Overeducation and Educational–Occupational Mismatch

Abstract

This article accounts for a renovation by enriching the existent literature regarding two major nowadays phenomena and their labor implications, which might require a rethinking regarding new career-development approaches: the overeducation phenomenon (academic graduates whose educational level exceeds the educational level required in their jobs) and the horizontal-mismatch phenomenon (the individuals’ fit between their educational and occupational fields). The article clarifies the difference between those two phenomena’s career repercussions—on three major individuals’ career outcomes: wage gaps, job mobility, and demand for further higher education. Two hundred and twenty-one participants took part to reveal that the more the individuals experience horizontal mismatch throughout their career, the less their earning level is, and the horizontally mismatched individuals have lower probability for pursuing further higher education compared to the horizontally matched individuals. These findings’ implications suggest individuals to try and modify their career planning, based on the ongoing technological and structural changes in Western-developed countries labor market.

Keywords

overeducation educational–occupational mismatch wage gaps job mobility higher education demand

The current article examines the implications of two central phenomena: (a) overeducation (vertical mismatch), which describes a case of a surplus education level comparatively to the required educational-level criteria of the individual’s job (Tsai, 2010), and (b) the match of the individual’s education field and occupational field (horizontal match; Nordin, Persson, & Rooth, 2010). Those two aforementioned concepts are actually subphenomena under the broader supraphenomenon of “underemployment,” which refers to a general underutilization of the individuals’ characteristics and competencies, such as potential income (underpayment), potential job extent and permanency (work-status congruence), occupational/professional experience (skills/experience underemployment), and so on (Thompson, Shea, Sikora, Perrewe, & Harris, 2013). The current article rational discusses the problem statement that will be addressed and clarified with this study: The possible differences in (a) wages, in the (b) demand for further higher education acquirement, and in the (c) job mobility, as possibly ascribed to the horizontal-mismatch and the vertical-mismatch phenomena.

The vertical-mismatch phenomenon was first presented by Freeman (1976) in the context of the American “baby boom” era, regarding those who couldn’t easily find jobs which fit their educational level in the local American market, inter alia due to the post-World War II higher education expansion in the U.S. youth population. As indicated in Romanov, Tur-Sinay, and Eizman (2008), the vertical-mismatch incidence in Israel and its possible outcomes in its labor force has been barely examined to this day: They found that the extent of vertical mismatch in Israel resides at the upper end of estimates of the phenomenon in Western countries. In Israel, as in several other developed countries, the population share that acquires academic schooling—toward first degrees and advanced degrees alike—has been rising steeply in recent decades (Romanov, Tur-Sinay, & Eizman, 2008).

Additionally to the vertical-mismatch phenomenon, the research about individuals’ horizontal match examines mainly its implications on individual’s wages, job satisfaction, and job stability (Leuven & Oosterbeek, 2011; Nordin et al., 2010). Even though those two aforementioned phenomena are independent, previous studies have indicated that vertically mismatched employees might also experience, during their career, a mismatch between their occupation and educational fields (Sicherman, 1991), what might possibly lead to several career outcomes, which some are examined in the article.

It has been mostly argued that vertical mismatch may translate into persisting wage penalties (Aleksynska & Tritah, 2013), as studies found that vertically mismatched individuals earn less than vertically matched ones (McGuinness & Bennett, 2007). Romanov et al. (2008) also showed that the wages of vertically mismatched employees in the Israeli market are some 11% lower and rise more slowly than the wages of the vertically matched ones. Notwithstanding, there are studies with some counter findings as well: Tsai’s (2010) results, for example, provided evidence that vertically mismatched status does not cause lower earnings, claiming that the significant wage differential found in other studies is simply a possible result of ignoring the nonrandom assignment of workers to jobs.

Despite the conflicting opinions in the literature regarding the vertical-mismatch’s impact on wages, several studies have indicated wage differentials among academic individuals regarding the match between their educational and occupational fields (Leuven & Oosterbeek, 2011; Ordine & Rose, 2011). Robst (2007) unequivocally showed that (both partially and completely) horizontally mismatched men and women earn less than horizontally matched ones, summarizing that “being employed in jobs unrelated to the degree field—lowers the rate of return to schooling” (p. 403). In his findings, Robst (2007) claimed also that most of those wage-differential effects vary accordingly to the individual’s educational field.

Moreover, Nordin et al. (2010) found that Swedish horizontally matched employees earn averagely about 20% more than the horizontally mismatched ones, suggesting that it is probably the wage penalty paid by those individuals for being horizontally mismatched. Due to the abovementioned unequivocal findings in the literature regarding the vertical-mismatch phenomenon’s impact on wages and earnings, the question whether and how is there a possible wage penalty caused by experiencing horizontal mismatch during the career still remains and therefore:

Hypothesis 1: The rate of horizontal-mismatch incidences throughout the individual’s career will negatively affect the individual’s wages level.

The second subject of this article, job mobility, is defined as the individuals’ ability to change, or the actual change rate, of their status/positions within the workplace or to change occupations (Borjas, 1981; Gicheva, 2012). Employees’ mobility often occurs in the first stages of their career: During their first 10 years in the market, they hold on average, about two thirds of the number of jobs they will have during their entire career (Topel & Ward, 1992). Romanov et al. (2008) had indicated that in the Israeli market, there is a negative correlation between two variables: an employee’s tenure and the number of past employers in the years after the completion of degree studies, and the probability of being vertically mismatched. Studies found that vertically mismatched workers (a) have a higher rate of occupational mobility and (b) their work histories typified by a higher rate of job switching and a shorter stay in one post, comparably to nonvertically mismatched workers (Alba-Ramirez, 1993; Sicherman, 1991). McGuinness (2003) fenced those findings claiming that most of the occupational mobility that is attributed to vertically mismatched workers takes place among those who remain vertically mismatched at their new jobs.

