Lithuania has made good progress in improving its system of active labour market policies and modernising its public employment service. Vocational training and employment subsidies have high positive effects on participants’ labour market outcomes, thus helping jobseekers to connect to good jobs. These measures have the potential to be even more effective by fine‑tuning their targeting. Training could be targeted more to low-skilled and older jobseekers, and combined with job matching and placement services. Employment subsidies are particularly important for older jobseekers and those living in non-urban areas. Furthermore, the system of active labour market policies has the potential to support a stronger Lithuanian labour market, particularly if increased funding is tied to systematic and rigorous impact evaluations, which can facilitate evidence‑informed policy making and effective policy design.
Impact Evaluation of Vocational Training and Employment Subsidies for the Unemployed in Lithuania
1. Assessment and recommendations
Abstract
1.1. Lithuania should continue to strengthen its system of active labour market policies
1.1.1. Lithuania has seen strong employment growth over the past decade but challenges remain
Lithuania has seen positive labour markets trends over the past years. At 73% in 2019, the employment rate among 15‑64 years old was higher than the OECD average of 68.7%. Moreover, employment has been resilient to the COVID‑19 pandemic, with the employment rate dropping by only 1.4 percentage points in 2020, versus 2.5 percentage points for the OECD on average, and returning close to its pre‑pandemic level by 2021 (at 72.4%). At the same time, the labour force participation rate increased faster in Lithuania than in the OECD on average and continued to grow in 2020, partly accounting for a relatively high unemployment rate (7.4% in 2021) and partly reflecting rising real wage levels.
One factor making the labour market tighter is the decreasing population of Lithuania. Between 1990 and 2020, Lithuania’s population shrank by 26% and the working-age population (15‑64 years old) shrank by 29%. Population decline is expected to continue. Lithuania is forecast to lose a higher share of its population by 2050 than any other OECD country (22% of its total population and 31% of its working-age population). To counteract population decline and the associated labour shortages, it is crucial that Lithuania provides employment support to people who are willing and able to work. This support should reach groups beyond jobseekers who are typically registered with the public employment service (PES) such as: discouraged workers and other inactive persons who would like and are able to work; people in low-paid jobs and at risk of job loss; and people at or beyond the retirement age who would like to continue working.
Despite the mostly positive trends in the Lithuanian labour market, disparities between population groups remain. While employment has increased strongly among persons with low education level, gaps with persons with higher and secondary education remain larger than in other OECD countries. There also large geographic disparities in the Lithuanian labour market, with the employment rate in Vilnius reaching 81.4% versus 50% in the remote and rural areas like Anykščiai, Ignalina, Lazdijai and Šakiai municipalities. In addition, geographic obstacles to jobs are particularly important in Lithuania, with about a quarter (23.5%) of 16‑64 year‑olds who were not in employment in 2019 living in a thinly populated area and in a household without a car.
1.1.2. A recent reform helped improve the institutional set-up and reach of active labour market policy provision
The labour market reform that was introduced in July 2017 in Lithuania centralised and modernised the organisational set-up of active labour market policy (ALMP) provision. The new social model introduced aimed to strengthen both labour market security and flexibility (flexicurity) and increase ALMP effectiveness and reach. The Lithuanian Employment Service (LES) which was established with the reform, has since experienced continuous and fundamental changes in its structure and management, its operating model, processes and infrastructure. The role of the social partners in ALMP design was strengthened with the inclusion of the relevant topics in the Tripartite Council discussions to provide strategic advice for the LES and the Ministry of Social Security and Labour. Such inclusion of the social partners has been assessed positively by all stakeholders involved in the provision of ALMPs.
Lithuania records one of the highest shares of jobseekers contacting the public employment service to find work across the OECD countries, reaching 86.4% in 2020. Even though the requirement to register with the LES to access health insurance may partly explain this high share, this is not the whole story, as similar requirements in other countries have not produced the same results. This high registration rate among jobseekers should be explored further by the LES to improve the reach of its services, especially among those persons who are further from the labour market and are usually the ones who require more comprehensive and intensive support.
Lithuania has intensified its efforts to engage more with employers since the 2017 reform by providing dedicated employers’ counsellors in the LES aiming to meet the employers’ needs. These efforts should continue and be strengthened to attract more vacancies requiring higher qualifications that are currently quite limited.
