2013–2014
ERSTE Foundation Fellowship for Social Research Labour Market and Employment in Central and Eastern Europe
Opening the “black box” of the labour market in Macedonia: Youth unemployment Viktorija Atanasovska Tijana Angjelkovska
Opening the “black box� of the labour market in Macedonia: Youth unemployment
Viktorija Atanasovska Tijana Angjelkovska
Abstract
This paper estimates the determinants of youth unemployment in Macedonian labour market. The motivation for it comes from high and persistent youth unemployment in the country in the past two decades, despites many active labour policies directed towards that segment of the labour market. The paper presents the two step conditional mixed processes model (CMP) estimation on binary and ordinal data obtained from the school to work transition survey (SWTS) by ILO, aimed towards obtaining relevant estimates. As socialeconomic issues cannot be always explained by pure quantitative estimation, the paper extensively explores the possible determinants of unemployment by conducting semistructured interviews and focus groups. The results suggest that main determinants of youth unemployment in Macedonia are age, level of education, region of living, gender and household financial situation.
JEL classification codes: J20, J23. Keywords: youth unemployment, Macedonia, job search.
1
1. Introduction Unemployment is very sensitive issue worldwide, but it matters the most in young democratic countries since raising unemployment reduces output and income, increases inequality and harms human capital (Kapstein and Converse, 2008). Youth unemployment rates compared to the adult unemployment rates are almost twice higher in most of the countries with available data on the labour force and also the youth unemployment rate is closely linked with adult unemployment rate (O'Higgins, 2001). Definitions of youth are different across counties; for the purpose of this paper we will use the one from European Commission that comprises young people between 15 to 29 years. Providing qualitative jobs for both young man and woman that who enter the labour market become an essential element in the “Europe 2020 Strategy� for progress towards wealthier economies, more stable societies and faster growth (ILO youth employment report 2011). However, the practice have shown that experienced workers are preferred over young – school leaver candidates; young people in the work place are first in line for losing their jobs on times of negative economic cycles, which on the long run may result in loss of skills, talents, creativity, enthusiasm and possibilities for innovations and growth. In the last 3 years, the youth unemployment rates are estimated to be around 40 percent of the total estimated unemployed world population. According to the latest ILO report from 2013 the global youth unemployment rate has proved oppressive, and remained close to its crisis peak. At 12.4 per cent in 2012 and projected at 12.6 per cent in 2013, the global youth unemployment rate remains at least a full percentage point above its level in 2007. In the global perspective, nearly 75 million youth are unemployed around the world, an increase of more than 4 million since 2007; the ratio of youth to adult ratio unemployment rate remained significant and in 2013 stands at 2.7, which suggest that the likelihood of youth to be unemployed is almost three times bigger than for adults, moreover the unfavorable
2
trends in the global unemployment make the situation with the youth unemployment more fragile and the whole group vulnerable. Figure 1: The trends of the youth unemployment and youth unemployment rate, 1991-2013
Source: ILO Global Employment Trends for Youth 2013-A generation at risk
The aim of this paper is to investigate the determinants of youth unemployment in the labour market in Macedonia, in particular individual characteristics which may impact the (un)employability of an individual such as education level, family financial situation, region of living and similar; which will help to draft policy recommendation aimed at reducing the unemployment rate. The paper is organized as follows: Section 2 presents some stylized facts for Macedonian labour market in general and participation of youth in the labour market in particular. Section 3 gives brief overview of the most prominent literature in the field of youth (un)employment. We pay particular attention to the theory relevant to the developing economies, in particular we explore market situations and developments that might influence the employment status of youth and from there we derive the potential determinants of youth unemployment. Section 4 gives overview on the methodology, data and the model used for the study, in favor of examining the youth labour market in Macedonia, using the available micro-data from 2012 for the quantitative part and additional qualitative analysis for further explanation of the obtained quantitative results; further it gives insight in the qualitative research from the analysis. Section 5 summarizes the main findings and their implications for 3
the country’s current employment and youth–friendly policies, followed by an overview of policy recommendations for improvement of the situation in the labour market and reducing the youth unemployment 2. Stylized facts about Macedonian labour market Macedonia is small and open economy with population of 2,022,547 inhabitants (State statistical office, census 2002) with 64.18% Macedonians, 25.17% Albanians, 3.85% Turks, 2.66% Roma, 1.78% Serbs, 0.84% Bosnians, 0.48% Vlachos and 1.04% of other nationalities, Macedonia faces many challenges in creating opportunities for social and economic development and growth. One of the main socio-economic tensions is related to the relatively high unemployment rates and double high youth unemployment rates that are reflected by either high idleness of the youth or the high brain drain. Unemployment started to rise drastically after 1995, when the transition started and the structural changes took place, hence contrary to the expectations for the labour market to become more competitive, the reality showed that human capital stock was not ready for the fast market changes, which resulted in high job loses. Despites the indications from latest figures suggesting decreasing trend in the unemployment, still the overall unemployment rate in Macedonia is significantly high, with striking figures in youth unemployment rates in particular. According to the latest Eurostat figures, Macedonia has highest youth unemployment rate after Greece, with high 58% unemployed youth, which is almost double the overall unemployment rate of 31%. Figure 2: Youth unemployment rates in selected countries and Euro area; Unemployment rates in Macedonia
Source: Eurostat, State statistical office and authors calculations.
4
Macedonia has a very low labour force participation rate which, remaining flat at 5556%, which is low for Central and Eastern Europe countries standard. According to Urdinola and Marcias (2006), after Turkey, Macedonia has the lowest LFP rate among the working population in the region. Young workers low participation rate can be explained by high rates of young people engaged in fulltime education. This can be perceived as positive investment and the return to it should be good job position, but in most of the cases, young people see the higher education as a “rescue plan” for not to stay unemployed or not to have the possibility for obtaining decent job, and also they see it as a possibility for obtaining visa for travelling and in some cases emigrating in the western countries. These two reasons for idle youth are alarming signs for addressing youth employment opportunities on policy level.
