6. Investing in the employability of jobseekers in Bulgaria

Kristine Langenbucher
Judd Ormsby
Nicola Düll
(Economix Research)

Employment opportunities for the out-of-work population in Bulgaria had been improving in the years prior to the COVID-19 pandemic, however many people, especially those from vulnerable groups, remained weakly connected to the labour market. With the onset of the COVID-19 pandemic, Bulgaria’s unemployment register has grown, highlighting the importance of the National Employment Agency (NEA) in (re-)connecting people with jobs. In this process the NEA acts as a job-broker offering services to both jobseekers and employers. For harder-to-place jobseekers the NEA needs to provide more intensive services and referrals to active labour market programmes (ALMPs).

The chapter begins by following jobseekers’ journeys with the public employment service (PES). Section 6.2 describes the process of registering as unemployed, discusses how the NEA segments its clients and provides services, and details the job-search requirements that are expected of jobseekers. Finding jobseekers sustainable employment requires an employer to hire them. Moreover, filling vacancies benefits employers in addition to jobseekers. Thus, employers are important clients of the PES and the role of the NEA as a job-broker and the services the NEA provides to employers are described in Section 6.3. During longer unemployment spells especially, referrals to Active Labour Market Programmes (ALMPs) can be used to improve employment prospects while keeping up good work habits through regular programme attendance. Section 6.4 provides an overview on Bulgaria’s ALMPs and discusses them in comparison with those offered by OECD countries.

A smooth and effective registration process should minimise the administrative burden for jobseekers and the PES while gathering the information necessary to efficiently and effectively tailor services to clients. Chapter 4 discussed how the NEA and other institutions reach out to the inactive and unemployed in order to begin the process of activating them. This section looks at how the journey of the NEA’s jobseekers continues after outreach ends, beginning with registration and segmentation and then discussing activation requirements for the registered unemployed.

The registration process collects the basic information needed to start assigning clients to different services. This section describes this process in Bulgaria.

Since March 2020 and the onset of the COVID-19 lockdown, registration has been possible online as well as by post, while in-person registration has also remained possible (e.g. for clients with low digital skills or without access to digital tools). Online registration can free up front-office staff and automatically digitise information, which can then be used for other purposes, including for statistical profiling to segment clients. Some clients with lower digital skills may need in person services or support with online registration (e.g.  in Iceland benefit applications must be made online but computers with in person support are provided at the PES office (OECD, 2015[1])). However, for many jobseekers online applications may be more convenient than in person appointments as the application can be done at a time and place of their choosing. Given these benefits, even prior to the COVID-19 crisis, many countries have used online registration as the primary mode of registration. For example, Australia, Canada, Finland, France, the Netherlands and the United States. Indeed, some countries have taken this further with Iceland and Italy, requiring all registrations to be made online (Immervoll and Knotz, 2018[2]).

As discussed in Chapter 3, registering as unemployed is a prerequisite for receiving unemployment insurance benefits and social assistance benefits. Indeed, around 45% of new NEA clients in 2019 were applying for one of these benefits at the time of registration with the NEA. For unemployment insurance, payments start after registration though they are backdated to when the person lost their job (Immervoll and Knotz, 2018[2]). For social assistance, recipients must be registered as unemployed for six months before they are able to claim (see Chapter 3). For both unemployment insurance and social assistance, jobseekers must also complete separate forms with other institutions. In the case of unemployment insurance, jobseekers must submit a separate application to the National Social Security Institute (NSSI). Since March 2020, the application to the NSSI can be submitted either electronically or in person at the local labour office. For social assistance a separate application must be made to the Social Assistance Agency (SAA). Alternatively, both the social assistance application and the unemployment registration form can be completed at the Centres for Employment and Social Assistance (CESA) the joint initiative of the NEA and the SAA described in Chapter 4.

It is important for a public employment service to differentiate among clients so that they can tailor services to clients in ways that will have the most impact. Following registration the NEA segments clients into three categories that guide service provision. These categories then help determine what services are delivered to clients and when. Jobseekers are assigned to categories either at the end of the registration process or, during the meeting when the Individual Action Plan (IAP) is drawn up. Counsellors can review the grouping over time and make changes when clients’ situations change (though in practice changes are rare). The three categories are:

  1. 1. Most job ready: People with good professional characteristics and active labour market behaviour. They will typically be expected to conduct independent job search and the NEA anticipates that these jobseekers will find a job within six months of registration.

  2. 2. People further from the labour market: NEA counsellors assess these jobseekers as harder to place due to their characteristics (for example, low education, living far from work, or having health and disability issues). The NEA aims to work with these clients more actively to match them with employers or offer them subsidised employment. The NEA anticipates that jobseekers in this group will find a job within nine months of registration.

  3. 3. People furthest from the labour market and least active. People in this group are considered to be far less active and motivated to find work. Counsellors do not believe these people are ready for immediate participation in employment or education. Instead, the vision is to work with clients in this group intensively to “activate” them group and increase their motivation and support their transition into Category 1 or 2 as quickly as possible. The NEA anticipates it will take up to 12 months for people in this group to find a job.

The NEA does have an IT tool which recommends phasing to counsellors. However, the tool was designed a decade ago and only produces recommendations for the first two of the three categories. Counsellors frequently deviate from these recommendations. For example, of the caseload on 31 December 2019, around 76 000 registered unemployed had been rated by the IT tool as Category 1 but only about half of these, 37 000 registered unemployed, were actually allocated to Category 1 by counsellors. While freedom to deviate from a simplistic model can be useful, these outcomes suggest the tool could be upgraded and better integrated with counsellors’ segmentation of clients.

To help better target services many countries now use sophisticated profiling tools to help differentiate their clients (Desiere, Langenbucher and Struyven, 2019[3]; OECD, 2018[4]). Such tools combine rich data on jobseeker characteristics with predictive modelling techniques – such as basic regression or more advanced machine learning – to forecast important labour market outcomes for jobseekers such as the probability of finding work or other important labour market outcomes. These tools can help standardise services and can be most useful to counsellors who face varied clients. A PES can use these tools to provide less services to those that are likely to find a job without support and to reallocate resources to those who need them. Some PES (e.g. Ireland and the Netherlands) use profiling tools to help anticipate budgeting needs. While others (e.g. the Netherlands) use profiling tools to identify clients who can have services delivered largely online. In case of contracting-out employment services, profiling tools are also important to guide assignment to payments groups, as is done in Australia and Sweden. Bulgaria already collects rich data on its clients so developing a better profiling tool may be possible without investing in further data collection, though this too could be considered.

The services provided to jobseekers are agreed between the jobseeker and their caseworker as part of the IAP. The IAPs set goals and identify what type of support jobseekers need. This includes determining the need for individualised services such as individual or group counselling sessions, or specialised counselling including referrals to psychologist or the need to work with Roma mediators. The IAPs are monitored for progress and adjusted over time in the client’s meetings with their caseworker. As a follow up to the IAP, after 12 months of unemployment there is an “Agreement for integration into employment” (AIE), for the long-term unemployed which further specifies services and actions for the jobseeker.

The three categories of job-readiness clients are placed in help caseworkers determine service delivery and shape the IAP. In addition, the IAPs are further tailored to individual characteristics. To assist counsellors, since mid-2019, seven standardised packages of services have been designed for people who fall into different groups. They include three packages for different groups of youth, and a package each for, people over 50, people with primary or lower education, single parents, and mother with children under five. These packages act as a guide to caseworkers to decide on the types of services and ALMPs jobseekers are referred to. Caseworkers can, at their own discretion, to further tailor service to clients’ needs. Referrals to ALMPs depend on jobseekers’ individual characteristics, the barriers to employment they face, as well as, at times, on jobseeker preferences. Various ALMPs are based on age such as several trainings and employment subsidies for younger workers, and (separate) training and employment subsidies for older workers. Further ALMPs are available to support specific barriers including for those with disabilities, support moving region, and support for people with young children. Section 6.4 gives more detail on what ALMPs Bulgaria provides, to whom, and how effective they are.

For clients who have just lost their job it is important to help them begin their search as fast as possible, as the longer jobseekers are unemployed the more challenging it is to find a job (Kroft, Lange and Notowidigdo, 2013[5]). Indeed, evidence supports early and frequent meetings with jobseekers as an effective way of placing jobseekers in sustainable employment. In an experiment in Denmark, early and frequent meetings with jobseekers increased employment over the next two years by up to five weeks (Maibom, Rosholm and Svarer, 2017[6]). Positive impacts on exits to employment have also been found for France, with evidence suggesting in particular that intensive counselling can improve the quality of job matches, thereby reducing unemployment recurrence (Crépon, Dejemeppe and Gurgand, 2005[7]).

The NEA aims to meet with clients and set up an action plan within a month of registration. Follow-up meetings can be scheduled at any interval and as frequently as weekly, with the NEA aiming to meet clients at least every two months. In practice, the meetings almost always occur at one, two, and sometimes, three month intervals (Figure 6.1, Panel A).