Feldman (1996) already suggested that underemployment might be inversely correlated with both early start to job hunting and an intensive job search effort. Neal (1999) suggested in his two stages theory that academic graduates who join the job market are expected to go through two main stages of matching in the job market: (a) career match and (b) employer match. This model assumes that as long as the market’s jobs supply is wide enough, graduates will try get interlaced on jobs that fit their career aspirations, while one of those aspired fits might possibly be the horizontal match. Busch (2009) explained that gaining occupational experience by training is vital, especially to those who are unable to find employment within their current educational field, in order to improve their future job employability.

Up to date, there is relatively little unambiguous knowledge about the possible differences in the actual job mobility between the horizontally matched and mismatched employees, especially in the subjective aspect of this phenomenon (Maynard, Joseph, & Maynard, 2006) and all the more so regarding the Israeli market. Those possible differences might be caused by attempts to fit the jobs to the educational field, which can be considered as the first of Neal’s (1999) two stages theory (career match). In order to expand the knowledge in this area, the second hypothesis of the current article stipulates that

Hypothesis 2: Among overeducated employees, the horizontally mismatched will experience greater job mobility rate than the horizontally matched ones.

The third subject of the article is the possible differences in pursuing further higher education among overeducated employees, between the horizontally matched/mismatched. Romanov et al. (2008) showed that even though about one third of first-degree recipients in Israel tends to immediately acquire master’s studies, the vertically mismatched workers have lower propensity to continue to advanced academic studies, and the following conspectus will discuss the schooling demand from the horizontal-mismatch standpoint: According to Feldman and Turnley (1995), involuntary underemployment (which, again, includes the horizontal mismatch) represents some kind of a violation of expectations; academic graduates expect, perhaps naively, to find challenging work which corresponds their degree’s field. As a result of their unfilled expectations, underemployed (including horizontally mismatched) individuals are more likely to feel frustrated with their career’s lack of success (Robinson, Kraatz, & Rousseau, 1994).

Nevertheless, there is no unequivocal agreement about the possible effect of the vertical/horizontal-mismatch phenomena on the demand for education. On the one hand, for many individuals, continuing their education may seem like an attractive alternative to the concept of underemployment (Winefield, Winefield, Tiggemann, & Goldney, 1991). Furthermore, many college graduates realize only after graduation that they have obtained a degree in a field with low jobs supply, or with a high supply of labor force with the same educational level in their field, and therefore may take pursuing further schooling as an attractive alternative to deal with underemployment (Robst, 2007). On the other hand, many underemployed graduates may view their schooling less favorably since the latter eventually has not been resulted in attractive job opportunities. As a result, underemployed graduates may be less likely to pursue further education since they might perceive it as less instrumental in obtaining satisfactory employment (Feldman & Turnley, 1995). From the aforementioned studies, it could be inferred that the horizontal mismatch might have more substantial impact regarding its implication on the individuals’ further schooling demand and therefore:

Hypothesis 3: The horizontally mismatched graduates will have a lower probability for a demand for higher education than the horizontally matched.

Method

Participants and Procedure

Two hundred and twenty-one Israeli participants had taken part in the research, as the inclusion criteria for participants were (a) a baccalaureate degree holders at least (BA/B.Sc./MD/LLB, etc.) and (b) 5 years (at least) of labor market experience since their graduation. The participants had been voluntarily recruited via various internet forums and social networks (such as alumni association’s official forums and members, via LinkedIn alumni members, online applications to relevant students for advanced degrees by e-mails, etc.). The research tool used in this study was designated by the article’s authors and had been based mainly on the survey of the (only) relevant study of this topic in the Israeli market (q.v. Romanov et al., 2008). The survey collected data regarding demographic, socioeconomic, and the educational characteristics, directly from its participants, a source which helped the authors to cull information regarding them, such as vertical/horizontal mismatch, tenure, wages, and so on.

Table 1 presents the distributions (in percentages) of the demographic, socioeconomic, and educational characteristics of all the participants. As can be seen among both baccalaureates and advanced degree graduates (Table 2), the participants’ distribution in different educational fields is quite extensive. Their relative sparseness in any category of area of study and the uneven distribution of participants in most of the areas of study have not allowed for a statistical analysis of any possible differences between groups in area of study (Tables 1 and 2).

Table 1.

Demographic, Socioeconomic, and Educational Characteristics of the Participants (in %).

Variable	Distribution in Percentage Among the Sample (N = 221)
Gender
Male	48
Female	52
Marital status
Married	68
Unmarried	32
Religious leaning
Secular	65
Nonsecular	35
Educational level
Baccalaureate	57
Advanced degree	43
Higher education demand
Has a demand	37
Doesn’t have a demand	63
Job size
Full time	77
Part time	23
Vertical match
Overeducated	58
Adequately educated	42
Horizontal match
Horizontally matched	57
Horizontally mismatched	43

Table 2.

The Distribution of the Participants by the Educational Fields (in %).

