Impact of work-related activities on the outflow to employment and education within guarantees JDG and JGY. A study of public employment offices

(1)

i

Ekaterina Sukhanova

Impact of work-related activities on the outflow to

employment and education within guarantees JDG

and JGY. A study of public employment offices

Effekten av arbetsplatsförlagda aktiviteter på utflödet till arbete

och utbildning inom garantierna JOB och UGA. En studie av de

offentliga arbetsförmedlingarna

Economics

Master’s Thesis

Autumn 2013 Supervisor: Katarina Katz

(2)

ii

Abstract

This study uses employment office-level panel data to investigate the impact of work-related activities on the outflow from unemployment to work and education. The two largest labor market programs in Sweden are studied: Job and Development Guarantee and Job Guarantee for the Youth. To examine how activity participation affects the outflow to

employment/education I had to examine differences between offices (office heterogeneity).

This heterogeneity could be differences in local labor market conditions, office’s managerial ability, quality of activities and other features that are non-varying for each individual office.

The method to answer the research question is fixed effects distributed lag regression analysis with lagged values of the independent variable “activity participation rate”. In addition to the one-way office-specific fixed effects model, a model with time-specific fixed effects was introduced to check for seasonal impacts and fixed macroeconomic conditions that influence all offices. According to the two-way fixed effects regression results, the seasonal variations seem to have a stronger impact on the outflow to work/education than the work-related activities participation rates. Particularly, when time-specific fixed effects were added to the model with office-specific effects, the overall effect of work-related activities on the outflow rate became statistically non-significant and brought inconclusive results. The results have shown that the month of September has the strongest and most positive influence on the outflow rate as more people find work and leave to education during this month. The months that have the lowest impacts on the dependent variable are March and October, which means that fewer individuals left to work/education during these months across offices, in

comparison to other months.

Key words: work-related activities, Job and Development Guarantee, Job Guarantee for the Youth, employment offices

(3)

iii

Sammanfattning

Denna studie använder paneldata på kontorsnivå för att undersöka effekterna av

arbetsplatsförlagda aktiviteter på utflödet från arbetslöshet till arbete och utbildning. De två största arbetsmarknadspolitiska programmen i Sverige studeras: Jobb-och utvecklingsgarantin (JOB) samt Jobbgarantin för ungdomar (UGA). För att undersöka hur aktivitetens deltagande påverkar utflödet till arbete/utbildning kontrollerade jag för skillnader mellan kontoren (kontors heterogenitet). Denna heterogenitet kan bestå av skillnader i lokala förhållanden på arbetsmarknaden, kontorets förvaltningsförmåga, kvalitet på verksamheten och andra

funktioner som är icke - varierande för varje enskilt kontor. Metoden som har använts för att besvara frågeställningen är "fixed-effects distributed lag regressions analysis" med

eftersläpande värden för den oberoende variabeln ”aktivitetsgrad”. Förutom kontors -

specifika "fixed effects", infördes en modell med tid - specifika effekter för att kontrollera för säsongsmässiga effekter och fasta makroekonomiska förhållanden som påverkar alla kontor.

Enligt de två-vägs "fixed effects" regressionsresultaten så verkar säsongsvariationerna ha en större inverkan på utflödet till arbete/utbildning än den arbetsplatsförlagda aktivitetsgraden.

Framför allt när tid-specifika fasta effekter inkluderades i modellen med kontors-specifika effekter, blev den totala effekten av arbetsrelaterade aktiviteter på utflödeshastigheten statistiskt icke-signifikant och gav osäkra resultat som utfall. Resultaten har visat att

september månad har den starkaste och mest positiva inverkan på utflödet, vilket betyder att fler människor hittar arbete eller återgår till utbildning just under denna månad. De månader som har visat sig ha lägst påverkan på den beroende variabeln är mars och oktober, vilket innebär att färre individer hittar arbete eller påbörjar utbildning under dessa månader i jämförelse med andra månader.

Nyckelord: arbetsplatsförlagda aktiviteter, jobb-och utvecklingsgarantin, jobbgarantin för ungdomar, arbetsförmedling

(4)

iv

List of abbreviations

APR Activity participation rate ALMP Active labor market program GLS Generalized Least Squares JDG Job and Development Guarantee JGY Job Guarantee for the Youth PES the Public Employment Service WG Within-group method

(5)

v

1. Introduction

1.1 Background

Recent economic crises and the overall economic decline during the past five years have considerably increased youth unemployment and long-term unemployment in Sweden. Active labor market programs (ALMPs) are an essential part in fighting unemployment and therefore it is important that they are being continuously evaluated and improved when needed.

Arbetsförmedlingen, the Swedish Public Employment Service (PES), is a public employment agency that matches job seekers with prospective employers and administers both

unemployment insurance and selection to labor market programs. The two largest types of active labor market measures that exist nowadays in Sweden are Job and Development Guarantee (the JDG) and Job Guarantee for the Youth (JGY). The guarantees are aimed at bringing the unemployed into employment or education, with the JDG targeting the long-term unemployed and JGY being directed at the unemployed youth. In 2012 the total number of unique individuals who participated in both guarantees was 285 600, which accounts for 75 percent of the total program operations of the Public Employment Service

(Arbetsförmedlingen 2013a). The guarantees consist of different labor market measures such as employment measures (e.g. job search activities with coaching) and work-related activities.

Work-related activities are active labor market measures that are designed after an individual and primarily located at a workplace (e.g. vocational labor market training, job practice, work training and enhanced work training). Work-related activities are very similar to regular employments, although participants receive either activity support or development allowance instead of a regular wage.

The effectiveness of the guarantees can be measured by the number of individuals who leave to work and education after they have been participating in them. The outflow to

work/education depends on the outside environment each employment office operates in. The outside environment can be location of an office, labor market demand and supply structure, big city vs. small town and composition of job seekers in each office (e.g. age, gender, education, disability and country of birth). There are other factors that influence the chances of job seekers to employment/education such as whether persons participated in any activity at all, the quality of the activities, job seekers’ motivation, job coaches’ ability to motivate job seekers to study, the level of unemployment insurance, the time in unemployment and more...

(8)

2

1.2 Problem formulation

Individuals, who participate in the guarantees, take part in job search and work-related activities which are aimed at connecting the job seekers closer to the labor market. The intention of the work-related activities is to improve the productivity of individuals, thereby making it easier for job seekers to find work. Participation in the guarantees may encourage job seekers to seek actively for work or motivate them to enter full-time studies. Individuals who leave for education will increase their chances to find work after they are done with studies. Therefore the guarantees may have a positive effect on labor force participation either in the short or in the long run.