1.1.3. Lithuania spends relatively little in ALMPs, which do not adequately respond to needs
Even though the 2017 reform aimed to make ALMPs more accessible, spending on ALMPs did not increase to support the change. Lithuania spent 0.21% of GDP on ALMPs in 2019, which is less than half of the average of OECD countries (0.45% of GDP). Allocations to the traditional package of ALMPs increased only marginally to reach 0.23% of GDP in 2020, despite the COVID‑19 crisis. As a result of this low spending on ALMPs, less than 18% of registered jobseekers entered into ALMPs in 2019. This corresponds to only 1% of the labour force participating in ALMPs in Lithuania in 2019, versus 5% in the OECD on average. As such, although the LES has the potential to help many jobseekers due to the high share of registered jobseekers, this support remains very limited.
The 2017 reform introduced changes to ALMP provision by targeting them according to jobseekers needs and reshuffling the ALMP basket. On the one hand, more emphasis was placed on training measures, including their workplace components and new possibilities for more varied support were introduced. On the other hand, ALMPs that were considered as non-effective, such as public works and job rotation schemes were dropped. Nevertheless, these changes are not visible yet and the ALMP basket remains mainly focused on employment incentives (particularly through “social enterprises” that however do not currently support well the people furthest from the labour market) and less so on training measures and PES support.
ALMPs are primary funded through European Union (EU) sources in Lithuania while national funding is mainly used as co-financing. This financing model results in volatile resources, strongly dependent on EU funding cycles and less responsive to changes in labour market conditions and needs for ALMP support. Nevertheless, looking ahead, additional ALMP funding through EU sources such as the Resilience and Recovery Plan and ESF+ is likely to improve both ALMP spending and the composition of the measures provided to jobseekers, employers and people at risk of unemployment. To ensure these additional resources are used effectively, it is important that budget allocation is driven by evidence on what works.
1.2. Evidence‑informed policy making is crucial to improve the system of ALMPs
1.2.1. Impact evaluations of ALMPs can help design more effective ALMPs and allocate funding more sustainably
Robust counterfactual impact evaluations (CIEs, which estimate the net effect of an intervention relative to the “counterfactual” situation of no intervention) of ALMPs can help secure sustainable and sufficient funding for such interventions which is better linked with labour market changes and emerging needs. On the one hand, the results of these evaluations can serve to adapt or terminate ineffective policies while providing evidence to boost those interventions that work, leading to more effective support for jobseekers and employers. On the other hand, such evaluations, when effectively communicated to the public and policy makers, can help the Ministry of Social Security and Labour and the LES attract the necessary resources for effective ALMPs and secure more substantive and sustainable funding.
Impact evaluations require rich data and thorough evaluation techniques, while their impact on actual policy making relies on making them an integral part of the system which designs and implements ALMPs. Evidence‑informed policy making needs to be systematic and involve the whole cycle of designing, monitoring and evaluation, generating knowledge, disseminating knowledge, and adjusting policies based on evidence. But it also involves process evaluations which assess how implementation of interventions corresponds to design and strategies and which can help design more efficient policy implementation practices. Furthermore, impact evaluations should not only concern evaluations of ALMPs but also evaluations of the tools, processes and approaches used by the PES.
Lithuania has made progress in evidence‑informed policy making. First, the introduction of the social model has created a strong legal basis for ALMP monitoring and evaluation and has assigned an important role to the LES for assessing the effectiveness of ALMPs. Second, Lithuania has improved its monitoring and evaluation framework by making use of more data from different registers and improving its IT infrastructure to support data management. Nevertheless, there is still progress to be made at least in four directions. First, it is important to move from a simple monitoring of ALMP outcome indicators to CIEs which are required to generate evidence on whether labour market outcomes are determined by ALMP participation. Second, it is necessary to ensure that aspects of job quality are included in the outcomes analysed. Third, further investments are needed to modernise the relevant IT systems in order to enable data extraction and sharing for research purposes. Finally, Lithuania could gain by communicating more effectively the results of the ALMP evaluations that are conducted to secure support and the necessary funding.
1.2.2. Robust evaluation techniques are required to establish whether ALMPs have the intended impact on participants
Evaluating the impact of an intervention requires comparing the labour market outcomes of participants with their theoretical labour market outcomes had they not taken part in the specific intervention. This is the counterfactual, that is what would have occurred to them in the absence of the intervention. This counterfactual cannot be observed, but must instead be estimated. Simply comparing the outcomes of participants with those of non-participants would not answer this question because these two groups of jobseekers may differ in ways that determine both their participation in the intervention and their subsequent labour market outcomes. It is therefore important to find ways to compare jobseekers who are as similar as possible by using data on their demographic characteristics (gender, age), their observed skills (e.g. foreign language, ICT skills, etc.) and qualifications, financial (dis)incentives to leave unemployment (unemployment benefits, other types of income), other employment barriers, their geographic area of residence and job search, and their labour market history. In addition, it is crucial to compare jobseekers with similar unemployment durations to minimise any role that unobservable characteristics can play in explaining the moment during an unemployment spell that people enter an ALMP as well as their subsequent employment outcomes.
