Samar Abdelmageed, The British University in Egypt (BUE)
Egypt is a young society with more than 20 million people within the age group (18-29) years, representing about 21% of the total population according to Egypt’s Central Agency for Public Mobilization and Statistics (CAPMAS) latest released statistics. Despite the increasing levels of educational attainment among the youth and the progressively narrowing gaps between genders in education, young Egyptian university graduates suffer from high unemployment rates, especially females. As of 2020, the unemployment rate for Egyptian youth between 15 to 24 years of age was 30% unemployment, with 25% among males and 42% among females (Danish Trade Development Unit, 2020).
This problem of youth unemployment goes hand in hand with other deficiencies embedded within the Egyptian labor market including the modest rates of job creation, stagnating productivity levels, increasing informalities, and low labor wages. A comparison of the youth and adult or prime-age employment patterns usefully differentiates between the problems exclusive to each age group and those which are inherent in the Egyptian economy and labor market. This classification would help design effective targeting policies. Furthermore, it is important to study the dynamics of employment and unemployment during crises, such as the Arab Spring and the revolutions witnessed by Egypt starting from 2011, to investigate their impacts compared to periods before and after the emergence of these crises and distinguish between problems and issues that are temporary and those which are inherited in labor markets.
Therefore, the main aim of this paper is to analyze youth employment and unemployment patterns in the Egyptian labor market over the years to detect any changes over time and analyze the causes behind these changes, if any. The paper attempts to answer three main research questions: what are the patterns of youth transitions among the employment, unemployment, and inactivity statuses? What are the characteristics that affect the probability of falling into a specific status? And how do transitions from one status to another differ between the youth (18-34) years and individuals in the prime-age category, identified similar to Flek and Mysíková (2016) as those who are between 35 and 54 years of age? The data used in my analysis comes from four rounds of Egypt Labor Market Panel Survey (ELMPS) conducted in 1998, 2006, 2012, and 2018. The survey offers detailed information on the development of the Egyptian labor market during this twenty-year period and is considered to be a rich source of data on a variety of topics related to the Egyptian labor market and the socioeconomic characteristics of the sampled individuals. Additionally, these waves enable the research to analyze the dynamics and transitions among the employment, unemployment, or inactivity statuses over an adequate period of time. To examine dynamics, the methodology of analysis depends on building a series of multinomial logistic models to analyze the characteristics, including age, that affect the probability of transitioning between different working statuses (employment, unemployment and inactivity), for each two consecutive rounds of ELMPS.
One key finding which emerges from the analysis is that age was not found to be related to moving from unemployment to employment in the Egyptian labor market. All cohorts had difficulty transiting from unemployment to employment, especially in the recent years that followed the eruption of the Arab Spring in Egypt in 2011. The analysis also revealed a steady movement of females out of the labor force and the tendency of workers with higher educational levels to step out of the job market compared to workers with no educational background. These results indicated the problems of mismatch between job supply and demand; the insufficient formal job creation in the Egyptian labor market; and the educational system’s inability to improve employability among its graduates. Moreover, young workers with better educational outcomes who also belong to families with higher educational backgrounds have bigger chances of leaving the labor force, which reflects the association between education and wealth that still persists in Egypt.
Literature Review
In general, youth face more struggles in labor markets compared to more experienced workers, especially when they first enter the market searching for jobs. The reasons behind these struggles vary by economy but may widely include mismatches between job supply and demand, and the insufficient formal job creation in the private sector and the associated queuing for formal public jobs especially in the developing countries (Flek and Mysíková 2016, Nilsson 2019). Moreover, young workers usually suffer from higher turnover compared to their prime-aged counterparts; however, on the other hand, the employability of older workers decreases over time since younger job seekers may accept jobs with worse-off job conditions such as low wages and work instability (Flek and Mysíková 2016).
