in an adjusted R2 of 0.23. There is
a significant positive effect of undereducation, which confirms the prediction that
holding a job for which the appropriate educational level is higher than the one
followed by the respondent results in higher wages. The predicted negative effect of
overeducation is also observed. In line with the predictions of assignment theory
and the results obtained in earlier research, the effects of overeducation are con-
siderably greater than those for undereducation. Each year of undereducation
(working above one’s level) yields a wage increase of some 4%. Each year of over-
education (working below one’s level) leads to a decrease in wages of 8%. There is
no significant effect of working in a job for which one’s own or a related field of
education is not required. It is interesting to note that taking educational mis-
matches into account increases the coefficients for own education. This reflects the
fact that the reference group, graduates of higher vocational education without any
additional postgraduate study, showed the lowest proportion of individuals work-
ing below their own level.
In model 3 we use skill mismatches instead of educational mismatches to explain
wage differences. Underutilisation, the ‘skills counterpart’ of overeducation, shows
the expected negative effect on wages. By contrast, a skill deficit appears to have no
effect at all on wages. Skill mismatches account for a good deal less wage variance
than do educational mismatches: the adjusted R2 amounts to 0.17, compared to
0.23 for model 2.
Model 4 combines both educational mismatches and skill mismatches. Both
kinds of mismatches have a significant effect on wages, even when controlling
for the other. However, about half of the effect of skill underutilisation disappears
when educational mismatches are taken into account. By contrast, only a small
part of the effects of over- and undereducation are accounted for by skill mis-
matches. In terms of additional explained variance, educational mismatches seem
to be much more important than skill mismatches. Whereas the adjusted R2 in
model 4 is just barely higher than that in model 2, it is clearly much higher than in
model 3.
The results presented in Table 4 are inconsistent with the explanation given by
assignment theory to the wage effects of over- and undereducation. This explana-
tion holds that such effects reflect differing levels of productivity as the match
between required and available skills is varied. In line with this expectation, we
444 educational vs skill mismatches
..........................................................................................................................................................................
8
These wage differentials are greater than those normally found for graduates one year after graduation,
indicating that the age-earnings profile for university graduates is steeper than for graduates from higher
vocational education.
do observe a significant negative wage effect of skill underutilisation. However,
contrary to what would be expected on the basis of assignment theory, skill mis-
matches account for only a small proportion of the wage effects of educational
mismatches. Skill deficit appears to have no effect at all. This might be due to the
lower effects of undereducation in general in combination with the low threshold
value of our measure of deficits.
As indicated at the start of this paper, educational mismatches have been found
to affect a broad range of labour market outcomes, not just wages. Outcomes such
as job satisfaction and on-the-job search are important not only to individual
workers, but also from the point of view of the workings of the labour market.
In particular, we would like to know to what extent and in what way dissatisfaction
with the match between schooling and/or skills and the characteristics of the job
constitutes a motivation for individuals to seek employment better suited to their
own capabilities. In the following two sections we analyse the effects of educational
and skills mismatches on the dependent variables job satisfaction and job quit
intention.
6. Mismatches and job satisfaction
In this section we discuss the effects of mismatches and other variables on job
satisfaction. Respondents were asked to rate their over-all job satisfaction on a scale
from 1 (very unsatisfied) to 5 (very satisfied). This variable has been recoded into a
dummy with value 1 if the respondent marked answer category 4 (satisfied) or 5
(very satisfied), and value 0 otherwise. To facilitate a comparison of effects, the
same independent variables have again been included as used for the wage esti-
mates. A number of indicators of job quality, including ln(hourly wage) itself, have
also been included as control variables. The other indicators comprise the gradu-
ates’ ratings of a number of aspects of their current job, namely the degree of
autonomy, the variety of work tasks, the prestige associated with the job, and
the opportunity to introduce their own ideas. These control variables are import-
ant, since the quality of the job might conceivably influence both job satisfaction
and the responses to statements 1 and 2. Table 5 presents the results of the logistic
regression analysis.
Model 1 shows that job satisfaction is strongly influenced by job characteristics.
As one might expect, wages have quite a strong effect on job satisfaction, and
self-employed respondents are also more satisfied. The respondents’ job quality
ratings, particularly for variety and room for own ideas, also show a strong
effect. Surprisingly, we find no significant effect of holding a temporary job.