Educational Fields	No. of Baccalaureates	No. of Advanced Degrees Holders
Social sciences and humanities	15	11
Logistics, business administration, management, industrial management, and organizational consulting	14	30
Chemistry, nursing, life/natural/health sciences	9	12
Computer science and science	13	6
Engineering	20	5
Anthropology, sociology, psychology, and criminology	15	9
Music, communications, and theaters	9	—
Accounting and economics	8	10
Education and physical education	15	10
Mathematics, statistics/biostatistics, and physics	10	8
Middle East studies and political science and law	10	—
Other disciplines	17	23

Materials

The research survey was a 21-items tool. The vast majority of the research on the vertical-mismatch and the horizontal-mismatch phenomenon are statistical correlative analyses (Dolton & Vignoles, 2000; McGuinness, 2006), based on accumulated data from relevant questionnaires/surveys which are written, distributed, and collected by public bureaus, such as the Americans Panel Study of Income Dynamics and the National Survey of College Graduates, the Canadian National Graduates Surveys, and etc. The same way, as aforementioned, the current survey is based on an Israeli Central Bureau of Statistics Council for Higher Education and Bank of Israel’s Research Department research survey (Romanov et al., 2008) relating to vertical mismatch, job mobility, and earning mobility among first degree holders in Israel.

In general, according to the vertical-mismatch literature, there are several main methods for measuring this phenomenon. Verhaest and Omey (2006) summarized four distinct ways to measure vertical mismatch, whereas one of those preferred objective approaches to measuring vertical mismatch include job analysis examining job titles and “realized matches”: An indirect self-assessment included asking respondents their schooling levels and then the required schooling level to get or to do their jobs. This objective method was used in this article by a comparative cross-check, likewise in Romanov et al. (2008).

The advantage of the current article’s measurement method is its objective individualized comparison per participant that is made and relies on the participants’ authentic report about their current schooling level and the required schooling level for their job. The more subjective measurement methods do not give much weight to the participants report about required schooling level for their jobs but rather a calculation is made to compare the averaged required schooling level for the participants’ jobs, relative to other positions in the participants’ occupational field (also known as the job analysis method). In addition, for the current research, it was assumed that all the participants who are not vertically mismatched are adequately educated, similar to previous studies on the subject (McGuinness & Bennet, 2007).

The 153 jobs that are listed in the questionnaire are mostly based on the local counterpart of the Department of Transportation and ISCO-08 classifications (International Standard Classification of Occupations), the Israeli Central Bureau of Statistics “Standard Classification of Occupations.” It generally brings together the range of the existing jobs in the market, by nomenclature (of positions) which is used for jobs in both the governmental and nongovernmental sectors. At the initial stage of the research, several more variables were created based on the initial collected data, and those aforementioned variables classified each observation as follows:

Vertical match is a nominal dichotomous variable which classifies each participant as overeducated or not overeducated. Participants whose current schooling level was higher than their present position’s required schooling level, were defined as overeducated or adequately educated otherwise. For possible regression analyzes, a dummy variable was created (overeducated and not overeducated, respectively).

Horizontal match is a nominal-dichotomous variable that classifies each participant as horizontally matched or as horizontally mismatched. Comparing between the participants’ current educational fields and current occupation, each participant was objectively classified as a horizontally mismatched (Value 1, mismatched) or horizontally matched (Value 2, matched). The intersection itself was made individually for each observation, based on data from each field of schooling comparatively to the relevant occupation. For possible regression analyses, a dummy variable was created: mismatched means that the participant is horizontally mismatched, and matched means that the participant is not horizontally mismatched (meaning horizontally matched).

The higher education demand is a nominal-dichotomous variable that classifies the participant as with or without a demand for further higher education. Participant who noted that during their current period of employment, they had not attained any additional academic degree, and in addition had not presently studied and/or planned to study in any framework, were classified as “doesn’t have a demand” (for further higher education). Moreover, participant who noted acquiring an academic degree during the current position, but currently was not studying and/or planned to study in the near future, for being horizontally mismatched and/or horizontally mismatched, was also classified as doesn’t have a demand. Conversely, participants who noted that they were currently studying and/or planned to continue studying in any framework, and/or during their current period of employment attained an additional academic degree, were classified as “has a demand” (for further higher education).

Job mobility is a variable that actually indicates the participant’s averaged seniority (in years) in each of his positions up till then, calculated by dividing the sum of the participants’ years since the end of the first academic degree, by the number of jobs employed during this period. The continuous variable was used as a dependent variable in the statistical processing for the second hypothesis. Job extent is a nominal dichotomous variable indicating the size of the participants’ current job, whether it’s a full-time job or a part-time job, as reported by the participant.

Career match is a variable that reflects the participants’ horizontal-match percentage throughout their career, comparing the occupational fields of each position in which they were employed to their (then) educational fields, in each career’s period. In other words, if the variable “number of positions in which the participant was employed” contained the Value 4, and it was stated that (for example) only two of them were compatible with the educational field at each point in time, then the participants horizontal-match percentage throughout the career is 2/4, that is, 50%. This continuous variable is an additional to the nominal “horizontal mismatch.” A possible problem in this variable might occur by allowing the participant to determine subjectively (and not objectively, like the case in the horizontal-mismatch variable) the match between the occupational field of each position the participant was employed to his educational field at each point of time during his career.

Education level is a nominal-dichotomous variable, which divides the observations into baccalaureates (the Value 0) and advanced degrees holders (the Value 1). Firstly, the variable contained three values (for baccalaureates, master degree holders, and doctoral degrees holders), but due to the pretty much equal division between the baccalaureates and the rest, the variable has been recoded into two categories.

Background variables are mainly sociodemographic variables as were reported by the participants, such as gender, marital status, age, income level, and job extent. Religious leaning was a nominal variable which included all the four relevant categories regarding the local Israeli population, both Jews and minorities (secular, traditional, orthodox, and other).

Since 65% of the participants reported themselves as seculars (i.e., nonorthodox), while the latter’s reported distribution in Israel is around 66%, then the values were recorded into two new different variables: seculars and nonseculars.