The NEA aims to meet its clients who are furthest from the labour market (i.e. in Category 3) more frequently. However, in practice, the NEA actually meets these clients less frequently than clients in Category 1 and Category 2 (Figure 6.1, Panel B). This result is driven in part by the NEA’s policy of meeting young jobseekers under 30 years at least monthly. Many young persons are classified as Category 1, which drives down the average time between the first and second meeting for Category 1 clients to a little over 40 days (Figure 6.1, Panel B). To increase the frequency of meetings with Category 3 clients, the NEA should consider setting a more frequent minimum meeting schedule for Category 3 clients, as it does with youth. This, however, is likely to require additional counsellor resources.

In general, not all clients will need frequent meetings with the NEA, and not all jobseekers will need a detailed or promptly prepared action plans. Some jobseekers will be able to effectively manage their job search independently. Identifying these jobseekers is thus important for freeing up resources for those who need them most, and as mentioned above, more sophisticated use of profiling tools may be of use in assisting with these decisions. For example, youth under 30 are met very frequently by the NEA. However even though this may be helpful for many youth, it could be that not all youth need this extra support as much as some, older, less-job ready clients who currently receive less support. A sophisticated statistical profiling tool could thus assist in such decisions.

The most important input into PES activities is its people. The NEA has around 2 900 staff. Of these about 1 350 are involved in providing labour mediation services. Staff satisfaction is considered high with around 80% of employees feeling personal satisfaction with their work according to a 2019 staff survey. About 38% of NEA staff are on temporary contract many of whom work on ESF financed projects where the funding is not permanent. The majority of NEA staff are highly skilled and have a bachelor’s or master’s degree.

In August 2020, the government granted a one-time permanent pay increase for NEA staff of 29.7% on average. This increase appears to have been mostly uniform across the organisation. The rational was to address a pay gap between the NEA and other state agencies, and also to reward front-line workers in the face of the COVID-19 crisis and was given in the context of a government review of pay for 24 state agencies. This substantial pay-rise will make the NEA a more attractive place to work, which may lower staff turnover (currently around 12%).

In 2019, during the July – September period, the NEA had around 142 jobseekers per caseworker. This calculation comes from an average of 172 689 jobseekers served by about 1 213 staff over this period (though this figure includes some employees who also provide services to employers as well as jobseekers).1 As the number of jobseekers rose in the wake of the COVID-19 crisis caseloads increased to around 217 per caseworker during the same period in 2020.

Lowering caseloads below these levels would allow for more personalised services (for example, increasing the frequency of meetings discussed above). Indeed, international evidence suggests improving jobseekers to case worker ratios below the levels seen in Bulgaria can result in improved labour market outcomes and to net financial savings. In an experiment in Austria, a local PES office in Vienna temporarily increased staff during the year 2015 to lower the client-to-staff to 100:1 instead of the usual 250:1 (Böheim, Eppel and Mahringer, 2017[8]). Similarly, an experiment in the German PES between 2007 and 2010 involved the hiring of additional caseworkers in 14 local PES offices to lower the client-to-staff ratio to an average of 70:1 (from the usual 80:1 to 250:1) to improve the quality of placement services (Hainmueller et al., 2016[9]).

The international evidence makes a strong case that hiring more staff to decrease workloads of caseworkers can achieve improved outcomes and reduce net costs to the government. Indeed, in both the Austrian and German experiments, the costs of hiring additional caseworkers was offset by decreased benefit expenditure within one year or less.2 Beyond hiring additional PES staff to complement existing PES services or address additional staff needs during periods of high unemployment, some countries contract out employment services to external service providers. Expanding PES capacity temporarily, without long-term commitments, can be achieved by contracting out employment services to the private sector. Two in five of the countries covered by the OECD-EU survey already contract out employment services to external parties, including both to for-profit and not-for-profit entities (OECD, 2021[10]). A number of countries foresee expanding the use of contracted out services in the near future, also to address higher levels of unemployment resulting from the COVID-19 crisis (see Box 6.1). Hiring of additional PES staff as well as contracted-out employment services, however, require large additional upfront financing, which may not be available, even so both measures are regarded as “invest-to-save” measures. An alternative approach is to free up the time of existing staff members with increased use of automation and digital services, which is discussed in the next sub-section.

Many PES have increased their service offering in recent years (Figure 6.2). Bulgaria is no exception but it still has room to go further in this direction. During the COVID-19 crisis the NEA was quick to roll out e-services. It is now possible for jobseekers to perform basic administrative tasks such as registration and deregistration online which many clients now prefer. IAPs can also now be drawn up remotely, whereas prior to the COVID-19 crisis these were all done in person. An increasing digital set up is important and emphasising online employment registration and services should continue beyond the COVID-19 crisis. Remote digital service provision, as with online registration discussed above, can offer more flexibility to jobseekers, can reduces time spent commuting to the employment service and waiting in ques, and, once the investments in setting up digital services have been made, may free up NEA staff time for other tasks.

The NEA also maintains an e-labour office (which existed prior to the crisis) that allows jobseekers to view job vacancies through a basic online-portal. Likewise, jobseeker’s basic profiles are included on the website and employers can search these to find prospective candidates. This website, however, could be modernised. For example, further functionality could also be added to the website to better match jobseekers and employers. Box 6.2 shows examples of such functionality used by the Public Employment Service (PES) in Flanders (Belgium).

An important limitation of the e-labour office, is that when jobseekers find a vacancy they are interest in, or employers find a potential candidate, they must contact the NEA to intermediate. This process takes up additional time for employers, jobseekers and NEA staff. Automating this process and allowing employers and jobseekers to directly contact each other could potentially save time for all involved. Indeed other countries (e.g. Austria, Germany, the Netherlands and Sweden) have fully open vacancy databases that offer all the information jobseekers need to apply for a position, in case they are registered within the database. While the NEA previously had concerns about data privacy, now has plans to modernise this aspect of its website and intends to allow such functionality in the future.

One example of the NEA taking the initiative with introducing more online and digital services is the “MyCompetence” website and its “personal profile of a jobseeker” tool (MyCompetence, 2019[12]). The “MyCompetence” website provides access to e-learning, allows jobseekers to view their competencies using the “personal profile of a jobseeker” online module, and provides information on the competencies required for various positions. The “personal profile of a jobseeker” application shows jobseekers the current data the NEA holds (e.g. on their skills, education, experience, and action plan) so that they can update it online. The e-learning courses offered on the “MyCompetence” website are wide ranging including courses on time management and business etiquette, conflict management, the use of decision-making tools, and digital competencies, as well as courses that support managers in hiring or leading. These resources are provided to jobseekers for free. The website’s development is also an example of co-operation across the public and private sector with the system developed in co-operation with the Bulgarian Industrial Association (BIA) and developed over several years with support from the European Social Fund (MyCompetence, 2019[12]). The NEA has promoted this tool, particularly during the COVID-19 pandemic and nearly 50 000 profiles had been created by early 2021(Figure 6.3). Statistics supplied by the NEA show that in 2020 an average of 74 new profiles were created per day and an average of 39 “development measures” (e.g. questionnaires, trainings) were planned each day. Profiles can continue to be used and updated after registration at the labour office ends.

During COVID-19, counselling sessions could be undertaken remotely and remote counselling was supported by the “MyCompetence” platform. Bulgaria should consider retaining this as an option for some customers where this is suitable post-COVID-19. Indeed, many countries now seek to provide their most job-ready clients with entirely online experiences. Already in 2014, the Dutch PES was using e-services as the primary means of servicing all clients in the first three months of registration, with almost all unemployment benefit recipients registering online and 85% using ongoing e-services to manage their benefit claim and automatically match with vacancies (European Commission, 2014[13]). As digital services are not appropriate for everyone, the Dutch PES now uses a statistical profiling tool to help segment clients into those that will initially use e-services and those that need in-person assistance.

While digital services can be convenient for certain jobseekers and save the PES time, they are not suitable for all jobseekers. In Bulgaria, just 3.5% of the registered unemployed on 31 December 2019 subscribed to e-mail notifications from the NEA. There can be many reasons for lack of uptake of digital services, including people’s awareness of the services, their willingness to use them, their trust in sharing data, and their access to appropriate equipment. One concern when providing digital services is people’s computer skills. This is particularly relevant in Bulgaria which has the lowest levels of digital literacy in the EU (Figure 6.4).

In order to mitigate the disincentive effects of benefits, countries impose eligibility requirements on benefit receipt that aim to activate jobseekers. This involves requiring jobseekers to be available for training and work and to take steps to search for a job. To help compare the stringency of eligibility requirements the OECD collects detailed information about each countries requirements (Immervoll, Knotz and Otmani, 2020[14]).3 The results are available on the OECD’s website (OECD, 2021[15]). Each countries’ rules are examined and coded on a 1-5 stringency score (with 5 as most strict). The OECD work assesses stringency across three categories:

  • Availability requirements: The reasons people are allowed to give for turning down jobs and refusing ALMP participation, including how mobile jobseekers are expected to be across place and occupation.

