Employment Effects of Skills Trainings in Sub-Saharan Africa: A Systematic Review of Recent Randomized Controlled Trials

Beber, Bernd; Lakemann, Tabea; Schnars, Regina; Lay, Jann Article — Published Version Employment Effects of Skills Trainings in Sub-Saharan Africa: A Systematic Review of Recent Randomized Controlled Trials De Economist Provided in Cooperation with: Springer Nature Suggested Citation: Beber, Bernd; Lakemann, Tabea; Schnars, Regina; Lay, Jann (2024) : Employment Effects of Skills Trainings in Sub-Saharan Africa: A Systematic Review of Recent Randomized Controlled Trials, De Economist, ISSN 1572-9982, Springer US, New York, NY, Vol. 173, Iss. 1, pp. 87-120, https://doi.org/10.1007/s10645-024-09442-6 This Version is available at: https://hdl.handle.net/10419/319176 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) De Economist (2025) 173:87–120 https://doi.org/10.1007/s10645-024-09442-6 ORIGINAL PAPER Employment Effects ofSkills Trainings inSub‑Saharan Africa: ASystematic Review ofRecent Randomized Controlled Trials BerndBeber1 · TabeaLakemann2,3 · ReginaSchnars2,3 · JannLay2,3 Accepted: 19 September 2024 / Published online: 8 October 2024 © The Author(s) 2024 Abstract This study provides a comprehensive systematic review of recent randomized controlled trials (RCTs) to evaluate the employment effects of skills training programs in sub-Saharan Africa. The review focuses on studies conducted between 2019 and 2024, a period marked by a significant increase in the number of RCTs of training interventions in this region, and we thus fill a gap left by earlier reviews that did not reflect this recent surge in experimental studies. We employ the standard SPIDER approach for defining search terms and the PRISMA procedure for search and selection to systematically analyze the impact of these training programs on employment outcomes. The findings reveal a shift towards a more optimistic assessment compared to previous analyses, suggesting that many recent skills training programs do contribute to improving employment-related outcomes. Keywords Systematic review· Skills training· Randomized controlled trials· SubSaharan Africa JEL Classification J24· O15· C93· I25 1 Introduction Skills trainings that target individuals or small businesses are an important element of contemporary development assistance, and they are a frequent feature of development cooperation in sub-Saharan Africa. The principal reason for this emphasis on skills development is the idea that a trained (and retrained) workforce is more * Bernd Beber ber[email protected] 1 RWI – Leibniz Institute forEconomic Research, Essen, Germany 2 German Institute forGlobal andArea Studies (GIGA), Hamburg, Germany 3 University ofGöttingen, Göttingen, Germany 88 B.Beber et al. productive and can help developing countries compete and grow economically, especially in an era of constant technological change and an increasingly global market for labor.1 Despite their prominence in sub-Saharan Africa, skills trainings in this part of the world were rarely evaluated with randomized controlled trials (RCTs) even a decade ago, a fact noted and lamented by earlier systematic reviews. However, this has changed drastically in the last few years, and this systematic review serves to summarize key insights from these recent RCTs and identify study gaps that remain. To our knowledge this is the first systematic review specifically of recent RCTs conducted in this research field in sub-Saharan Africa. We focus on studies published in the last half-decade, because several systematic reviews and meta-analyses were conducted in 2019 and before, prior to the recent surge in experimental studies implemented in sub-Saharan Africa. We will refer extensively to this prior literature throughout this article, highlighting differences and new directions. This paper focuses on employment outcomes that could in principle be attributed to skills trainings and related development interventions, and hence speaks to Sustainable Development Goal (SDG) 8. SDG 8 aims to promote sustained, inclusive, and sustainable economic growth, full and productive employment, and decent work for all. One way in which skills trainings are often seen as contributing to these objectives follows a straightforward logic: Trainings lead to more productivity, which leads to more growth and better jobs. But historically many trainings have not fulfilled this promise, and we suggest—as have other reviews—that high-quality and often multipronged programs that are responsive to market demand perform better. Unlike many previous reviews, we see signs for cautious optimism: Among recent rigorously evaluated trainings in sub-Saharan Africa, positive effects on employment-related outcomes are relatively common, perhaps in part because of differences and improvements in program siting, conception, and delivery. In this sense we believe our systematic review contributes to our understanding of how skills training can help fuel progress towards achieving SDG 8, in addition to earlier review studies. Methodologically, we use two standard tools of systematic reviews: the SPIDER tool for defining search terms and the PRISMA procedure for search and selection. Our systematic search is conducted across three databases: Web of Science (WoS) and the Bielefeld Academic Search Engine (BASE) as principal sources, and Google Scholar as a supplementary source to ensure comprehensive coverage. The screening process resulted in the selection of 33 reports and peer-reviewed studies, covering research from fourteen countries within the region, from an initial pool of 2905 records. We proceed as follows. We first briefly lay out the principal theory of change associated with training programs in the literature, in which the creation of human capital serves as the link between interventions and outcomes, and which requires high-quality programs, often with multiple components, and responsiveness to 1 See e.g. the World Banks description of the objective of “jobs-rich economic growth” for its activities around skills and workforce development at https:// www. world bank. org/ en/ topic/ skill sdeve lopme nt. 89 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… market demands. We then turn to a description of the methodological approach we take in carrying out this systematic review of RCTs on skills training in sub-Saharan Africa. We present the results of the review, and summarize key insights, while linking back to prior literature and earlier reviews. Finally, we identify a set of research gaps that remain. 