The length of youth unemployment

142 G. Lassibille et al. Economics of Education Review 20 2001 139–149

3. The length of youth unemployment

As noted previously, the survey permits measurement of the length of unemployment of both cohorts of school leavers on a discrete scale only, due to data collection observations on both cohorts being right-censored at 18 months. Table 3 reports the distribution of young people according to the length of unemployment prior to the first job. Results show that 39 percent of school leavers obtain a first job 6 months after completion of formal schooling, 22 percent are unemployed for 6 to 18 months, and about 40 percent are still unemployed 18 months after leaving the education system. Analysis by gender indicates that among young men and women dif- ferences in length of unemployment are quite large: women are considerably less likely than men to find a job 6 months after completing formal schooling 33 per- cent vs. 44 percent, and they are much more likely to be still unemployed 18 months after leaving the education system 45 percent vs. 34 percent. While it would be useful to compare the situation in Spain with those of other European countries, we are only able to emphasize differences in the length of youth unemployment between France and Spain 6 . By relating our results with the data from the French Enqueˆte Emploi, conducted approximately at the same period, we find that Spanish school leavers are on average in a worst situation compared to their French counterparts. In fact, 50 percent of French school leavers obtain a first job 6 months after leaving the school; this percentage is 12 points higher than in Spain. Furthermore, while less than 20 percent of French school leavers are still unemployed 12 months or more after completion of formal schooling, 40 percent of Spanish school leavers are still unem- Table 3 Length of youth unemployment percent Females Males Total Less than 6 months 33.41 43.74 38.56 Between 6 and 18 21.92 22.05 21.98 months More than 18 44.67 34.21 39.45 months Total 100.00 100.00 100.00 6 The paper by Andrews and Bradley 1997 gives interest- ing insights into transition from school in the United Kingdom. However, their results are not comparable because the labor market outcomes considered by these authors differ from the labor market situation we analyze in this paper. Furthermore, Andrew and Bradley’s article focuses only on the situation of school leavers at the end of compulsory education. ployed 18 months after leaving the education system. Differences between the two countries may arise for at least three reasons: 1 overall conditions of the labor market are different in both countries, 2 the French educational system is organized in a way which makes the transition from school to work easier see Lassib- ille Navarro Go´mez, 2000; 3 the relative size of youth cohorts differ in both countries see Korenman Neumark, 1997. Several studies have shown that the burden of unem- ployment is not shared equally among young people see, for example, Blau Kahn, 1997; Edin, Forslund Hol- mund, 1996; Franz, Inkmann, Pohlmeier Zimmerm- ann, 1997. In the following, we contribute to the exist- ing literature on the topic by explaining the probability of young Spanish people finding a first job in 6 months or between 6 and 18 months, compared with that of find- ing a job in more than 18 months. Table 4 reports the results of these estimations using a multinomial logit specification; 7 we follow tradition by computing the mar- ginal effects of the explanatory variables see, for example, Greene, 1997. The length of unemployment is explained by the following individual attributes: age, gender, highest level of education completed, partici- pation in non-formal programs, father’s occupation, place of residence and cohort. Apart from these personal characteristics, we control the transition process from school to work for local labor market conditions, includ- ing explicitly in the regression model the unemployment rate of the entire population in the home region. We take also into account the structure of employment at the regional level, using the ratio between the number of workers employed in the service sector and the number of workers employed in the industrial sector. The results show that, with all else being the same, females are less likely to find their first job in less than 18 months compared with males; an observation which clearly indicates a discrimination against women at a very early stage of their working life. 8 As formulated by Rees and Gray 1982, youth unemployment may depend on the contacts or the influence parents bear on the labor market. In this case, the greater the parents’ influence the lower the probability of being unemployed. In our analysis, father’s occupation is viewed as a proxy for influence. However, regression results show that, with everything else constant, social class background is not statistically significant. In other words, at the beginning 7 We do not model the decision to stay on at school prior to the labor market status of individuals because the Encuesta Socio Demografica contains no information on the academic performance of pupils, which is a major determinant of the decision to stay on at school. 8 For an analysis of gender differences in the Spanish labor market, see Lassibille 1998. 