To increase the number of job seekers who leave to employment or education, the Swedish Public Employment Service has a need of identifying effective working measures in

employment services offices. These working measures can be the different types of activities, quality of activities or managerial ability of offices to get job seekers to employment and education. Much attention has lately been given to work-related activities (in Swedish arbetsplatsförlagda aktiviteter) in the guarantees the JDG and JGY as some of them have been found to lead the unemployed to work or full time education (Forslund et al. 2013;

Forslund & Vikström 2011; Carling & Richardsson 2001). Forslund & Vikström summarize that labor market measures which are very similar to regular jobs and current labor market training are effective at getting the unemployed to work (see their discussion from 2011).

Furthermore the Public Employment Service (the PES) has increased work-related activity participation rate for offices over the past years as they are assumed to get job seekers to employment/education (Arbetsförmedlignen 2013a). With regard to the above it becomes important to study the impact of work-related activities on the outflow to work and education.

Furthermore the chances of job seekers to find employment or leave to education do not only depend on the number of individuals in work-related activities or the type of unemployed individuals each office has, but also on how different employment offices operate. Some offices are simply better than others at getting job seekers to jobs or education because of the features that are specific to the offices themselves (e.g. managerial ability of offices to get job seekers to work or quality of work-related activities specific to an office). Therefore it is important to separate such differences in offices to be able to assess the overall impact of work-related activities on the outflow to work or education.

(9)

3

1.3 Purpose

The purpose of this thesis is to study the effect of work-related activity participation in the programs JDG and JGY on the outflow from unemployment to work/education with regard to differences across employment offices. The final research question of this paper is: will an office that has more job seekers in work-related activities have more individuals leaving the guarantees for work or education.

1.4 Method

Because offices do differ a lot in terms of the number of individuals per office and, therefore, in the number of individuals in activities, we need a method that could compare these offices on an equal scale. To compare the differences in offices we introduce all data in shares instead of numbers. The statistical data sets for JGY and the JDG were collected from the Swedish National Employment Office statistics (Datalagret). We have statistical data per employment office per month on shares of participants who have had at least one work-related activity during this specific month, shares of participants who are men/women, shares of participants who have secondary/elementary/postsecondary education, shares of participants in different age groups, shares of participants with disability/ without disability, and finally shares of participants who are born in Sweden/overseas. The dependent variable is called “outflow rate”

and represents an outflow from a specific guarantee (the JDG or JGY) to employment or education. Cross-sectional unit is an office and the time-series dimension includes 48 month periods (from January 2009 to December 2012). In JGY data set there are in total 233 offices and in the JDG there are 240 offices.

The method to answer the research question is chosen to be fixed effects distributed lag regression analysis with lagged values of the independent variable “activity participation rate”

(henceforth APR). Fixed effects accounts for the differences among offices in how they get job seekers to work and education. There can be things about the offices that make them have more activities or have more people leaving to employment or education. The things that affect both activities and outflow to employment/education would make the results biased if we did not control for office heterogeneity. Thus to see how activity participation affects the outflow to employment/education we must control for heterogeneity of offices. This

heterogeneity is assumed to be time-invariant. It could be the location of an office (big or small city), office’s managerial ability, quality of activities and other features that are non-

(10)

4

varying for each individual office. Another very important aspect of office heterogeneity is differences in local labor market conditions which also accounts for an office’s ability both to get people to work and to find places for work-related activities. Of course, labor market conditions such as the local business structure are not time-invariant but in the analysis we assume that they are approximately the same over the relatively short time period that we are studying (mainly between January 2009 and December 2012).

In addition to the models with only office-specific effects, I also include models with time- specific fixed effects (monthly effects) and compare them. Time-specific effects are included to control for time-varying influences common to all offices. They control for macroeconomic changes that affect all the offices and that vary with each month (on a time scale from January 2009 to December 2012). This so called two-way fixed effects regression method is used because there could be seasonal impacts on the outflow rates that affect all offices.

Besides the fixed effects and APR, the data sets include participant composition independent variables that may influence the outflow from unemployment to work/education.

Furthermore, in this method it is believed that there is a correlation between the fixed effects and explanatory variables. In this study we think that in addition to a specific office being (perhaps) more effective at helping unemployed to find jobs (having a higher share of individuals who leave to work/education), that could also be correlated with other

independent variables (e.g. share of participants with disabilities, share of participants with higher education).

(11)

5

1.5 Limitations

This study is done on two guarantee programs Job and Development Guarantee and the Job Guarantee for the Youth. Only the first two phases of the JDG are studied because the 3^rd phase is different from the first two (as will be explained later in the Background section).

The first two phases are also more important to study than the third one because the intention of the guarantee is to get people to work as soon as possible. In this study we assume that some local labor market conditions, which are included in the office specific effects, are fixed (they do not vary over time). The differences in labor market conditions can be a dominance of a particular industry, number of vacancies, the skills and composition of the population in their working ages, closeness to a larger labor market, etc. Of course such labor market conditions change over time but we assume that some fixed part of such conditions doesn’t vary over the short period of time we are studying (January 2009-December 2012). The changes in macroeconomic conditions that affect the local labor market conditions, and hence offices, will be captured by time-specific effects (time dummies) in the regressions.

Furthermore the effect of activity participation rate on the outflow to work and education may not happen immediately. For that reason lagged explanatory variables are introduced into the models. A maximum of 6 lagged variables of the activity participation rate are included because many of the work-related activities are supposed to last for a maximum of 6 months.

This is, however, a big limitation of the thesis because we cannot know for sure the maximum length of the lag. Another reason for not including more than 6 lagged variables is that the more lagged variables there are, the fewer degrees of freedom left, which makes statistical results insecure.

An important limitation is that we study individuals who leave unemployment to

work/education, but we do not study what happens to them after they leave the programs (we do not follow them up). It is true that some individuals who leave the guarantees may come back after some period. In this work we do not study the actual results of the activity

participation. We only study how the activity participation rates influence the general outflow to work/education. Furthermore we do not have data on how many individuals start an activity each month. We only calculate the share of individuals who are in activities each month.

Some individuals may have just started participation; some may have been there for a long time. We do not follow individuals but offices’ activity participation rates.