Model 2 shows no improvement in model fit. Overeducation—working in a job
that requires a lower level of education than one’s own—has a negative effect on
the job occupant’s satisfaction, but the effect is not significant. Undereducation has
no significant effect on job satisfaction. The same applies to working in a job that
does not match the respondent’s field of education.
jim allen and rolf van der velden 445
In model 3, skill mismatches are introduced in the place of educational mis-
matches. In contrast to educational mismatches, skill mismatches appear to exert a
strong influence on job satisfaction. The model fit is greatly improved. Skill under-
utilisation has a strong negative effect on satisfaction. The effect of skill deficits is
also negative, although not significant. Skill mismatches account for a considerable
part of the effects of job quality indicators. Notably, the effects of wages and
prestige are no longer significant.
446 educational vs skill mismatches
Table 5 Results of logistic regression-analyses with dependent variable job satisfac-
tion
Dependent variable: Model 1 Model 2 Model 3 Model 4
(very) satisfied
.......................... .......................... .......................... ..........................
with current job B SE B SE B SE B SE
Human capital
Level of education (reference group HVE)
University þ 0.225 0.224 0.323 0.237 0.292 0.238 0.218 0.253
postgraduate study
University 70.051 0.170 0.057 0.178 0.060 0.181 0.038 0.190
HVE þ 0.199 0.162 0.273 0.170 0.217 0.172 0.202 0.181
postgraduate study
Tenure in current job 70.032 0.016 70.035 0.016 70.038 0.017 70.040 0.017
Experience before 70.007 0.013 7 0.008 0.011 70.002 0.012 70.003 0.012
current job
Job characteristics
In (hourly wage) 0.830* 0.208 0.638* 0.220 0.458 0.224 0.499 0.236
temporary job 0.049 0.195 0.037 0.200 70.076 0.208 70.037 0.213
self-employment 0.797* 0.300 0.774 0.303 0.794 0.315 0.787 0.318
autonomy 0.204* 0.062 0.187* 0.063 0.197* 0.066 0.190* 0.067
variety 0.570* 0.062 0.572* 0.063 0.424* 0.068 0.424* 0.068
prestige 0.196* 0.062 0.181* 0.063 0.104 0.068 0.112 0.069
room for own ideas 0.500* 0.066 0.486* 0.067 0.339* 0.071 0.353* 0.072
Education–job match
overeducation (years) 70.134 0.055 0.020 0.061
undereducation (years) 0.124 0.129 0.029 0.136
job outside own field 70.190 0.143 0.098 0.152
Skill–job match
skill underutilisation 70.760* 0.059 70.755* 0.062
(statement 1)
skill deficit 70.129 0.053 7 0.141* 0.053
(statement 2)
Constant 7 0.756 7 0.814 7 0.884 7 0.928
77.062* 76.381* 72.230 72.429*
Model chi-square 405.5* 416.8* 577.5* 571.7*
d.f. 12 15 14 17
change in chi-square 7 11.3 172.0 166.2
relative to model 1
change in d.f. 7 325
relative to model 1
* significant at 1% level.
Model 4 incorporates both educational and skill mismatches. This model fits the
data less well than model 3, despite the additional three degrees of freedom used.
Interestingly, in model 4 the effect of skill deficits is now (just) significant. These
results show that skill mismatches really matter to workers. Even after controlling
for a range of job quality indicators, a poor match between available and required
skills has a strong negative effect on job satisfaction. In the following section, we
attempt to determine to what extent this effect constitutes a motivation on the part
of workers to seek alternative employment.
7. Mismatches and on-the-job search
We have seen that, although skill mismatches only have rather moderate effects on
wages, they have a strong impact on job satisfaction. An important question now is
whether this has any real behavioural consequences for employees. Are workers
who experience a poor match between their own skills and those required in their
current job motivated by this to quit their job in favour of other work? In this
section we examine effects of mismatches on the likelihood that respondents are
looking for alternative employment. We asked the respondents whether they had
actively sought other work in the past four weeks. Table 6 presents the results of the
logistic regression analysis.