An approach that has been used in the related literature and is also used in this report is based on a “dynamic selection-on-observables” methodology. This approach compares the labour market outcomes of jobseekers who enter an ALMP (specifically vocational training and employment subsidies) in a given month of their unemployment spell with those who have not (yet) entered one of those ALMPs at a similar unemployment duration. In addition, this approach also compares individuals who have the exact same characteristics along a number of additional dimensions: calendar month and year of entry into the programme, age group, and whether individuals are receiving unemployment benefits.
1.2.3. Evaluations should look at outcomes beyond the probability of employment to account for aspects of job quality
Most CIEs of ALMPs in the international literature examine as main outcomes the probability of employment and, when possible, earnings. Even though the impact of an intervention on one’s probability of job finding is the first indicator to look at, it does not provide any information about the type of job obtained, any job-quality related characteristics or sustainability of employment. To address this, this report relies on Lithuania’s rich data available to evaluate the impact of ALMPs on a wider set of outcomes. More specifically, it is possible to look at the cumulative employment duration, as an indicator of job sustainability. In addition, the evaluation examines wages and cumulative earnings over the observation period, as well as cumulative earnings net of subsidies or training cost. This later indicator is used to compare the benefits of an intervention expressed as cumulative earnings over a three‑year horizon with the direct cost of the intervention.
In addition to these indicators, this report proposes an innovative way to assess the impact of ALMPs on occupational mobility. The main reason for analysing this outcome is because it is often the case that jobseekers who return to employment tend to enter lower-skilled occupations or return in occupations that pay lower wages because of scarring effects of (long-term) unemployment. The key question that the analysis in this report aims to answer is whether ALMP participation can in fact counteract these effects, offering a boost not only in terms of the likelihood of finding a job but also in terms of career progression in case a job is found. The measure of occupational mobility relies on an occupational index which is based on observed wages. Based on data on wages of all employed people in Lithuania during the 2018‑20 period, a wage index is calculated for each detailed occupational code. Increases and decreases in that index can be interpreted, respectively, as positive and negative changes in an individual’s occupation. Changes in this index are then used as an additional outcomes indicator in the CIE of vocational training and employment subsidies conducted in this report.
1.2.4. Lithuania’s rich linked administrative data represent an invaluable resource for evaluating the impact of ALMPs
Rich data available in Lithuania’s administrative registers can be linked and used to conduct CIEs of ALMPs, as well as evaluations of other labour market and social policies. As in many other OECD and EU countries, Lithuania can collect and link information on jobseekers characteristics, their participation in ALMPs and their employment outcomes by linking public employment service data (the LES register) with social security data (the Board of the State Social Insurance Fund, SODRA) and the business register (State Enterprise Centre of Registers). The dataset that is constructed by merging all these registers provides rich information on the personal characteristics of jobseekers (such as their age, education and possible barriers to become employed), their labour market outcomes (notably employment, unemployment, earnings, days worked, occupation) and their participation in various ALMPs. In addition, the data enable constructing the employment history of jobseekers before becoming unemployed as well as their previous unemployment spells history.
While the data are comprehensive and detailed, they could be further enriched along several dimensions. One variable that would be useful to include in the analysis but is currently missing in the data is that of hours worked. The impact evaluation in this report documents the generally positive effects of the programmes studied on a number of outcomes, including employment probability, but also hints at some trade‑offs in terms of occupational mobility. A similar trade‑off could conceivably be present in terms of hours worked. It would also be useful to obtain information on other sources of income received by jobseekers, including data on social assistance or disability benefits and income data from the State Tax Inspectorate. These could be useful in accounting for (or examining) the role of financial incentives in exiting to employment from unemployment. Finally, in terms of measuring training outcomes, it would be useful to have information on the target occupations of vocational training programmes. This could help with a systematic assessment of whether individuals are being hired in the occupations for which they have trained.
1.3. Vocational training and employment subsidies help to connect people with jobs, but could be fine‑tuned further
1.3.1. Vocational training and employment subsidies are two of the main ALMPs in Lithuania, representing half of the total ALMP budget
Vocational training and employment subsidies are two of the main ALMPs in Lithuania, accounting for about one‑third of participants in ALMPs and half of expenditures on ALMPs during the 2014‑20 period (excluding counselling and job brokerage by the LES and COVID‑19 employment support measures).