Results

First, a χ² test of independence was performed to examine the relationship between all the relevant nominal background variables, most of which were later used as independent variables in the various statistical tests performed to test the three hypotheses. These variables were gender, marital status, religious leaning, job size, vertical mismatch, and horizontal mismatch. It was found that the relations between the variables religious leaning and vertical mismatch, χ²(1) = 10.13, p < .01, religious learning and horizontal mismatch, χ²(1) = 13.33, p < .01, job size and gender, χ²(1) = 10.32, p < .01, religious leaning and gender, χ²(1) = 5.86, p < .05, horizontal mismatch and gender, χ²(1) = 7.29, p < .01, and vertical mismatch and horizontal mismatch, χ²(1) = 4.83, p < .05, were significant. Namely, secular people were more likely to be adequately educated than nonsecular; secular people were more likely to be horizontally matched than nonsecular; both males and females were more likely to work in full-time jobs than in part-time jobs; females were more likely to be secular than males; females were more likely to be horizontally matched than males; and horizontally mismatched people were more likely to be overeducated than horizontally matched. After these preliminary tests, more specific tests were conducted for each of the hypotheses.

Wage Gaps Hypothesis

Hierarchical multiple regression (stepwise method) was performed to investigate relevant variables’ (gender, vertical mismatch, horizontal mismatch, education demand, marital status, job scope, age, education level, career match, and job mobility) ability to predict the income level. Preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, and homoscedasticity. Additionally, the correlations among the continuous variables (age, career match, job mobility, and income level) included in the study were examined (Table 3), and the results showed that correlations were weak to moderate, ranging between r = .09, p < .2 and r = .45, p < .01, which indicates that multicollinearity was unlikely to be a problem (see Tabachnick & Fidell, 2007). All predictor variables were statistically correlated with the income level, indicating that the data were suitably correlated with the dependent variable for examination through multiple linear regression to be reliably undertaken. The correlations between the continuous predicting variables and the dependent variable (income level) were all weak to moderate, ranging from r = .09, p < .2 to r = .40, p < .01 (Table 3).

Table 3.

Pearson Correlation Matrix for the Sample.

Variables	Variable’s description	1	2	3	4
1	Career match		.40**	.09	.09
2	Income level	.40**		.26**	.19**
3	Age	.09	.26**		.45**
4	Job mobility	.09	.19**	.45**	—

Note. N = 221.

**Statistically significant at p < .01.

Since the hierarchical multiple regression’s method was stepwise, all the relevant variables were entered right from the first step and step by step—only the variables that significantly contributed to the model were entered, while the others were eventually excluded. In the first step of the regression, only career match was found significant to enter to the model. This latter was statistically significant, F(1, 219) = 37, p < .001, and explained 14.5% of variance in income level (Table 4). After entry of job extent at Step 2, the total variance explained by the model as a whole was 21%, F(2, 218) = 28.82, p < .001, whereas the introduction of job extent explained additional 6.4% variance in income level, R ² change = .065, F(1, 218) = 17.78, p < .001.

Table 4.

Hierarchical Regression Model of Income Level.

Steps	R	R ²	R ² Change	B	SE	β	t
Step 1	.38	.14***
Career match				1.57	.26	.38***	6.08
Step 2	.46	.21***	.06***
Career match				1.5	.25	.36***	6
Job extent				1.05	.25	.25***	4.22
Step 3	.5	.25***	.04***
Career match				1.24	.25	.3***	4.85
Job extent				1.14	.24	.27***	4.64
Education level				0.73	.22	.21***	3.4
Step 4	.51	.26***	.01**
Career match				1.18	.25	.29***	4.64
Job extent				1.13	.24	.27***	4.63
Education level				0.62	.22	.18**	2.8
Age				0.03	.01	.13*	2.17

*Statistically significant at p < .05. **p < .01. ***p < .001.

After the entry of education level at Step 3, the total variance explained by the model as a whole was 25%, F(3, 217) = 23.96, p < .001. The introduction of education level explained additional 4% variance in income level, R ² change = .04; F(1, 217) = 11.49; p < .001. In the final model, the participants’ age was found statistically significant, F(1, 216) = 4.7, β = .13, p < .05, and after its entry to the model, the total variance explained by the model as a whole was 26.5%, F(4, 216) = 19.45, p < .001, meaning that the introduction of age explained additional 1.6% variance in income level (R ² change = .016; Table 4).

The Demand for Higher Education Hypothesis

For testing a probabilistic connection of the demand for higher education and schooling continuation (the third hypothesis), two logistic regression analyses were conducted (Table 5). In the first logistic regression analysis, all the relevant predictive nominal variables were entered: gender, job extent, marital status, religious, and religious leaning, while the continuous predictive variables were income level, age, job mobility, and career match. In the second logistic regression analysis, the same predictive variables were entered (both nominal and continuous) as in the first one, while the two main studied variables—vertical/horizontal mismatch—were also entered in order to assess whether there was a contribution of these two variables to the explained variance of the model.

Table 5.

The Logistic Regression Steps.

Variable	Exp(B)	SE	B
Step 1
Job extent	0.47*	0.38	−0.75
Career match	15.17**	0.43	2.72
Age	0.96*	0.02	−0.04
Step 2
Career match	5.36**	0.53	1.68
Age	0.95	0.02	−0.05
Horizontal-mismatched	0.22**	0.45	−1.52

Note. R ² = .291 for Step 1; R ² = .349 for Step 2.

*p < .05. **p < .01.