(12)

6

Another limitation is that I had to delete those people who left the guarantee for an unknown reason from the final data set because we cannot know whether these individuals left to work/education or not. It is a weakness in this study, but generally the number of such individuals is very small compared to the total number of individuals who left to

work/education and would not impact the estimates radically. It should also be noted that calculations made by me may give different results from the Public Employment Office official results, mainly because not all the offices were used in the final data sets. The reason why not all the offices were used is that some offices had too many missing values and thus, were excluded from the data sets. The offices in the data sets have at least 12 monthly observations of the dependent and independent variables. In the JDG data 28 offices were excluded in total and in the JGY 18 offices were taken away.

1.6 Disposition

The second section will present general information about the guarantees and the kinds of work- related activities. In section 3 different theories of unemployment and impacts of participation in the guarantees on the job seekers are studied. Section 4 includes previous studies on the Job and Development Guarantee, the Job Guarantee for the Youth and other earlier studies that are related to the topic.

In the fifth section a set of independent and dependent variables that are used in the study are presented and explained in detail. Section 6 presents the econometric method of fixed effects and discusses assumptions for fixed effects. In the section Descriptive statistics the reader can find summary of the variables used in the JDG and JGY data, broken down by year.

Descriptive statistics will present a summary of average share of individuals who have left the guarantees to employment/education, a summary of independent variables such as average share of individuals in activities in each of the job guarantees and other important variables that identify a jobseeker composition in all the offices, per each year. The results are presented in section 8 and discussed in the 9^th section. A conclusion is made in the last section.

(13)

7

2. Background

2.1 Job and Development Guarantee

Guarantees are quite different from the regular labor market programs. A guarantee includes a set of various Employment Service measures such as job search activities, subsidized

employment and other work-related activities. The Job and Development Guarantee (the JDG) is a labor market program which was introduced in July 2007 in order to assist long- term unemployed individuals in finding jobs. The purpose of the program is to get the participants to the labor market as quickly as possible. The number of participants in the program has increased over the last four years with the outbreak of financial crisis and the overall weakening of the labor market. The total number of participants was around 100 000¹ in April 2013 (Liljeberg et al. 2013). The JDG is considered to be the largest labor market program within the Public Employment Service, with a share of 53 % of all the programs with activity support (Arbetsförmedlingen 2013a). In 2012 the programs costs added up to 13 109 million crowns, of which 7 619 million was for the phases 1 and 2 and 5 490 was for the phase 3.

The Job and Development Guarantee includes three different phases. The first phase consists primarily of employment mapping, job coaching and preparatory measures. The second phase includes work-related activities, such as Job practice, vocational labor market training and On-the-job rehabilitation. An individual who is assumed to have less need of coaching can be transferred to the second phase earlier in the program. However the time before transition into the phase 2 varies with individuals and the availability of work-related activities. There have been cases when individuals had coaching after the first 150 days (during the 2^nd phase) in the program mainly because there have been problems to arrange activities (Arbetsförmedlingen 2011a, p. 5). All the activities are to comprise 75 percent of time during the first two phases of the guarantee and the remaining time is devoted to individual job seeking. In the third phase activities should be the accounted for 100 % of a job seeker’s time supply. In the 3rd phase individuals are referred to mandatory subsidized employment with a so called “provider”. The 3^rd phase is also called “the employment phase”.

Job and Development Guarantee (JDG) is a program for long-term unemployed who have either run out of days of unemployment insurance (300days), have been registered at PES

1 Total number of participant in all the 3 phases of the program. In April 2013 there were 64 700 remaining participant in the 1^st and 2^nd phases.

(14)

8

offices for 14 consecutive months, or participated in The Job Guarantee for the Youth for 15 months(see Figure 1). Those who have the right to unemployment insurance (in Swedish A- kassa) can start participating in the JDG when the first 300 benefit days are over. Even those individuals who didn’t have a right to unemployment insurance, but have been registered at the Public Employment Service for 14 months can start participating in the Job and

Development Guarantee. The maximum days that can be spent in Job and development guarantee, phase 1 and 2 are 450 days. During the first two phases participants can get either activity support or development allowance from the Swedish Social Insurance Agency (Försäkringskassan), which actually means a decrease in compensation level compared to unemployment insurance from A-kassa. Benefits can vary from 223 to 680 crowns per day, depending on the presence of membership in an unemployment insurance fund (A-kassa), previous work and unemployment history, and are given as long as individuals participate in the program. The 3^rd phase follows right after the first 450 days in the program. The Figure 1 below shows a flow of unemployed over time for both the Job and Development Guarantee and the Job Guarantee for the Youth.

Figure 1. Flow of unemployed over time.

Note: the figure is adopted from the Public Employment Service

There is a possibility that an individual comes back to the program after leaving it. An individual can be included in the program until he/she will find employment, leave to education or leave the PES for another reason (e.g. retirement or death). Furthermore a job seeker can participate in the guarantee several times if he/she gets eligible again. Under the legislation, an individual has the right to reenter the program if he or she has not earned a new

(15)

9

“work condition” (worked in such extent that she/he is again eligible for unemployment insurance) during deregistration period (Arbetsförmedlingen 2011a).

Table 1 shows the total number of remaining participants in all the 3 phases. We can see that the number of participants increased by roughly 79 000 between 2007 and 2012.The number of remaining job seekers (yearly average) in all the phases in the JDG has increased by approx. 4 700 people in 2012 compared with 2011.

Table 1. Remaining job seekers per month in the JDG, yearly average²

Year 2007 2008 2009 2010 2011 2012

Total 18 427 41 938 58 193 79 192 93 072 97801

Of which in employment phase 0 0 4 712 19 769 27 370 32160

The number of job seekers in the JDG has increased steadily since its introduction in 2007.

Figure 2 shows the development of the number of remaining job seekers in the JDG’s first two phases over the period January 2009-April 2013. We can see that it has increased during the last year.

Figure 2. Number of remaining job seekers in Job and Development Guarantee (phases 1&2). Development over the period January 2009- April 2013.