Vocational training offer unemployed people in Lithuania the possibility to select formal training (which leads to an accreditation or certificate) or non-formal training from accredited training providers. Roughly three‑quarters of training taken up was formal training. The average duration of vocational training was 2.8 months, with the average durations amounting to 3.4 months for formal training and 1.3 months for non-formal training. During the 2014‑20 period, approximately 94 000 episodes of such vocational trainings took place, with individuals entered vocational training in roughly 4% of unemployment spells during this period. Furthermore, about one‑quarter of training measures were accompanied by a tripartite agreement between the jobseekers, the LES and an employer who commits to employ the worker after the end of vocational training for at least a period of six months. Relatively small firms with less than 50 employees employed 58% of vocational training participants who entered into a tripartite agreement over the period 2018‑20.
Employment subsidies paid by LES subsidise 50% of participant’s wage costs and up to 75% for individuals with disabilities, with a ceiling amounting to twice the statutory minimum wage during the 2017‑19 period and one and a half statutory minimum wages thereafter. The employment subsidy is paid for six months (and indefinitely for people with severe disabilities or low work capacity). From 2017, employers are obliged to keep the subsidised worker for at least six months after the end of the subsidy. Approximately 80 000 unemployed people benefited from employment subsidies in the period between 2014‑20, representing approximately 4% of unemployment spells during this period.
Small firms and new firms make a disproportionately large use of employment subsidies (as a share of their total employment). During the 2018‑20 period, small firms’ annual intake of employment subsidy programme participants amounted to 1.5% of their average employment and they accounted for 75% of all participants. The annual intake of employment subsidy participants amounted to 0.8% of the average employment of new firms (those younger than two years), compared to 0.5% of those aged 5‑9 years. Evidence from OECD countries suggest that new firms are less likely to be making a profit and thus have a stronger incentive to seek out employment subsidies. Several sectors stand out for their extensive use of vocational training and employment subsidies, including agriculture, manufacturing, as well as wholesale and retail trade.
1.3.2. Younger jobseekers, men and high-skilled jobseekers are more likely to enter the vocational training or employment subsidy programmes
Jobseekers under the age of 30, particularly men, are disproportionally likely to enter vocational training or the employment subsidy programme. Women aged 30 or over, on the other hand, are disproportionally less likely to enter vocational training or the employment subsidy programme, with women 50 and over four times less likely to enter vocational training relative to their share of the unemployed. Men in general are more likely to enter either programme, but particularly vocational training, where they accounted for 78% of participants, even though men and women were roughly evenly represented amongst the registered unemployed during the 2014‑20 period. The gender disparities in training could partly reflect the types of courses that are available and which are the most popular courses, like obtaining a licence to drive a commercial vehicle, has mostly male participants.
One feature that may encourage women to undergo vocational training is the presence of a tripartite agreement, which obliges the employer to employ the worker after the end of vocational training for at least a period of six months. Across all age groups, the share of women entering training with tripartite agreements is considerably higher than the share entering without tripartite agreements.
Low-skilled jobseekers are disproportionally less likely to enter either of the two ALMPs studied. This likely reflects the fact that people without any qualification were not eligible for training programmes for some periods within the timeframe analysed – rather, such individuals were to be referred to formal education programmes before potentially being eligible for vocational training targeted towards the unemployed. Interestingly, in terms of the urban location of jobseekers entering the ALMPs examined, individuals from non-urban areas are slightly more likely to participate, even though, in the case of vocational training, consultations with stakeholders have indicated that finding a suitable training provider can be more of a challenge in practice in non-urban areas than in the larger urban regions More generally, vocational training tends to be taken up earlier during the registered unemployment period, while long‑term unemployed are more likely to enter subsidised employment.
The characteristics of participants in vocational training and employment subsidies indicate that groups that are close to the labour market are more likely to receive this support, such as young and prime‑age men and high-skilled jobseekers. This can raise concerns for creaming (support is provided to those people who enter the labour market quickly in any case) and can be particularly problematic if there is no other support available for the people further from the labour market. These results also highlight the importance of continuously assessing jobseeker needs and revisiting their individual action plans, not only in the beginning of the registered unemployment period but subsequently one year later.
1.3.3. Vocational training helps individuals become employed, especially in the short term, without adversely affecting occupational mobility or wages
The CIE results in this report show that vocational training has a positive and statistically significant effect on the probability of employment. The effect is initially modest but reaches a peak at around nine months after beginning the training programme. At this point, the likelihood of being employed was 21 percentage points higher for individuals who participated in training (the treatment group) than for those who had not entered an ALMP (the comparison group). The initially lower magnitude of the effect reflects the so-called “lock-in” effects, which arise because individuals in training are generally not engaging in intensive job search and may not be willing to accept a job until the conclusion of their training. After nine months, the effects of training on the probability of employment diminish but remain positive through the 3‑year evaluation period, amounting to 5 percentage points at the end of the period. The lock-in effect is also reflected in the estimation of the effect of training on days in employment. Initially up to four months (similar to the average duration of training), the impact of vocational training is negative and becomes positive at six months. Over the longer term, the effect of vocational training amounts to approximately 75 additional days of employment.