Studies on employment dynamics in the Middle East and North Africa (MENA) region are generally scant due to the lack of comprehensive longitudinal data that detect transitions into and out of employment and the related socio-economic characteristics of the working-age populations over time in the region’s labor markets. However, the recent availability of a group of panel labor market surveys that have been conducted by the Economic Research Forum (ERF) in few MENA countries over time offered an opportunity for a number of papers to study the dynamics of employment focusing on youth in MENA. Assaad and Krafft (2016) employed the data of the labor market panel surveys conducted in 1998, 2006 and 2012 for Egypt; in 2010 for Jordan; and in 2014 in Tunisia to study the dynamic movements of youth into and out of employment in these countries. The paper highlighted some of the chronic issues that face the youth in their pursuit for work, including their common behavior of seeking a formal job at first, then resorting to informal employment when their efforts fail, if they are desperate to work, or giving up their place in the job market and turning to inactivity, if they and their families can afford that, which is the usual case for young women in MENA.
Young Egyptian workers who enter the labor market for the first time tend to face very long unemployment (two years or more); therefore, this insertion dilemma is one of the main contributors to the problem of unemployment in Egypt. This conclusion was also reiterated in Assaad and Krafft (2021) and AlAzzawi and Hlasny (2020), which used data from the different ELMPS rounds in their analyses and emphasized the high unemployment probability of young Egyptian university graduates, who mainly wait for formal employment, and the difficult job mobility for those who take over informal jobs. Other studies also highlighted the slow transition from school to work which prevails among young Egyptian workers (Angel-Urdinola and Semlali 2010, Assaad 2007), and their increasing vulnerable employment as a result of the insufficient formal job generation in the Egyptian economy (Gadallah 2011, Assaad et al 2016). Therefore, it is important to track the unemployment/employment dynamics among young Egyptians and examine any relevant changes over time. It would also be interesting to compare these dynamics among youth to older workers to disentangle the dynamics related to Egyptian young workers from those of older and perhaps more experienced workers who might have better opportunities in the labor market.
Data and Methods
This paper aims to detect unemployment/employment transitions in the Egyptian labor market over time especially among young workers. Data used in analysis mainly come from the ELMPS rounds of 1998, 2006, 2012 and 2018. This survey is a longitudinal study of the Egyptian labor market that provide rich datasets of the socio-economic characteristics of its representative samples of respondents, which have been used extensively in the literature and can be employed by this paper to track the transitions and dynamics of employment/unemployment over time. In each round of ELMPS, a refresher sample of between 2,000 to 3,000 households are surveyed to ensure the representative of the samples throughout the years. The final sample of ELMPS 2018 included 61,231 individuals coming from 15,746 households representing people who live in different parts of Egypt coming from various backgrounds and characteristics (Krafft, Asaad and Rahman 2018). The presence of such longitudinal data offers an opportunity to analyze employment dynamics and compare these dynamics during crises, if existed (Kelly et al 2014).
To examine employment transitions, the paper starts with a descriptive analysis of the dynamics among young workers compared to prime-aged workers over time. Youth are defined as those belonging to the (18-34) years of age category, while prime-aged individuals in the prime-age category are identified similar to Flek and Mysíková (2016) as those who are between 35 and 54 years of age. These two groups are chosen to compare young labor market entrants to those who are more experienced, including mid and senior level workers.
The paper follows its descriptive analysis with a series of multinomial logistic regression models to study factors that affect the transition from unemployment to other statuses including employment and inactivity (out of labor force) between each two consecutive ELMPS rounds.
Similar to Assaad and Krafft (2016), the study will depend on broad unemployment, which does not require the active search for work but the availability and readiness to work, also using the market definition that excludes subsistence workers. Furthermore, the independent variables selected for analysis, which are commonly used in the literature to analyze the dynamics of unemployment, include age, gender, education, area of residence (urban/rural) and the father’s level of education (Cincǎ and Matei, 2018; Assaad and Krafft, 2016, 2021; AlAzzawi and Hlasny, 2020). Data used in the survival analysis and examining factors that affect unemployment exit mainly come from ELMPS 2018, which is the only survey that includes a right-censoring variable to identify unemployment vs. exit from unemployment as well as the unemployment duration estimated based on the employment history for those who have worked before and are new and currently unemployed, in addition to those who have never worked even if not currently unemployed. For a full specification of the model, see the Appendix.