  • Job-search requirements: How frequently jobseekers must provide evidence of job-search and how well-documented this evidence must be.

  • Sanctions: How severe sanctions are, including for refusing valid job offers and ALMPs as well as for voluntary resignation.

Tougher activity-related eligibility requirements can speed up transitions back to work (Abbring, Berg and Ours, 2005[16]; Lalive, van Ours and Zweimüller, 2005[17]; van den Berg and van der Klaauw, 2006[18]) but policy makers must balance this with equity considerations (Immervoll, Knotz and Otmani, 2020[14]). Indeed, some studies find sanctioning can lead to lower quality employment such as less stable and lower paid job-matches (Arni, Lalive and Van Ours, 2012[19]) or higher take up of part-time instead of full-time jobs (van den Berg and Vikström, 2014[20]), though overall evidence on the effects of eligibility requirements on job quality is mixed (Tatsiramos and van Ours, 2012[21]; Le Barbanchon, 2016[22]). However, if jobseekers do end up in lower quality employment this could potentially contribute to skill erosion.

Overall, Bulgaria’s activation requirements are relatively light – ranking 12th most lenient out of 39 OECD and EU countries covered (Figure 6.5). Like several other eastern European countries, Bulgaria opts for an approach that combines strict sanctions on unemployment insurance recipients with lenient job-search and availability requirements. The rest of this section examines in more detail Bulgaria’s activation requirements.

To allow candidates to take time to find not just any job, but a good job, many countries allow candidates to decline “unsuitable” jobs without loss of benefit. Suitability can take into account jobseekers education, skill and previous occupations as well as how far away the job is. Jobseekers are usually expected to consider a broader range of jobs the longer they are unemployed.

In Bulgaria, jobseekers are allowed to decline jobs if they do not match their education, qualifications or profession and experience during the first 12 months of registration. After this period, these are no longer valid reasons to refuse a job. Regardless of registration length jobseekers are compelled to accept a job in the same locality or within 50 kilometres of it provided there is adequate transportation, but jobseekers are not compelled to accept job offers that are further afield. Similarly, job seekers can always turn down jobs that are unsuitable given their health. Further when participating in ALMPs there are no requirements to be available for jobs.

In an international context these rules are relatively lenient scoring only a two out of five for each of occupational mobility and geographic mobility and a maximally lenient one out of five for ALMP participation (Figure 6.6, Panel B). Out of 35 OECD and EU countries for which there is data in the OECD’s 2020 survey, 30 require jobseekers on tier one unemployment benefits to be available for work when participating in at least some ALMPs while 11 countries report that jobseekers must be available for and actively searching for work when participating in any ALMP (OECD, 2021[15]). In terms of occupational mobility, examples of more strict approaches include countries that only allow jobseekers to turn down jobs that do not match their experience and qualifications for a shorter period, e.g. six months or, or even to never allow such protections as is the case for Australia, Denmark, Hungary, Japan, New Zealand and Poland (OECD, 2021[15]).

Interestingly, Bulgaria allows the same protections relating to education and qualifications for jobseekers on social assistance as it does for unemployment insurance. Social assistance benefits are targeted more toward the long-term unemployed who have exhausted any unemployment insurance they may qualify for. Some other countries require those on lower tier social assistance benefits to be available for a wider range of jobs than those on unemployment insurance – for example Austria, Canada, Italy, Japan and the United Kingdom operate stricter overall activation criteria for lower tier compared to first tier unemployment benefits (Immervoll and Knotz, 2018[2]).

One area of particular interest would be geographic job mobility requirements. One option would be to make geographic mobility requirements stricter, however this may be difficult to enact in practice as moving far from existing networks and family may not be feasible for all. Perhaps for this reason only two countries in the OECD’s 2020 survey (Croatia and Korea) report requiring participants to move for a job, though three further countries (Canada, Japan and Malta) are assessed as requiring commutes of more than four hours, and 16 countries report requiring commutes of up to four hours (OECD, 2021[15]). As discussed in Chapters 1 and 2, there are large differences in labour market opportunities within and across regions in Bulgaria. Hence, improving incentives may have the potential to help many. Indeed, the OECD has previously identified high-levels of home ownership in Bulgaria as a factor that may contribute to Bulgaria’s low levels of residential mobility (OECD, 2021[23]). Instead of tightening eligibility rules, subsidies that support moving are an alternative (or additional) way to change incentives. Bulgaria does offer two programmes that provide subsidies to support moving. The first subsidises only transport costs while the second provides broader support including for child care costs and even rent for people taking a job outside of 50km. Both these programmes however only last for 12 months and have very few participants numbers (see Section 6.4). Such programmes are important given the Bulgarian context. Bulgaria should consider how well these programmes incentivise longer-term moves and assess whether or not there is a case for a more generous one-time subsidy conditional on a permanent move to a different location.

As discussed above, meetings with the NEA are not set very frequently and meeting frequencies vary by person. Further, jobseekers need only provide verbal information about their job seeking activities outside of the employment agency (OECD, 2021[15]). Of the same 39 countries in Figure 6.5 24 countries (62%) are assessed as having stricter documentation requirements and 16 (41%) are assessed as performing more frequent checks (OECD, 2021[15]).4 In some cases these checks of documented job-search are as frequent as every four weeks or less (e.g. Malta or the United Kingdom) but in most of the cases assessed as having more frequent checks than Bulgaria the frequency is every nine weeks or less. Examples of more strict documentation include requirements that jobseekers specify the specific actions they took to find work, to specify the names and addresses of employers jobseekers have applied to, or even, in the most strict case of Malta, to regularly produce declarations from an employer stating that they have applied for work. To perform more frequent checks with stricter documentation, while minimising any increased demands on staff, the NEA could consider greater use of online tools to monitor jobseekers such as the digital submissions of evidence of job search.

Bulgaria’s sanctions for failing to participate in an ALMP or for turning down a suitable job are among the harshest in the OECD and EU. Rejecting a suitable job or ALMP resulting typically results in a six month removal from the unemployment register and associated loss of benefit entitlement and in the case of being dismissed from subsidised employment it is 12 months. Many other countries opt for a less strict approach, especially for first refusals.

A tough sanction approach provides strong incentives for jobseekers to comply with rules. Nevertheless, there are drawbacks to this approach. First, harsh sanctions reduce income to those in need. This reduces the role benefits play in abating income inequality and can induce hardship. One study in the United Kingdom using a “fixed-effects” or “difference-in-difference” methodology, found an association whereby local authorities that had relatively higher increases in sanctions over time suffered increased mental health problems (as measured by increased anti-depressant prescriptions) compared to local authorities with relatively lower increases in sanctions over time (Williams, 2021[24]). Another study in the United Kingdom, with a similar methodology, found higher sanctions were associated with higher food insecurity (Loopstra et al., 2018[25]).

Second, due to the magnitude of the sanction counsellors may exercise discretion in actually applying them especially for more minor setbacks. It is difficult to tell the extent this occurs in Bulgaria. Of jobseekers that were registered as unemployed at the end of 2019, about 11% left the register via a sanction by February 2021.5 While international comparisons are rare, this does not appear to be a low number. Gray (2003[26]) calculated the number of sanctions per year and divided these by the average stock of unemployment beneficiaries for a sample of 14 OECD countries in the 1990s. Gray found the median was about 7%. For another recent data point using a similar methodology, annual sanctions rates in Slovenia in 2015 were around 6% in 2015 (OECD, 2016[27]).

Third, sanctions in Bulgaria not only result in the loss of benefit entitlement they also imply exclusion from all PES services. These disengaged clients thus need to manage their job search independently and without the support of the PES. This includes clients registered with the PES who are not on benefits as they too can be sanctioned. Other countries, including Luxembourg, Latvia and Greece, still allow sanctioned benefit recipients to access at least some services – sometimes with a delay, but one shorter than the full sanction period for their benefit (OECD, 2016[27]).

Finally, there is some evidence that sanctions may be less effective than stricter availability or job-search requirements. Knotz (2020[28]), analysing cross-country panel data, finds that both more strict job-search and more strict availability requirements have a statically controlled positive association with the employment rate but that stricter sanctioning (on its own) does not. Unfortunately, such studies are not definitive as it is difficult to be certain that these statistically controlled associations are causal.

To encourage job-to-job transitions and discourage unneeded government spending on unemployment benefits, most OECD and EU countries impose sanctions on claiming unemployment insurance for those who leave their jobs voluntarily. Bulgaria, however, is rated as one of the least strict countries in the OECD and the EU for sanctioning of people who voluntarily resign (Figure 6.6, Panel A). This is because Bulgaria imposes no up-front restrictions on people who voluntarily leave their job. Bulgaria instead places sanction on voluntary resignation – if at all – at the end of the unemployment insurance period: with the maximum unemployment insurance period reduced to four months for all people who voluntarily quit their job.