In contrast to the impact on employment probability and days in employment, vocational training has no significant positive effects on wages or occupational mobility. Individuals becoming employed early on after entering training experience a small wage cut relative to their pre‑unemployment wages. However, both groups recoup the wage gap one year after entering training, when they earn a slightly higher wage than they had before becoming unemployed. The estimated effects of training on occupational mobility are found to be generally insignificant, with some estimates pointing to negative effects from month 12 onward. This result implies that jobseekers who became employed after training on average entered occupations that paid slightly less than those who had not engaged in training. The sub-group analysis indicates that the results are largely driven by the effects observed for men under 30, who do not “climb the occupational ladder” as quickly as their peers who exit unemployment without first undertaking vocational training. The magnitudes of the effects in the index are not particularly large, but they do indicate that for training, participants are indeed more likely to be employed than non-participants but they often enter into lower wage occupations. On the whole, however, the positive employment effect far outweighs these two factors: participation in training has positive effect on cumulative earnings from nine months onwards, which is when most participants have completed their training.
The estimated effects of vocational training in Lithuania compare favourably with those in other countries and previous evaluations in Lithuania. The effects of training for Lithuania are generally much larger in the short term (16.9 percentage points for Lithuania versus 2 percentage points found in other studies) and in the lower range of estimates over longer time horizons (5.3 percentage points in Lithuania versus 6.7 percentage points on average in other studies). When compared to the results of less recent studies in Lithuania, the estimated effects are considerably more positive than the effects of vocational training offered around 2010 and are more similar to those of a more recent study that analysed the effects for jobseekers entering vocational training in 2016.
1.3.4. The voucher system for vocational training may partly explain the positive effects
The positive results in Lithuania may indicate that the design and implementation of training is producing more effective labour market outcomes than training programmes in other countries on average (although this report evaluated only part of training provision in Lithuania). Implementing vocational training through a voucher system – with the LES counsellor first assessing the individual needs and agreeing with the jobseekers on relevant support, and subsequently the jobseeker choosing a training provider – offers two main potential benefits, while keeping the administrative burden low.
First, the possibility of having several providers offer similar types of training can lead to competitive pressures improving the quality of training provision. While also a traditional public procurement process assures a competitive process and allows LES to sign contracts with several providers, the voucher system enables co‑operating with many providers more efficiently. The finding that the estimated benefits of vocational training are slightly, although not statistically significantly, greater in large urban areas than elsewhere is consistent with this interpretation: large urban areas have a more competitive market with both a greater number of providers and clients.
Second, the voucher system could allow for a wider array of training programmes to be offered with lower administrative burden, which may help to address local skills shortages more quickly and effectively. During the 2014‑20 period, jobseekers enrolled in 2 000 different types of courses, indicating that meeting the demand for training might be faster and more versatile using vouchers than in a traditional public procurement system. The simplified procurement procedures in the voucher system with providers not having to undergo competitive tenders but a more simplified accreditation system also lowers the administrative burden of offering training for both the LES and the training providers.
The voucher based system is in stark contrast to the system of public procurement which had been in place in Lithuania prior to 2012, when there were longer public tenders for the purchase of training services, and the procurement procedure for acquiring training providers taking three months or more. The finding that Lithuania’s vocational training programmes are particularly effective in the short term compared with programmes in other countries is consistent with this rationale.
1.3.5. Men, older (especially older women), low-skilled and long-term jobseekers gain larger benefits from participating in vocational training
The positive effects of vocational training are not the same for all population groups. Men tend to benefit slightly more from training than women particularly in terms of cumulative days in employment and cumulative earnings and especially in the short to medium term. Two years after the start of training however, the effect of training is very similar for men and women. In addition, women experience a positive effect on occupational mobility in the short term (during the first 12 months after completing training), while men experience a negative effect for most of the periods observed after entering training. Despite this result, men experience a large, positive effect on cumulative earnings (including cumulative earnings net of the direct training cost) which suggests that the positive impact of training on days worked for men offsets any negative effects on occupational mobility. For women, on the other hand, the positive effect on earnings is not large enough to offset the direct costs of the training. Part of the reason might lie in the gender wage gap in Lithuania, which makes it more difficult for women to achieve a higher wage even after up-skilling. Decreasing the gender wage gap should be continuously addressed by wider employment and social policy responses in Lithuania. Furthermore, women’s choices might be more limited on the labour market in case of insufficient support to address care responsibilities (such as childcare, care for older people and people with disabilities, which tend to be more commonly the responsibility of women than men).