Model 1 shows that on-the-job search is strongly influenced by the variety of
work tasks and the room to introduce own ideas. Other characteristics that showed
an effect on satisfaction, such as wages, don’t appear to constitute a sufficient
motivation to actively seek other work.
Model 2, in which educational mismatches have been introduced, does not fit
the data significantly better than model 1. A poor match between formal schooling
and that considered appropriate to the job does not increase the probability that a
worker will seek other employment.
Model 3 shows that skill mismatches, in particular underutilisation of skills, do
have real consequences, in terms of on-the-job search behaviour. The model fit is
improved considerably by the introduction of these two variables. Workers who
report an underutilisation of skills are significantly more likely to look for alter-
native employment than those who report little or no underutilisation. Skill deficits
also have a positive effect, although this is not significant. Interestingly, on con-
trolling for the effects of skill mismatches, wages now also show a significant
positive effect on on-the-job search. By contrast, the effects of variety and
opportunity to introduce own ideas are no longer significant.
As was the case in the analyses of job satisfaction, model 4 resembles model 3
greatly. Taking educational mismatches into account increases the effect of skill
deficits to the point of significance, but negates the significant effect of wages. The
results establish that skill mismatches clearly have behavioural consequences. In
fact, of all the variables, only skill mismatches show a significant effect on on-the-
job search.
jim allen and rolf van der velden 447
8. Conclusion
In this paper we have explored the relation between educational mismatches
and skill mismatches. Educational mismatches are indicated by comparing the
acquired level and field of education with the level and field of education con-
sidered most appropriate for the job. Skill mismatches are indicated by worker’s
responses to the statements ‘My current job offers me sufficient scope to use my
knowledge and skills’ and ‘I would perform better in my current job if I possessed
448 educational vs skill mismatches
Table 6 Results of logistic regression-analyses with dependent variable looking for
other work
Dependent variable: Model 1 Model 2 Model 3 Model 4
looking for
.......................... .......................... .......................... ..........................
other work B SE B SE B SE B SE
Human capital
Level of education (reference group HVE)
University þ 0.078 0.235 0.067 0.249 0.091 0.240 0.160 0.256
postgraduate study
University 0.167 0.187 0.155 0.195 0.101 0.192 0.148 0.201
HVE þ 0.141 0.179 0.106 0.187 0.140 0.184 0.150 0.193
postgraduate study
Tenure in current job –0.006 0.018 –0.002 0.018 –0.004 0.018 –0.001 0.018
Experience before 0.000 0.012 0.001 0.012 –0.003 0.013 –0.002 0.013
current job
Job characteristics
In (hourly wage) –0.169 0.213 –0.117 0.226 0.047 0.221 –0.005 0.233
temporary job 0.471 0.189 0.428 0.194 0.556* 0.193 0.478 0.198
self-employment –0.259 0.281 –0.251 0.282 –0.206 0.283 –0.196 0.285
autonomy –0.089 0.065 –0.089 0.065 –0.072 0.067 –0.079 0.067
variety –0.190* 0.065 –0.191* 0.065 –0.089 0.068 –0.093 0.069
prestige –0.023 0.066 0.031 0.067 0.069 0.068 0.064 0.069
room for own ideas –0.246* 0.069 –0.237* 0.070 –0.133 0.073 –0.138 0.074
Education–job match
overeducation (years) 0.016 0.056 –0.042 0.060
undereducation (years) –0.213 0.140 –0.137 0.141
job outside own field –0.151 0.148 –0.069 0.152
Skill–job match
skill underutilisation 0.380* 0.059 0.380* 0.061
(statement 1)
skill deficit 0.130 0.052 0.138* 0.053
(statement 2)
Constant 0.873 0.754 0.750 0.821 –2.169 0.889 –1.917 0.933
Model chi-square 62.5 66.5 102.5* 103.3*
d.f. 12 15 14 17
change in chi-square 7 4.0 40.0* 40.8*
relative to model 1
change in d.f. 7 325
relative to model 1
* significant at 1% level.
additional knowledge and skills’. In assignment theory the two concepts
educational and skill mismatches are assumed to be closely related: educational
mismatches imply skill mismatches which in turn have an effect on productivity
and wages.