This is very unusual internationally. Most countries, either, do not pay unemployment insurance at all to people who voluntarily quit their job (which is also very harsh), or they impose up-front sanctions. Countries may opt for this up-front deterrent because people tend to place more weight on costs in the near term than those in the future. To the extent this is true, it means that sanction periods at the beginning of the unemployment spell will act as a greater deterrent to voluntary resignations, than a sanction at the end. In theory, this could potentially allow the government to use smaller up front sanctions compared to larger end-of-period sanctions in order to achieve the same deterrent effect thus minimising the impact of sanctioning on poverty. Bulgaria could therefore consider reforming the way it sanctions voluntary quits. However, in addition to moving sanctions from the end to the beginning of the unemployment spell, Bulgaria also should make sure that people with entitlement to unemployment insurance who voluntarily end employment (excepting those with valid reasons) face at least some sanction. At present Bulgaria’s sanction for voluntary quits only affects those with more than four months unemployment insurance. This means those who have contributions of less than three years (potentially many youth) face no disincentive for quitting.

Finding jobseekers suitable employment requires matching jobseekers with employers who will hire them. Thus, in order to achieve higher living standards for PES clients through greater labour market participation, it is important to work with jobseekers – the labour supply side – and with employers – the labour demand side. Indeed, both jobseekers and employers are PES clients and the PES provides useful services to both groups. Whereas the previous section focussed on the NEA’s work with jobseekers, this section details the work the NEA does with employers.

The NEA works with employers in a number of ways. These include, listing vacancies on the NEA’s online job-board – the e-labour office described in the previous section; referring candidates to employers; offering employment mediation; and the organising of job-fairs. On its website, the NEA publishes information for employers on issues such as support for dismissed employees to register with the employment agency; employers’ obligations for mass layoffs (including the timing and structure of consultations); information on the COVID-19 job-retention scheme; and information on hiring foreign workers. The provision of many ALMPs including employment subsidies and trainings involve further interactions with employers including both providing information on what is available in addition to working with employers during the implementation of these programmes.

As with PES jobseekers clients, perhaps the first step in working well with employers involves outreach activities. As part of its outreach strategy, the NEA has teams that identify and work with important large employers in each region. At the other end of the spectrum, to reach out to employers in smaller settlements and more remote areas the NEA uses its mobile labour offices (see Chapter 4).

To deliver on its work with employers, particularly in gathering vacancies to share internally and list on its e-labour website, the NEA has staff who specialise in working with employers. While all staff are expected to, when needed, work with clients from both the employer and jobseeker side, Bulgaria has, 60 counsellors who specialise in working with employers compared with about 1 170 counsellors in specialising in working with jobseekers. International comparisons of the number of PES staff working with employers are hard to come by, but, as one data point, Slovenia, which has high levels of engagement with employers, has a much larger share of employer counsellors than Bulgaria: with about one-in-five counsellors at the Employment Service of Slovenia (ESS) specialising in working with employers or about 80 in total (see Box 6.3).

The NEA’s investments in working with employers appear to be well received. As part of this review, the OECD conducted semi-structured interviews with a small (and not representative) sample of employers who work with the NEA. These employers gave largely positive responses regarding their interactions with the NEA (see Box 6.4). In the NEA’s own surveys of its clients it receives high and growing rates of satisfaction from both employers and jobseekers (Figure 6.7, Panel A). Further the NEA manages to match more than 80% of job vacancies it receives from employers (Figure 6.7, Panel B). The challenge for Bulgaria going forward will be to maintain these high levels of satisfaction from employers while matching more hard-to-place and disadvantaged jobseekers with vacancies.

Pre-pandemic economic growth and increased labour demand have enhanced the chances for unemployed and inactive persons further away from the labour market to find employment. However, for the most vulnerable jobseekers opportunities have only slightly increased. For those with longer spells of unemployment, referrals to active labour market programmes (ALMPs) can help improve employment prospects and keep up work habits through regular participation in programmes. This section first compares Bulgaria’s overall investments into ALMPs with those of other EU and OECD countries. After that, it zooms into the different programmes offered by the Bulgarian public employment service (PES) and concludes with reviewing the targeting of ALMPs to vulnerable groups.

In 2019, spending on ALMPs as a share of GDP was 0.16% in Bulgaria, less than half the EU-average of 0.39% (Figure 6.8, Panel A). Spending on ALMPs fell after the global financial crisis in 2009 and 2010. It increased afterwards in a context of high unemployment, reaching a peak in 2013, and decreased again as unemployment fell. Likewise, spending on active labour market measures6 per unemployed as a share of GDP per capita was considerably below OECD and EU averages in 2019 (Figure 6.8, Panel B), ranking as the sixth lowest EU country. Moreover, Bulgaria’s expenditure on active labour market measures has substantially fallen since the mid-2000s, when Bulgaria ranked about mid-field in a European wide comparison (OECD, 2019[31]). In addition, when considering the number of participants in active labour market measures, Bulgaria ranks relatively low. Participation in active measures per 100 persons wanting to work (LFS concept) was one of the lowest in the EU (Figure 6.8, Panel C).

The National Employment Action Plan (NEAP) defines the objectives, priority measures and target groups of active labour market programmes (ALMPs) financed both through the national budget and EU sources. The state budget earmarked in the NEAP (see Box 6.5) amounted to BGN 73 million (EUR 37.4 million) in each year between 2017 and 2020 and increased to BGN 83 million in 2021 (EUR 4.3 million; (Ministry of Labour and Social Policy, 2021[32])). EU funding for ALMPs through the Human Resource Development Operational Programme (HRD OP) of the European Social Fund (ESF), the Youth Employment Initiative (YEI), and the Youth Guarantee plays an important role. While funding through the state budget has been rather stable, EU funding showed stronger variations.7

Direct job creation has been the main type of ALMPs in Bulgaria for many years (except in 2016) (see Figure 6.9, Panel B). Since its peak in 2013, spending on job creation programmes has significantly fallen and has been partly replaced with somewhat higher spending on training and employment incentives. Nevertheless, it was still above the spending on training and employment incentives measures from 2017 to 2019. The share of ALMP spending used for job creation measures was largely above the EU average in 2019 (Figure 6.9, Panel A).

Conversely, spending on employment incentives and training measures has increased since 2015, but remains at low levels, despite the NEAP objective to improve the skills of jobseekers and reduce skills mismatches. In particular, spending on training measures is very low in Bulgaria, at only 8% of total spending on active labour market measures (categories 2-7), against 40% on average in the EU (Figure 6.9, Panel A).

A number of job creation measures are part of mixed measures, which combine subsidies for temporary, non-market jobs with training.8 This is in principle the right approach, as it helps to link training to the demand for skills and has the potential to improve the effectiveness of job creation measures. However, in practice, the effectiveness of the measures depends on concrete implementation conditions such as the quality of employment and the training received.

The very low share of participants in supported employment and vocational rehabilitation in Bulgaria as compared to the EU average is due to the fact that the larger national programmes for employment and training (mainly categories 2 and 6 of the LMP database) also target people with disabilities. In contrast to OECD and EU good practices, except employment incentives, there are no specific vocational rehabilitation programmes managed by the PES addressing specific skills needs of people with disabilities and managing up-skilling and re-skilling, measures for adapting workplaces and sheltered workshops for people with severe disabilities. Some specialised programmes are organised by the Agency for People with Disabilities, including through the financing of projects of people with disabilities wishing to open a business and the financing of social projects aiming to improve the working conditions of workers with disabilities (MLSP, 2021[33]). In many EU countries, PES co-operate with other institutions in charge of people with disabilities and provide vocational rehabilitation measures (see examples in Box 6.7).

Other EU countries spent on average 6% of the active expenditure on start-up incentives, while in Bulgaria these are very small programmes in Bulgaria; a new programme has, however, been introduced in 2020.

Bulgaria’s HRD OP aims to boost employment and reduce social exclusion as well as reduce poverty levels. ALMPs that benefit from EU funding are mainly direct job creation programmes as well as programmes for young people, including NEETs. Other target groups of HRD OP include the long-term unemployed, older people, people with disabilities and minority groups such as those belonging to Roma communities (EC, 2020[34]). Annex Table 6.A.1 provides details on expenditure and participants of Bulgaria’s programmes, while Annex Box 6.A.1 provides a short description of the seven largest state-funded programmes.9 The majority of programmes have very small budgets and numbers of participants (see Annex Table 6.A.1). Most of them have been maintained for a long period of time. The 13 largest ALMPs out of a list of 54 ALMPs, absorbed 86% of funding and 88% of participants in 2019. Running small programmes involves a high administrative burden per participant and the availability of small programmes may not be known by jobseekers and employers. Therefore, it might be beneficial to consider streamlining the programmes to improve their management and increase the take-up of those that have proved to be successful. Some of the small programmes are nevertheless relevant, in particular the ones that address the needs of jobseekers belonging to vulnerable groups and facing specific labour market barriers (see Chapter 2). In these cases, it would be more effective to raise the number of participants and improve their design.