In the significance tests of the first model, the logistic regression’s equation emerged significant at p < .01. The only variables which contributed to the explanation of the dependent variable—demand for education—and entered the regression equation in the final step were job extent (p < .05), career match (p < .01), and age (p < .05). The coefficient values of the regression of the variables, full-time job (B = −.75) and age (B = −.05), were negative, while the regression coefficient value of the variable career match was found positive (B = 2.72). The explained variance percentage by this model is R ² = .29.

In the second logistic regression’s significance tests, the regression equation emerged significant at p < .01. The only variables which contributed to the explanation of the dependent variable—demand for education—and significantly entered the regression equation in the final step are horizontal mismatch (p < .01), career match (p < .01), and age (p < .05). The regression coefficient values of the variables, horizontal mismatch (B = −1.52), and age (B = −.05) were negative, while the regression coefficient value of the variable, career match, was found positive (B = 1.68). This regression equation is Demand for education = 3.75 + .22 × horizontally_mismatched + 5.36 × career match + 0.95 × age.

The value of the regression coefficient of the variable career match is positive, so an increase of the variable career match increases the individual’s probability of having a demand for higher education by 18% (= 1/Exp (B) = 1/5.36), assuming that the rest of the model’s equation factors remain constant. For the other two significantly predictive variables whose regression coefficient values are negative, it can be determined that being horizontally mismatched reduces the individual’s probability to have a demand for higher education by a factor of 4.6 (= 1/Exp [B] = 1/0.22), compared to a horizontally matched individual, whose probability to have a demand for higher education is 4.584 factors (=Exp [B]) higher.

Furthermore, as each year passes, the individual’s probability to have a demand for higher education will be decreased by a factor of 1.05 (= 1/Exp [B] = 1/0.95), assuming that the rest of the model’s equation factors remain constant. The explained variance percentage of this model is R ² = .35, thus that the main variable—horizontal mismatch—of the two which were added in this analysis (for its significance in the model’s equation) alone contributed nearly 6% to that final explained variance percentage (Table 5).

Job Mobility Hypothesis

Hierarchical multiple regression (stepwise method) was performed to investigate relevant variables’ (gender, vertical mismatch, horizontal mismatch, education demand, marital status, job scope, age, education level, career match, and income level) ability to predict the individual’s job mobility rate. Preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, and homoscedasticity. Additionally, the correlations among the continuous predicting variables were already examined and presented above in Table 3. All correlations were weak to moderate, ranging between r = .098, p < .2 and r = .404, p < .01, which indicates that multicollinearity was unlikely to be a problem (Tabachnick & Fidell, 2007).

In this hierarchical multiple regression, the total variance explained by the model was 20.7%, F(1, 204) = 51.93, p < .001. Already during the first step, all the independent variables (besides the age) were excluded, with the age recording a relatively high β value (β = .111, p < .001). The latter means that with each year passes, the individual’s averaged job stability rate increases by 0.111, that is, the individual’s averaged seniority in each job/position increases by 0.111 per year as the time passes (Table 6).

Table 6.

Hierarchical Regression Model of Job Mobility.

Variable	R	R ²	B	SE	β	t
Age	.45	.203***	.111	.015	.45***	7.2

***Statistically significant at p < .001.

Discussion

The first and the third hypotheses were confirmed by the results: The horizontal-mismatch incidences throughout the individual’s career affects negatively the individual’s income level. Also, the horizontally mismatched individuals have less probability for a demand for further higher education than their horizontally matched counterparts.

Wage Gaps Hypothesis

The positive significant correlation that was found in the hierarchical regression between the income level and career match showed that for each 1% increase in the individuals’ horizontal-mismatch rate throughout the career (i.e., the more they are employed in horizontally matched jobs during their career), the income level increases by 1.18 (B = 1.18), assuming all other variables in the regression equation are held as constant. This finding is in line with Shaw (1984), who indicated that enhancing throughout the career vocational skills that are related to specific occupation usually increases future wage-level expectancy. Following this notion, Ordine and Rose (2011) suggested that possible wage “penalties” related to horizontal-mismatch incidences during the career might stem from inefficient self-selection in education, as well as from asymmetric information about the individual’s abilities, due to unoptimal utilization of those skills acquired in his studies. Other studies (Fu, 2011; Neumann, Olitsky, & Robbins, 2009) explained that the longer the employees experience a horizontal mismatch, the lower their probability to gain any meaningful vocational experience (which might affect the average income-level expectancy during their career).

The significant correlation that was found in the hierarchical regression between the variables age and income level means that with a 1 year increase of the individual’s age (can also be interpreted by 1 year increase in the individual’s seniority in the job/market), the income level increases by 0.03 (B = 0.03), assuming all other variables in the regression equation are held as constant. This finding might be explained by the individual’s accumulated vocational experience over the years, whether this experience is more diverse as a consequence of on-the-job-training or a job-specific experience (Sicherman, 1991). Employees who gain greater occupational experience can be rewarded by greater wages increase over the years, since the individual’s accumulative experience is strongly related to a higher level of wages (Becker, 1993). Also, the significant correlation that was found in the hierarchical regression between the variables job extent and income level is explained by the pretty-much obvious wage differences that were found between full-time job and part-time job employees: Full-time job employees income level is 1.13 times higher (B = 0.62), comparatively to part-time job employees with the same age, educational level, and career-match rate.