Our findings in this paper have important implications for research into the
effect of education on labour market outcomes. They establish beyond reasonable
doubt the importance of the distinction between schooling and skills. The results
provide strong support for the assumption that the match between individual
human capital and the characteristics of the job matters. In line with the
predictions of assignment theory and with findings from earlier research,
educational mismatches were found to have a strong effect on wages. Contrary
to the assumptions of assignment theory however, educational mismatches are
neither a necessary nor a sufficient condition for skill mismatches. Furthermore,
only a small proportion of the wage effects of educational mismatches is accounted
for by skill mismatches. Underutilisation of skills also exerts a negative effect on
wages distinct from the effects of overeducation, but the effects are very small.
However, skill mismatches do have a strong effect on job satisfaction and on-
the-job search, after controlling for job quality, whereas educational mismatches
lack any effect on these outcomes.
What do these findings mean? One explanation could be that our operational-
isation of skill mismatches is inadequate. In particular the indicator for skill
deficits seems to be somewhat lacking in discriminatory power, and its effects
throughout are weak. Although this indicator is not entirely lacking in content
and prediction validity, improving this measure is clearly an important aim for
future research. Fortunately, the results pertaining to overeducation and under-
utilisation, which are in terms of both the theory and empirical findings by far the
most important, are in no way affected by this indicator. The results on these
variables hardly change at all when skill deficits and undereducation are omitted
from the analyses.
There is no reason to doubt the basic validity of the indicator for skill under-
utilisation. There is undoubtedly some noise in this variable, as in every measure
obtained through survey research, However, it seems implausible that respondents
who are working below their own level to the extent that this has a strong negative
effect on their wages would fail to report that their skills were not being fully
utilised, if this was in fact the case. The lower wages associated with overeducation
must have a different cause.
Human capital theory provides us with one possible explanation for the observed
wage effects, namely that graduates with the same level of education but differing
abilities are sorted in the market, with the most competent obtaining jobs ‘above
their own level’ and the least competent obtaining jobs ‘below their level’. Accord-
ing to this explanation, workers are rewarded according to individual productivity,
which is not dependent in any major extent on the job. In other words, the
apparent effects of over- and undereducation are spurious, masking unmeasured
jim allen and rolf van der velden 449
ability differences that are the real determinants of productivity and thence wages.
This human capital explanation does not provide an explanation for our finding
that a considerable proportion of workers report quite serious underutilisation of
skills. This could be accounted for by a watered-down version of assignment
theory, whereby the effects of educational mismatches are primarily due to differ-
ences in unmeasured abilities, but where there is still a distinct effect of skill
mismatches.
Screening theory offers an alternative explanation for the observed effects. It has
in common with the human capital explanation the assumption that individuals
are sorted in the labour market. However, in contrast to human capital theory,
screening theory asserts that individuals are sorted—and rewarded—primarily on
the basis of easily observable proxies for productivity, rather than on the basis of
productivity itself. One such proxy is education, but there are many others such as
work experience, gender and social background. Because these other indicators are
differentially distributed within educational categories, and because different
employers assign different weights to each indicator, a considerable proportion
of workers end up in jobs which don’t match their education. Those who as a
result of this sorting process end up in a job below their own level will earn less
than those working at their own level, regardless of their actual level of skill, and
regardless of the degree to which their skills are being utilised. In a highly institu-
tionalised system of wage bargaining as in the Netherlands, rules about the level of
the job and the educational level of the occupant are frequently incorporated into
wage scales.
Our results relating to job satisfaction and in particular on-the-job search are
important, since they shed light on the mechanisms through which adjustments
take place in the market. Skill mismatches are an important cause of job dis-
satisfaction, which provide an incentive for workers to look for other work,
presumably work which is better suited to their own abilities. This shows that
adjustments in the labour market are strongly driven by the relation between job
content and individual abilities, and less by the material and social rewards
provided by work.
9
Acknowledgements
This article is a revised version of a paper prepared for the conference ‘Skill Measurement
and Economic Analysis’, 27–29 March 2000, University of Kent, Canterbury. We would like
to thank Lex Borghans, Lia Potma, Timo Huijgen, and two anonymous reviewers for their
comments on earlier versions of the article.
450
educational vs skill mismatches
..........................................................................................................................................................................
9
This need not mean that such rewards are not important to workers, but could simply mean that the
differences between current rewards and those available elsewhere are not sufficiently large.