Source: the Swedish National Employment Office statistics (Datalagret)

2 In 2007 the average of 6 months is taken, in 2009 in the employment phase the average of 10 months is taken

(16)

10

2.2 Job Guarantee for the Youth

The Job Guarantee for the Youth (JGY) is an active labor market program that has a purpose of countering youth unemployment and decreasing the youth unemployment duration. The Job Guarantee for the Youth was introduced in December 2007 and replaced the earlier programs Municipality youth program and Youth guarantee program (Swedish: Kommunallt ungdomsprogram and Ungdomsgarantin). In 2011 22% of participants in all programs with activity support were participants in JGY. The purpose of this guarantee is to offer its

participants effective labor market measures in the earlier stage of unemployment so that they could find a job or leave to education as soon as possible (Arbetsförmedlingen 2013a).

Individuals in the age group 16-24 can register in the guarantee if they have been openly unemployed and registered at the PES for 90 days (3 months). Figure 1 shows the flow of unemployed individuals in the two studied guarantees over time. Young individuals can begin their participation in Job and Development Guarantee after 15 months of participation in the Job Guarantee for the Youth.

The Job Guarantee for the Youth itself lasts for a maximum of 15 months and during (at least) the first 3 months participants will have:

- Extensive employment mapping - Job search activities with coaching - Study and career guidance

Later these activities can be combined with work-related activities such as Business start-up grants, On-the-job training, Vocational rehabilitation or other activities listed in Table 3, p. 12 (Arbetsförmedlingen 2013b). These activities should ideally be combined with a minimum of 4 hours job seeking activities per week. Furthermore individuals may participate in work- related activities already in the first months in the Job Guarantee for the Youth if they are assumed to be in need. During the time in the program participants get either activity support as little as 320 crowns per day (if they are entitled to unemployment insurance, A-kassa) or development allowance at 223 crowns per day. Those individuals who reject to participate in the guarantee do not receive any allowance or activity support. Any individual who left JGY for job/education can come back there any time if he/she is eligible again. Furthermore a job seeker can participate in several different activities and there is no limit on how long in total an individual can be enrolled in activities (within the program participation period, of course).

For JGY the total number of unique individuals was 108 600 in 2012 (Arbetsförmedlingen

(17)

11

2013a). However on average only 40 000 people participated in the program each month in 2012 (see Table 2).

Table 2. Remaining job seekers per month in JGY, yearly average

Year 2008 2009 2010 2011 2012

Total 11 120 34 816 46 838 39 094 40 019

The number of remaining unemployed young individuals has increased considerably during 2008-2010 as a result of the financial crisis. To be more specific, the number of remaining unemployed youth in the program increased drastically between September 2009 and April 2010 (Figure 3), and we can see that the number of participants remaining in JGY in April 2013 was still very high (42 100 individuals).

Figure 3. Number of remaining job seekers in the Job Guarantee for the Youth.

Development over the period January 2009 - April 2013.

Source: the Swedish National Employment Office statistics (Datalagret)

The objective of the guarantee is not only to help young people to find jobs, but also to motivate them to study because education is a very important factor that increases chances to find employment. Therefore the PES offers a variety of activities with a purpose of motivating individuals to enter full-time studies (e.g. studies in the Swedish Folk High School,

Folkhögskola). According to the PES, 28.6 % of participants in JGY had at least one work- related activity in 2012. The average monthly activity participation rate has increased with 7.2 percentage points in 2012 when compared to 2011(Arbetsförmedlingen 2013a).

(18)

12

2.3 Work-related activities

In the second phase of the JDG and in the later period of JGY job seekers participate in work- related activities, even though job search activities are also present. Work-related activities are activities that are organized by an employer which can be offered by both public and private sector. There is a variety of work-related activities within Job and Development Guarantee and the Job Guarantee for the Youth. Altogether there are 15 different types of activities, but the activities Project, Work training and Enhanced work training are absent in the Job Guarantee for the Youth.

Table 3. Work-related activities in the guarantees. Swedish – English translated Arbetsplatsförlagda aktiviteter Work-related activity

Lyft Lyft

Samhall Samhall

Utbildning Training (e.g. Labor market training)

Arbetspraktik On-the-job training (also called Job practice) Praktisk kompetensutveckling Practical skills development

Kulturarvslyftet Development employment

Praktikantprogrammet Practical training program Start av näringsverksamhet Business start-up grants

Arbetslivsinriktad rehabilitering Employability rehabilitation program

Timanställning Hourly employment

Tillfälligt arbete i önskad omfattning Temporary employment

Deltidsarbete Part-time work

Projekt* Project*

Arbetsträning* Work training*

Förstärkt arbetsträning* Enhanced work training*

Note: *activities are present only in the Job and Development Guarantee

It is important to understand that work-related activities within the JDG and JGY guarantees and other active labor market programs that have the same names mean slightly different things, e.g. On-the-job training (Arbetspraktik), Labor market training

(Arbetsmarknadsutbildning), Hourly employment (Timanställning). Although they have the same names and contents, one of them is a labor market program and the other is a work- related activity within JDG/JGY. On-the-job training within JDG/JGY is carried out under supervision at a workplace with a purpose of vocational guidance, vocational training or practical work training. On-the-job training may last for up to six months in the JDG (3 months in the JGY) and participants get activity support during that time. Business start-up grants are a form of subsidized self-employment. These grants can be given to unemployed who want to start their own businesses and they consist of unemployment insurance which can be received for up to six months. Work training and Enhanced work training are

(19)

13

subsidized employment programs and a form of vocational rehabilitation for persons with disabilities and those with long-term sicknesses. Enhanced work training can last for up to 12 months. “Lyft” is another type of work-related activities which is aimed at activating a job seeker at a workplace and has duration from 3 up to 6 months (Arbetsförmedlingen 2010a).

Generally the average share of work-related activities increased over time and now it is as large as the share of job seeker activities with coaching in all the 3 phases of the JDG program (34,5 % and 34,6% in 2012). According to the PES, the share of work-related activities in the JGY has increased by 7 percentage points over the year 2012 and the number of participants in the JDG who didn’t have a registered activity decreased by 63 percent between 4^th quarter of 2012 and 4^th quarter of 2011 (Arbetsförmedlingen 2013a). Activities that contributed to an increase between the years 2011 and 2012 are Labor market training and Preparatory

education.

Table 4 shows that on average the amount of people in activities in the JDG has more than doubled between the years 2009 and 2012. The number of individuals remaining in activities has increased for both JGY and the JDG over these years, as well as the number of remaining participants in general (see Table 1 and Table 2).

Table 4. Monthly averages of remaining job seekers in activities, in the guarantees the JDG and JGY.