NEAPs have underlined the need to develop skills of the workforce to meet the employers’ needs. This objective does respond to the challenge of tackling skills mismatch and overcoming labour shortages (see Chapter 1). Moreover, the main priorities of the Human Resources Development Programme 2021-27 include, reforming the lifelong learning system, supporting the acquiring of digital competences and the development of skills including the sustainable (“green”) skills, encouraging adaptation capacity, improving working conditions, and achieving a work-life balance for the workforce.

In principle, the content of the training provided by the NEA is embedded in the vocational education and training (VET) system of Bulgaria.10 However, while training has a high level of quality, many of the NEA’s clients have low levels of skills or obsolete skills. In addition, a poor quality of vocational training received by those entering the labour market, poor supply of lifelong learning, as well as a lack of literacy and other basic skills courses for people with low levels of education are severe weaknesses of the education system that the NEA needs to cope with (European Parliament, 2017[38]).

In 2016, 79% of participants in training measures received such training in combination with subsidised employment in the context of direct job creation measures. In this case, participants attend a training measure before they are placed in subsidised employment, which, however, are non-market jobs.

In addition to training programmes for the unemployed, the NEA also offers training to upskill workers in employment through the Programme “Training Voucher for Employed”. The budget was, however, fairly low in 2019 (see Annex Table 6.A.1) and the measure has been suspended since the start of the COVID-19 pandemic. There is a co-payment of 15% by the employee to ensure selected participants are motivated to take and finish the training courses. Since the start of the programme in 2017, the number of applications was more than twice the number of vouchers available (25 946), reflecting strong demand for this support. 67% of the vouchers were issued for obtaining a vocational qualification or improving vocational skills, 23% for training communication in a foreign language and 8% for digital skills.

While training programmes include low-qualified and other disadvantaged jobseekers among the target groups, the majority of training participants typically have at least mid-level education and few have a low educational level only (10% in 2020). Training programmes that are in particular targeting low-qualified exist, but are small-scale. One of them is offering low-skilled unemployed people the possibility to get vocational training through the dual training system under the Vocational Education and Training Act. The expenditure for the programme combines a training subsidy for the employer and funding for the training institution and for mentors. However, in 2018 the programme had only 14 participants and 35 in 2019. Another small, but promising scheme called “Consultation and mentoring after starting work”, which has recently been introduced, offers follow-up services for younger recent training programme participants. It provides support to both employers and employees to adapt to the workplace and achieve sustainable employment. The new initiative has received positive feedback from all stakeholders involved, as it supports the early identification of potential risks in the new workplace and supports the newly employed youth to gain stability in their new workplace.11

Given the overall low expenditure on training programmes in Bulgaria, there are limited opportunities to offer training to NEA clients. Nevertheless, while the number of training seats is low, the range of training programmes covers many different types and levels of training, in small training programmes. International evidence suggests that additional expenditure on training programmes can produce positive outcomes, especially in the medium to long run and should be part of an activation strategy supporting more inclusive and resilient labour markets (OECD, 2015[1]). Hence, there is a business case for investing more into training programmes. If additional investments are not feasible, Bulgaria should consider streamlining its basket of training programmes to have a limited number of training programmes and increase the focus on basic skills training. It is also important to ensure that information on available training is clear and easily available, that jobseekers receive guidance to choose suitable training (OECD, 2021[39]) and that training meets employers’ needs and addresses labour market shortages (see Chapter 1).To address the latter point, social partners’ trainings are preceded by a preliminary study of labour market needs, distinguishing by profession and region. Against the background of limited funding, some form of employers’ funded training is important, as is e.g. done by the Austrian “Implacement Foundations” (see Box 6.7).

Going forward, it would be important that one major programme could focus on basic skills training for low-skilled jobseekers only. Bulgaria currently has no major programme for illiterate jobseekers, whereas other OECD countries – e.g. Australia, Finland, the Netherlands, Ireland, and France (Windisch, 2015[40]) – have implemented measures for providing basic literacy and numeracy skills for this group of jobseekers.

Another challenge in Bulgaria is accessibility of training for highly disadvantaged groups in remote areas. Interviews conducted by the OECD team in the Montana region, a rural area with high unemployment, point to the problem that VET training centres are often located at great distance to potential training participants, even though in case of mass dismissals, the Bulgarian-German Vocational Training Centre State Enterprise may sometimes be able to organise trainings in remote settlements where the unemployed live. This reduces the opportunities for jobseekers to participate in this type of training. Moreover some interviewees highlighted that the support for transportation costs offered by the NEA is insufficient for trainees from remote settlements. To address this issue, supplements for transportation as well as board and lodging should be sufficiently high the cover the costs participants incur.

Employment incentives provide subsidies to employers for “open market jobs” for wages and social security contributions. If time-limited, well-designed and targeted, such employment incentives can be a cost-effective way to support jobseekers back into employment and strengthening their employability (Kluve, 2010[41]; Brown, 2015[42]). In Bulgaria, the Employment Promotion Act (EPA) provides the possibility to grant subsidies to support wages and social security contributions for specific target groups. The level of the subsidy is determined in the National Employment Action Plan. Usually it equals the minimum wage level, but it is higher for unemployed people with higher education. After the amendment of the Act on Employment Promotion in 2016, two state aid schemes were implemented. One of the schemes grants subsidies to support 50% of wages and social contributions for the employment of unemployed persons who meet specific characteristics. In particular, they are available for unemployed people with a continuous registration with the NEA for at least six months, for unemployed under 24 years old, for unemployed over 50 and for unemployed with lower secondary education or below. The second scheme subsidises 75% of the wage costs for employed people with permanent disabilities. (Ministry of Labour and Social Policy, 2019[43]).

The largest employment incentives programme in terms of the number of participants is the programme “Parents in employment” (see Annex Table 6.A.1), funded under HRD OP, which started in 2017 and replaced previous programmes which offered comparable support. Employers get wage costs subsidies for a period of 18 months and pays for the childcare support of the participants. It targets unemployed and inactive people aged below 29 years with children aged up to five years and not attending nurseries, kindergarten or pre-school.12 The babysitter chosen by the parents is appointed under an employment contract for a maximum period of 18 months or for the period up to five years until the beginning of pre-school education of the child / children. The project is foreseen to last until 2023. Since its start in September 2017, 7 801 employment contracts have been concluded with babysitters.

In terms of budget spent in 2019, the “Youth Employment Scheme”, implemented under the Youth Employment Initiative, is the largest employment incentive programme (see Annex 6.A).13 It is targeted at young people below the age of 29 outside the capital Sofia, who have no income from other economic activities and are not enrolled in education. The central element of the programme are internships and on-the-job training (the latter are classified as training programmes). A requirement for internships is that the young people have completed secondary or higher school and are lacking work experience.14

Another employment incentive programme – “Training and employment of disabled persons” – is targeted at registered unemployed people with permanent disabilities15 or jobseekers who have successfully passed a course of treatment for drug addiction. The employer is required to offer a job for a period of 24 months. Employment incentives are also used to promote the employment of older workers. The programme “Support in retirement”, provides wage subsidies for employing older unemployed aged 58 and over in order to bridge the time to early retirement. Employers are required to offer employment contracts for a period of not less than three months and not more than 24 months. In parallel a smaller programme for older workers (“Incentives for employer to hire older unemployed”) provides wage cost subsidies for older workers aged 55 years and above for a period of 12 months.

As in the case of training programmes, merging different employment incentives programmes targeted at overlapping groups of registered unemployed could be considered for further streamlining Bulgaria’s ALMPs. This was manifested, on the one hand, in the large amount of documents that employers need to prepare both when applying under a certain programme and during the monthly reporting submission, and, on the other hand, in the delay of approving recruited job candidates “labour contracts” by the NEA central office. However, stakeholders also raised the importance of using employment incentives for hiring older workers, which are seen to produce positive outcomes also after the end of the subsidised period, which, however, sometimes was assessed as too short. Given the low hiring rates of older workers in comparison to their younger peers and the demographic pressure, additional efforts are needed to retain older workers in the work force and re-connect them quickly with the labour market, once they become unemployed.

In addition to the before mentioned larger schemes, there are some smaller scale employment incentive programmes. One employment incentive called, “Subsidies for new jobs in micro-enterprises”, is targeted at micro-enterprises. In this case employment incentives can be received for a period of 18 months, for the first five employees the micro-enterprise hires. Another small-scale programme is targeted at promoting employment in green jobs (Duell, Anghel and Ziminiene, 2021[44]). These programmes are relevant, but have probably little impact on the stabilisation and growth of micro enterprises and the greening of the economy, as they are too small.

Furthermore, the NEA also runs a small-scale programme to support mobility for the unemployed, which has been in place for a number of years. In 2019, only 134 unemployed benefitted from the measure.16 Support for commuting and relocation are in place in a number of OECD countries. Evaluations have shown that they tend to reduce reservation wages and increase the regional radius for job-search (Guglielminetti et al., 2010[45]). A recent evaluation of measures supporting distant job-search activities, commuting and relocation in Germany has shown that the existence of these measures shifts individuals’ search effort from local to distant regions without affecting the total number of job applications. The increase in search radius causes a higher geographical mobility and hence higher employment probabilities and wages (Caliendo, Künn and Mahlstedt, 2017[46]). A recent evaluation of Latvia’s regional mobility programme which offers support with taking up distant job offers (at least 20 km from the current residence) or with attending distant training measures, by reimbursing costs for transport or housing, has found positive effects on job-related mobility of unemployed (OECD, 2019[47]).