2 Demand‑Responsive Human Capital Development The primary way in which skills training programs tend to be linked to improvements in employment and job quality outcomes focuses on human capital development. At the level of individual workers and potential employees, many programs aim to improve technical (hard) skills, which can be either general (e.g. literacy, basic quantitative or financial reasoning) or specialized (e.g. trade-specific competencies or knowledge concerning production processes). Many programs additionally or alternatively target life (soft) skills such as the ability to communicate effectively and work well with others. At the entrepreneurial (or firm) level, programs similarly strive to improve technical capacity (e.g. accounting techniques, product development, or marketing) as well as process and soft skills (e.g. decision-making and effective team leadership).2 A longstanding concern is that “managerial capital” is often missing in developing countries and that suboptimal business practices can help explain the productivity gap between poor and rich countries (Bloom etal., 2010), so this is where business training interventions try to make a difference. Most trainings have traditionally been classroom-based, but innovation and pluralism in techniques characterize the field today. Today’s training programs could feature visits by role models (Lafortune etal., 2018), peer learning (Dalton etal., 2019), or individual coaching sessions (Bardasi etal., 2021). But they all have in common that they claim to increase valuable human capital, which then affects employment-related indicators. In turn this implies a first expectation for effective training programs: They must be of sufficient quality and intensity for skills to be transferred to beneficiaries. However, this is only the first step. For job outcomes to improve, transferred skills also must be valued in the labor market, and the value of newly created human capital often depends on the specific context in which beneficiaries find themselves. Hence, for interventions to work, implementers and participants must know local markets and respond to market demands in identifying value-generating training opportunities. Market awareness and demand-responsiveness are crucial drivers of program efficacy. In fact, programs can have negative or no effects on economic outcomes even if they successfully and comprehensively transfer target skills to beneficiaries (e.g. Cho etal., 2013). This can be the case if the value of newly acquired skills fails to 2 See McKenzie and Woodruff (2014) for an overview. 90 B.Beber et al. offset any earnings and career opportunities lost during the duration of the course itself, or if a course induces beneficiaries to switch into a sector that turns out to yield inferior earnings.3 The latter can easily produce a situation in which a training program can be a success in terms of teaching skills while failing to create value for the individual participant.4 These requirements—high program quality and intensity for effective human capital development, and training content that is responsive to market demand—inform our reading of the literature summarized in this systematic review, and we return to this theme when we discuss the results of the included studies.5 3 Method In order to conduct a systematic review of recent studies, we follow two standard sets of procedures: The SPIDER tool for the determination of search phrases (Cooke etal., 2012; Methley etal., 2014), and the PRISMA search and selection procedure for systematic reviews (Pageet al., 2021). The purpose of this paper’s systematic search is to review recent randomized controlled trials of skills trainings on employment effects in sub-Sahara Africa, and we derive our search terms accordingly. First, we restrict the sample to the relevant set of countries. Second, the phenomenon of interest is any kind of training, skills development, or vocational education intervention. Third, we limit this search to study designs that are randomized controlled trials or field experiments. Fourth, we are only interested in studies where the outcomes to be evaluated are broadly related to employment. Fifth, we are interested in effects and estimates as the currency of the target research. Qualitative researchers typically use this last element of the SPIDER tool to declare their interest in “qualitative” work, but the alternative term “quantitative” would not ordinarily appear in RCTs, while “estimate” or “effect” would. We restrict our search to the five-year period from January 1, 2019 to January 1, 2024. A set of well-known reviews was published in 2019 or in the immediately preceding years, with the dearth of RCTs in sub-Saharan Africa a key insight of those reviews. Since then, a substantial number of precisely such RCTs have been published, and we are setting out to review this body of work. 5 Two secondary mechanisms linking skills trainings to employment outcomes that also feature in this literature and later in our review are certification and signaling, and access to formal and informal networks. Both can help beneficiaries be successful on the labor market, even without skills transfers. They generally still require market awareness to know what is needed to compete effectively in the labor market. 3 This raises the question why individuals would join such a course or work in a job that matches their newly acquired skills in such a situation in the first place. One possibility is that potential beneficiaries are uncertain about the long-term returns to their labor market decisions and may defer to others such as training providers that they believe to be better-informed (Del Carpio and Guadalupe 2022). 