3. Literature Review There has been an extensive literature devoted to the youth roles in society and in the family through time. The evidence shows existence of discouraging changes that have taken place over time and across cohorts in perceptions about youth’s roles or identities. According to O’Higgins (2010) the entry of the youth to the labour market is highly related to the general happenings in the economy as a whole. Particularly, the youth employment rates are in correlation to the aggregate labour supply and the labour market regulations, policies and legislations. Recent global economic and financial crisis resulted in high unemployment rates due to the big amounts of job losses. The impact of it was mostly felt by the young people willing to enter or already in the labour market. According to Fares and Tiongson (2007), if the situation with the unemployment in recent future, it can have long-term impact of the life’s of today’s youth. As, indicated in the ILO (2006) report, that can lead to social exclusion within the society and even increase in deviant behaviors. Neoclassical economics explains the decision for individual to enter the labour market using utility maximization over the work-leisure choice. If leisure is considered as a “good”, 5
the net effect of an increase in the wage rate over the quantity of labour supplied will be: (i) increase if the substitution effect is larger than the income effect and (ii) decrease if the income effect is larger than the substitution effect. In the case of the net effect is due to their own-wage elasticity and the household income (including their family background) elasticity. According to Freeman (1979) the dimensions of the youth unemployment problem is that young workers usually showed lower levels of work activity than older workers. He further finds that the LFP and employment rates to population rates are lower for youth, compared to those of the adults, while the rates of unemployment are higher. This is also confirmed in the Shimer (2001) study, namely he found a negative relationship between youth population size and youth unemployment rates; He explains: “labour markets containing substantial numbers of young people are likely to be more flexible than those dominated by older workers and therefore in such markets employers are more willing to create jobs.� Given the indications from Shimer’s study, relative increase in the size of the youth population leads to decrease in unemployment rates; hence the unemployment is the result of the capability of the country to handle the increases in the supply on the youth labour market. Freeman (1979) suggest two basic reasons for high youth unemployment: from the demand side, the main reason for high youth unemployment rates is due to lack of adequate demand for young labour force, as a result of the economic cycles and slow growth; and on the supply side, the main reason for high unemployment is due to inappropriate training, lack of skills and experience on the behalf of the youth. As, at some point the increase in the unemployment is due to the structural changes of the economy, as exogenous shocks (financial crisis, recession), according to Peterson and Vorman (1992) significant impact on the rise of the unemployment rate can have the mismatches that can occur on the demand side of the market, and the supply is not able to respond immediately to the change. They further explain that trough the technological advancements that occur and require more skilled and more educated workers. 6
According to the O’Higgins (2001) study on OECD countries it was found that unemployment was decreasing proportionally to the increase in the education level in the country. However, in the same study he shows that this doesn’t hold for developing countries, as it was found that there is high unemployment among educated groups. According to the neoclassical theory of individual labour supply the opportunity cost of being out of labour market increases with each additional finished level of education. There are findings indicating that transition school-to-employment is very important factor in youth unemployment (Osterman, 1980). According to Osterman (1980) the relationship of the age seems to be linear with the labour force participation, as the young workers quitting rate falls as their age increases, showing that those individuals are more dedicated and the attitude towards employment is changing. According to MacDonald (1988) regional characteristics (urban-rural) have impact of the labour participation of the youth. In 1993, O’Grady in his study also found significant results for the importance of the regional factors in relation to school to work transition period. In 1996, DeLamatre found that the perception of the person for the area/region of living also have impact of the employability and success in the labour market. O’Regan and Quigley’s (1996) study findings suggest that the perception and attitude of an individual for the region one lives, along with the poverty level of the neighbourhood, have significant impact on the employment success. O’Higgins (2001) study indicates that in developing countries, regardless the gender, there is a difference in rural vs. urban unemployment rates. Female tend to participate less in the labour market compared to males, both in rural and urban areas, mainly because they became part of the informal labour (taking care of the household), while the young males continued with education (O’Higgins, 2001). According to O’Higgins (2001) in many countries with more ethnicities, ethnic origin appeared to be very significant determinant of (un)employment, indicated by the large differences in the labour participation on national level by across different ethnic. Therefore, empirical model will comprise two broad categories of determinants: (i) personal 7
characteristics (education, age, health condition) and (ii) household characteristics (household wellbeing, size, parents education, number of children – if married) 4. Analysis The research will include two broad stages of empirical analysis. The first stage will make use of conditional mixed processes methods. Considering the complex nature of youth unemployment, the quantitative research above will be complemented by qualitative analysis in the second stage, aiming at further enrichment of the empirical findings. The qualitative analysis will address socio-cultural dimensions of determinants of the youth unemployment which cannot be addressed or fully explained by the quantitative part. The mixture of quantitative and qualitative approach to the topic is of particular importance for this topic because qualitative data can provide us with more detailed explanations of the findings obtained from the quantitative part of the research. It can explain socio-cultural and traditional reasons and relationships that were detected with the quantitative approach. Further, there are factors such as traditions, expectations, and common discrimination on the basis of ethnicity, age, political determinants, that are rarely mentioned in the surveys and quantified. The next section of the papers is organized such that we first present the model, that we will use for the regression analysis; then we present the data that is used in the regression; than one part will focus on the quantitative analysis of determinants of the youth unemployment, while in the second part we will present the findings from the qualitative analysis, based on semistructured interviews and focus groups. 4.1. Estimation technique and model specification, data and empirical findings The empirical model is based on the theoretical review presented in the Section 3. For the purpose of the study aiming at obtaining relevant results indicating the reasons for youth (un)employment conducted on the micro data “school to work transition� conducted by the State Statistical Office of Macedonia for the purposes of the International Labour Organization (ILO), we apply the conditional mixed processes (CMP) model. The CMP 8
model, firstly introduced by Roodman (2009), can handle a variety of models using a maximum likelihood approach. Its flexibility allows for the model to vary depending on the type of observations used. For example, this model allows combination of equations that relay on different samples with matching observations. According to Roodman (2009) this procedure is appropriate in the following two scenarios: (1) “those in which a truly recursive data-generating process is posited and fully modelledâ€? and 2) “those in which there is simultaneity but instruments allow the construction of a recursive set of equations, as in two-stage least squares (2SLS). CMP is fundamentally an SUR estimation program. But it turns out that the ML SUR can consistently estimate parameters in an important subclass of mixed-process simultaneous systems: ones that are recursive, with clearly defined stages; and that are fully observed, meaning that endogenous variables appear on the right hand side only as observed “, (Roodman, 2009, p.2). According to the nature of the variables included in the CMP model we will use probit and orderit probit equations. The probit model is based on assumption that value of the unobserved variable Y, is determined by the explanatory variable Xi (Gujarati, 2004): đ?‘Œđ?‘–=đ?›˝0+
� +� Where:
represents the latent propensity to be either 0 or 1, employed or
unemployed, highest level of completed education, financial status,
(Where i = 1,2...n)
are explanatory variables that stand for individual supply factors, such as age, sex, marital status, number of children, regional characteristics and similar.
- is the error term (with
mean = 0 and variance = 1) The explanation of the probability (Pr) of the young person to be unemployed is obtained by the following model: Prâ Ą( =1|
)=ÎŚ(
) 9
Where: Y=1 if the young person is unemployed and 0 otherwise; ÎŚ is the Cumulative Distribution Function of the standard normal distribution and explanatory variables
are the parameters of the
that will be estimated by maximum likelihood.
The orderit probit inspects the probability that an alternative will be selected: đ?‘?