Main Findings
Figure 1 shows that, according to the data of different ELMPS rounds over time, youth unemployment rates are stable and constantly higher than those of prime-age workers. However, total unemployment rates increased by 11.24% between ELMPS 1998 and ELMPS 2018 and prime-age workers witnessed an increase in their unemployment rates from 1.64% in ELMPS 1998 to 4.49% in ELMPS 2018. This may refer to problems of lay-offs faced by prime-age workers after the eruption of the Arab Spring in Egypt in 2011. The descriptive analysis of the evolution of current unemployment durations between ELMPS 1998 and 2018 presented in Table 1, highlights their constant increase over the years for all age groups, especially in 2012 and 2018, implying that it is getting more and more difficult to obtain a new job for someone who loses his/her job. It is also worth mentioning that the unemployment duration for the prime-aged has substantially increased above that of young workers in ELMPS 2018.
Figure 1. Unemployment rates, total, youth (18-34 years) and prime-aged (35-54 years), ELMPS 1998-2018
Source: based on ELMPS 1998, 2006, 2012 and 2018
Table 1. Current unemployment durations (in months) by ELMPS round (1998-2018), total, youth (18-34 years) and prime-aged (35-54 years) *
(months)
Mean | Std dev. | Min | Max | |
ELMPS 1998 | ||||
Total | 31.67 | 31.39 | 1.00 | 132.00 |
Youth (18-34 years) | 33.74 | 31.50 | 1.00 | 132.00 |
Prime-aged (35-54 years) | 27.85 | 35.73 | 1.00 | 132.00 |
ELMPS 2006 | ||||
Total | 36.91 | 31.66 | 2.23 | 132.00 |
Youth (18-34 years) | 37.84 | 31.69 | 2.23 | 132.00 |
Prime-aged (35-54 years) | 38.65 | 34.97 | 2.69 | 120.00 |
ELMPS 2012 | ||||
Total | 57.03 | 48.44 | 3.00 | 188.00 |
Youth (18-34 years) | 57.98 | 47.82 | 3.00 | 188.00 |
Prime-aged (35-54 years) | 52.04 | 55.47 | 3.00 | 178.00 |
ELMPS 2018 | ||||
Total | 68.00 | 68.88 | 1.00 | 258.00 |
Youth (18-34 years) | 62.44 | 58.08 | 1.00 | 242.00 |
Prime-aged (35-54 years) | 88.74 | 92.09 | 1.00 | 258.00 |
*Excluding observations that are below the 5th or above the 95th percentiles of the unemployment durations
Source: based on ELMPS 1998, 2006, 2012 and 2018
Examining movements between the different working statuses (employed, unemployed and out of labor force) show the decreasing transitioning from unemployment to employment among the youth over time, especially between 2012 and 2018 compared to the period between 2006 and 2012 [Figure 2]. Transitions from unemployment to inactivity prevails among young females [Figure 3]. Figures 4 and 5 show the reduction in the percentages of prime-aged workers who transform from unemployment to employment between 2006 and 2012 and between 2012 and 2018 compared to 1998 and 2006; additionally, unemployed females in this age category are finding it more difficult over time to transition from unemployment to employment. Furthermore, the percentage of females who move from unemployment to out of labor force is higher among prime-aged workers compared to the young women.
Figure 2. Transitions between different working statuses (employment (emp.), unemployment (unemp.) and out of labor force (OLF)) for each two consecutive ELMPS rounds, youth (18-34 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Figure 3. Transitions between different working statuses (employment (emp.), unemployment (unemp.) and out of labor force (OLF)) for each two consecutive ELMPS rounds, female youth (18-34 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Figure 4. Transitions between different working statuses (employment (emp.), unemployment (unemp.) and out of labor force (OLF)) for each two consecutive ELMPS rounds, prime-aged (35-54 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Figure 5. Transitions between different working statuses (employment (emp.), unemployment (unemp.) and out of labor force (OLF)) for each two consecutive ELMPS rounds, female prime-aged (35-54 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Exploring the vulnerability of jobs obtained by those who transitioned from unemployment to employment between each two consecutive ELMPS rounds, in terms of informality identified by the lack of a working contract, highlights the dominance of informal jobs among both young and prime-aged workers over time [Figures 6, 7 and 8]. The percentage of unemployed prime-aged workers who obtain informal jobs between ELMPS 1998 and 2006 and between ELMPS 2006 and 2012 is higher than that of young workers. This implies the spread of informality among workers in the Egyptian labor market. Moreover, unemployed prime-aged workers, who could be the main breadwinners in their households, might not afford to wait until obtaining a formal job compared to young workers who can depend on their families for financial support.