References
Allen, J. and van der Velden, R.K.W. (forthcoming). The labour market position of Dutch
tertiary graduates: an international comparative analysis, Research Centre for Education and
the Labour Market, Maastricht.
Cohn, E. and Khan, S.P. (1995). ‘The Wage Effects of Overschooling Revisited’, Labour
Economics, 2, 67–76.
Duncan, G. and Hoffman, S. (1981). ‘The Incidence and Wage Effects of Overeducation’,
Economics of Education Review, 1, 75–86.
Groot, W. and Maassen van den Brink, H. (1996). ‘Overscholing en verdringing op de
arbeidsmarkt’, Economisch Statistische Berichten, 81, 74–7.
Hartog, J. (2000). ‘Over-education and Earnings: where are we, where should we go?’,
Economics of Education Review, 19, 131–47.
Hartog, J. and Oosterbeek, H. (1988). ‘Education, Allocation and Earnings in the
Netherlands: Overschooling?’, Economics of Education Review, 7, 185–94.
Hersch, J. (1991). ‘Education Match and Job Match’, Review of Economics and Statistics, 73,
140–4.
Sattinger, M. (1993). ‘Assignment Models of the Distribution of Earnings’, Journal of
Economic Literature, 31, 851–80.
Sicherman, N. (1991). ‘‘Overeducation’’ in the Labour Market’, Journal of Labour
Economics, 9, 101–22.
van Smoorenburg, M.S.M. and van der Velden, R.K.W. (2000). ‘The Training of
School-leavers: Complementarity or Substitution?’, Economics of Education Review, 19,
207–17.
Thurow, L. (1975). Generating Inequality, Basic Books, New York, NY.
Topel, R. (1986). ‘Job Mobility, Search and Earnings Growth’, Research in Labor Economics,
8, 199–233.
Tsang, M. and Levin, H. (1985). ‘The Economics of Overeducation’, Economics of Education
Review, 4, 93–104.
van der Velden, R.K.W. and van Smoorenburg, M.S.M. (1997). The Measurement of
Overeducation an Undereducation: Self-Report vs. Job-Analyst Method, Research Centre for
Education and the Labour Market, Maastricht.
Viscusi, W. (1979). ‘Job Hazards and Worker Quit Rates: An Analysis of Adaptive Worker
Behavior’, International Economic Review, 20, 29–58.
jim allen and rolf van der velden 451
Appendix
Descriptive statistics of variables used
452
educational vs skill mismatches
Table A1 Interval variables
Variable mean standard deviation
Hourly wage (guilders) 33.0 10.3
Tenure (years) 3.8 3.6
Age at start of current job (years) 31.1 5.1
Years of overeducation 0.6 1.1
Years of undereducation 0.2 0.5
Skill underutilisation (5-point scale) 2.2 1.1
Skill deficit (5-point scale) 3.3 1.2
Autonomy in job (5-point scale) 3.7 0.9
Variety in job (5-point scale) 3.9 0.9
Prestige of job (5-point scale) 3.3 0.9
Room for own ideas in job (5-point scale) 3.9 0.9
Table A2 Dummy variables
Variable Percentage
Job satisfaction 73
Looking for other work 18
Level of education:
University þ postgraduate study 12
University only 36
Higher vocational education þ postgraduate study 37
Higher vocational education only 16
Temporary job 9
Self-employed 6
Work outside own field 20
... Thus, many of the earlier studies show employee turnover is the result of job dissatisfaction. Lambert (1991) and Allen and Van Der Velden (2001) in their studies identify On-The-Job search theory which explains the behavior of employed individuals who search for a better alternative job. This behavior of the employees they label as relative dissatisfaction (Alam and Hasan, 2015). ...
Article
Full-text available
This study was conducted at the Department of Human Resource Management, Jatiya Kabi Kazi Nazrul Islam University, Trishal, Mymensingh, Bangladesh during August 2020 to May 2021 to at showing the influence of relative job dissatisfaction on employees' job quitting from organizations. The study conducted a field survey of 140 respondents who previously leave their job and reviewed available literature in order to find out the major factors that create relative job dissatisfaction of an employee. The study discussed four major factors of relative dissatisfaction from previous literature as: pay & working conditions, supervision & work pressure, retirement benefit and job challenges. The findings of the study show that among the four factors of relative job dissatisfaction, pay & working conditions and supervision and work pressure have a significant influence on the job quitting of employees. On the other hand the study does not find any significant influence of the other two factors retirement benefit and job challenges on the job quitting of employees from the organizations.