= (đ?‘Ś = )=
đ?&#x203A;ź â&#x2C6;&#x2019;1<
â&#x2C6;&#x2014; â&#x2030;¤
=đ??š
â&#x2C6;&#x2019;đ?&#x2018;Ľâ&#x20AC;˛
â&#x2C6;&#x2019;
â&#x2C6;&#x2019;1â&#x2C6;&#x2019; â&#x20AC;˛
Where for the ordered probit F is the standard normal cdf; the model with j alternatives will have one set of coefficients (j-1) intercepts and j sets of marginal effects. According to the basic human capital theory and the empirical findings parents tend to shape the lifestyle of their children depending on their educational, social and material status. Namely, if the parents are with higher education it is expected that their children will also stay longer in education to obtain their level of education or above. That also relates to the wealth of the household, the expectations are that parents with higher education, tend to earn more, and have better network, hence their children have shorter transition school-to-work period. And the opposite is with parents with lower education level and lower household wealth. It is expected that there is on-going competition between siblings, so if one of the siblings is in education it is highly likely that the others will also participate in education. So the impact is expected to be positive. However, it is expected to vary across regions (more and less developed) and across gender. The relationship of the household characteristics and the youth employability is varying. Ceteris paribus, household wealth may impact the probability for being employed both positively and negatively; Namely, it is expected that regardless the level of education if the household is poor the young person can be less picky when it comes to accepting any kind of job, or coming from poor family in small communities can reduce the chances for obtaining a job. It is expected that the employability of the youth in relation to the regional characteristics to be inverse to the size of the youth population in the region and the economic development of the region.
10
4.2. Data The school to work transition survey used in this analysis was conducted in Macedonia in 2012 on total sample of 2,544 individuals between the ages of 16 to 29. In our empirical model we include 1239 (employed and unemployed individuals-actively looking for a job). Dependent variable of the model is the probability of young person being unemployed. In the dataset is based on the questions: “what is the employment status of the person?” and “Is the person looking for a job?”
Table1: Description of the dependent variable Employment status of Respondents: Employed
733
Unemployed (actively looking for a job)
506
The dependent variable is unemployed defined as follows: “0” represents employed individuals,
“1” represents unemployed individuals.
We further present the list of explanatory variables for Macedonia and their expected signs.
11
Table 2. Description of the variables Variable
Explanation
Age
15-29 year
Expected sign -
Gender
“0” for female, “1” for male
-
Single - dummy variable
What is your current marital status?
+
nb_of_children Poor – dummy varaible
“0” for married, “1” for single How many children do you have? + How would you describe your household's overall + financial situation? “1” for poor, “0” for other This is a dummy variable that takes the value “1” if + the individual has no health problems at all, and “0” otherwise
Health
Highestlevel_compdummy variable Region - dummy Father_edu Mother_edu
unempregion
4.3.
What is your highest level of completed formal education/training? Elementary: no education, elementary level east southwest southeast pelagonia polog northeast skopski and What is the highest education that your parents have? Elementary: no education, elementary level Secondary: vocational school, secondary level and post-secondary level University: university and post-graduate, postdoctoral level Callculated regional unemployment rates and assigned to the region dummies
+ +
+
Empirical results The dependent variable in the model is “unemployed”, which is equal to 1 if the
respondent is unemployed and 0 if employed; whereas the independent variables are level of education, gender, age, age squared, marital status, number of children, household financial situation, health and unemployment rate in Macedonia by region. The variables are either binary or ordinal choice. According to the theory, we consider the level of education as endogeneous. These variables are instrumented on the second specification with fathers’ level
12
of education and in the third specification with the mothersâ&#x20AC;&#x2122; level of education. The CMP regression results are presented on Table 8. Table 3: CMP regression results
EQUATION
VARIABLES
CMP1 estimates employed
unemployed
elementary poor age agesq gender single nb_of_children health east southwest pelagonia polog northeast skopski unempregion _cons
0.601*** 1.788*** 0.354*** (-)0.008*** (-)0.228*** 0.095 -0.046 0.180 -0.159 0.191 0.368*** 0.3509*** -0.470 0.343*** 0.029*** (-)5.799***
Marginal effects employed 0.018 0.000 0.006 0.006 0.002 0.391 0.469 0.254 0.254 0.288 0.002 0.008 0.034 0.003 0.000 0.000
0.167*** 0.498*** 0.098*** (-)0.002*** (-)0.063*** 0.026 -0.013 0.050 -0.044 0.053 0.102*** 0.098*** (-)0.131*** 0.095*** 0.008***
0.017 0.000 0.006 0.006 0.002 0.390 0.468 0.254 0.255 0.287 0.002 0.008 0.034 0.003 0.000
The estimates obtained from the regression are suggesting that household wellbeing together with personal characteristics can significantly influence the (un)employment status of an individual. Namely, among significant variables that determine the unemployment were found to be: gender (If female the probability of being unemployed goes down by 0.06 comparing to male), age (For each additional year of age the probability of being unemployed goes up by 0.09), wealth of the family (if the household is poor the probability of being unemployed increases by 0.5), education level (if the individual has only elementary education the probability of being unemployed increases by 0.17) and regional characteristics (If living in Polog, Pelagonia and Skopski regions the probability of being employed goes up
1
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
13
by 0.1,in North the probability of being unemployed goes down by 0.13 comparing to youth living in Vardarski region). 4.4. Qualitative analysis In the following section we present the findings from the qualitative analysis that involves analysing and interpreting focus groups and interviews aiming to detect meaningful explanations the researched area that wasnâ&#x20AC;&#x2122;t explained by the quantitative analysis (Auerbach and Silverstein, 2003). The qualitative analysis of this paper includes processing data gathered from both focus groups and semi-structured interviews conducted in two regions in Macedonia. The focus groups were organized in two cities (Tetovo and Skopje); each group was consisted of 10 participants on the age range 15-29; gender and ethnic balanced groups. Both focus groups had same aim, to give space to the participants to share their experiences related to the process of making the first step towards finding first job, i.e. labour market penetration. It is important that the groups were both gender and ethnically mixed, in order to get the full picture for the youth participation in the labour market, are there same opportunities/obstacles for all, how hard/easy is to enter the labour market and get well paid and secure job. In view of the semi-structured interviews the questions were directed towards their personal everyday experiences during the process of looking for a job, on the working spot, if a job was found and the working environment within the organization; possible cases of positive/negative discrimination with regard to: level of education, gender, ethnicity, location, age, experience and similar factors. This section is structured as follows: aim of the investigation in Section 1.1. Data analyses are provided in Section 1.2. Section 1.3 presents the findings.
4.4.1 Aims of the Investigation The aim of the qualitative analysis is to address questions related to youth employability conditions such as: How do they perceive the current situation in terms of 14
employment? What are main challenges for finding the first job? Do they believe that higher education gives more possibilities for obtaining good job? To what extend the situation in the labour market affects their education choices?
