«DISCUSSION PAPER SERIES IZA DP No. 87 Educational Mismatch and Wages in Germany Thomas K. Bauer December 1999 Forschungsinstitut zur Zukunft der ...»
DISCUSSION PAPER SERIES
IZA DP No. 87
Educational Mismatch and Wages in Germany
Thomas K. Bauer
zur Zukunft der Arbeit
Institute for the Study
(GXFDWLRQDO 0LVPDWFK DQG
:DJHV LQ *HUPDQ\
,=$ %RQQ 8QLYHUVLW\ RI %RQQ DQG &(35 /RQGRQ
Discussion Paper No. 87
P.O. Box 7240
This Discussion Paper is issued within the framework of IZA’s research area *HQHUDO /DERU (FRQRPLFV Any opinions expressed here are those of the author(s) and not those of the institute. Research disseminated by IZA may include views on policy, but the institute itself takes no institutional policy positions.
The Institute for the Study of Labor (IZA) in Bonn is a local and virtual international research center and a place of communication between science, politics and business. IZA is an independent, nonprofit limited liability company (Gesellschaft mit beschränkter Haftung) supported by the Deutsche Post AG. The center is associated with the University of Bonn and offers a stimulating research environment through its research networks, research support, and visitors and doctoral programs. IZA engages in (i) original and internationally competitive research in all fields of labor economics, (ii) development of policy concepts, and (iii) dissemination of research results and concepts to the interested public. The current research program deals with (1) mobility and flexibility of labor markets, (2) internationalization of labor markets and European integration, (3) the welfare state and labor markets, (4) labor markets in transition, (5) the future of work, (6) project evaluation and (7) general labor economics.
IZA Discussion Papers often represent preliminary work and are circulated to encourage discussion. Citation of such a paper should account for its provisional character.
IZA Discussion Paper No. 87 December 1999 $%675$&7
This paper contributes to the literature considering the wage effects of educational mismatch in Germany. It uses a large German panel data set for the period from 1984 to 1997 and stresses the importance of controlling for unobserved heterogeneity when analyzing the labor market effects of over- and undereducation. Using pooled OLS, the estimation results confirm those found in the existing literature. The estimated differences between adequately and inadequately educated workers become smaller or disappear totally, when controlling for unobserved heterogeneity. In contrast to the common results using only cross-sectional data, the human capital model could not be rejected once controlling for unobserved heterogeneity.
JEL Classification: I21, J31 Keywords: Overeducation, undereducation, human capital model, job competition model Thomas K. Bauer IZA P.O. Box 7240 53072 Bonn Germany Tel.: +49-228-3894-305 FAX: +49-228-3894-210 Email: Bauer@iza.org * I want to thank Kurt Brännäs, Don DeVoretz, John Haisken-DeNew, Magnus Lofström, Anja Thalmaier, Rainer Winkelmann and Melanie Ward for their helpful comments on earlier drafts of this paper.
,,1752'8&7,21 Similar to most other developed countries, Germany has experienced an increased supply of highly educated workers in the last decade. In the period from 1985 to 1995 the enrollment of students in universities increased by more than 24%. In the same period, the share of the labor force with a university degree increased by 3.4 percentage points from 9.7% to 13.1%. The share of unemployed with an university degree increased by 2.6 percentage points from 3.8% in 1985 to 6.4% in 1995. At the first glance, these numbers indicate that the increased supply of high skilled workers has been accompanied by an increased demand. It might be, however, that the increased human capital of German workers is not utilized to the full due to an increased misallocation. If the growth of the supply of highly skilled workers outpaces the growth in the demand for these workers, overeducation of the work force might increase, with highly educated workers substituting the lower skilled.
The incidence and the labor market effects of educational mismatch have received increasing attention in recent years. In this literature workers are differentiated into workers which are adequately educated for the job they perform, overeducated if they have an educational attainment which is higher than that required for the job and undereducated if they have less schooling than required. Among the issues discussed in this literature are the effects of inadequate education on job and career mobility (Sicherman, 1991; Hersch, 1991; Robst, 1995a; Alba-Ramirez, 1992; Tsang, Rumberger and Levin, 1991; Sloane, Battu and Seaman, 1999) and on job satistifaction (Hersch, 1991; Tsang, Rumberger and Levin, 1991). Most of the existing studies study the wage effects of inadequate schooling (Hartog and Oosterbeck, 1988; Groot, 1993, 1996; Kiker, Santos and Oliveira, 1997; Alba-Ramirez, 1992; Hersch, 1991; Cohn and Kahn, 1995;
Duncan and Hoffman, 1981; Verdugo and Verdugo, 1989; Rumberger, 1987; Sloane, Battu and Seaman, 1999). In general, the latter studies find that the returns to years of overschooling are positive, but significantly lower than the returns to adequate education and that the returns to undereducation are negative.1 See Groot (1993) and Kiker, Santos and Oliveira (1997) for a survey of the wage effects of inadequate education.