While the employment effects of vocational training are positive for all age groups, they are progressively stronger for older groups of jobseekers and this holds particularly for women. For women over 50 years, the estimated effect on employment is 10.7 percentage points at 24 months after entering training, while for younger women this is only 5.4 percentage points. Low-skilled jobseekers appear to benefit slightly more than high-skilled jobseekers, while there do not appear to be systematic differences between large urban areas and other areas. Long-term unemployed benefit slightly more from being included in training than short-term unemployed, consistent with some findings in the literature.
1.3.6. Employment effects of vocational training are stronger when there is a tripartite agreement between jobseekers, the LES and employers
In about one‑quarter of vocational trainings undertaken, employers commit to hiring a worker who successfully completes the training. The counterfactual impact analysis shows that, as expected, signing a tripartite training agreement with an employer in advance of receiving the training boosts the observed employment effects of the training considerably. The presence of such an agreement results in an 11.5 percentage point increase in employment probability 24 months after entering training, compared with a 6.3 percentage point effect among those who did not have such an agreement before their training. As expected given the commitment of employers, the differences in the short-term effects are even more pronounced: six months after entering training, individuals with tripartite agreements experience a 28.3 percentage point increase in employment probability, compared with 11 percentage points for those entering training without a tripartite agreement. The point estimates associated with having a tripartite agreement remain consistently higher than those for training without such an agreement throughout the period observed, although the differences are not as large after the first 12 months (corresponding essentially to the minimum period that employers are obliged to retain the employment relationship), averaging five percentage points in the ensuing period. In addition, the effect of training on cumulative employment duration is considerably more positive among individuals undergoing training with a tripartite agreement.
These positive results of tripartite training agreements highlight the importance of combining ALMPs like training with job matching and placement services to increase the effectiveness of LES support. While people close to the labour market could find an employer willing to commit to hire them after a short up-skilling measure, people further from the labour market are more likely to need more support from the LES to help them find job opportunities, or even solicit them to the potential employers. Only relying on jobseekers’ initiative to find employers for the tripartite training agreements might lead to creaming (mostly job-ready jobseekers benefitting from the service provision) and deadweight loss (jobseekers that would have been hired anyway benefitting from the measure). The dedicated employers’ counsellors put in place in 2017 in the LES have the potential to reach out and co‑operate with employers to facilitate signing tripartite training agreements also with jobseekers further from the labour market.
The stronger positive results of training with tripartite agreements also highlight the importance of employer engagement in training provision. Evaluation results from other countries indicate for example that involving employers in training design can contribute to policy design that correspond better to their labour market needs. Involving employers in training implementation enables providing jobseekers with up-to-date knowledge and skills for the job and enables them to make direct contact with prospective employers, increasing their chances for employment. The tripartite training provision in Lithuania might enable to engage jobseekers in training that are (in the short term) needed by employers, as well as enable them practice the new skills and gain experience shortly after classroom training.
1.3.7. The positive results of training suggest that Lithuania should invest more in this measure and find solutions to increase access to it
Although there was only limited take up of online training at the beginning of the COVID‑19 pandemic, it has potential to offer additional solutions for scaling-up training provision and to provide a wider selection of courses, especially in remote areas. While the pandemic hit training provision unexpectedly, providers were not prepared to move the training content and methods from classroom training to digital channels quickly and keep the level of quality. As the digital channels have gained more prominence during the pandemic throughout public and private services, the possibilities and necessary skills to use these channels have likely improved.
In addition, online courses for independent learning can be an option to make some upskilling available to a wider share of jobseekers. Furthermore, these learning modules can be used more flexibly as they are not dependent on the trainer’s availability. Nevertheless, in order to support particularly the low-skilled jobseekers, training in digital skills might be required before they can benefit more extensively from such online upskilling resources.
1.3.8. Subsidies have significant positive effects on labour market outcomes
Employment subsidies have large and persistent effects on individuals’ employment probability. The effects are particularly strong during the first year of employment, although these largely reflect the fact that employers receive subsidies for six months and are then required to retain workers for another six months. Nevertheless, after 12 months, when employers no longer have an obligation to retain the workers, individuals who entered subsidised employment were 26.7 percentage points more likely to be in employment than those who had not entered an ALMP 12 months prior. Even though the observed benefit of having been in subsidised employment becomes more modest over time, it remains statistically significant three years after the start of subsidised employment (with a 10.9 percentage point difference in employment compared to jobseekers who did not enter an ALMP at the beginning of the observation period).