Year 2009 2010 2011 2012

Remaining in activities JDG

9 695 12 973 16 002 20 838

Remaining in activities JGY

2 913 6 662 7 219 10 976

Source: the Swedish National Employment Office statistics (Datalagret), own calculations

(20)

14

3. Theoretical background

3.1 Unemployment and its causes

The unemployment rate and the speed at which job seekers find employment vary greatly with the economic condition of a country. Recent economic crises (financial crisis and European debt crisis) and the overall economic decline during the past five years have

significantly increased the youth unemployment and the long-term unemployment in Sweden.

Recessions and especially depressions affect negatively the amount of individuals who move to employment. This occurs because of a decline in the aggregate demand (e.g. the forest industry in Sweden) which forces firms to lay off and discharge workers. When an economy is in a recession there are generally fewer work places and fewer employers are willing to hire new workers. There is an excess supply of workers and wages are sticky and cannot adjust downwards. Such a temporary cyclical negative shock as the financial crisis has caused the unemployment in Sweden to remain on a high level.

Structural unemployment arises if job seekers do not fit the jobs available on the labor market.

An example of structural unemployment in Sweden is that the manufacturing sector has declined substantially over the past years while service sector has increased. Workers who were laid off from the manufacturing sector cannot quickly move to the growing sectors because of the lack of appropriate skills and qualifications. This structural change following the recession, higher minimum wages, together with the fact that jobs become more and more knowledge-intensive (meaning that employers demand more educated personal) creates a problem for some of the job seekers to find work. For relatively weaker groups of job seekers (regarding the chances to find employment) such as youth, long-term unemployed, persons with disabilities and foreign-born the situation is much worse. Long-term unemployed individuals have a large risk to become structurally unemployed. Long-term unemployed can suffer from structural unemployment because the job seekers may lose their skills and abilities or lose motivation for job search over the unemployment period and thus never become employed again. Employers themselves may not be willing to hire a long-term unemployed as they may be seen as less productive. As a result, the unemployment spells of these individuals may last for a long time and structural unemployment will sustain.

Youth unemployment is very responsive to the business cycle and it has increased steadily in Sweden over the past five years. Youth unemployment is usually higher than unemployment

(21)

15

in other age groups. We can compare the youth unemployment rate (those who are younger than 25 years) with unemployment among older people (those over 24). Among Swedes over 25 years, the unemployment rate was at 5.7 percent in 2012, while youth unemployment rate was 23.7 percent (Eurostat 2013). Employers are reluctant to hire young individuals because they are unsure of the young peoples’ competencies. Youth can be seen as less productive than persons in their middle ages because it takes time for the young to acquire knowledge and skills necessary for a certain job. Many of the educational programs in Sweden are not so good at providing practical training or other ways to connect students with the labor market.

Because of that young individuals have lower chances to find employment than those with labor market experience.

As it was already said , youth unemployment can be higher than unemployment in other age groups, but the unemployment spells for the young unemployed are usually shorter than unemployment spells for unemployed in their late ages (Zetterberg 2012). This is because young persons can switch in between jobs and education more frequently and they are willing to change in between temporary employments or move to a different workplace location. The so called “job shopping” has also a positive effect on the employment probability for the youth as young individuals may receive new skills and knowledge from different market industries. Educational background, motivation and personal ability influence the time in unemployment for the youth. Many of the young tend to find employment quickly. These groups of individuals usually have completed upper secondary education, are motivated and search for work more frequently than others. However there are some exposed groups of unemployed youth that have longer spells of unemployment. The groups that have a

“disadvantage” are persons with no completed upper secondary school grades, foreign-born and persons with disabilities (SOU 2007:18).

Recent inflows of unemployed are made up to a large extent of weaker groups of job seekers who have a difficulty to find work. These groups need extra help in getting a footing in the labor market. In order to improve the skills of job seekers and thereby to decrease the time in unemployment the Swedish government introduced active labor market programs Job and Development Guarantee for the long-term unemployed and Job Guarantee for the Youth.

(22)

16

3.2 Effects of participation in the guarantees on job seekers and labor

market actors

Job search and work-related activities are very important in the Job and Development

Guarantee and Job Guarantee for the Youth as they are aimed at connecting job seekers closer to the labor market. Participation in the guarantees may not only affect the participants

themselves, but also other labor market actors, such as open unemployed and employers. We will first look at a guarantee’s framework and then study how the guarantees will impact diverse actors with the help of economic theories. This theoretical background is based on a work by Calmfors et al. (2002).

Forslund & Vikström (2011a) separate different types of labor market measures into three main groups:

1. Placement measures. These measures include job coaching, extensive employment mapping, job search activities, preparatory actions, group activities, study/career guidance and are aimed at increasing the matching efficiency of the labor market.

2. Educational measures. These measures can be either preparatory or vocational (e.g.

Labor market training) and are aimed at educating the job seekers so that they leave to full-time studies or find work.

3. Employment measures. These measures are aimed at facilitating the job seekers’

transition to employment via stimulation of employers’ demand. Examples are subsidized employment (e.g. Work training, Business start-up grants) and other subsidy for job seekers when employed by an outside employer (e.g. Job practice).

Labor market’s functioning depends on the ability of the labor market to match unemployed workers with the available jobs. Characteristics of job seekers, types of jobs available, location of the labor market and the search behavior of unemployed will all affect the matching efficiency. What the guarantees can do to increase the matching efficiency is to offer the job seekers effective job search activities with coaching and study/career guidance.

These activities teach job seekers how to write CV and cover letters, prepare for job interviews, and teach about other ways to find employments. Job search activities help to understand the presence of frictional unemployment: they are aimed at increasing the

matching efficiency and decreasing the frictional unemployment. Frictional unemployment is a type of unemployment that occurs because both workers and employers need time to locate each other and learn about each other. Imperfect information may worsen the frictional

unemployment if the job seekers are unaware of available jobs. Figure 4 depicts the aggregate

(23)

17

labor market in an economy. The supply curve gives the total number of working hours that individuals are willing to supply the market at any given wage (Borjas 2006, p. 148). Each worker has a reservation wage, which is a minimum wage required to make them willing to take a job rather than remain unemployed. As real wages rise, more and more workers are willing to enter the labor market. Hence we get an upward sloping labor supply curve. Labor demand depends negatively on the real wage. The supply curve intersects the demand curve at an equilibrium, generating a competitive wage W* and a number of workers E*. Figure 4 also shows a vertical line that represents the total number of persons in the labor force (Ef).