This paper expands the existing literature in two respects. First, using a panel data set - the German Socioeconomic Panel (GSOEP) - the paper provides the first investigation of the extent and the wage effects of over- and underschooling in Germany using two different measures of inadequate education. Second, existing studies of the labor market effects of educational mismatch are based solely on cross-section data. It may be the case that the results of these studies are biased due to unobserved heterogeneity. Controlling for unobserved heterogeneity might for example be important when individuals with lower innate ability need more education to attain a job for which they are formally overeducated. Therefore, this paper utilize the panel structure of the data set by estimation random and fixed effects panel estimators to analyze whether previous results concerning the wage effects of inadequate schooling remain when controlling for unobserved individual effects.
Section II gives a brief description of the data set together with a discussion of possible measures of over- and undereducation. This section further outlines the incidence and development of educational mismatch in Germany. The econometric approach used in the analysis is described in Section III. The estimation results are presented in Section IV. The estimation results for pooled OLS are presented first to in order to compare the wage effects of educational mismatch to the results of studies for other countries. Then the results from random and fixed effects models are presented and compared to the results from the pooled OLS estimations to analyze the importance of controlling for unobserved heterogeneity. Section V concludes.
,, 29(5 $1' 81'(5('8&$7,21,1 *(50$1 The sample used in this study is drawn from the German Socio-Economic Panel (GSOEP), a panel data set for the years 1984 to 1997 consisting of about 13,500 individuals and 7,000 households living in Germany. All full-time employed, prime-aged males and females of German nationality were selected. Foreigners and East Germans were excluded in order to study just those individuals who received their education in a similar schooling system. In addition, self-employed and those currently in education or training were not considered. After eliminating all observations with missing values, a final sample of 18,647 person-year observations consisting of 2,945 individuals remained for the analysis.
Educational mismatch is usually measured as the difference between the educational attainment of an individual and the skill requirements of the job the individual performs, i.e., a worker is considered to be overeducated (undereducated) if he has more (less) education than is required for his or her job and adequately educated if his or her education just meets the job requirements. In order to classify a worker to be inadequately educated one has to estimate the amount of schooling required for a particular job. Different approaches have been discussed in the literature to determine the amount of required education. Kiker, Santos and Oliveira (1997) and Sicherman (1991) give a detailed discussion of the advantages and disadvantages of different measures of over- and undereducation.
Several studies have used an exogenous definition of schooling requirements for the particular occupation the individual is working in order to determine required schooling (Rumberger 1987; Hartog and Oosterbeck, 1988; McGoldrick and Robst, 1996; Kiker, Santos and Oliveira, 1997; Sloane, Battu and Seaman, 1999). A major criticism of this measure arises through the potential measurement errors resulting from translating categorical values of educational requirements into a single continuous variable of years of schooling (Sicherman, 1991). Furthermore, if such a measure is very rigid, misclassifications of workers as over- or undereducated might increase over time, if adjustments of schooling requirements occur due to technological progress or changes in workplace organization (Kiker, Santos and Oliveira, 1997). In Germany, there is no equivalent of an exogenous definition of schooling requirements for different occupations available, such as the 'LFWLRQDU\ RI 2FFXSDWLRQDO 7LWOHV in the US. Hence, this type of measurement of educational mismatch cannot be used.
A second approach is to ask workers directly whether they are over- or undereducated for the job they perform (Sicherman, 1991; Hersch, 1991; Robst, 1995a, 1995b;
McGoldrick and Robst, 1996). The main advantage of this measure is that it is jobspecific. It has been criticized, however, due to its subjective nature. Furthermore, the type of question which is used to determine educational mismatch usually refers to the education needed to get a job, not to the education necessary to perform a job. Even though the GSOEP provides a question about the skills required to perform the job an individual is employed in, this information is not suitable for the empirical analysis of the wage effects of inadequate education in Germany either. The categorical values provided within this question cannot be converted into a single measure of years of schooling which is comparable with the coding of actual years of schooling provided in the data set.
The following empirical analysis makes use of two other possible measures of required schooling. The first measure, which follows from Verdugo and Verdugo (1989), defines required schooling as an one-standard-deviation range around the mean level of schooling within occupations disaggregated at a 3-digit level. Workers are considered to be adequately educated if their actual education falls within this range, overeducated if their actual education is greater than one standard deviation above the mean for the specific occupation, and undereducated if their actual education is more than one standard deviation below the mean education. The main criticism of this measure has also been its subjectivity, since there is no rationale behind the arbitrary choice of one standard deviation. Furthermore, this type of measure results in estimated levels of inadequate schooling, which are far below those of other measures. As Kiker, Santos and Oliveira (1997) note, this measure is more sensitive to technological changes and changes in workplace organization than the others, potentially suggesting misleading conclusions about the development of inadequate schooling over time.
The second measure of required schooling used in this study is suggested by Kiker, Santos and Oliveira (1997). They suggest using the modal value within a given occupation to measure required schooling. Hence, workers are considered to be adequately educated if their actual education equals the mode value of education within their occupation. Those workers whose actual educational attainment is above the modal education of their specific occupation are considered to be overeducated, those whose actual education falls below the mode are considered to be undereducated. Similar to the measure of Verdugo and Verdugo (1989) the mode-index is determined within 3-digit occupations. The main advantages of this measure over the one proposed by Verdugo and Verdugo (1989) is that it is less sensitive to technological changes and that it provides estimates of inadequate schooling, which are more in line with the first two approaches.