How easy/difficult is to find the first
employment? How "friendly" is the selection process? How transparent is the selection process? To what extent, does their employment status affect their degree of social inclusion? What are the perceived barriers to their labour participation? What are their most immediate needs and priorities that would help in minimizing the gender gaps in the labour market participation? What are the possibilities of facilitating their integration in the labour market?
4.4.2 Sampling Methodology and Data Analysis A sample for the focus groups was consisted of 20 young people aged 15-29 in two different cities: Tetovo and Skopje (the capital), in order to get more information about the situations in different regions from different categories of young people. The focus groups were conducted by pair of researchers in two occasions; each focus group took between 60-80 minutes. The semi structured interviews were conducted with 20 young people in the same age range as for the focus groups. Each interview was conducted individually (young person and researcher), with youth form urban and rural areas, for more in-depth insights about the challenges and issues young people from different regions and backgrounds face, while searching for employment. The participants for the interviews were approached trough local youth clubs, NGOs, schools. The questionnaire applied consists of two main parts: (i) demographic part; and (ii) open ended questions - regarding the obstacles they face as young people in search for job, and suggestion on the necessary supportive instruments for young people to penetrate the labour market. In the process of transcribing the focus groups and interview recordings, it was taken care that the respondents are not identifiable in any part of the paper. After the qualitative analysis part of the report was written, the recordings from the focus groups and interviews were destroyed. 15
4.4.3 Findings 4.4.3a Focus groups data analysis The findings presented in this chapter are based on the responses obtained by the young people participating in the focus groups, summarised in several subcategories: How easy / difficult is to find the first employment? When young person tries to enter the labour market in most of the cases one can offer is enthusiasm and energy to work and learn, but it is very difficult to compete with elderly candidate applying for the same position. Usually what they find as a biggest obstacle is the working experience, and in the very beginning of the looking for a job process it seems as a vicious circle. Namely, even though the advertisement does not indicate that particular years of working experience is one of the selection criteria, as most of the jobs they apply are entry level positions, once they are at the interview they get stuck with the questions: “what is your previous experience?” and “How do you know that you can satisfactorily perform if you have never worked?”. So it becomes more like a frustration whether one day they will manage to enter any job, as with working experience being one of the main criteria for obtaining the entry level job that would be very difficult, almost impossible. That also brings further implications on the age and personal and professional development; the longer one candidate is waiting of a first job position, the less are the chances for getting good job related to the educational background. One of the participants shared that on a particular interview the interviewer was to some level rude, commenting on the age and no experience of the candidate. How "friendly" and transparent is the selection process? Most of the participants in the focus groups said that the applications process is hidden. There are plenty of positions that are advertised one day or the day of the dead line, so one do not even get a chance to participate in the competition. They say most of the job advertisements do not publish whole selection criteria, and the exact time of selection. In about 80% of the time they did not received any feedback about the stage of the selection process, in 10 % of 16
the cases they received the confirmation that their application was received, and in only 10% of the times they were contacted and preceded further with the selection process. They said that it is very discouraging the fact that in most of the cases especially in the advertisements about openings in the public administration, they receive the information that the position is already filled and the advertisement is only published so that they can give or prolong the existing contract of a specific employee. There were situations where in the advertisement wasnâ&#x20AC;&#x2122;t published the selection criteria, besides only the minimum education level, and after few weeks they would receive feedback that their profile do not correspond to the selection criteria for that particular position. Access to information regarding labour market demands Most of the focus group participants said that there is very limited information about the situations in the labour market. They also pointed out that there is no management in the education, which reflects in hyper production of some professions, issue that should be solved on policy level by restraining the openings at some specific faculties, i.e. humanities, and incentivise students to shift to more technical fields. The interviews also suggested that by making the secondary education compulsory for all leads to situation where most of the pupils continue to the university level education and after they do not want to accept lower scale jobs, even though they have skills and knowledge for them. So there are many young people deciding to stay at home and wait for a job in accordance to their education background and personal affiliations, than accepting the offered post. They said that maybe students should be contacted during their secondary education by explaining to them the labour market situations. In that context one of the participants mentioned that it would be very useful if we can develop some system similar to the Dutch, where people around their 18th year are contacted by scouts, they are offered good and secure, in the same time well-paid jobs. Along with that they receive very good loan offers, with low interest rates, so they easily get to the first housing. 17
Access to information regarding vocational training courses Most of the participants said they are aware of the vocational trainings and courses offered for free by the state employment agencies, but they said that they are provided on such a low level that are not much of a help during the job search process. They also said that the higher levels of trainings are usually very exclusive to specific group of people that look for a job. For example courses like: data mining, programming, My SQL are usually offered only to a computer sciences or technical university graduates. So an economist or legal studies student cannot prequalify. Also there are cases of inappropriate prequalifying offers, i.e. undergraduate and graduate students are offered to become sewers or cooks. This again can be linked to the lack of management in the education system in terms producing university educated candidates that cannot be placed and free low level positions that cannot be filled. Recommendations regarding increase in youth employment Most of the participants have agreed that there should be more cooperation between educational institutions and the real sector. Companies should offer traineeships and the university curricula should be tailored in accordance to the labour market demand. Some of the suggestions regarding how to increase the employability of the candidates were: •
More restricting university entry selection criteria;
•
More labour market specific trainings and courses;
•
More communication with pupils in secondary education;
•
Education management;
•
Cooperation between companies and universities;
•
Transparency in the selection criteria;
•
Standardised entry exams;
18