143 G. Lassibille et al. Economics of Education Review 20 2001 139–149 Table 4 Multilogit estimates of the probability of employment a marginal effects In less than 6 months Between 6 and 18 months Coefficient t-statistics Coefficient t-statistics Intercept 0.1246 0.93 20.0062 20.05 Age 0.0019 0.26 20.0040 20.66 Sex 0.0843 3.33 0.0163 0.76 Highest educational degree Vocational education 20.0375 20.90 0.0800 2.20 Upper secondary education 20.1959 24.10 0.0246 0.59 Higher education Escuela Universitaria 20.0259 20.38 0.1114 1.91 Facultad or ETS 20.1403 21.93 0.1348 2.17 Non-certified years of schooling 0.0334 0.97 0.0392 1.29 Participation in non-formal 20.0676 22.15 0.0302 1.32 education program Father’s occupation Self-employed 20.0243 20.70 20.0352 21.18 Managers and professionals 20.0262 20.73 20.0281 20.93 Skilled workers 20.0179 20.50 0.0052 0.17 Living in town of population Between 20,000 and 100,000 0.0775 2.02 20.0070 20.22 Over 100,000 0.0645 1.88 20.0178 20.63 Characteristics of the labor market in the home region Rate of unemployment 20.0026 20.98 20.0054 22.31 Size of the service sector 20.0258 22.42 20.0014 20.16 Cohort 8990 20.0629 21.44 0.0602 1.74 Number of observations 1638 Chi 2 32 134.42 Log-likelihood 21675.29 a Individuals unemployed for more than 18 months is the comparison group. of their professional career young people from lower socio-economic cohorts are not faced with more restric- ted access to networks for job recruitment compared with their counterparts from a higher socio-economic group. To appreciate better the effect of education on the length of employment we use regression results in Table 4 to simulate the employment status of school leavers at the three points in time in question. Table 5 displays the results of these simulations. The results show that, with all else equal, the level of education has a strong effect on the duration of unemployment. Compared with other individuals, young people who have left school with an upper secondary education diploma are in a worse situ- ation. Their probability of being unemployed for more than 18 months is respectively 42 percent and 34 percent higher than those who have a vocational or a compulsory education diploma. It is noteworthy that those people having a higher education diploma have less difficulties in finding a first job in less than 18 months, compared with young people with compulsory or upper secondary education; among individuals who leave the schooling system with a higher education diploma, those who have Table 5 Simulated probability of employment a In less than Between 6 In more than 6 months and 18 18 months months Compulsory 0.39 0.15 0.46 education Vocational 0.39 0.23 0.38 education Upper secondary 0.12 0.08 0.80 education Higher education Escuela 0.42 0.28 0.30 universitaria Facultad or ETS 0.39 0.15 0.46 a Based on regression results in Table 4. Simulations for the following individuals: 24-year-old male, father unskilled worker, living in a town between 20,000 and 100,000, regional rate of unemployment = 16.47, relative size of the tertiary sec- tor = 2.55. 144 G. Lassibille et al. Economics of Education Review 20 2001 139–149 an escuela universitaria diploma are in a better situation compared with those who have a facultad or an ETS dip- loma. Another feature to emphasize is that people who leave the schooling system with a vocational diploma have a higher probability of getting an acceptable job offer in 18 months or less, compared with individuals who have just a compulsory level or an upper secondary school diploma. The most remarkable findings so far are: 1 after con- trolling for personal characteristics and overall con- ditions of the regional labor market, individuals with an upper secondary school diploma have problems in find- ing acceptable job offers; 2 young people with a vocational or a higher education diploma have the short- est duration of unemployment; and 3 individuals who have just a compulsory level diploma are in a better situ- ation compared with those who have an upper secondary school diploma; however, they have more difficulties entering the labor market compared with those who have a vocational or an escuela universitaria diploma. The regressions results in Table 4 show that non-certi- fied years of schooling have no significant impact on the length of unemployment; a result which might confirm the predominance of signaling in the labor market. As this effect probably differs according to the level of edu- cation, we include in a second specification an interac- tion term between the highest diploma obtained and the non-certified years of schooling an individual may have. 9 The results indicate that the labor market rewards sig- nificantly non-certified years of schooling for those indi- viduals who have started higher education after either a vocational education diploma or upper secondary edu- cation. Simulations not displayed here show that non- certified years of higher education reduce by respectively 13 percent and 18 percent the probability of these indi- viduals finding a job in more than 18 months. According to Table 4, participation in non-formal edu- cation programs has no negative impact on the length of unemployment. To what extent does this effect depend on the level of formal education of each individual? To answer this question we re-estimate the probability of employment adding an interaction term between formal and non-formal education. 10 The results indicate that non-formal education programs reduce significantly the length of unemployment of young people with upper sec- ondary or vocational education. Finally, regression results in Table 4 indicate that the higher the unemployment rate in the home region, the greater the length of youth unemployment—a common result in studies on youth unemployment. We observe also that the larger the size of the services sector in a region, the higher the length of youth unemployment; 9 Results available from the authors on request. 10 Results available from the authors on request. otherwise stated, regions where employment is concen- trated in services industries generate less jobs.

4. Mismatch in the youth labor market