Figure 6. Transitions from unemployment to employment (unemp-emp) between ELMPS 1998 and 2006 and the formality of obtained jobs (existence of a working contract), total, youth (18-29 years) and prime-aged (35-54 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Figure 7. Transitions from unemployment to employment (unemp-emp) between ELMPS 2006 and 2012 and the formality of obtained jobs (existence of a working contract), total, youth (18-29 years) and prime-aged (35-54 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
Figure 8. Transitions from unemployment to employment (unemp-emp) between ELMPS 2012 and 2018 and the formality of obtained jobs (existence of a working contract), total, youth (18-29 years) and prime-aged (35-54 years)
Source: based on ELMPS 1998, 2006, 2012 and 2018
To inspect factors that may affect the transition from unemployment, in particular, to other working statuses, the study employs a series of multinomial regression models whose aim is to predict the probabilities of leaving unemployment based on a group of independent variables that include gender, age, area of residence, educational attainment and father’s level of education. Tables 1, 2 and 3 in the appendix present the results of these models for the transitions between each two consecutive ELMPS rounds. The analysis highlights the significant continuous impact of gender over the years on the probability of exiting unemployment with lower odds ratios of transitioning to employment and higher odds ratios of moving out of the labor force compared to males. Results of the models also show the effect of higher educational levels which are associated with higher probabilities of transiting from unemployment to out of the labor force compared to illiterates. Additionally, age does not seem to have a significant impact on leaving unemployment.
The next step in analysis is to investigate unemployment durations and examine the hazards of exiting unemployment focusing on young workers and using the data of ELMPS 2018. This ELMPS round allows for estimating unemployment durations using retrospective employment history data for those who have worked before and are new and currently unemployed, in addition to those who have never worked even if not currently unemployed. Graphs of the Kaplan-Meier survival function show the higher probabilities of spending longer durations of unemployment among young females compared to male workers who exit unemployment faster, and the tendency of higher educational levels to exit unemployment to other working statuses compared to workers with no educational degree [Figure 9].
Figure 9. The Kaplan-Meier Survival Function (proportion remaining unemployed), youth (18-34 years), total and by gender, area, educational level, and father’s education, ELMPS 2018
In addition, fitting a Cox proportional hazards model by using the data of workers from all age groups shows that young compared to prime-aged workers; females compared to males; and workers with no educational degree relative to those who hold one, all tend to spend longer times in unemployment [Table 4 in the Appendix]. Figure 10 shows the Kaplan-Meier Survival estimates by age group based on the results of the Cox proportional hazards model and refers to the higher survival rates in the state of unemployment among young compared to prime-aged workers.
Figure 10. The Kaplan-Meier Survival estimates by age group based on the results of the Cox proportional hazards model for predicting the probability of exiting unemployment, ELMPS 2018
Since the survival analysis of unemployment here examines discrete time units, the Cox proportional hazards model might not be very appropriate to analyze the unemployment dynamics. Therefore, the study fits a complementary log-log model to examine factors that affect the probability of separation from first unemployment (exit to employment or out of labor force) among young workers. The model is first estimated using Gamma frailty [Table 5 in the Appendix], which accounts for individual heterogeneity, if existing in the data; however, results of the test of rho = 0 indicates that frailty is not significant.
Table 6 in the Appendix shows that, for young workers, spending more time in unemployment is associated with higher probability of remaining unemployed. Females have higher probabilities of remaining unemployed compared to males. In addition, young workers with intermediate or above levels of education experience higher probabilities of remaining unemployed compared to having no education. Furthermore, young workers whose fathers are university graduates have higher probabilities of remaining unemployed compared to young workers whose fathers are illiterates. The margins plots associated with the model are also displayed in Figure 11.
Figure 11. The margins plots of the categorical independent variables included in the complementary log-log model for predicting the probability of exiting unemployment among young workers, ELMPS 2018
Conclusion
The main aim of this paper was to analyze the employment dynamics among young workers in Egypt and compare these dynamics to workers of older age groups. Data used in analysis came from the different ELMPS rounds carried out in the years of 1998, 2006, 2012 and 2018. Results highlighted the difficulty of transiting from unemployment to employment among all age cohorts especially in the recent years that followed the eruption of the Arab Spring in Egypt in 2011. Age is also not related to moving from unemployment to employment in the Egyptian labor market and older workers are not offered better jobs compared to the young ones. This was shown through the high incidence of informality among prime-age workers who lost their jobs during one ELMPS round and went back to employment in the next round. While some young workers do not easily accept employment in the informal sector as they can resort to their families for financial support, prime-age workers, who could be the main breadwinners in their households, might not afford to wait until obtaining a formal job.