Because labor force consists of both unemployed and employed individuals, the distance in between Ef and E* is the number of unemployed individuals (both open unemployed and participants in ALMPs). As we can see the total labor force remains always higher than the employed labor force. The resulting unemployment is called the “natural rate of

unemployment” and is unemployment that consists of frictional and structural unemployment.

Even if there were no fundamental imbalances between the supply and demand for workers, there would always be frictional unemployment because it takes time for job seekers to match for the available job vacancies. Efficient job search activities would improve the chances of job seekers to find work and the employment would increase (labor supply shifts to the right).

Figure 4. Labor market equilibrium, real wage and unemployment

Consequently one of the main objectives of the Public Employment Office is to improve the matching process between job seekers and employers. Participation in activities may motivate job seekers to look for jobs more than if they were to look on their own and only receive allowance. Job search activities and guidance have been found to have positive effects on the employment probability (see Hägglund 2009; Kluve & Schmidt 2002; Liljeberg et al. 2012b).

(24)

18

An individual will leave unemployment once she/he accepts a particular job offer. Acceptance of a job offer depends on this individual’s reservation wage. Reservation wage makes a job seeker indifferent between working and not working. One of the effects of the guarantees is the effect of the unemployment insurance on the job search and the duration of

unemployment. The job search theory explains that more generous unemployment insurance decrease incentives for job search and increase the unemployment duration (Björklund et al.

2006). After 300 days of unemployment benefits with A-kassa are exhausted, when individuals enter the JDG, their unemployment insurance is lowered in comparison to the previous level.³ Particularly, in the JDG activity support is 80 % of earlier income during day1-day200 of unemployment period, 70% during day201-300 and 65 % from day 301. On the other hand in JGY activity support is based on earlier income (with a minimum of 320 and a maximum 680 SEK) and decreases over the unemployment period (individuals get 80% of earlier income during unemployment period of day1 to day100, 70% during day101-200 and 65% from the day 201) (Arbetsförmedlingen n.d.). As we can see, during the unemployment period in JGY job seekers benefit level decreases faster than benefit level during the

unemployment period in the JDG. Higher unemployment benefits increases a job seeker's reservation wage because they reduce the marginal cost of search (the cost of an additional search incurred by a job seeker) (Borjas 2010). When unemployment benefits are lowered so is the reservation wage (also called asking wage), which in turn increases the probability of an individual to accept a job offer. In the earlier stage of unemployment the insurance is

realtively high, but when participants enter any of the guarantees, it drops. This may

contribute to the question why individuals have higher chances to leave unemployment in the beginning of participation in the guarantees. It is also known that a relatively small amount of unemployment benefits can actually be beneficial for individuals as they assist in covering the costs of job search (Björklund et al. 2006). When used effectively unemployment benefits may in fact increase the probability of becoming employed. Nevertheless unemployment benefits should never be too extensive so that unemployed would prefer them as a source of income.

The programs may have a positive effect on the labor force participation if the different offices succeed at quickly recognizing weak and strong individuals and distributing persons into right activities. Work-related activities will create more competition for jobs since the

3 Only for those individuals who are eligible for A-kassa insurance

(25)

19

participants will become more competitive on the labor market after being given e.g.

vocational labor market training or Job practice. The outcome for the labor market can be explained with the help of Figure 4. Labor supply schedule shifts to the right because of increased competition (work force increases). As a result of the increased competition the real wage bargaining is worsened and the real wage drops. However the overall regular

employment is now increased (as a result of a shift of the labor supply). On the other hand this competition aspect may weaken the employment prospects of the open unemployed

individuals and thus the program participation could be viewed as less efficient.

One of the positive and desired effects of the guarantees, and in particular work-related activities, is an increased productivity of job seekers. Human capital grows when a person participates in vocational labor market training or any other activity that upgrades and adapts the skills of the unemployed to the current labor market. Human capital corresponds to any stock of knowledge and skills that a worker has that contributes to his or hers productivity.

Consequently such activities would increase job seekers’ chances to find employment. Work- related activities are aimed at reducing another type of unemployment, mainly the structural unemployment.

A disadvantage of work-related activities is that they may instead increase the time in unemployment because of the so called locking-in effect. During the period when an

unemployed person participates in an activity, he or she does not look for a job with the same intensity as an open unemployed person because this individual is busy with the activity, such as labor market training (Arbetsmarknadsutbildning), Job practice (Arbetspraktik) or Practical training program (Praktikantprogrammet). Therefore individuals may not exit the programs before certain activities are completed. Work-related activities may instead increase the chances of outsiders to find work while the insiders are still participating in these activities.

However this may have a positive effect on the employment because the outsiders will experience shorter unemployment spells and thus may never enter the guarantees in the first place. The result will depend on whether treatment effect will be larger than the locking-in effect (see the discussion by Calmfors et al. 2002). It will also depend whether more open unemployed than program participants will become employed as a result of program participation.

(26)

20

Calmfors et al. (2002) questions the large-scale youth programs as they seem to have large displacement effects. They conclude that subsidized employment and youth programs seem to have caused displacement effects during the 1990’s in Sweden. One of the negative effects of subsidized employment, such as Work training, is displacement (crowding out of the labor demand). Displacement effect assumes that same individuals could have been hired elsewhere in the absence of these subsidies. In this case jobs that are created by the guarantees are at the expense of other jobs, and consequently the regular employment is reduced. Several studies have concluded that subsidized employment is not an effective measure to bring down the unemployment (Calmfors et al. 2002, Boone & van Ours 2004). However others point to that subsidized employment could lead to regular employment (Forslund et al. 2004). Kluve &

Schmidt (2002) write that private sector subsidized employment is better than public sector programs and that active labor market programs should in general be targeting the

“economically disadvantaged” individuals. Other types of subsidized employment such as Business start-up grants can cause competition in the product market which may cause other employers to lay off workers. However this argument is not very big and is quite hard to study. Work training (Arbetsträning), where employers receive compensation, can be

successful in getting unemployed to work, but it should not be used on a larger scale because of its substitution effect. By substitution effect it is meant that an employer would prefer to hire a jobseeker in a e.g. the JDG guarantee than an openly unemployed because he receives an income for hiring the jobseeker in the guarantee. Subsidized employment increases

incentives of employers to hire unemployed individuals, which will shift the labor demand up by an amount of the subsidy (Figure 4). The outcome is a shift of the equilibrium level and an increase in the number of employed.