In line with the positive effects on their employment probability, jobseekers entering subsidised employment were employed for a considerably longer period than jobseekers who did not enter subsidised employment (269 days over the three‑year observation period, 60% more than the six months of subsidised employment). In addition to providing a boost to their employment probability and duration, jobseekers entering subsidised employment experience a short-term boost in their occupational mobility observed in months 9 to 18 after starting the employment subsidies. The analysis does not find any statistically significant effects on wages but finds a positive effect on cumulative earnings. The effects of the employment subsidies on cumulative earnings are positive also after subtracting the direct costs of the employment subsidies.
The estimated effects of employment subsidies on labour market outcomes in Lithuania are higher than those in the international literature in the short term and somewhat lower in the long term (even though many studies find statistically insignificant effects in the long term). These differences between the short-term and the long-term effects may reflect the requirements for employers in Lithuania to retain the subsidised employees for six months following the end of the programme or else the employer is temporarily excluded from using the scheme. The estimated effects in this report also compare favourably to previous evaluations of employment subsidies in Lithuania in 2010 and 2016.
1.3.9. Employment subsidies have positive effects across the different groups of beneficiaries, although the channels and magnitudes of these effects vary
Men and women tend to benefit slightly differently on average from employment subsidies. While both men and women experience a similar boost in terms of average cumulative earnings, the underlying channels are somewhat different: women experience a greater increase in employment probability, while men experience a more positive effect on occupational mobility. The effects of employment subsidies on wages are inconclusive.
The positive impact of employment subsidies on employment probability at 24 months is highest among older women (21 percentage points for women older than 50 and 18.2 percentage points for those aged 30‑50, relative to 12 percentage points for those below 30 years). The effects are also higher for non-urban jobseekers (15.8 percentage points) relative to urban jobseekers (12.2 percentage points). There are no other significant differences across characteristics such as jobseeker age among men, skill level, or unemployment duration. These findings suggest that employment subsidies could be targeted even more extensively to older women and jobseekers living in non-urban areas to support the labour market integration of these groups.
In terms of cumulative earnings, women above 30 years’ experience an especially large boost, although both older men and women benefit more compared with younger jobseekers. In contrast to the effects on employment probability, urban jobseekers experience a larger boost in earnings than their non-urban counterparts, as do high-skilled individuals relative to low-skilled individuals.
1.3.10. Subsidised jobs are not replacing unsubsidised jobs
Employment subsidies in Lithuania show no evidence of inducing displacement effects, which would exist if subsidised workers were replacing unsubsidised ones within a given firm. This finding is based on an examination of the pattern of replacement flows – analysing the job positions (occupations) occupied and vacated by individuals entering and leaving firms – which does not suggest any systematic differences for subsidised job positions. The finding is remarkable particularly in light of the strong estimated effects of employment subsidies on the probability of becoming employed.
The absence of displacement effects may arise due to the conditions tied to the receipt of the subsidy, which may provider a disincentive for employers to engage in such strategic behaviour. From July 2017 onwards, employers who dismiss a worker in the six months from the last subsidy payment for that worker are not eligible for further employment subsidies for six months. However, this analysis does not rule out the presence of any deadweight effects, which would arise if firms receive employment subsidies for individuals they would have hired even in the absence of the subsidy.
1.3.11. ALMP targeting could be improved
Especially in the context of relatively low ALMP funding in Lithuania, it is crucial that these interventions are targeting those who need them the most and for whom they are most effective. This highlights how crucial it is to carry out CIEs of the existing ALMPs to identify which measures have an impact on participants’ subsequent labour market outcomes. This is one of the aims of the evaluations conducted in this report and the related recommendations provided so that Lithuania can further invest in the use of its rich linked administrative data to build more evidence‑informed policy making.
While the results of impact evaluations inform policy design and implementation guidelines, it is important that the implementation of ALMP targeting is supported by tools to assess jobseekers’ individual needs for support. At the end of 2021, Lithuania adopted a new digital jobseeker profiling tool that uses statistical methods, machine learning and administrative data to predict jobseekers’ probability of long-term unemployment and needs, which is likely to contribute to a better ALMP targeting. This tool replaced the previously used profiling tool which was based on a number of questions related to the jobseeker’s characteristics, barriers to employment and motivation, and which segmented jobseekers into three groups according to their distance to employment and five groups of support needs. Both with the new and the old profiling tools, the final decision on support needs is taken by the counsellor and the jobseeker in a mutual agreement, not necessarily fully adhering to the suggestions by the profiling tool, enabling counsellors to take into account further individual circumstances of the jobseeker. To ensure the new profiling tool is used sufficiently and helps counsellors better support the jobseekers, it will be important to evaluate its use and impact.