Across OECD countries, direct job creation programmes are targeted at disadvantaged groups. These jobs are created under the assumption that the jobseekers would not find employment in the regular labour market. The effectiveness of direct job creation programmes in bringing participants back to open market jobs is questionable.A meta analysis of evaluations of ALMPs by Card et al. (2018[48]) finds that these direct job creations programmes are generally ineffective in the short, medium and longer term. A number of OECD countries therefore do not use these programmes anymore, e.g. Denmark, Estonia, Israel, Norway and Switzerland. Over the past decade, many other countries have shifted spending from these programmes towards ALMPs which they deem more effective, such as training and employment incentives (OECD, 2021[10]). In contrast, Bulgaria still spends two-thirds of its active labour market programme expenditure on direct job creation (Figure 6.9).

One of the largest direct job creation programmes is the employment component of the “Training and employment Scheme – Providing employment/training to unemployed people -project under HRD OP”, targeted at inactive and registered unemployed persons over the age of 29 (over 30 years of age) and people with disabilities. Subsidies under this scheme can be received by private employers or institutions of local government. Training should be provided according to the needs of employers. The programme “Jobs in Public administration for youth (<30)” is targeted at young people under the age of 29 without relevant work experience and who have graduated from higher education. Another comparatively large programme is “Training and employment for LTU – Providing employment to unemployed people”. Target groups of this programme also includes young unemployed, older unemployed (50+) and people receiving social assistance17 (see Annex Table 6.A.1 for participant stock figures and expenditure of the programmes). The effectiveness of direct job creation measures depends on the way they are implemented. Key success factors are the type of work that is carried out and the counselling and coaching of programme participants with the objective to make them ready for the regular labour market after the end of the programme. For example, the French pilot project “local areas with zero long-term unemployment” (Territoires zéro chômeur de longue durée), which has been implemented in small municipalities in rural areas and economically weak urban areas, puts a strong focus on identifying and defining the “useful activities” that are carried out by subsidised employment. Such “useful activities” must address an unmet need, may not be in direct competition with the private market and have to take account of the competencies of participants (TZCLD, 2021[49]). To this end, local committees are set up, which consist of a variety of key actors at the local labour market. Intermediary evaluations suggest that the programme is effective in supporting jobseekers to take up employment (see Chapter 4). In Austria, social enterprises offer subsidised fixed-term “transition jobs” to vulnerable groups, as well as targeted skills training and a holistic care and support package. Each participant’s time with a social enterprise is structured in the following phases: preparatory phase, introduction phase, training and employment phase, job seeking phase, concluding phase and follow-up (AMS, 2018[50]). In Bulgaria, impact evaluations of the programmes and measures which are funded by the State budget (thus excluding many direct job creation measures funded through HRD OP) were carried out in 2015, 2017 and 2019. They suggest that most of state funded ALMPs have positive net effects on employment (MLSP, 2019[51]).

Direct job creation measures should be strictly targeted at long-term unemployed jobseekers that have no prospect of integration in the primary labour market. For young people, the priority should be to find employment in the primary labour market through support in the form of wage subsidies, if necessary. This is why many OECD and EU countries expanded the use of employment incentives rather than direct job creation measures also in response to the COVID-19 crisis. Twelve countries use these measures especially to support the employment of young jobseekers (Australia, Chile, France, Greece, Hungary, Ireland, Korea, Luxembourg, New Zealand, Portugal, Romania and the United Kingdom) and six countries to support the long-term unemployed (Flanders and Wallonia regions of Belgium, Greece, Hungary, Korea, Portugal and Sweden) (OECD, 2021[10])).

In Bulgaria, unemployed persons participating in direct job creation programmes are no longer on the unemployment register. As a result, during participation, they are not referred to jobs in the primary labour market by NEA counsellors (even though they might register again with the NEA after the end of the programme). This rule should be revised because it creates lock-in effects and reduces the effectiveness of direct job creation measures, as placement into “open market” jobs should always be the priority.

Beyond direct job creation measures that are seen as ALMPs, Social Assistance (SA) recipients also have a “public works” obligation. While the aim of the public works is to support SA recipients registered at the labour offices to develop work habits and discipline as set out in their individual action plan, the requirement may hinder participants from finding employment in the open market economy and could create a disincentive to claim SA assistance, thus reducing the likelihood of disadvantaged groups being activated through the benefit system (see Chapter 3).

The MLSP and the NEA are responsible for the entrepreneurship policy of both unemployed and some minority groups such as Roma (OECD, 2020[52]). Although entrepreneurship support is an objective fixed in the NEAPs, there is very little support offered to unemployed and inactive people to boost entrepreneurship and business creation. The start-up incentive measure Employment Through Business Support JOBS (which was not financed through state budget) was terminated in 2010. More recently, the HRD OP has offered measures to support self-employment and entrepreneurship, but take up has been very low. The NEA grants financial support to unemployed people who have entrepreneurial intentions. The start-up of a business activity can be financially supported on the basis of an approved business plan where the unemployment benefits are received at once, or a lump sum is provided for starting a business, which in 2020 was increased to up to BGN 4 000. Funding can also be provided for other activities such as entrepreneurship training and/or counselling and training for managing an approved business project. Unemployed persons entitled to unemployment benefits who wish to set up a business either individually or together with others may receive additional funds provided that they employ another unemployed person with no entitlement on benefits, and can receive further support such as external advisory services and a credit for training (OECD, 2020[52]). In parallel to this possibility, there has been for years the option for unemployed people to get support for setting up a private economic activity such a micro-enterprise under the SME Law with the requirement that the business project is approved by the territorial division of the NEA. Beneficiaries are also granted means for external advisory services, additional resources for qualification, as well as means to cover the insurance costs for 12 months.

Apart from these programmes, entrepreneurship support is mainly provided through NGOs, which play a substantial role in stimulating entrepreneurship for women, youth, and unemployed in Bulgaria (OECD, 2020[52]).

More women than men participate in ALMPs, and the share of female participants has increased over time reaching around 60% in 2019, while the share of women among registered unemployed decreased from 62.5% in 2008 to 55.9% in 2018 (Ministry of Labour and Social Policy, 2019[43])). Thus the likelihood for women to be referred to an ALMP has slightly increased over time. Women participate more often in training measures (69%), employment incentives (60%) and direct job creation programmes (59%) than men, which may be warranted to address higher employment barriers (Figure 6.10).

The share of registered unemployed young people under the age of 25 was 5.1% in 2018, down from 9.9% in 2012. In contrast, their share among participants in ALMPs was twice as high in 2019, reflecting the ALMPs targeting young people (Ministry of Labour and Social Policy, 2021[32]).

The share of long-term unemployed among LFS unemployed (i.e. not registered jobseekers as above, but people who do not work, but are available for work and look for employment) in Bulgaria was 56.6% in 2019, well above OECD average of 25.7% (OECD, 2021[53]). Among jobseekers who were registered with the NEA at the end of 2019, 24.8% had been registered for at least one year. The share of long-term unemployed people among ALMP participants is lower, at 19.3% in 2018. While it is desirable that people are referred to ALMPs before they become long-term unemployed, the overall comparatively low participation rate in ALMPs implies that the long-term unemployed are rarely referred to ALMPs.

While people belonging to Roma communities represent 15% of registered jobseekers (see Chapter 2), their share among participants in training measures (3.5%), employment incentives (5.5%), direct job creation measures (8.7%) and start-up incentives (0.6%) is significantly lower (Figure 6.10). In the view of Roma organisations interviewed by the OECD team in the Montana region, the subsidised employment measures available to Roma are mainly provided in sectors that require low qualification, such as waste collection, maintenance, etc. While many unemployed Roma have a low education, the interviews indicate that also unemployed Roma with higher education are in many cases not referred to other types of ALMPs. Participation in these measures does not contribute to a sustainable labour market integration of Roma, as the same people are repeatedly beneficiaries of such employment support programmes.

Pre-pandemic economic growth and increased labour demand in Bulgaria have enhanced the employment opportunities also for those furthest from the labour market. However, opportunities for the most vulnerable have increased only slightly. Since the start of the COVID-19 pandemic unemployment and inactivity have increased again, contributing to a substantial increase in the NEA’s caseload. International evidence suggests that early and frequent meetings with jobseekers can support an effective (re-)integration of jobseekers into the labour market. With a higher caseload, there is, however, a risk that the NEA cannot support more vulnerable jobseekers with comprehensive support and services. NEA data also shows that NEA counsellors meet the most disadvantaged clients less frequently already before the pandemic.