Results also revealed some chronic issues inherent in the Egyptian labor market over time, including the steady movement of females out of the labor force and the tendency of workers with higher educational levels to step out of the job market compared to workers with no educational background. The analysis highlighted the continuously significant impact of gender over the years on the probability of exiting unemployment with lower odds ratios of transitioning to employment and higher odds ratios of moving out of the labor force for female workers compared to males. Higher educational levels were also found to be associated with higher probabilities of transiting from unemployment to out of the labor force. In addition, young workers with intermediate or above levels of education experience higher probabilities of remaining unemployed compared to those with no education. Furthermore, young workers whose fathers are university graduates have higher probabilities of remaining unemployed compared to young workers whose fathers did not receive any education. These results emphasize the problems of mismatch between job supply and demand; the insufficient formal job creation in the Egyptian labor market; and the lack of competency suffered by the educational system to improve the employability among its graduates. Moreover, young workers with better educational outcomes who also belong to families with higher educational backgrounds can afford staying unemployed or leaving the labor force, which reflects the strong link between education and wealth that still persists in Egypt.
References
AlAzzawi, Shireen, and Vladimir Hlasny. “Vulnerable Employment of Egyptian, Jordanian, and Tunisian Youth: Trends and Determinants.” Helsinki, Finland, 2020.
Angel-Urdinola, Diego, and Amina Semlali. “Labor Markets and School-to-Work Transition in Egypt: Diagnostics, Constraints, and Policy Framework.” Economic Policy, 2010. http://mpra.ub.uni-muenchen.de/27674/.
Assaad, Ragui. “Unemployment and Youth Insertion in the Labor Market in Egypt,” 2007. https://doi.org/10.5743/cairo/9789774161544.003.0005.
Assaad, Ragui, Rana Hendy, Moundir Lassassi, and Shaimaa Yassin. “Explaining the MENA Paradox: Rising Educational Attainment, Yet Stagnant Female Labor Force Participation.” Demogr Res. 43 (2020): 817–50. https://doi.org/10.4054/demres.2020.43.28.
Assaad, Ragui, and Caroline Krafft. “Excluded Generation: The Growing Challenges of Labor Market Insertion for Egyptian Youth.” Journal of Youth Studies 24, no. 2 (2021): 186–212. https://doi.org/10.1080/13676261.2020.1714565.
———. “Labor Market Dynamics and Youth Unemployment in the Middle East and North Africa: Evidence from Egypt, Jordan and Tunisia.” Working Paper 993. Giza, Egypt, 2016.
Ciucǎ, Vasilica, and Monica Matei. “Survival Analysis for the Unemployment Duration.” Proceedings of the 5th WSEAS International Conference on Economy and Management Transformation I (2010): 354–59.
Cox, D. R. “Regression Models and Life-Tables.” Journal of the Royal Statistical Society. Series B (Methodological) 34, no. 2 (1972): 187–220.
Danish Trade Union Development Agency. “Egypt Labour Market Profile 2020/2021,” 2020. https://www.ulandssekretariatet.dk/wp-content/uploads/2020/09/LMP-Egypt-2020-final1.pdf.
Economic Research Forum, and Central Agency For Public Mobilization & Statistics (CAPMAS). “Egypt Labor Market Panel Survey, ELMPS (2012),” 2013. http://www.erfdataportal.com/index.php/catalog.
———. “Egypt Labor Market Panel Survey, ELMPS (2018),” 2019.
Flek, Vladislav, and Martina Mysíková. “Youth Transitions and Labour Market Flows- Who Moves & How?” STYLE Working Papers, WP5.2. CROME. Brighton, 2016. http://www.style-research.eu/publications.
Gadallah, May. “Factors of Career Mobility in Egypt by Gender.” Cairo, Egypt, 2011.
Hosmer, David W, and Stanley Lemeshow. Applied Logistic Regression. Second Edi. John Wiley & Sons, Inc., 2000.