From the theory we have seen that what can be effective for an individual may not be very effective for the aggregate level of unemployment. However the final effect of the program participation is very hard to study as it can last for a very long period of time.

(27)

21

4. Previous studies

4.1 Empirical literature review on the JDG and work related activities

When a labor market program has such a large volume as the JDG, there comes a question whether this program is effective in the short or in the long run. The probability to leave the program (the JDG) varies with the time in unemployment (in the program), type of activities, activity levels and also the participants’ composition which vary over time. In a study by the Public Employment Office (Arbetsförmedlingen 2011a) it was found that generally the probability to leave the program decreases the longer time an individual is written in the program. Individuals who stand closer to the labor market leave the programs earlier than the ones who are in weaker groups. The result from the study has shown that persons in ages 55- 59, foreign-born and those who have disabilities have lower chances to leave the program. On the other hand young persons have had a higher chance to leave the program than persons in ages 30-49. Liljeberg et al. (2013) study results also show two groups that seem to have a significant problem to leave the program: people who are in ages 55-59 and people with disabilities that decrease working ability. Liljeberg et al. (2013) studied the program participants between July 2007 and April 2012. In their report it was found that the

participants in the phase 1 and 2 were older than in the other groups (e.g. open unemployed), with an average age of 42 years old. The share of individuals who have a disability that reduces work performance was 25% compared to 17% of the open unemployed (page. 14).

More participants had upper secondary school education compared to open unemployed, but fewer individuals had postsecondary education. Participants in the Job and Development Guarantee have a weaker position in the labor market because they have been unemployed for longer periods than others. Liljeberg et al. (2013) find that before entering the JDG, job seekers have been unemployed for about two years on average.

The probability to leave the program increases in the very early stage of the program participation, which could happen because individuals allowance level decreases and/or the fact that they start to participate in activities (Arbetsförmedlingen 2011a; Liljeberg et al.

2013). There are different explanations for why the chances to find employment depend on the duration of unemployment. Human capital continues to deteriorate the longer an

individual is unemployed. Long-term unemployed individuals are seen by employers as less productive than other job seekers and therefore employers are hesitant to hire them. The longer an individual is unemployed the less are the incentives to look for jobs because

(28)

22

individuals may give up on their job search because they could not find any employment (discouraged worker effect).

Probability of entering employment depends on the type of activities each job seeker has had previously. Ideally each activity needs to be individualized so that each job seeker would get an activity that best fits him/her and would lead to employment. In their studies Hägglund (2009) and Kluve & Schmidt (2002) have found that job-search assistance decrease the time in unemployment. The Swedish National Audit Office (Riksrevisionen 2009) concluded that job-search activities with coaching of the JDG are effective provided that their extent is limited and that they target only those individuals who are in great need of them. Another study by Liljeberg et al. (2012b) found that job coaching seems to help more short-term unemployed than long-term unemployed. According to their work, job coaching increases the chances of women, persons with postsecondary education and persons born outside of Europe to find employment. When it comes to work-related activity participation rate, it was

relatively low according to a work by Martinson and Sibbmark (2010). Liljeberg et al. also conclude in their study on the Job and Development Guarantee from 2013 that the work- related activity level among the participants was quite low. Liljeberg et al. (2013) studied 250 000 participants in the JDG phase 1 and 2 between July 2007 and April 2012. They found in their survey that 3 out of ten persons did not have a work-related activity or labor market training before they entered phase 3. Furthermore 6 out of ten participant in the 3^rd phase said that they didn’t get enough help with job search or with finding practical work training in the first two phases. Twelve percent of participants in phases 1 and 2 have at least once had Work training and 10 % have had Enhanced work training. A small share of participants had Part- time job (2%) or Hourly employment (6%). They conclude that the share of individuals in job search activities (79%) is considerably higher than the share of individuals who have had work-related practical training. Furthermore, participants who were involved in the activities of the Public Employment Service seem to leave unemployment quicker than those who had been placed in activities with a complementary agent (kompletterande aktör) at the same time.

A complementary agent is any market agent that is not a PES agent that offers job-seekers help to find a job.

Liljeberg et al. 2013 emphasizes the importance of work-related activities on the outflow to work. Research shows that particularly work-related activities have a possibility to lead to employment, while job search activities with coaching are assumed to be less effective for the

(29)

23

long-term unemployed (Liljeberg et al. 2012b). De Luna et al. (2008) found that vocational labor market training (which is also a type of a work-related activity) decreased the time in unemployment for job seekers who entered the program in 2002-2004 with 20%. Also de Luna et al. (2008) presented results that job seekers in weaker groups (with low education and those who are born overseas) have a beneficial influence from labor market training. Okeke (2006) studied individuals 180 days after they left labor market training and found that the labor market training had a positive effect on the probability of receiving a job. Other studies also confirm that labor market training is an effective measure to fight unemployment (Boone

& van Ours 2004; Forslund et al. 2011b). However earlier studies of effects of labor market training on the outflow to work showed no positive results (Calmfors et al. 2002). Calmfors et al. explain in their study (2002) that to be effective, active labor market programs should be used on a smaller scale. They write: “There should be a greater emphasis on holding down long-term unemployment in general and a smaller emphasis on youth programmes.” They also conclude that ALMPs should not be used as a means to renew unemployment benefit eligibility (like it was during the 90’s) because it would “distort the incentives for labor participation and result in very large program volumes”. As Forslund & Vikström (2011a) explain it, earlier studies of labor market training has given a bad image of the effect of the active measure on employment because of the bad labor market condition during the 1990’s, but a combination of new UI rules, lower program volumes and a better economic situation seem to give better results for the labor market training. The training results in the short term will depend on whether there are jobs to train the job seekers for.

According to another study by Forslund, Fredriksson and Vikström (2011b), labor market training should be used during recession despite their locking-in effect. Job seekers in labor market training suffer from the so called locking-in effect. Forslund et al. (2011b) write that if

“programs with large locking-in effects should ever be used, they should be used during a downturn” because the opportunity cost of search time is low in recession. That is individual and the society as a whole would experience higher costs if the individual would seek for work on his own. Therefore Forslund et al. (2011b) suggest that because the locking-in effect is high in labor market training, it is best to use this measure in a recession rather than during boom economy. They also find that labor market training is more effective than Job practice during recession.