Bulgaria divides clients into three categories of job-readiness, which guide service provision. While the NEA has an IT tool to help support segmenting its clients, the IT tool recommendations are often not followed by case workers. Furthermore, the IT tool was designed about a decade ago and does not use have the same level of sophistication as profiling tools used in other countries. Many PES across the EU now make extensive use of digital tools, also to free up staff time to provide more intensive counselling to harder-to-place clients. Bulgaria has recently introduced new digital services and the use of pre-existing digital tools increased in the wake of COVID-19. Some other countries, however, go further and have a “digital-first” approach for jobseekers with sufficient digital skills who initially mostly self-manage their job search, while reserving more intensive and costly face-to-face services for jobseekers who are more difficult to place.

Active labour market programmes – including, training, employment incentives, supported employment and rehabilitation, direct job creation and start-up incentives – can play an important role in re-integrating unemployed into employment, when they do not find employment early in the unemployment spell. International evidence, suggests that not all types of programmes are effective and some types of programmes are more suitable than others to support different types of jobseekers. In an international comparison, Bulgaria spends relatively little on active labour market measures and spending is highly dependent on EU funding. With regard to the mix of different types of measures, Bulgaria puts lots of emphasis on direct job creation measures and less on training than other OECD/EU countries do. Evidence from international evaluations, however, suggest that the effectiveness of direct job creation programmes, in bringing participants back to open market jobs is questionable. Such programmes are also unlikely to address existing skills shortages in Bulgaria. Furthermore, specialised programmes that provide supported employment and rehabilitation for jobseekers with disabilities and health issues are limited in Bulgaria. Bulgaria runs a high number of ALMPs, many of which have a small number of participants. It is questionable whether running very small programmes is efficient, as it induces administrative costs and potential participants may not be aware of the programmes. A consolidation of small programmes would be worthwhile.


[16] Abbring, J., G. Berg and J. Ours (2005), “The Effect of Unemployment Insurance Sanctions on the Transition Rate from Unemployment to Employment”, The Economic Journal, Vol. 115/505, pp. 602-630, https://doi.org/10.1111/j.1468-0297.2005.01011.x.

[30] Algan, Y., B. Crépon and D. Glover (2020), Are active labor market policies directed at firms effective? Evidence from a randomized evaluation with local employment agencies, J-PAL, Working paper.

[50] AMS (2018), Bundesrichtline für die Förderung Sozialökonomischer Betriebe (SÖB), https://sozialplattform.at/files/inhalte/downloads/richtlinien/AMF%2015-2018_BRL%20S%C3%96B_oN.pdf.

[19] Arni, P., R. Lalive and J. Van Ours (2012), “HOW EFFECTIVE ARE UNEMPLOYMENT BENEFIT SANCTIONS? LOOKING BEYOND UNEMPLOYMENT EXIT”, Journal of Applied Econometrics, Vol. 28/7, pp. 1153-1178, https://doi.org/10.1002/jae.2289.

[8] Böheim, R., R. Eppel and H. Mahringer (2017), Die Auswirkungen einer Verbesserung der Betreuungsrelation für Arbeitslose in der Arbeitsvermittlung des AMS. Ergebnisse eines kontrollierten Experiments des AMS Österreich in der Beratungszone der RGS Esteplatz in Wien (The Effects of Raising the Number of Caseworkers for the Unemployed. Evidence from a Randomised Controlled Trial of the AMS Austria in the Counselling Zone of the Regional AMS Office Vienna, Esteplatz), Österreichisches Institut für Wirtschaftsforschung, http://www.wifo.ac.at/wwa/pubid/61297.

[42] Brown, A. (2015), “Can hiring subsidies benefit the unemployed?”, https://doi.org/10.15185/izawol.163.

[35] Bundesagentur für Arbeit (2021), Merkblatt 12 - Förderung der Teilhabe am Arbeitsleben für Arbeitnehmerinnen und Arbeitnehmer, https://www.arbeitsagentur.de/datei/merkblatt-12-teilhabe_ba015371.pdf.

[46] Caliendo, M., S. Künn and R. Mahlstedt (2017), “Mobility Assistance Programmes for Unemployed Workers, Job Search Behaviour and Labour Market Outcomes”, IZA Discussion Paper, Vol. 11169, https://www.iza.org/publications/dp/11169/mobility-assistance-programmes-for-unemployed-workers-job-search-behaviour-and-labour-market-outcomes.

[48] Card, D., J. Kluve and A. Weber (2018), “What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations”, Journal of the European Economic Association, Vol. 16/3, pp. 894-931.

[7] Crépon, B., M. Dejemeppe and M. Gurgand (2005), “Counseling the Unemployed: Does It Lower Unemployment Duration and Recurrence?”, IZA Discussion Papers, No. 1796, IZA, https://www.iza.org/publications/dp.

[3] Desiere, S., K. Langenbucher and L. Struyven (2019), “Statistical profiling in public employment services: An international comparison”, OECD Social, Employment and Migration Working Papers, No. 224, OECD Publishing, Paris, https://dx.doi.org/10.1787/b5e5f16e-en.

[44] Duell, N., L. Anghel and N. Ziminiene (2021), Greening of the labour market – impacts for the Public Employment Services - Small scale study, European Network of Public Employment Services, https://op.europa.eu/en/publication-detail/-/publication/a5ce471b-f0dd-11eb-a71c-01aa75ed71a1/language-en.

[34] EC (2020), Operational Programme Human Resources Development: Bulgaria, https://ec.europa.eu/regional_policy/en/atlas/programmes/2014-2020/bulgaria/2014bg05m9op001.

[37] European Commission (2020), Towards an inclusive labour market: ambitions of the Dutch Public Employment Service . Host Country Discussion Paper - the Netherlands, https://ec.europa.eu/social/main.jsp?langId=en&catId=1047&newsId=9528&tableName=news&moreDocuments=yes.

[29] European Commission (2018), Promising PES Practice: PES Offices for employers, Publication Office of the European Union, Luxembourg, https://ec.europa.eu/social/BlobServlet?docId=19258&langId=en.

[13] European Commission (2014), Blended service delivery for jobseekers, Peer review comparative paper by Willem Pieterson, European Commission, Brussels, https://ec.europa.eu/social/BlobServlet?docId=11833&langId=en.

[38] European Parliament (2017), Bulgaria: Recent Developments in Employment and Social Affairs. In-depth Analysis, https://www.europarl.europa.eu/committees/en/supporting-analyses/sa-highlights.

[26] Gray, D. (2003), “National versus regional financing and management of unemployment and related benefits: The casse of Canada”, OECD Social, Employment and Migration Working, Vol. 14, https://doi.org/10.1787/023874261242.

[54] Grubb, D. (2000), “Eligibility criteria for unemployment benefits”, OECD Economic Studies, No. 31, https://www.oecd.org/social/labour/2724341.pdf.

[45] Guglielminetti, E. et al. (2010), Home Sweet Home? Job Search with Commuting and Unemployment Insurance, https://www.eui.eu/Documents/DepartmentsCentres/Economics/Seminarsevents/Lalive-Commuting-Revision-Sept10.pdf.

[9] Hainmueller, J. et al. (2016), “Do Lower Caseloads Improve the Performance of Public Employment Services? New Evidence from German Employment Offices”, Scandinavian Journal of Economics, https://doi.org/10.1111/sjoe.12166.

[2] Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers”, OECD Social, Employment and Migration Working Papers, No. 215, OECD Publishing, Paris, https://dx.doi.org/10.1787/2bdfecca-en.

[14] Immervoll, H., C. Knotz and I. Otmani (2020), Results 2020: Activity-related eligibility conditions for receiving unemployment benefits, OECD, Paris, https://www.oecd.org/els/soc/Activity-related%20eligibility%20conditions_2020.pdf.

[41] Kluve, J. (2010), “The effectiveness of European active labor market programs”, Labour Economics, https://doi.org/10.1016/j.labeco.2010.02.004.

[28] Knotz, C. (2020), “Does Demanding Activation Work? A Comparative Analysis of the Effects of Unemployment Benefit Conditionality on Employment in 21 Advanced Economies”, European Sociological Review, Vol. 36/1, pp. 121-135, https://doi.org/10.1093/esr/jcz041.

[5] Kroft, K., F. Lange and M. Notowidigdo (2013), “Duration dependence and labor market conditions: Evidence from a field experiment”, The Quarterly Journal of Economics, Vol. 128/3, pp. 1123-1167, https://doi.org/10.1093/qje/qjt015.

[17] Lalive, R., J. van Ours and J. Zweimüller (2005), “The Effect of Benefit Sanctions on the Duration of Unemployment”, Journal of the European Economic Association, Vol. 3/6, pp. 1386-1417, https://doi.org/10.1162/154247605775012879.

[11] Langenbucher, K. and M. Vodopivec (2022), “Paying for results:  Contracting out employment services through outcome-based payment schemes in OECD countries”, OECD Social, Employment and Migration Working Papers, No. 267, OECD Publishing, Paris, https://dx.doi.org/10.1787/c6392a59-en.