Jenkins, Stephen P. Discrete Time Proportional Hazards Regression. Edited by H. Joseph Newton, Francis X. Diebold, Joanne M. Garrett, Marcello Pagano, James L. Powell, and J. Patrick Royston. Stata Technical Bulletin, 1997.
———. “Estimation of Discrete Time (Grouped Duration Data) Proportional Hazards Models: Pgmhaz,” 2004.
———. “Survival Analysis with Stata,” 2008. http://www.iser.essex.ac.uk/survival-analysis.
Kelly, Elish, Seamus McGuinness, Philip J. O’Connell, David Haugh, and Alberto González Pandiella. “Transitions in and out of Unemployment among Young People in the Irish Recession.” Comparative Economic Studies 56, no. 4 (2014): 616–34. https://doi.org/10.1057/ces.2014.23.
Krafft, Caroline, Ragui Assaad, and Khandker W. Rahman. “Introducing the Egypt Labor Market Panel Survey 2018.” ERF Working Paper Series. Giza, Egypt, 2019. https://erf.org.eg/publications/introducing-the-egypt-labor-market-panel-survey-2018/.
Menard, Scott. Applied Logistic Regression Analysis (Vol. 106). Sage, 2002.
Nilsson, Björn. “The School-to-Work Transition in Developing Countries.” Journal of Development Studies 55, no. 5 (2019): 745–64. https://doi.org/10.1080/00220388.2018.1475649.
OAMDI. “Labor Market Panel Surveys (LMPS).” Egypt: Economic Research Forum (ERF), 2013. http://www.erf.org.eg/cms.php?id=erfdataportal.
———. “Labor Market Panel Surveys (LMPS).” Egypt: Economic Research Forum (ERF), 2016. http://erf.org.eg/data-portal/.
Stock, James H., and Mark W. Watson. Introduction to Econometrics. Boston: Pearson Addison Wesley, 2003.
Williams, Richard. “Multinomial Logit Models – Overview,” 2019. https://www3.nd.edu/~rwilliam/.
Table 1. Results of the multinomial logistic regression model for predicting the probability of exiting unemployment between ELMPS 1998 and 2006
VARIABLES | unemp98_emp06
relative risk ratios |
unemp98_OLF06
relative risk ratios |
Sex: ref. Male | ||
Female | 0.134*** | 12.69*** |
(0.0377) | (5.017) | |
Age_group_06: ref. youth(18-34) | ||
prime-age(35-54) | 1.622 | 1.002 |
(0.582) | (0.352) | |
Urban/Rural: ref. Urban | ||
Rural | 1.325 | 1.356 |
(0.408) | (0.412) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | 464,382 | 685,126 |
(5.653e+08) | (8.341e+08) | |
Less than Intermediate | 415,731 | 870,581 |
(2.786e+08) | (5.833e+08) | |
Intermediate | 0.127** | 0.123** |
(0.133) | (0.131) | |
Above Intermediate | 0.119* | 0.119* |
(0.135) | (0.136) | |
University | 0.116* | 0.0500*** |
(0.128) | (0.0570) | |
Post-Graduate | 1.390e+06 | 0.220 |
(3.886e+09) | (683.9) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | 1.324 | 2.049** |
(0.497) | (0.745) | |
Less than Intermediate | 1.472 | 1.034 |
(0.573) | (0.404) | |
Intermediate | 0.942 | 0.799 |
(0.411) | (0.351) | |
Above Intermediate | 1.831 | 3.385 |
(1.603) | (2.795) | |
University | 2.994 | 2.610 |
(2.237) | (1.917) | |
Constant | 40.33*** | 1.941 |
(42.82) | (2.164) | |
Observations | 684 | 684 |
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 2. Results of the multinomial logistic regression model for predicting the probability of exiting unemployment between ELMPS 2006 and 2012
VARIABLES | unemp06_emp12
relative risk ratios |
unemp06_OLF12
relative risk ratios |
Sex: ref. Male | ||
Female | 0.0823*** | 21.57*** |
(0.0208) | (10.36) | |
Age_group_12: ref. youth(18-34) | ||
prime-age(35-54) | 1.040 | 1.546 |