... The potential mismatch of graduates is a long-standing and ongoing issue (e.g. Freeman, 1976;Allen and van der Velden, 2001;McGuiness and Bennett, 2007;Green and Zhu, 2010;Kiersztyn, 2013;Verhaest and Van der Velden, 2013;Meroni and Vera-Toscano, 2017;Green and Henseke, 2016a;Wu and Wang, 2018;Turmo-Garuz et al., 2019). A graduate is said to be 'matched' if they are doing a graduate job, and to be 'under-employed' (or 'overeducated' or 'overqualified') if they are doing a non-graduate job. ...
Article
We distinguish between “task-warranted” and “task-unwarranted” graduate jobs. For both types, a degree is required but task-warranted graduate jobs involve carrying out typical graduate-level tasks. We operationalize the distinction, using representative surveys of resident Singapore workers. We find that the ongoing fast expansion of higher attainment between 2013 and 2017 was met by a similarly-strong growth in task-warranted graduate jobs. Compared with the matched graduates, the graduates in task-unwarranted graduate jobs and in non-graduate jobs both perceive lower skills utilization. There is a negative wage gap of 18% for graduates in task-unwarranted graduate jobs, and of 31% for underemployed graduates.
Article
This paper analyses the wage effects of educational mismatch by workers’ origin using a sizeable, detailed matched employer–employee dataset for Belgium. Relying on a fine-grained approach to measuring educational mismatch, the results show that over-educated workers, regardless of their origin, suffer a wage penalty compared to their well-matched former classmates. However, the magnitude of this wage penalty is found to vary considerably depending on workers’ origin. In addition, the estimates show that origin-based differences in over-education wage penalties significantly depend on both demographics (workers’ region of birth, education, gender and tenure) and employer characteristics (firm size and collective bargaining). To our knowledge, the role played by these different moderating variables has been either little or not explored in this context before.
Article
The paper investigates into the link between education mismatch and wages in Mauritius. It adds to the existing literature on education mismatch and labour market outcomes by using Mauritius as a case study in Africa, where overeducation is a growing concern, with graduates struggling to find jobs corresponding to their education level and field, the so-called vertical and horizontal mismatches. This supply-demand inadequacy ascends from a mix of factors and one of them can be termed as the ‘diploma disease’. Using the Continuous Multi-Purpose Household Surveys from 2013 to 2017, the pooled estimation results find evidence for a wage penalty of 1.3% for overeducated individuals compared to perfectly matched workers. A wage penalty of around 9.6% is also noted for undereducated workers compared to their perfectly matched counterparts. Heckman two step estimation approach is adopted to account for sample selection bias whilst the instrumental variable two stage least squares is applied to cater for the endogeneity of overeducation. The wage penalty for overeducation and undereducation under Heckman estimation method compares to ordinary least squares estimates with slight differences in the magnitude of the wage penalty. Once we account for endogeneity, the wage penalty for overeducated workers rose to 11% while that of overeducation did not change.
Article
This study supplements the existing conceptualisation of skills mismatch based on cognitive evaluations (being underskilled or overskilled) with an affective aspect that captures how workers cope with skills (mis)match situations (feeling overchallenged or underchallenged) and an analysis of skills mismatch situations’ influence on job satisfaction of workers with higher education and vocational education and training (VET). Using the German BIBB/BAuA Employment Survey 2018, the results indicated that underskilling by itself was not negatively associated with job satisfaction; however, underskilling in combination with feeling overchallenged exerted a significant negative influence on job satisfaction. This corroborates the approach of differentiating challenging (i.e. non-detrimental) jobs from overchallenging jobs. Overskilling was associated with penalties regarding job satisfaction, whereas overskilling in combination with underchallenging jobs produced a remarkably larger negative impact on job satisfaction. Moreover, overskilled HE workers received larger penalties regarding job satisfaction than those received by overskilled VET workers.