4.4.3b Interviews - data analysis This section presents the findings of the qualitative analysis based on the conducted interviews both in rural and urban area of Macedonia. The general opinion gathered from the conducted interview about the situation in the labour market in general and youth labour market in particular is that the labour market is much politicized and not very transparent. People believe that there is no much connection between the education institutions and the companies. Most of the people are very sceptic that will get a good job in the recent period and some consider leaving the county in the next 2 years. It is interesting that young people from rural areas are less picky when it comes to jobs as they are used to more manual work, whereas their counterparts from urban areas are less likely to accept some manual or agricultural job. Most of the interviewees’ responded that they choose staying home over accepting any kind of job. Almost all of the interviewees said that they feel discriminated in the selection process based on several criteria (age, gender, ethnicity, political thinking, and personal connections). On average, most of the respondents felt discriminated more than once during selection process. The findings of this qualitative analysis is that main criteria for obtaining the job position is belonging to a political party, second is personal connections, and after that are the educational background, non-formal knowledge and volunteer experience. However there are still positive examples about qualitative and competitive selection procedures. The most promising statements by interviewees for a good employment opportunities and fair selection process are the following: “... I believe that if I have better education I can get better employment opportunities, that is why I decided to study hard in the school but also to enter extracurricular activities where I can learn more than what is written in the book and share opinions with my school mates.” “I believe that my volunteering experience brought me closer to obtaining my first entry level jobs. They are not my dream jobs, but I still learn a lot, get working 19
habits and I know that once I get a chance to have interview in the real sector they will value my experience.” “I am not afraid of work, and I believe that there is no bad and good job, but it is more about preferences. I am nearly finishing my faculty and if I do not get employed immediately I will work the farm with my parents.” The quotations above indicate that young people are not easily discouraged by the bad situations in the labour market in Macedonia, and they believe that education is important and can help them in obtaining better jobs in future. Also, when a person want to work and have a responsibility there is no way that that person will stay idle and jobless. Even though in long run can be little bit disappointing that when one was entering the university level studies no one haven’t warned that with that particular field of study will be hard to find a placement in the labour market. Still there are situations when people felt offended and undervalued by the offer: “At the labour office i was suggested that I should prequalify in sewer, well If I wanted to spend my life in the factory I wouldn’t dedicate 4 years of my life to studies, but I would go there strait away.” “In one bank I was invited to an interview, and apparently they weren’t looking for “a specific major, but more for a person” so they invited me but will not proceed to the next stage even though I comply with all the selection criteria. I do not understand why they waste my time if they already know the person that should be in that position.” “After 2 weeks of selection process I was told: “you are best candidate, but we were making the selection more like a lottery”... than why 2 weeks of selection procedure?” People face awkward interview situations that discourage them for future applications in the same organisation that in a way limits and reduces the job opportunities. They sometimes lose 20
self-confidence and after few interviews they just decide to stay idle and not apply for any job. There are plenty of advertisements about FDIs in the country, but the interviewees commented that they do not see the change in the labour market situations: “Even they come from more developed countries; I have a feeling that once they enter Macedonia, their human resource departments just stop working.” Most of the companies, even though legally are requested to have human resource department, in practice they do not pay attention to the applicants and people that showed interest. In most of the cases they do not receive feedback that the application was received and if they should expect to be invited to the interview or not. On the other hand, the general opinion of the respondents already in the job position is bit more enthusiastic. They detect many irregularities in the selection process, bad wages, little protection even in the NGO sector, but they say that “any job is better than no job”. In general they believe that the state employment offices will not help you in finding a better job, but you can get informed about some openings in the region. Even though most of them are in the informal market, family business or honorary based contract, they believe the experience they obtain there will be valued: “I prefer to tell that I have worked in a small shop, than that I was not working at all.” But there are opposite experience: “I thought volunteering and taking trainee ship programs will ease my employment process, but it turned out no one counts that as working experience.” In a nutshell, according to the qualitative analysis, the most important factors that affect youth unemployment are: high polarisation of the society, lack of consistency between education and real sector. Some of the possible solutions are: better schooling system, better linkage between education and labour market demand, more transparent selection procedure and less 21
political and family connections (especially in the public sector) as all of the above either lowers the chances for getting employment or discourages young people to actively search for an employment. 5 Conclusions and Policy Recomendations This paper examined the determinants of the unemployment by employing conditional mixed processes method on the data from the school-to-work transition collected by ILO. Additionally it includes extensive qualitative research on the topic that can explain some social factors that can explain the unemployment but that are not included in the quantitative indicators. To this end, in Macedonia, the youth unemployment rate reduces less than the overall unemployment rate. The estimates showed that gender, age, the wealth of the household, the education level and regional characteristics tend to influence the probability of oneâ&#x20AC;&#x2122;s unemployment status. Additionally to this, the qualitative analysis suggested that given the socio-economic situation in Macedonia, it is likely that most of the youth live with their families from where they cover their expenses (and probably study) and have less necessity to search for employment, however in long run can influence the personal choices. In general, there is long transition school to work period, along with highly politicized employment channels, which discourages young people to actively participate in the labour market. Accounting for the findings, we propose two policy measures aimed at reducing the youth unemployment rate. Each policy is aiming at addressing determinants of the unemployment that are specific for Macedonian market. Three key reasons for these policy recommendations come from both qualitative and quantitative data. First, the unemployment rate is significantly higher for youth compared to overall rate; second, the unemployment rate is persistently high despites some active labour market policies and reforms in the educational system; and finally, transition from school to first formal employment is most difficult for most of the Macedonian young people.
22
First recommendation is a higher cooperation between the educational institutions and the real sector economy. We propose that in the last year of education every young leaver in the last year of study (high school or university) should have consultation meeting with the professional with whom everyone can extensively discuss the employment opportunities with a specific level of education. After that the real sector entity together with the educational institution should offer a tailored internship/apprentice program that will enable the individual to gain the required skills that along with the knowledge will enable them to get to the first employment easily. This measure should be mandatory for each educational institution, meaning that they will have to improve the educational program they offer and ease the path for the future employees. The budgetary implications are low, as the main role of the ministry of education will be to coordinate the process. The second policy recommendation is about the wage subsidy/social package for school leaver that will be conditional on the active job search and personal development. The proposal is that each individual that is not continuing to a next level of education should be entitled to social package that will ease the transition from school to work. The proposed system is that each individual that is non-working and not in the education, but actively searching for a job should receive the national minimum wage and social package for a period up to one year. In that period he should attend the internship/apprentice program, and attend specific courses that will increase its employability skills. When the aprentice program will finish, the social package should be extended for another year but the amount should be reduced to half national minimum wage up to maximum, and in the same time the job seeker should have access to all employability programs. The amount is not big, so it is not creating a risk that the individual will be reluctant to search a job and the money are initially director towards coverage for the costs related to attending interviews and examinations since all of them are money consuming and if the one has not any means may be impossible to attend them (sometimes they are in different town or require specific clothing). The budgetary 23