[22] Le Barbanchon, T. (2016), “The effect of the potential duration of unemployment benefits on unemployment exits to work and match quality in France”, Labour Economics, Vol. 42, pp. 16-29, https://doi.org/10.1016/j.labeco.2016.06.003.

[25] Loopstra, R. et al. (2018), “Impact of Welfare Benefits Sanctioning on Food Insecurity: a Dynamic Cross-Area Study of Food Bank Usage in the UK”, Journal of Social Policy, Vol. 47/3, pp. 437-457, https://doi.org/10.1017/S0047279417000915.

[6] Maibom, J., M. Rosholm and M. Svarer (2017), “Experimental Evidence on the Effects of Early Meetings and Activation”, The Scandinavian Journal of Economics, Vol. 119/3, pp. 541-570, https://doi.org/10.1111/sjoe.12180.

[32] Ministry of Labour and Social Policy (2021), Национални планове за действие по заетостта (“National employment action plans”), https://www.mlsp.government.bg/natsionalni-planove-za-deystvie-po-zaetostta.

[43] Ministry of Labour and Social Policy (2019), Elaboration of a subsequent assessment of the effect of the active labour market policy financed by state budget resources at individual level (net effect) of the programs and measures, included in the national employment action plan for 2017, E-Research Consortium, European Commission, European Social Fund.

[33] MLSP (2021), ОТЧЕТЕН ДОКЛАД ЗА ДЕЙНОСТТА НА АГЕНЦИЯ ЗА ХОРАТА С УВРЕЖДАНИЯ ЗА 2020 Г., https://ahu.mlsp.government.bg/home/.

[51] MLSP (2019), Elaboration of a Subsequent Assessment of the Effect of the Active Labour Market Policy financed by State Budget Resources at Individual Level (net effect) of the Programs and Measures, including in the National Employment Action Plan for 2017, https://www.mlsp.government.bg/uploads/1/net-effect-final-report-en1.pdf.

[12] MyCompetence (2019), MyCompetence, http://www.mycompetence.bg/en/ (accessed on 23/06/2021).

[39] OECD (2021), Career Guidance for Adults in a Changing World of Work, Getting Skills Right, OECD Publishing, Paris, https://dx.doi.org/10.1787/9a94bfad-en.

[15] OECD (2021), How demanding are activation requirements for jobseekers?, OECD, Paris, https://www.oecd.org/social/strictness-benefit-eligibility.htm.

[23] OECD (2021), OECD Economic Surveys: Bulgaria 2021: Economic Assessment, OECD Publishing, Paris, https://dx.doi.org/10.1787/1fe2940d-en.

[10] OECD (2021), OECD Employment Outlook 2021: Navigating the COVID-19 Crisis and Recovery, OECD Publishing, Paris, https://dx.doi.org/10.1787/5a700c4b-en.

[53] OECD (2021), OECD Labour Force Statistics, https://data.oecd.org/unemp/long-term-unemployment-rate.htm.

[52] OECD (2020), Inclusive Entrepreneurship Policies, Country Assessment Notes, Bulgaria 2020, OECD, Paris, https://www.oecd.org/cfe/smes/Inclusive-Entrepreneurship-Policies-Country-Assessment-Notes.htm.

[47] OECD (2019), Evaluating Latvia’s Active Labour Market Policies, Connecting People with Jobs, OECD Publishing, Paris, https://doi.org/10.1787/6037200a-en.

[31] OECD (2019), “LMP interventions for the long-term unemployed: In-depth evaluation”, OECD, Paris, https://www.oecd.org/els/emp/LMP%20interventions%20for%20LTU%20-%20in-depth%20evaluation.pdf.

[4] OECD (2018), “Profiling tools for early identification of jobseekers who need extra support”, Policy Brief on Activation Policies, OECD, Paris, https://www.oecd.org/els/emp/OECD-Profiling-policy-brief-DEC-18.pdf.

[27] OECD (2016), “Investing in the employability of jobseekers in Slovenia”, in Connecting People with Jobs: The Labour Market, Activation Policies and Disadvantaged Workers in Slovenia, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264265349-8-en.

[1] OECD (2015), OECD Employment Outlook 2015, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2015-en.

[36] rehapro (2021), Modellvorhaben rehapro, https://www.modellvorhaben-rehapro.de/DE/Home/home_node.html.

[21] Tatsiramos, K. and J. van Ours (2012), “Labor market effects of unemployment insurance design”, Journal of Economic Surveys, Vol. 28/2, pp. 284-311, https://doi.org/10.1111/joes.12005.

[49] TZCLD (2021), Territoire zéro chômeur de longue durée, https://www.tzcld.fr/.

[18] van den Berg, G. and B. van der Klaauw (2006), “Counseling and monitoring of unemployed workers: Theory and evidence from a controlled social experiment*”, International Economic Review, Vol. 47/3, pp. 895-936, https://doi.org/10.1111/j.1468-2354.2006.00399.x.

[20] van den Berg, G. and J. Vikström (2014), “Monitoring Job Offer Decisions, Punishments, Exit to Work, and Job Quality”, The Scandinavian Journal of Economics, Vol. 116/2, pp. 284-334, https://doi.org/10.1111/sjoe.12051.

[24] Williams, E. (2021), “Unemployment, sanctions and mental health: the relationship between benefit sanctions and antidepressant prescribing”, Journal of Social Policy, Vol. 50/1, pp. 1-20, https://doi.org/10.1017/S0047279419000783.

[40] Windisch, H. (2015), “Adults with low literacy and numeracy skills - A literature review on policy intervention”, OECD Education Working Papers.

This Annex provides additional information on expenditure and participants in Bulgaria’s active labour market measures in 2019 (Annex Table 6.A.1), as well as a more detailed description of the major programmes (Annex Box 6.A.1). Annex Table 6.A.2 provides supplementary information on the targets groups of the National Employment Action Plans.


← 1. This calculation takes into the average number of labour mediators over this period of 631 plus a further 583 employees appointed on contracts through the National HRD OP programme.

← 2. Böheim, Eppel and Mahringer (2017[8]) provide a cost-benefit analysis that shows how hiring costs for additional caseworkers are off-set by benefit savings and increased revenues from tax and social security.

← 3. This approach compares countries statutory rules though actual enforcement may differ for countries with the same rules (Grubb, 2000[54]).

← 4. These data are for 2020 except for Iceland, Ireland, Norway, and Portugal, which are from the 2017 survey.

← 5. In almost all cases the reason given for termination is “failure to comply with the Individual Action Plan” making it difficult to understand the precise offence.

← 6. Active labour market measures refer to programmes included in categories 2-7 in the OECD/EC LMP database, including training (Cat. 2), employment incentives (Cat. 4), sheltered and supported employment and rehabilitation (Cat. 5), direct job creation (Cat. 6) and start-up incentives (Cat. 7).

← 7. The share of state budget in ALMP spending was only 20.8% in 2013, when the ALMP spending was at its highest level since 2004. Subsequently, as overall expenditure fell because of lower EU funding, the share of state financing increased to 46.5% in 2016 (EUR 68.4 million) (European Parliament, 2017[38]). Since then, the share of national budget has decreased and the share of ESF funding increased.

← 8. The different elements are individually captured in the corresponding ALMP categories of the European Commission/OECD labour market policy data base (mainly category 2 “training measures” and category 6 “direct job creation measure”).

← 9. In 2019, the largest programmes were: the regional employment programmes funded from the national budget, the national programme for employment and training of people with permanent disabilities, the national Program “Assistants to People with Disabilities”, the National Retirement Assistance Program; the training and employment programme for the long-term unemployed, the Training and employment for youth (<30) (or “Career Start” programme) and the National programme “Activation of inactive persons”.

← 10. The Minister of Labour and Social Policy together with the Minister of Education and Science develops and co-ordinates the state policy for adult education. The activities related to adult education are carried out by NEA, National Agency for Vocational Education and Training, the vocational training centres, other institutions specified in a law or an act of the Council of Ministers. Vocational training is provided by vocational high schools, vocational colleges, art schools and vocational training centres (VTCs).

← 11. Information collected as part of fact-finding meetings in the Montana region.

← 12. The programmes also includes families with children attending a childcare facility up to the age of 12 in case of large families with at least 3 children.

← 13. The Youth Employment Scheme is implemented in the whole country, but in the South-western region its financing is not under the Youth Employment Initiative, but from ESF.

← 14. Supported internships in the public administration under this programme are only available for high-skilled. In contrast, there are no educational level requirements for the on-the-job training component.

← 15. This includes those with 71% of reduced work capacity, military invalids with 50% or more work incapacity, people with sensory disabilities and people with mental disabilities.

← 16. The 134 refer to the total inflow of participants into the programme over the course of 2019. Annex Table 6.A.1 in contrast shows the annual average participant stock for programmes, which would be misleading for this type of measure.

← 17. The programme is a mixed measure and also has a training component. However, since 2018 the training component has not been used anymore.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2022

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.