Moreover, the significant correlation that was found in the hierarchical regression between the variables educational level and income level shows that advanced degrees holders’ income level is 0.62 factors higher (B = 0.62), comparatively to baccalaureates with the same age, job extent, and career match rate. That might be simply explained by the return for the human capital investment: A recent Bank of Israel’s research revealed that an extra schooling year increases wages in the local market approximately by 8.6%, and that wage gaps between the educated employees and the less-educated ones related almost entirely to the educational-level gaps, what rejects the possible argument that those wage differentials are due to other features such as differences in intelligence quotient and motivation among the groups (Frish, 2007). In the wider aspect, after witnessing major increases in the spread of wages since the early 1980s, some Western decision-makers have already portrayed schooling as the best tool to erode the supposedly globalization-related forces that increase wage inequality (Martins & Pereira, 2004). As Ashenfelter and Rouse (2000, p. 111) put it, “The schooling is a promising place to increase the skills and incomes of individuals.”

Job Mobility Hypothesis

The second hypothesis postulated that overeducated individuals who are horizontally mismatched will experience higher job mobility than overeducated individuals who are horizontally matched, but the results did not fortify the hypothesis. Researches (Arthur & Rousseau, 1996; Ng & Feldman, 2007; Sullivan & Arthur, 2006) indicate that in the modern labor market, individuals tend to change jobs more frequently and abandon the traditional career model of holding one stable job (Cheramie, Sturman, & Walsh, 2007; Hall, 1996). This emphasizes that job mobility isn’t necessarily caused by horizontal mismatch; turnover and job mobility might be influenced and caused by the employees’ geographical location changes (Kan, 2002), the need to resolve work-family conflicts (O’Neill et al., 2009), job satisfaction (Green, 2010), and occupational burnout (Miller, 2010) as well.

The significant correlation found in the hierarchical regression between the age and job mobility means that with a 1 year increase of the individual’s age, the individual’s averaged seniority in the job increases by 0.111 years (B = 0.111). The aforementioned result is in line with previous findings that younger employees are less committed to a specific job in the long run, and that they usually tend to change jobs more frequently than older employees (Kirkpatrick-Johnson & Monserud, 2012). Moreover, Ng and Feldman (2009) reported a significant negative correlation, albeit a relatively weak one, between age and voluntary personnel turnover. Hedge, Borman, and Lammlein (2006) try to explain this tendency by the employers’ trends to fire younger employees during labor cutoffs in their organization. Therefore, younger employees might tend to resign more and try to look for better jobs that can provide them with relative stability. Feldman (2003) attributes this tendency to the nondecisiveness of younger employees dealing with career decisions, while older employees tend more to develop their “career anchors.”

The third hypothesis of the demand for higher education was confirmed. An explanation for this finding might emerge from the research about the investment in human capital under uncertainty (Olson, White, & Shefrin, 1979). Since individuals acquire from the outset a specific educational field in order to engage in it throughout their career (Robst, 2007), then once individuals work for long periods in an occupational field that differs from their educational field, and especially when this mismatch is involuntary, it might yield an occupational uncertainty since their educational investment has “paid off” in the market. This uncertainty might increase the risk aversion that may possibly affect the individual’s motivation for further investments in higher education (Bilkic, Gries, & Pilichowski, 2012; Henderson, 2007).

Further Higher Education Hypothesis

It was also found that the demand for higher education decreases as the individual’s age rises. The human capital theory (Becker, 1964, 1975) claims that higher education acquisition at an early age is much more worthwhile than at a later age, since the individual can utilize the acquired skills for a longer time. Moreover, it is suggested that seniority in the market and accumulated occupational experience can “behave” as substitute goods with the educational level, as individual’s cumulative work experience is used at his later career stages as a substitute for further investment in his education level (Wilson, Zozula, & Gove, 2011). The findings also revealed that the most critical factor associated with the demand for higher education is the horizontal match, rather than the wage level, for example. In line with the aforementioned human capital theory explanation, academic graduates might even settle for a lower income at the beginning of their career in order to gain experience relevant to their educational field (Leuven & Oosterbeek, 2011).

Methodological Considerations

Most of the vertical-mismatch research is correlational and leans on the basis of existing large databases (e.g., Dolton & Vignoles, 2000; Robst, 2007; Tsai, 2010). Due to procedural limitations and constraints, the current study was based on a limited sample size, and while a list of 153 jobs was given to the participants to choose their job, it perhaps might have been helpful to add a blank field for jobs that participants felt were not accurately described. Even though the statistical analyses used in the current article included all the possible and relevant variables, there might be some more variables that may explain the wage gaps (such as marital status change, geographical location—in central areas or in the periphery, specific employment terms at any position, etc.) and the demand for higher education (variables such as satisfaction from previous studies, loss aversion of investment under uncertainty, marital status change, etc.). Future research might investigate the possible effect of these aforementioned variables as explanatory variables.

Given society’s assignment of women as the primary caregivers, their disproportionate likelihood of being laid off, career disruptions, reentry into the workforce after breaks, and their tendency to settle for lower salaries and positions to balance family demands, some studies (De Jong & Madamba, 2001; Kalleberg, Reskin, & Hudson, 2000) cited that women were more likely to experience higher levels of underemployment comparatively to men, and therefore suffer disproportionately diverse career consequences. However, studies on underemployment and gender have not supported this assertion clearly and definitively, since the relationship of underemployment and gender and its outcomes are complex and somewhat counterintuitive (Thompson et al., 2013): Some studies found mixed results, near zero correlations, or nonsignificant relationships with gender (e.g., Sum & Khatiwada, 2010), while others even found higher underemployment among men than among women (Tam, 2010). Therefore, it could be well understood why no gender differences were found regarding the current article’s hypotheses, albeit the fact that the frequency between the two sexes was almost even in the sample.