implication may be higher, but high and persistent youth unemployment is more harmful for the economic growth and social peace. Reference list Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 95 –120. DeLamatre, M. S. (1996). Investigating the antecedents of youth unemployment: Individual, family, and neighborhood factors. Unpublished Masters Thesis. Louisiana State . University, Baton Rouge. Fares, J., & Tiongson, E. R., (2007).Youth unemployment, labor market transition, and scarring: Evidence from Bosnia and Herzegovina, 2001–04. Policy Research Working Paper Series 4183, The World Bank. Fields, G. S. (2008). Segmented labor market models in developing countries. Retrieved from http://works.bepress.com/gary_fields/27. Gujarati, D.N. 2003. Basic Econometrics. New York: McGraw Hill Book Co. ILO (2004, 2006, 2008, 2012, 2013). Global Employment Trends for Youth. International Labor Office: Geneva. Jimeno, J.F., Rodríguez-Palenzuela, D., 2002. Youth unemployment in the OECD: demographic shifts, labour market institutions, and macroeconomic shocks. ECB Working paper 155. Kapstein, E. and N. Converse, 2008, “The Fate of Young Democracies,” (Cambridge: Cambridge University Press). MacDonald, R. F. (1988). Schooling, training, working and claiming: Youth and unemployment in local rural labour markets. Unpublished Ph.D. Dissertation. The University Of York. United Kingdom. Neumark, D., & Wascher, W. (2004). Minimum wages, labor market institutions, and youth employment: A cross-national analysis. Industrial and Labor Relations Review, 57(2), 223–248. Retrieved from: http://digitalcommons.ilr.cornell.edu/ilrreview/ O’Grady, W. L. (1993). Coming of age on the periphery: Youth unemployment and the transition to adulthood in Newfoundland. Unpublished Ph.D. Dissertation. University of Toronto. Canada. O'Higgins, N., & International Labour Office. (2001). Youth unemployment and employment policy: A global perspective. International Labour Office. O'Higgins, Niall, 2010. "The impact of the economic and financial crisis on youth employment : measures for labour market recovery in the European Union, Canada and the United States," ILO Working Papers 462129, International Labour Organization. 24
O'Higgins, Niall, 2010. "Youth Labour Markets in Europe and Central Asia," IZA Discussion Papers 5094, Institute for the Study of Labor (IZA). Osterman, P. (1980). Getting Started: The Youth Labor Market. Cambridge, MA: MIT Press. Rees, A. & Gray, W. (1982). Family effects in youth employment. The Youth LaborMarket Problem: Its Nature, Causes and Consequences, Freeman and Wise,(Eds.) (Chicago: NBER and University of Chicago Press). Roodman, D (2009). Estimating fully observed recursive mixed-process models with cmp. Working Paper 168. Washington, DC: Center for Global Development. Shimer, R. (1999). The impact of young workers on the aggregate labor market. Working Paper 7306. NBER. Cambridge. Retrieved from http://www.nber.org/papers/w7306. Tiongson, E. R., & Fares, J. (2007). Youth unemployment, labor market transitions, and scarring: Evidence from Bosnia and Herzegovina, 2001â&#x20AC;&#x201C;-04 The World Bank, Policy Research Working Paper Series: 4183. Retrieved from: http://www.worldbank.org/servlet/WDSContentServer/WDSP/IB/2007/03/27/000016 406_20070327134051/Rendered/PDF/wps4183.pdf Verick, S. (2009). Who is hit hardest during a financial crisis? The vulnerability of young men and women to unemployment in an economic downturn. International Labour Organization (ILO) and IZA. Discussion Paper No. 4359.
25
APPENDICES A1: YOUTH EMPLOYMENT - FOCUS GROUPS Guiding questions 1. Presentation and employment status. 2. How easy / difficult is/was to find the first employment? 3. How "friendly" is the selection process? 4. How transparent is the selection process? 5. Regarding youth employment, who is more predisposed to find a job, males or females? 6. Are you encourage apply for employment or you expect someone to find you a job? Why? 7. Is there enough access to information regarding labour market demands? 8. Is there enough access to information regarding vocational training courses? 9. Are there policies regarding employment in your area? Do you have information on employment policies? Have you ever addressed to employment offices? 10. What would help regarding youth employment? 11. What legal adjustments do you think are necessary to increase youth employment?
A2: SEMI-STRUCTURED INTERVIEWS 1. Residence location _____________________________________________ 2. Education ____________________________________________________ 3. Civil status ________________________________________________________ 4. Did you have any work experience?_____________________________________ 5. If so, how many years did you work?____________________________________ 6. What job did you do? How do you assess that experience? __________________________________________________________________________________________ __________________________________________________________________________________________ __________________________________________________________________________________________ __________________ 7. How did you find it? __________________________________________________________________________________________ __________________________________________________________________________________________ ____________________________________ 8. How is the employment situation in your area? __________________________________________________________________________________________ __________________________________________________________________________________________ ____________________________________ 9.
What are the main activities in your area?
a) b) c) d) e) f) g)
Private enterprise Utilities Family business Agriculture/farming Stateâ&#x20AC;&#x2122;s institutions/public Trade Other (specify) _____________________________________________________
26
10. Is it easy to find a job there? _________________________________________________________________ 11.
If not so, what are the reasons there is no labour availability in your area?
a) b) c) d) e)
Undeveloped area Enterprises/businesses are in distance from residence location There is no appropriate infrastructure There are only man-related jobs Other (specify) ____________________________________________________
12. If you are seeking employment, what kind of job would you prefer to do? __________________________________________________________________________________________ __________________________________________________________________________________________ ____________________________________ 13. What kind of efforts have you done to seek employment? __________________________________________________________________________________________ __________________________________________________________________________________________ ____________________________________ 14. What reasons do you think prevented you from finding a job up to now? __________________________________________________________________________________________ __________________________________________________________________________________________ __________________________________________________________________________________________ __________________ 15. In your opinion, how do you evaluate the role of public and private institutions? (labour offices, city hall/municipalities, other public agencies, NPOs, ecc.) __________________________________________________________________________________________ __________________________________________________________________________________________ __________________________________________________________________________________________ __________________________________________________________________________________________ 16. What is your opinion about youth unemployment in Macedonia? Do you think there should be some plan for support of youth employability?