Lithuania could improve its targeting of ALMPs by further fine‑tuning its tools for assessing jobseekers’ individual needs for support. For example, Lithuania could invest in an extension of the jobseeker profiling tool that, in addition to the profiling and segmentation exercise, provides recommendations on ALMPs that could support jobseekers based on their own characteristics and the labour market outcomes of similar jobseekers who benefits from these measures. That could be initially implemented in a randomised manner to facilitate a robust evaluation.
Key policy recommendations
Increase spending on ALMPs and ensure funding sustainability
Increase spending on ALMPs, with emphasis on programmes that support upskilling and reskilling and promote employment in the primary labour market.
Plan strategically ALMP funding in the years to come to reduce dependency exclusively on EU funding and ensure budget sustainability.
Expand the reach of ALMPs
Strengthen further LES engagement with employers, including to enable mediating high-skill vacancies for registered jobseekers.
Reach out to employers to engage in tripartite training agreements that secure employment opportunities for jobseekers that have lower chances to engage with employers themselves.
Consider the high share of jobseekers registering with the LES as an opportunity to engage with them and offer to persons furthest from the labour market comprehensive support that combines employment services with other services they may need, such as social, health and education services.
Ensure support is provided according to clients’ needs and improve targeting
Target training measures and employment incentives according to people’s needs and in line with the measures’ effectiveness for different groups of jobseekers.
Assess needs for services and revisit individual action plans regularly to provide appropriate support to jobseekers that have not been able to integrate into the labour market quickly.
Monitor and evaluate the use and impact of the new profiling tool to assess whether this informs decision making in the LES while identifying ways for continuous improvement.
Consider extending the profiling tool to include recommendations on ALMPs considering the individual characteristics of a jobseeker and matching these with similar jobseekers who have benefitted from these measures. This could be initially implemented in a randomised manner to enable robust evaluation.
Expand upskilling and reskilling opportunities, particularly for people who need them the most
Strengthen vocational training accessibility, targeting in particular those groups which benefit from it the most, notably older jobseekers aged 50 and above, low-skilled persons and long-term unemployed.
Promote access to online training, possibly in modular form to support upskilling and reskilling, including for jobseekers in remote areas who face limited choice of courses available locally in‑person. In addition to live virtual courses, online modules for independent learning have the potential to reach wider groups of jobseekers at a lower cost.
Ensure that the employment subsidies reach groups that are further from the labour market
Fine‑tune employment subsidies that aim at integrating jobseekers with lower job opportunities to the primary labour market to target even more closely those that have the potential to benefit from this measure the most, such as older jobseekers and people living in non-urban areas.
Continue re-designing the employment subsidy scheme for people with disabilities and long-term unemployed to reach those groups that are the furthest from the labour market. Complement this scheme to involve training, job search assistance and other relevant support corresponding to their specific individual needs, and strengthen the integration of these vulnerable groups into the primary labour market as a longer-term objective.
Invest in evidence‑informed policy making
Establish a mechanism for counterfactual impact evaluations (CIE) of ALMPs which goes beyond the monitoring of gross labour market outcomes of ALMP participants to generate knowledge on the effects induced by ALMPs.
Build analytical capacity in the LES or build good co‑operation practices with external experts and researchers to ensure continuity of rigorous and systematic ALMP impact evaluations.
Further enrich the linked administrative data available for CIEs of ALMPs by including data on hours worked and benefits received by jobseekers beyond unemployment benefits.
Complement CIEs of ALMPs with process evaluations (assessments how implementation corresponds to design and strategies), as well as impact evaluations of the tools and approaches used by the LES.
Use the results of CIEs to conduct systematic cost-benefit analyses to demonstrate the cost-effectiveness of ALMPs and make the LES business case.
Integrate impact evaluations into the policy making cycle by disseminating the results of the evaluations, using them to drive policy design and implementation and to scale up funding for effective ALMPs.
Invest in IT data management systems and step up data use in policy making and implementation
Continue expanding data sharing between the LES and other relevant institutions for operational purposes to achieve more accurate assessments for support and holistic service provision, while strengthening data availability across administrative registers for research.
Modernise the IT infrastructure in the LES to support data analytics and knowledge dissemination, such as data warehouse or data lake solutions linked to user-friendly business intelligence tools.