(0.323) | (0.430) | |
Urban/Rural: ref. Urban | ||
Rural | 1.506* | 0.936 |
(0.369) | (0.214) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | 0.173 | 0.113 |
(0.235) | (0.215) | |
Less than Intermediate | 2.140 | 1.523 |
(3.145) | (2.460) | |
Intermediate | 0.447 | 0.119* |
(0.480) | (0.144) | |
Above Intermediate | 0.692 | 0.0941* |
(0.796) | (0.120) | |
University | 1.259 | 0.0634** |
(1.386) | (0.0785) | |
Post-Graduate | 0.226 | 0 |
(0.398) | (3.77e-07) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | 0.823 | 0.590* |
(0.271) | (0.178) | |
Less than Intermediate | 0.942 | 0.702 |
(0.314) | (0.226) | |
Intermediate | 0.700 | 0.689 |
(0.250) | (0.230) | |
Above Intermediate | 5.374 | 1.303 |
(5.946) | (1.501) | |
University | 0.482* | 0.714 |
(0.211) | (0.312) | |
Post-Graduate | 69,049 | 15.15 |
(5.820e+07) | (15,276) | |
Constant | 14.87** | 1.628 |
(15.73) | (1.928) | |
Observations | 876 | 876 |
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 3. Results of the multinomial logistic regression model for predicting the probability of exiting unemployment between ELMPS 2012 and 2018
VARIABLES | unemp12_emp18
relative risk ratios |
unemp12_OLF18
relative risk ratios |
Sex: ref. Male | ||
Female | 0.0701*** | 4.897*** |
(0.0165) | (1.403) | |
Age_group_18: ref. youth(18-34) | ||
prime-age(35-54) | 1.227 | 0.864 |
(0.271) | (0.158) | |
Urban/Rural: ref. Urban | ||
Rural | 0.917 | 0.834 |
(0.193) | (0.148) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | 1.240 | 1.082 |
(1.571) | (1.318) | |
Less than Intermediate | 0.302 | 0.368 |
(0.234) | (0.273) | |
Intermediate | 0.345 | 0.206** |
(0.230) | (0.130) | |
Above Intermediate | 0.364 | 0.178** |
(0.277) | (0.125) | |
University | 0.492 | 0.109*** |
(0.340) | (0.0713) | |
Post-Graduate | 0.927 | 0.0190*** |
(0.968) | (0.0252) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | 1.207 | 1.202 |
(0.393) | (0.317) | |
Less than Intermediate | 0.951 | 0.882 |
(0.294) | (0.228) | |
Intermediate | 1.034 | 0.755 |
(0.314) | (0.192) | |
Above Intermediate | 1.324 | 1.089 |
(0.959) | (0.668) | |
University | 0.669 | 1.112 |
(0.282) | (0.412) | |
Post-Graduate | 0.214 | 2.00e-06 |
(0.272) | (0.00213) | |
Constant | 19.55*** | 4.281** |
(13.38) | (2.843) | |
Observations | 1,118 | 1,118 |
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 4. Results of the Cox proportional hazards model for predicting the probability of exiting first unemployment, ELMPS 2018
(1) | (2) | |
VARIABLES | analysis time when record ends
Model coefficients |
analysis time when record ends
hazard ratios |
Age_group: ref. prime-aged(35-54) | ||
youth(18-34) | -0.170*** | 0.844*** |
(0.0487) | (0.0411) | |
otherwise | 0.165** | 1.179** |
(0.0752) | (0.0886) | |
Sex: ref. Male | ||
Female | -1.341*** | 0.262*** |
(0.0500) | (0.0131) | |
Urban/Rural: ref. Urban | ||
Rural | 0.0432 | 1.044 |
(0.0458) | (0.0478) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | 0.276 | 1.318 |
(0.201) | (0.265) | |
Less than Intermediate | 0.613*** | 1.847*** |
(0.167) | (0.308) | |
Intermediate | 0.882*** | 2.416*** |
(0.156) | (0.376) | |
Above Intermediate | 1.092*** | 2.981*** |
(0.179) | (0.534) | |
University | 1.280*** | 3.595*** |
(0.160) | (0.575) | |
Post-Graduate | 1.601*** | 4.960*** |
(0.231) | (1.146) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | 0.0289 | 1.029 |