27
A3: CMP MODEL REGRESSION RESULTS . cmp (unemployed=elementary poor age agesq gender single nb_of_children health east southwest southeast pelagonia polog northeast skopski unempregion) (elementary=father_edu mother_edu age agesq gender single nb_of_children health east southwest southeast pelagonia polog northeast skopski unempregion) (poor=inc_required age agesq gender single nb_of_children health east southwest southeast pelagonia polog northeast skopski unempregion), nonrtolerance ind($cmp_probit $cmp_oprobit $cmp_oprobit) Fitting individual models as starting point for full model fit. Note: For programming reasons, these initial estimates may deviate from your specification. For exact fits of each equation alone, run cmp separately on each. note: southeast dropped because of collinearity Iteration Iteration Iteration Iteration Iteration
0: 1: 2: 3: 4:
log log log log log
likelihood likelihood likelihood likelihood likelihood
= = = = =
-837.89684 -716.7421 -714.54464 -714.54132 -714.54132
Probit regression
Number of obs LR chi2(15) Prob > chi2 Pseudo R2
Log likelihood = -714.54132
= = = =
1239 246.71 0.0000 0.1472
-----------------------------------------------------------------------------unemployed | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------elementary | .2822691 .1106189 2.55 0.011 .0654601 .4990781 poor | .5312606 .0942189 5.64 0.000 .3465949 .7159263 age | .4434053 .1460861 3.04 0.002 .1570818 .7297287 agesq | -.010178 .0031562 -3.22 0.001 -.016364 -.003992 gender | -.1850142 .0819681 -2.26 0.024 -.3456688 -.0243597 single | .0098967 .1161742 0.09 0.932 -.2178005 .2375939 nb_of_chil~n | -.0086912 .068078 -0.13 0.898 -.1421217 .1247393 health | .2128226 .1788675 1.19 0.234 -.1377512 .5633964 east | .0150864 .1609429 0.09 0.925 -.300356 .3305287 southwest | .2400947 .1901778 1.26 0.207 -.132647 .6128363 pelagonia | .3044795 .1323396 2.30 0.021 .0450987 .5638603 polog | .4631297 .1449022 3.20 0.001 .1791266 .7471327 northeast | -.0801583 .2463944 -0.33 0.745 -.5630824 .4027658 skopski | .3516487 .1231463 2.86 0.004 .1102864 .593011 unempregion | .0391405 .0067244 5.82 0.000 .025961 .0523201 _cons | -6.544885 1.676113 -3.90 0.000 -9.830007 -3.259763 -----------------------------------------------------------------------------Warning: regressor matrix for unemployed equation appears ill-conditioned. (Condition number = 2595.4062.) This might prevent convergence. If it does, and if you have not done so already, you may need to remove nearly collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance option to the command line. See cmp tips for more information. note: southeast dropped because Iteration 0: log likelihood = Iteration 1: log likelihood = Iteration 2: log likelihood = Iteration 3: log likelihood = Iteration 4: log likelihood = Iteration 5: log likelihood = Ordered probit regression
Log likelihood = -701.41423
of collinearity -922.885 -734.67458 -705.18699 -701.47104 -701.41424 -701.41423 Number of obs LR chi2(15) Prob > chi2 Pseudo R2
= = = =
2544 442.94 0.0000 0.2400
--------------------------------------------------------------------------------
28
_cmp_y2 | Coef. Std. Err. z P>|z| [95% Conf. Interval] ---------------+---------------------------------------------------------------father_edu | -.576793 .0561202 -10.28 0.000 -.6867865 -.4667995 mother_edu | -.1303904 .04148 -3.14 0.002 -.2116897 -.0490911 age | .3279737 .1086886 3.02 0.003 .1149479 .5409995 agesq | -.0067658 .0024684 -2.74 0.006 -.0116038 -.0019277 gender | -.0336403 .0784753 -0.43 0.668 -.187449 .1201684 single | -.4978798 .1099826 -4.53 0.000 -.7134417 -.2823179 nb_of_children | .1763061 .05732 3.08 0.002 .063961 .2886513 health | -.2296982 .1524016 -1.51 0.132 -.5283998 .0690034 east | .1108011 .1443148 0.77 0.443 -.1720506 .3936529 southwest | .9353306 .2272204 4.12 0.000 .4899869 1.380674 pelagonia | .1958312 .1556938 1.26 0.208 -.1093232 .5009855 polog | .2627622 .1621818 1.62 0.105 -.0551082 .5806326 northeast | .535069 .3041817 1.76 0.079 -.0611161 1.131254 skopski | .5888454 .1488013 3.96 0.000 .2972002 .8804905 unempregion | -.0169059 .0081889 -2.06 0.039 -.0329558 -.0008559 ---------------+---------------------------------------------------------------/cut1 | 3.149063 1.192826 .8111673 5.486958 -------------------------------------------------------------------------------Note: 3 observations completely determined. Standard errors questionable. Warning: regressor matrix for _cmp_y2 equation appears ill-conditioned. (Condition number = 3046.5328.) This might prevent convergence. If it does, and if you have not done so already, you may need to remove nearly collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance option to the command line. See cmp tips for more information. note: southeast dropped because Iteration 0: log likelihood = Iteration 1: log likelihood = Iteration 2: log likelihood = Iteration 3: log likelihood = Ordered probit regression
Log likelihood = -665.44831
of collinearity -716.9228 -665.80266 -665.44851 -665.44831 Number of obs LR chi2(14) Prob > chi2 Pseudo R2
= = = =
1252 102.95 0.0000 0.0718
-------------------------------------------------------------------------------_cmp_y3 | Coef. Std. Err. z P>|z| [95% Conf. Interval] ---------------+---------------------------------------------------------------inc_required | -.0383072 .0835472 -0.46 0.647 -.2020567 .1254423 age | -.0423801 .1344133 -0.32 0.753 -.3058252 .2210651 agesq | -.0003883 .0029591 -0.13 0.896 -.006188 .0054115 gender | .2119596 .0855976 2.48 0.013 .0441913 .3797279 single | -.0667899 .1230142 -0.54 0.587 -.3078933 .1743136 nb_of_children | .0492504 .0733682 0.67 0.502 -.0945486 .1930494 health | -.0271576 .1818195 -0.15 0.881 -.3835172 .3292021 east | .4654553 .1408012 3.31 0.001 .1894899 .7414206 southwest | -.1809033 .1999188 -0.90 0.366 -.572737 .2109304 pelagonia | -.4026056 .1516379 -2.66 0.008 -.6998104 -.1054009 polog | -.0439289 .1552206 -0.28 0.777 -.3481557 .2602978 northeast | .5728414 .2480189 2.31 0.021 .0867333 1.058949 skopski | -.2250782 .1347479 -1.67 0.095 -.4891793 .0390229 unempregion | .0124074 .0067373 1.84 0.066 -.0007974 .0256122 ---------------+---------------------------------------------------------------/cut1 | -.1363932 1.503311 -3.082829 2.810043 -------------------------------------------------------------------------------Warning: regressor matrix for _cmp_y3 equation appears ill-conditioned. (Condition number = 2683.1112.) This might prevent convergence. If it does, and if you have not done so already, you may need to remove nearly collinear regressors to achieve convergence. Or you may need to add a nrtolerance(#) or nonrtolerance option to the command line. See cmp tips for more information.