(30)

24

Carling & Richardssons study (2001) showed that Work training (Arbetsträning) works better and quicker at getting the job seekers to work than educational measures. Subsidized

employment is also included in the Job and Development Guarantee (e.g. work-related activities such as Work training and Business start-up grants). Employers, who provide e.g.

work training and enhanced work training, receive employment support (subsidy) for the risk of employing long-term unemployed individuals. Several studies have been made on effects of subsidized employment on the employment probability. For example Carling & Gustafson (1999) compared participants in self-employment grants and participants with recruitment subsidy (rekryteringsstöd). Their results show that the risk to become unemployed again is twice as high for those who were employed with a recruitment subsidy than those who started their own businesses. Forslund et al. (2004) estimate a treatment effect of employment

subsidies and also the treatment effect over the time in unemployment. They concluded that program participation increases the outflow from unemployment to work. However these results had time dependency: during the first seven months of entering a subsidized

employment program the job seekers were locked into them. After 7 months the probability of getting a job was larger for those with subsidized employment than for those who didn’t have a subsidized employment.

Forslund & Vikström (2011a) compare a so called profiling and targeting with guarantee programs. Profiling can be explained as searching for individuals who are at risk of becoming long-term unemployed and for those who can find jobs faster. If it is possible to identify individuals who are in risk of becoming long-term unemployed, a set of specific actions can be applied to these individuals and there will be no need to spend a lot of money on

individuals who can find employment relatively fast. Targeting is providing measures for persons so that each individual would get an activity that would be designed just for her/him.

In the Job and Development guarantee individuals get their activities at already predetermined dates (generally in phase 2) and not in the beginning of the program participation. Because a large share of individuals in the JDG is assumed to be far from the labor market and long-term unemployed benefit less from job seeking activities with coaching (Liljeberg et al. 2012b), attention of the guarantee is switched on the work-related activities instead of job seeking activities. That it is why it is so important for the PES to increase the share of individuals in work-related activities.

(31)

25

A large drawback of the work-related activities for the long-term unemployed is that they may stimulate workers to reduce their job search activities and let them instead to participate in e.g. Labor market training. There is a risk of participant to be locked in the activities as they can last for long periods (Riksrevisionen 2009). Forslund et al. 2013 estimated the average time in job practice and labor market training for job seekers which were registered with the PES between 1999 and 2006. The average duration time of job practice was 96 days (3 months) whereas for labor market training it was considerably higher (139 days). Forslund et al. study results also show that participants in job practice have been unemployed for 50 days after they start to participate in this program. As Forslund et al. note the locking-in effect seems to be rather short. However in the long run, after the locking-in effect, the treatment effect of the program is positive (participants in job practice start to leave to work at a faster rate than the open unemployed or participants in other ALMPs). They also found that the program participation in general decreased the time duration to work by 7 % compared to the control group, given a period of 700 days. Also, in their study participants gained more from the program during the years 2003-2006 than during the other period (1999-2002) because during 1999-2002 the locking-in effect seems to play a larger role in the outcome.

Another interesting study was conducted by Rosen (2010). She surveyed 2 075 participants in the JDG’s first two phases and asked questions about their perception of the guarantee. On the question “How the participation in the guarantee influenced your chances to find work?” 49 percent answered that participation in the guarantee did not increase their chances to get a job, while 43 percent of participant believe that the participation did increase their chances to get a job. In conclusion Liljeberg et al. (2013) and Forslund & Vikström (2011a) emphasize that it is very hard to evaluate the JDG program, mainly because the guarantee cannot be compared to any other similar guarantees. Since all individuals become eligible for participation in the JDG guarantee at the same point in time, there is no other control group to compare the JDG with.

4.2 Empirical literature review on the JGY

Since the middle of 1980s there has existed several different youth programs aimed at

decreasing the unemployment duration of the young population in Sweden. A number of other programs for youth that existed before JGY (e.g. municipality youth programs and youth guarantee (Swe. Ungdomsgaranti) was evaluated by several studies and shown to be

(32)

26

ineffective (Carling & Larsson 2005, Forslund & Nordström Skans 2006). Forslund &

Nordström Skans (2006) found that municipality youth programs were worse at getting individuals to work than the PES programs for both youth and older groups. Carling &

Larsson (2005) studied the Youth Guarantee (Swe. Ungdomsgaranti) and found no overall effect of the guarantee on the flow from unemployment to employment. They found that pre- program effect was a positive one (more young people found jobs before entering the

program) perhaps because people didn’t want to participate in the Youth Guarantee. Forslund

& Vikström (2011a) call this pre-program effect as an advertisement effect, which occurs when individuals tend to change their job search behavior already before they start

participation in the program. The Youth Guarantee program was seen more as a threat than an opportunity to find work, and therefore, individuals may have spent more time on job seeking activities before entering the program (during the first 100 days of open unemployment).

However the effect during the time in the program was a negative one. An earlier study by Carling & Larsson from 2002 found that the UVG-program (utvecklingsgarantin, which was later renamed to ungdomsgaranti) didn’t include enough job seeking activities and it was not very effective in the long run. The UVG-program increased the total number of program participants making more individuals to be “locked in” into a passive job search. Municipality youth program and youth guarantee program were abolished in 2007 and instead the Job Guarantee for the Youth was introduced as a measure to fight youth unemployment. One of the latest studies on the Job Guarantee for the Youth by Hall & Liljeberg (2011) finds that during the study year 2008 employment agency measures, such as job search activities, positively influenced the outflow to work and education. Furthermore, Hall & Liljeberg also find a positive pre-program effect when the guarantee indirectly influenced the job seekers behavior before the program entrance during a different study year, 2009.

The aim of the Job Guarantee for the Youth is not only to help job seekers to find work, but also to motivate them for further studies. Young individuals who lack completed grades from Swedish upper secondary school should be given opportunity to study in the Swedish Folk High School (Folkhögskola) and receive final grades. JGY offers a time period of intensified support and in-depth analysis, study and career guidance, job search activities with coaching as well as work-related activities such as Practical training program (Praktikantprogrammet) and shorter vocational labor market training. An important factor in helping the unemployed to find jobs is to offer them good contact network with local employers such as private education companies, companies that provide practical training, municipalities and other

Impact of work-related activities on the outflow to employment and education within guarantees JDG and JGY. A study of public employment offices

Ekaterina Sukhanova