(0.0646) | (0.0664) | |
Less than Intermediate | 0.102 | 1.108 |
(0.0661) | (0.0732) | |
Intermediate | -0.0554 | 0.946 |
(0.0720) | (0.0681) | |
Above Intermediate | 0.0939 | 1.098 |
(0.140) | (0.154) | |
University | 0.0484 | 1.050 |
(0.0894) | (0.0938) | |
Post-Graduate | -1.254* | 0.285* |
(0.710) | (0.203) | |
Observations | 3,546 | 3,546 |
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 5. Results of the complementary log-log model for predicting the probability of exiting first unemployment among young workers with Gamma frailty, ELMPS 2018
(1) | (2) | |
VARIABLES | Exiting unemployment
Model Coefficients |
Exiting unemployment
Odds Ratios |
Log (time) | 0.250*** | 1.284*** |
(0.0309) | (0.0398) | |
Sex: ref. Male | ||
Female | 0.229*** | 1.257*** |
(0.0685) | (0.0860) | |
Urban/Rural: ref. Urban | ||
Rural | 0.0935 | 1.098 |
(0.0634) | (0.0696) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | -0.0316 | 0.969 |
(0.296) | (0.286) | |
Less than Intermediate | 0.252 | 1.287 |
(0.216) | (0.278) | |
Intermediate | 0.595*** | 1.814*** |
(0.185) | (0.335) | |
Above Intermediate | 0.485** | 1.624** |
(0.236) | (0.383) | |
University | 0.913*** | 2.493*** |
(0.192) | (0.480) | |
Post-Graduate | 1.079*** | 2.942*** |
(0.288) | (0.847) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | -0.0311 | 0.969 |
(0.103) | (0.0996) | |
Less than Intermediate | 0.168* | 1.183* |
(0.0968) | (0.115) | |
Intermediate | 0.436*** | 1.547*** |
(0.0831) | (0.129) | |
Above Intermediate | 0.0358 | 1.036 |
(0.205) | (0.212) | |
University | 0.351*** | 1.421*** |
(0.116) | (0.164) | |
Post-Graduate | 0.382 | 1.466 |
(0.384) | (0.563) | |
Constant | -6.633*** | 0.00132*** |
(0.234) | (0.000308) | |
Observations | 125,404 | 125,404 |
Number of groups | 2,100 | 2,100 |
LR test of rho = 0: chibar2(01) = 1.3e-03 Prob >= chibar2 = 0.486 |
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 6. Results of the complementary log-log model for predicting the probability of exiting first unemployment among young workers, ELMPS 2018
(1) | (2) | |
VARIABLES | Exiting unemployment
Model Coefficients |
Exiting unemployment
Odds Ratios |
Log (time) | 0.250*** | 1.284*** |
(0.0389) | (0.0500) | |
Sex: ref. Male | ||
Female | 0.229*** | 1.257*** |
(0.0677) | (0.0851) | |
Urban/Rural: ref. Urban | ||
Rural | 0.0935 | 1.098 |
(0.0642) | (0.0705) | |
Educational Attainment: ref. Illiterate | ||
Reads & Writes | -0.0316 | 0.969 |
(0.297) | (0.288) | |
Less than Intermediate | 0.252 | 1.287 |
(0.218) | (0.280) | |
Intermediate | 0.595*** | 1.814*** |
(0.184) | (0.335) | |
Above Intermediate | 0.485** | 1.624** |
(0.236) | (0.384) | |
University | 0.913*** | 2.493*** |
(0.192) | (0.479) | |
Post-Graduate | 1.079*** | 2.942*** |
(0.290) | (0.855) | |
Father’s Level of education: ref. Illiterate | ||
Reads & Writes | -0.0311 | 0.969 |
(0.103) | (0.0996) | |
Less than Intermediate | 0.168* | 1.183* |
(0.0970) | (0.115) | |
Intermediate | 0.436*** | 1.547*** |
(0.0842) | (0.130) | |
Above Intermediate | 0.0358 | 1.036 |
(0.206) | (0.213) | |
University | 0.351*** | 1.421*** |
(0.116) | (0.165) | |
Post-Graduate | 0.382 | 1.466 |
(0.384) | (0.562) | |
Constant | -6.633*** | 0.00132*** |
(0.257) | (0.000338) | |
Observations | 125,404 | 125,404 |
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1