29
Fitting constants-only model for LR test of overall model fit. Fitting full model. Likelihoods for 1178 observations involve cumulative normal distributions above dimension 2. Using ghk2() to simulate them. Settings: Sequence type = halton Number of draws per observation = 69 Include antithetical draws = no Each observation gets different draws, so changing the order of observations in the data set would change the results. Iteration Iteration Iteration Iteration Iteration Iteration Iteration Iteration
0: 1: 2: 3: 4: 5: 6: 7:
log log log log log log log log
likelihood likelihood likelihood likelihood likelihood likelihood likelihood likelihood
Mixed-process regression Log likelihood = -2063.9909
= = = = = = = =
-2087.8407 -2078.6874 -2066.5851 -2064.8789 -2064.0394 -2063.9931 -2063.991 -2063.9909 Number of obs LR chi2(44) Prob > chi2
= = =
2544 729.79 0.0000
-------------------------------------------------------------------------------| Coef. Std. Err. z P>|z| [95% Conf. Interval] ---------------+---------------------------------------------------------------unemployed | elementary | .6011066 .254501 2.36 0.018 .1022938 1.099919 poor | 1.788204 .110618 16.17 0.000 1.571397 2.005011 age | .353827 .1289379 2.74 0.006 .1011134 .6065406 agesq | -.0076298 .0027991 -2.73 0.006 -.0131159 -.0021437 gender | -.2280057 .0737012 -3.09 0.002 -.3724574 -.083554 single | .0946181 .1102957 0.86 0.391 -.1215574 .3107937 nb_of_children | -.0463763 .063976 -0.72 0.469 -.1717669 .0790143 health | .1803879 .1580965 1.14 0.254 -.1294755 .4902513 east | -.1586607 .1392092 -1.14 0.254 -.4315056 .1141843 southwest | .1905087 .1791762 1.06 0.288 -.1606702 .5416877 pelagonia | .3681013 .1212675 3.04 0.002 .1304214 .6057812 polog | .3508713 .1333081 2.63 0.008 .0895922 .6121503 northeast | -.4703056 .2216096 -2.12 0.034 -.9046523 -.0359588 skopski | .342689 .1145402 2.99 0.003 .1181943 .5671836 unempregion | .0285118 .0064227 4.44 0.000 .0159235 .0411 _cons | -5.798954 1.46939 -3.95 0.000 -8.678906 -2.919002 ---------------+---------------------------------------------------------------elementary | father_edu | -.5576404 .0555499 -10.04 0.000 -.6665162 -.4487646 mother_edu | -.1230279 .0405298 -3.04 0.002 -.2024649 -.0435909 age | .333207 .1076507 3.10 0.002 .1222155 .5441986 agesq | -.0069454 .0024456 -2.84 0.005 -.0117387 -.0021522 gender | -.0302794 .0781086 -0.39 0.698 -.1833694 .1228106 single | -.5015302 .110219 -4.55 0.000 -.7175554 -.2855049 nb_of_children | .1841194 .0579022 3.18 0.001 .0706332 .2976057 health | -.2248957 .1509271 -1.49 0.136 -.5207074 .070916 east | .1653797 .1446663 1.14 0.253 -.118161 .4489204 southwest | .9383202 .2270178 4.13 0.000 .4933735 1.383267 pelagonia | .2145826 .1558617 1.38 0.169 -.0909006 .5200658 polog | .2941687 .1620153 1.82 0.069 -.0233756 .6117129 northeast | .5349439 .3043169 1.76 0.079 -.0615062 1.131394 skopski | .6114721 .1486719 4.11 0.000 .3200805 .9028638 unempregion | -.0154985 .008289 -1.87 0.062 -.0317448 .0007477 ---------------+---------------------------------------------------------------poor | inc_required | -.065111 .0703807 -0.93 0.355 -.2030546 .0728327 age | .0471486 .126597 0.37 0.710 -.2009768 .2952741 agesq | -.0021908 .00278 -0.79 0.431 -.0076396 .0032579 gender | .1751602 .0828477 2.11 0.034 .0127817 .3375387
30
single | -.0429824 .1224082 -0.35 0.725 -.282898 .1969332 nb_of_children | .0543203 .0752332 0.72 0.470 -.0931342 .2017747 health | -.1555307 .1729652 -0.90 0.369 -.4945363 .1834749 east | .5203645 .1384318 3.76 0.000 .2490431 .7916858 southwest | -.1611637 .1982224 -0.81 0.416 -.5496725 .227345 pelagonia | -.3137744 .1404726 -2.23 0.026 -.5890956 -.0384532 polog | .0291196 .1518782 0.19 0.848 -.2685561 .3267954 northeast | .542278 .2432735 2.23 0.026 .0654707 1.019085 skopski | -.1571566 .131013 -1.20 0.230 -.4139374 .0996243 unempregion | .0125621 .0066872 1.88 0.060 -.0005447 .0256688 ---------------+---------------------------------------------------------------/atanhrho_12 | -.375618 .1674704 -2.24 0.025 -.7038539 -.0473821 /atanhrho_13 | -1.4392 .3092193 -4.65 0.000 -2.045259 -.8331416 /atanhrho_23 | .2613853 .0610196 4.28 0.000 .141789 .3809816 /cut_2_1 | 3.290352 1.18089 2.79 0.005 .9758492 5.604855 /cut_3_1 | .8538028 1.420696 0.60 0.548 -1.93071 3.638316 ---------------+---------------------------------------------------------------rho_12 | -.3588959 .1458991 -.6068083 -.0473466 rho_13 | -.8935367 .0623362 -.9670895 -.6821593 rho_23 | .2555908 .0570334 .1408464 .3635596 --------------------------------------------------------------------------------
A4:CMP MODEL MARGINAL EFFECTS RESULTS . margins, dydx(*) predict(pr) force nochainrule Warning: cannot perform check for estimable functions. (note: prediction is a function of possibly stochastic quantities other than e(b)) Average marginal effects Model VCE : OIM
Number of obs
=
1252
Expression : Pr(unemployed), predict(pr) dy/dx w.r.t. : elementary poor age agesq gender single nb_of_children health east southwest pelagonia polog northeast skopski unempregion father_edu mother_edu inc_required -------------------------------------------------------------------------------| Delta-method | dy/dx Std. Err. z P>|z| [95% Conf. Interval] ---------------+---------------------------------------------------------------elementary | .1672962 .0703082 2.38 0.017 .0294947 .3050977 poor | .4976817 .0257207 19.35 0.000 .44727 .5480934 age | .0984749 .0358129 2.75 0.006 .0282829 .168667 agesq | -.0021235 .0007776 -2.73 0.006 -.0036475 -.0005995 gender | -.0634571 .0204119 -3.11 0.002 -.1034636 -.0234506 single | .0263335 .0306543 0.86 0.390 -.0337478 .0864149 nb_of_children | -.0129072 .0178014 -0.73 0.468 -.0477972 .0219829 health | .0502044 .0440184 1.14 0.254 -.03607 .1364789 east | -.0441574 .0387604 -1.14 0.255 -.1201264 .0318115 southwest | .0530212 .0498435 1.06 0.287 -.0446703 .1507127 pelagonia | .1024476 .0337489 3.04 0.002 .036301 .1685942 polog | .0976523 .0370394 2.64 0.008 .0250563 .1702483 northeast | -.1308925 .0616579 -2.12 0.034 -.2517398 -.0100451 skopski | .095375 .031853 2.99 0.003 .0329442 .1578059 unempregion | .0079352 .0017867 4.44 0.000 .0044334 .011437 father_edu | 0 (omitted) mother_edu | 0 (omitted) inc_required | 0 (omitted) --------------------------------------------------------------------------------
31