Practical Implications

From the findings of the current article, it is highly recommended for academic studies candidates to make profitability and advisability analyses of the job supply offered by the labor market for their desired educational field, before investing in resources such as time, money, and so on. This should be undertaken in order to make an informed decision about investing in human capital, just as in any physical or financial investment, since, as was found in the current article, the horizontal-mismatch incidences might decrease the income-level expectancy during the career.

The Article’s Contribution, Further Research, and Limitations

First, the current research has indicated that the wage gaps (among both overeducated and adequately individuals) are attributed mainly to horizontal mismatch and not to the vertical mismatch itself. Since the last decades’ conspicuous increase in the supply of academically educated workers in the labor force of the developed-Western countries, academic schooling has become less “screening tool” for employers regarding potential candidates/applicants (Shavit & Muller, 1998).

As for the demand for higher education, the current article revealed that the most critical factor associated with the demand for higher education is the horizontal match, even more than the individual’s wage level, for example. That implies that, at least in the beginning of their career, the horizontal match might moderate the individuals’ possible uncertainty about the advisability of further higher education investments (Olson et al., 1979). Further studies might examine whether the individual’s horizontal mismatch does not directly affect the demand for higher education attainment. It might affect it through possible mediator variables, such as (a) the occupational uncertainty that the individual might experience (Fitzsimons, 2007) or (b) the level of the individuals’ risk propensity (Wong, 2005) toward further investment in their education, since horizontally mismatched might consider schooling as a nonworthwhile investment post factum.

The current career match variable actually reflects the individuals’ subjective interpretation for horizontal-mismatch experiences during their career rather than the objective matter (see horizontal match variable). During recent studies, the perception of underemployment is believed to be an important influence on employees’ attitudes and behaviors, that is, job dissatisfaction, low level of job involvement, feelings of being unappreciated, low mental health, and absence of career routes (Lee, 2005). Therefore, another research direction might continue investigating the possible subjective aspect of horizontal mismatch, as perceived by the individuals, and its possible antecedents and outcomes regarding the individuals’ career development.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Alba-Ramirez

(1993). Mismatch in the Spanish labor market: Over-education. Journal of Human Resources, 28, 259–278.

Aleksynska

Tritah

(2013). Occupation–education mismatch of immigrant workers in Europe: Context and policies. Economics of Education Review, 36, 229–244.

Arthur

M. B.

Rousseau

D. M.

(1996). The boundary-less career. New York, NY: Oxford University Press.

Ashenfelter

Rouse

(2000). Schooling, intelligence and income in America. In Arrow

Bowles

Durlauf

(Eds.), Meritocracy and economic inequality (pp. 89–117). Princeton, NJ: Princeton University Press.

Becker

G. S.

(1964). Human capital: A theoretical and empirical analysis, with special reference to education. New York, NY: Columbia University Press, National Bureau of Economic Research.

Becker

G. S.

(1975). A theory of social interactions. Journal of Political Economy, 82, 1063–1093.

Becker

G. S.

(1993). The economic way of looking at behavior. Journal of Political Economy, 101, 385–409.

Bilkic

Gries

Pilichowski

(2012). Stay in school or start working? The human capital investment decision under uncertainty and irreversibility. Labor Economics, 19, 706–717.

Borjas

G. J.

(1981). Job mobility and earnings over the life cycle. Industrial and Labor Relations Review, 34, 365–376.

10.

Busch

(2009). What kind of intercultural competence will contribute to students’ future job employability? Intercultural Education, 20, 429–438.

11.

Cheramie

R. A.

Sturman

M. C.

Walsh

(2007). Executive career management: Switching organizations and the boundary-less career. Journal of Vocational Behavior, 71, 359–374.

12.

De Jong

G. F.

Madamba

A. B.

(2001). A double disadvantage? Minority group, immigrant status, and underemployment in the United States. Social Science Quarterly, 82, 117–130.

13.

Dolton

Vignoles

(2000). The incidence and effects of over-education in the UK graduate labor market. Economics of Education Review, 19, 179–198.

14.

Feldman

D. C.

(1996). The nature, antecedents, and consequences of underemployment. Journal of Management, 22, 385–407.

15.

Feldman

D. C.

(2003). The antecedents and consequences of early career indecision among young adults. Human Resource Management Review, 13, 499–531.

16.

Feldman

D. C.

Turnley

W. H.

(1995). Underemployment among recent business college graduates. Journal of Organizational Behavior, 16, 691–706.

17.

Fitzsimons

(2007). The effects of risk on education in Indonesia. Economic Development and Cultural Change, 56, 1–25.

18.

Freeman

R. B.

(1976). The overeducated American. New York, NY: Academic Press.

19.

Frish

(2007). The causal effect of education on earnings in Israel (Discussion Paper 03). Jerusalem, Israel: Bank of Israel Research Department.

20.

(2011). Training, search and wage dispersion. Review of Economic Dynamics, 14, 650–666.

21.

Gicheva

(2012). Worker mobility, employer-provided general training and the choice of graduate education. Labor Economics, 19, 232–240.

22.

Green

(2010). Well-being, job satisfaction and labor mobility. Labor Economics, 17, 897–903.

23.

Hall

D. T.

(1996). Protean careers of the 21st century. Academy of Management Executive, 10, 8–16.

24.

Hedge

J. W.

Borman

W. C.

Lammlein

S. E.

(2006). The aging workforce: Realities, myths and implications for organizations. Washington, DC: American Psychological Association.

25.

Henderson

(2007). Valuing the option to invest in an incomplete market. Mathematics and Financial Economics, 1, 103–128.

26.

Kalleberg

Reskin

Hudson

(2000). Bad jobs in America: Standard and nonstandard employment relations and job quality in the United States. American Sociological Review, 65, 256–278.