Thesis
Full-text available
The dissertation deals with the impact of automation on the labor market. The dissertation examines the most frequently mentioned risk of automation - mass (technological) unemployment - and concludes that the most likely consequence of automation, which could lead to technological unemployment or otherwise limited job flexibility, is an educational mismatch; a discrepancy between the education with which the employee performs his job and the education required to perform that job. The aim of the dissertation is to study the educational mismatch and to identify the reasons that cause this mismatch. The dissertation examines the educational mismatch in the context of employees of industry and the Slovak Republic. Industry employees were chosen because this group of employees is most at risk due to automation, and it is not sufficiently researched both in the literature on the impact of automation and in the literature on educational mismatch as one of the consequences of automation. The Slovak Republic was chosen as it is one of the countries most at risk of automation. International labor market analyzes estimate that the Slovak Republic has 33% to 44% of automated jobs. Domestic studies dealing with the impact of automation assume that the share of jobs at risk of automation in the Slovak labor market is in the range of 20% to 47%. The research of the dissertation was carried out by a quantitative questionnaire survey on a sample of 370 industrial employees with work experience with one of the selected seventeen jobs that are most endangered by automation. Research has shown that 64.9% of employees at risk of automation at risk are in objective educational mismatch and 50.5% of employees at risk are identified in subjective educational mismatch. An analysis of the reasons influencing an educational mismatch has shown that women and employees who have spent fewer years in their current job have a higher chance of being in an objective educational mismatch. Subjective educational mismatch more often identifies women, young people, people who have experienced unemployment, employees who have spent fewer years in their current job, workers in larger companies, employees who have received less training, employees whose work does not require special qualifications and employees who do not use previous work experience in their current job. Despite the high share of educational mismatch, we believe that employees at risk of disability at risk are not significantly handicapped on the labor market, as they have the skills and experience to perform other jobs.
Article
Full-text available
The aim of this study is to identify the returns to education in Kyrgyzstan, with special reference to employment type and gender differences. The empirical analysis of this study is based on Life in Kyrgyzstan (LiK) survey data collected in 2016. The sample for analysis is constructed with employees and self-employed persons aged 18-65, who indicated their monthly income from employment. According to the empirical outputs, there is a wage premium for higher education such that the marginal return to education for women is higher than men.
Article
Background One way of evaluating immigrants’ labor market outcomes is to assess the extent to which immigrants are able to enter into jobs that are commensurate with their education and experience. An imperfect alignment between workers’ educational qualifications and these required for their current job, or education-job mismatch, has implications for both the broader economy and individual workers. In this study, we investigate the factors associated with education-job mismatches among US workers by immigrant generation. Methods We analyzed the data from the US sample of the Program for the International Assessment of Adult Competencies (PIAAC) 2012/2014. Our analytic sample included 4022 employed (full and part-time) individuals between the ages of 20–65 years. We documented the distribution of education-job mismatches across selected independent variables and estimated the relationship between the individual characteristics of workers such as race, gender, presence of children, location, time in the country and knowledge of English for first-generation immigrant workers, and education-job mismatch using multinomial logistic regressions for the full sample and for the sample of first- and second-generation workers. Results We found that on average, immigrant workers in the US labor market were more likely to hold jobs which required less education that they had (being overmatched for the job), with first-generation workers being overmatched more frequently than second-generation workers. The probability of being overmatched for immigrant workers declines with the length of stay, and workers who are proficient in English are less likely to be overmatched. Our results also suggest that there may be labor market disadvantages to immigrant status that persist beyond the first-generation. Conclusions Previous research demonstrated that over-education depresses wages and lowers workers’ standards of living and their abilities to accumulate wealth. Our findings confirm that this dynamic may be particularly acute for first- and second-generation workers who are finding it difficult to become fully integrated into US labor markets, even though the factors behind the mismatch differs between the two immigrant generations.
Article
Full-text available
Resumo O objetivo deste estudo é investigar a relação de satisfação profissional dos jovens e a adequação educacional, mensurada a partir da compatibilização da escala ISCO e ISCED. Para tanto, foram utilizados dados da Pesquisa de Transição da Escola para o Trabalho (SWTS) da Organização Internacional do Trabalho (OIT), uma base de dados específica para o público jovem de 15 a 29 anos, no Brasil. Utilizou-se um modelo próbite ordenado, sendo a variável dependente o grau de satisfação profissional dos jovens. Foi verificada uma relação negativa e significativa da qualificação excessiva sobre a satisfação no trabalho. Além disso, ter um trabalho formal e maior salário/hora está positivamente relacionado com a satisfação no trabalho.