27.

Kan

(2002). Residential mobility with job location uncertainty. Journal of Urban Economics, 52, 501–523.

28.

Kirkpatrick-Johnson

Monserud

M. A.

(2012). Work value development from adolescence to adulthood. Advances in Life Course Research, 17, 45–58.

29.

Lee

C. H.

(2005). A study of underemployment among self-initiated expatriates. Journal of World Business, 40, 172–185.

30.

Leuven

Oosterbeek

(2011). Over-education and mismatch in the labor market. Handbook of the Economics of Education, 4, 283–326.

31.

Martins

P. S.

Pereira

P. T.

(2004). Does education reduce wage inequality? Quantile regression evidence from 16 countries. Labor Economics, 11, 355–371.

32.

Maynard

D. C.

Joseph

T. A.

Maynard

A. M.

(2006). Underemployment, job attitudes and turnover intentions. Journal of Organizational Behavior, 27, 509–536.

33.

McGuinness

(2003). Graduate over-education as a Sheepskin effect: Evidence from Northern Ireland, Applied Economics, 35, 597–608.

34.

McGuinness

(2006). Overeducation in the labor market. Journal of Economic Surveys, 20, 387–418.

35.

McGuinness

Bennett

(2007). Over-education in the graduate labor market: A quantile regression approach. Economics of Education Review, 26, 521–531.

36.

Miller

J. F.

(2010). Burnout and its impact on good work in nursing. Journal of Radiology Nursing, 30, 146–149.

37.

Neal

(1999). The complexity of job mobility among young men. Labor Economics, 17, 237–261.

38.

Neumann

Olitsky

Robbins

(2009). Job congruence, academic achievement and earnings. Labor Economics, 16, 503–509.

39.

T. W. H.

Feldman

D. C.

(2007). Organizational embeddedness and occupational embedment across career stages. Journal of Vocational Behavior, 70, 336–351.

40.

T. W. H.

Feldman

D. C.

(2009). Re-examining the relationship between age and voluntary turnover. Journal of Vocational Behavior, 74, 283–294.

41.

Nordin

Persson

Rooth

D. O.

(2010). Education–occupation mismatch: Is there an income penalty? Economics of Education Review, 29, 1047–1059.

42.

Olson

White

Shefrin

H. M.

(1979). Optimal investment in schooling when income is risky. Journal of Political Economy, 87, 522–539.

43.

O’Neill

J. W.

Harrison

M. M.

Cleveland

Almeida

Stawski

Crouter

A. C.

(2009). Work–family climate, organizational commitment, and turnover: Multilevel contagion effects of leaders. Journal of Vocational Behavior, 74, 18–29.

44.

Ordine

Rose

(2011). Inefficient self-selection into education and wage inequality. Economics of Education Review, 30, 582–597.

45.

Robinson

S. L.

Kraatz

M. S.

Rousseau

D. M.

(1994). Changing obligations and the psychological contract: A longitudinal study. Academy of Management Journal, 37, 137–152.

46.

Robst

(2007). Education and job match: The relatedness of college major and work. Economics of Education Review, 26, 397–407.

47.

Romanov

Tur-Sinai

Eizman

(2008). Over-education, job mobility, and earnings mobility among holders of first degrees in Israel. Jerusalem: Israel Central Bureau of Statistics.

48.

Shavit

Muller

(1998). (Eds.). From school to work: A comparative study of educational qualifications and occupational destinations. Oxford, England: Clarendon Press.

49.

Shaw

K. L.

(1984). A formulation of the earnings function using the concept of occupational investment. Journal of Human Resources, 19, 219–340.

50.

Sicherman

(1991). Over-education in the labor market. Journal of Labor Economics, 9, 101–122.

51.

Sullivan

S. E.

Arthur

M. B.

(2006). The evolution of the boundary-less career concept: Examining physical and psychological mobility. Journal of Vocational Behavior, 69, 19–29.

52.

Sum

Khatiwada

(2010). The nation’s underemployed in the “Great Recession” of 2007-2009. Monthly Labor Review, 133, 3–15.

53.

Tabachnick

B. G.

Fidell

L. S.

(2007). Using multivariate statistics (5th ed.). Boston, MA: Pearson Education.

54.

Tam

(2010). Characteristics of the underemployed and the overemployed in the UK. Economic and Labor Market Review, 4, 8–20.

55.

Thompson

K. W.

Shea

T. H.

Sikora

D. M.

Perrewe

P. L.

Ferris

G. R.

(2013). Rethinking underemployment and over-qualification in organizations: The not so ugly truth. Business Horizons, 56, 113–121.

56.

Topel

R. H.

Ward

M. P.

(1992). Job mobility and the careers of young men. Quarterly Journal of Economics, 10, 439–479.

57.

Tsai

(2010). Returns to over-education: A longitudinal analysis of the U.S. Labor market. Economics of Education Review, 29, 606–617.

58.

Verhaest

Omey

(2006). The impact of over-education and its measurement. Social Indicators Research, 77, 419–448.

59.

Wilson

J. A.

Zozula

Gove

W. R.

(2011). Age, period, cohort and educational attainment. Social Science Research, 40, 136–149.

60.

Winefield

A. H.

Winefield

H. R.

Tiggeman

Goldney

R. D.

(1991). A longitudinal study of the psychological effects of unemployment and unsatisfactory employment on young adults. Journal of Applied Psychology, 76, 424–431.

61.

Wong

K. F. E.

(2005). The role of risk in making decisions under escalation situations. Applied Psychology: An International Review, 54, 584–607.