Article
Full-text available
This paper, based on data from Survey of Family Income and Expenditure of Taiwan, shows that the recent trends of job match in Taiwan labor market have been marked by increasing proportion of overeducated workers due to the higher education expansion policy, while the incidence of undereducation continues to decline. Furthermore, workers¡¯ economic position is not completely determined by their educational levels. Working experience also plays an important role in workers¡¯ job placement and their wages. Workers with relatively less working experience are more likely to be overeducated, while workers with relatively more working experience are more likely to be undereducated. Overeducated (Undereducated) workers would earn more (less) than their co-workers with adequate education but less (more) than the workers having the same educational level with adequate education for jobs. However, the rewards (penalties) to adequate education and overeducation (undereducation) decline as more experience accumulated. Evidence also shows effect of bumping down from overeducation on the wages and employment of lower educated workers.
Article
Full-text available
Data gathered from a recent national sample of workers on educational requirements and attainments are used to examine the extent and economic effects of overeducation. Nearly 40 percent of the U.S. work force-and about 50 percent of black males-have more education than their jobs require. But we also find that “surplus” education does have economic value. The individual return to an additional year of surplus education was positive and significant for all major demographic groups. The estimated return is, however, only about half the size of the return to an additional year of required education.
Article
This study replicates models developed by Verdugo and Verdugo (1989) and Sicherman (1991) to study the wage effects of overschooling. Using the 1985 wave of the Panel Study of Income Dynamics, our results confirm earlier work showing that the rate of return to required schooling exceeds the rate of return to overschooling, and that the rate of return to underschooling is negative. At the same time, our results also confirm that, on the average, persons whose schooling exceeds (is less than) the required schooling for their occupation or job, respectively, receive lower (higher) wages than workers with similar levels of schooling in occupations or jobs having the required schooling. These results remain robust for alternative definitions of required, over- and underschooling, as well as for alternative specifications of the wage equations.
Article
In theoretical discussions about the relation between education and training, the question of complementarity or substitutability between these two different forms of human capital is raised. If initial education and industrial training are substitutes, overeducated workers will participate less in additional training than workers who are adequately educated. It could explain the persistence of overeducation and implies that the social wastage of overeducation will be less. On the other hand, if initial education and industrial training are complements, existing differences in human capital will only increase by industrial training, implying the risk for some workers of `missing the boat'. Supplementary to Groot we not only look at the impact of over- and undereducation (level) but also at non-matching fields of studies and the `narrowness' of types of education. A sample of labour market entrants was used, so we did not have to cope with the disturbing influence of other forms of human capital: life and labour market experience. The paper gives evidence in support of both substitutability and complementarity between initial education and firm training.
Article
Drawing on empirical studies from five countries spanning an interval of two decades, regularities in the incidence of over- and undereducation are outlined, as well as consequences for individual earnings. The results are confronted with three theoretical models (search, human capital and assignment), but none of these is convincingly related to the specification of the earnings function. Directions for further work are suggested. (C) 2000 Elsevier Science Ltd. All rights reserved. JEL classification: J24; J31.
Article
This paper documents the increased participation in higher education in the Netherlands and its consequences for the relation between levels of education and job levels. Undereducation has been reduced, overeducation has been increased. This does not imply private or social inefficiency, as even years of “overeducation” earn a positive rate of return. A general specification of the earnings function is derived from allocation models of the labor market. It contains the human capital specification and the job competition specification as special cases, and proves superior to both.
Article
This article examines the reasons for the observed discrepancy between workers' actual and required levels of schooling and the resulting differences in returns to schooling. "Overeducated" workers are found to be younger and to have lower amounts of on-the-job training than workers with the required level of schooling. They also have higher rates of firm and occupational mobility, characterized by movement of higher-level occupations. The findings suggest that overeducation can be explained by the trade-off between schooling and other components of human capital and by the mobility patterns of overeducated workers. Copyright 1991 by University of Chicago Press.