4 For example, Novella etal. (2018) analyze the Chilean Bono Trabajador Activo (BTA) voucher program, which funded training courses for formally employed individuals. They find that employment probabilities and formal-sector income are lower for those who enrolled in a training opportunity, even years later and especially for those who were willing to change sectors. 91 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Table 1 summarizes the search logic. We deploy our SPIDER search in three engines, and in doing so follow the recommendations by Gusenbauer and Haddaway (2020). First, we use the Web of Science (WoS), which includes the Social Sciences Citation Index (SSCI) in its core collection.6 Second, we use the Bielefeld Academic Search Engine (BASE).7 Full-text searches in these two engines, WoS and BASE, are our principal means to identify citations. Third, we use Google Scholar, as a supplementary source, given its very broad coverage. Google is not suitable as a principal source for the execution of our SPIDER search for two reasons. First, Google does not allow fully articulated Boolean logic, in particular nesting or multiple OR-groups between different AND operators.8 Second, the extent of Google search results appears to vary with server load, and results are not exactly replicable. It is very unlikely that this would have resulted in a missed relevant article in our case, given that searches always returned results well in excess of those actually relevant, but it is a reason why Google Scholar is not suitable as the principal source for a systematic review.9 3.1 Eligibility Criteria We include any English-language peer-reviewed article or non-refereed policy report or discussion paper in any kind of series, published between 2019 and 2024, that analyzes a randomized controlled trial of a skills development or training intervention aimed at changing employment-related outcomes in a sub-Saharan African country. We exclude articles that focus on different primary outcomes, such as mental health indicators (Evans etal., 2022), antisocial behavior (Vicente & Vilela, 2022), women’s empowerment (Karimli et al., 2021), or social inclusion (Kimou et al., 2022), if the paper does not also discuss employment outcomes. We also exclude studies that focus on lab-in-the-field outcomes (such as competitiveness in the case of Almås etal., 2020) if they do not report outcomes related to actual employment. In general, our search terms cast a fairly wide net. We do not narrow our search to titles or abstracts or specific job-related outcomes such as income but identify any studies that refer to employment or the workplace anywhere in their full text. Skills 6 We use the following advanced exact full-text search phrase, cycling through all applicable country names: ALL = ("[country]" AND ("training" OR "skills development" OR "vocational education") AND ("randomized control*" OR "field experiment") AND ("employment" OR "job" OR "workplace") AND ("effect*" OR "estimate")). 7 We use the following search term in BASE’s full-text document search: "[country]" AND ("training" OR "skills development" OR "vocational education") AND ("randomized control*" OR "field experiment") AND ("employment" OR "job" OR "workplace") AND ("effect*" OR "estimate"). 8 For Google Scholar, we use the following alternative search term to produce a result that is reasonably specific while at the same time including all text segments shown in Table1: "[country]" "training" "randomized control*" "employment" "effect*" intitle:"[country]"|intitle:"training"|intitle:"skills development"|intitle:"vocational education"|intitle:"randomized control*"|intitle:"field e xpe rim ent"|intitle:"e mployment"|intitle:"job"|intitle:"workplace"|intitle:”effect*"|intitle:"estimate". 9 Note that current Google Scholar searches do not appear to truncate searches at 256 characters, as is widely believed, so this is not an ongoing concern. 92 B.Beber et al. trainings need not focus on technical skills, but can involve, e.g., soft skills such as self-efficacy or a “growth mindset” (Morris etal., 2023). We do require trainings to involve some kind of skill transfer, so exclude exercises like a “self-training intervention “ (Demerouti, 2023). We also exclude “teaching” interventions if the treatment consists purely of an instance of information provision, such as a presentation of legal facts (Aberra & Chemin, 2023; Bertrand & Crépon, 2021). 3.2 Screening, Selection, andCross‑Verification Figure1 shows the PRISMA search procedure. We first retrieve 2905 records with the SPIDER-based electronic search.10 These were screened based on their titles and abstracts / executive summaries to remove irrelevant or duplicate records.11 A total of 44 articles were left and read in more detail, of which 28 studies matched the inclusion criteria and objectives of this review. This includes 17 peer-reviewed articles, and 11 discussion papers or technical reports. Table 1 SPIDER specification a While a typical SPIDER search will only require (D OR E), we require both D and E in order to limit searches to randomized controlled trials SPIDER tool Search terms Sample Angola OR Benin OR Botswana OR “Burkina Faso” OR Burundi OR Cameroon OR “Cape Verde” OR “Central African Republic” OR Chad OR Comoros OR Congo OR “Côte d’Ivoire” OR “Ivory Coast” OR Djibouti OR Eritrea OR Eswatini OR Swaziland OR Ethiopia OR Gabon OR Gambia OR Ghana OR Guinea OR Kenya OR Lesotho OR Liberia OR Madagascar OR Malawi OR Mali OR Mauritania OR Mauritius OR Mozambique OR Namibia OR Niger OR Nigeria OR Rwanda OR “Sao Tome” OR Senegal OR Seychelles OR “Sierra Leone” OR Somalia OR “South Africa” OR “South Sudan” OR Sudan OR Tanzania OR Togo OR Uganda OR Zambia OR Zimbabwe Phenomenon of Interest “training” OR “skills development” OR “vocational education” Design “randomized control*” OR “field experiment” Evaluation “employment” OR “job” OR “workplace” Research Type “effect*” OR “estimate” SPIDER search: S AND P of I AND D AND E AND Ra 10 Google Scholar produced 2841, WoS 36, and BASE 28 records to be examined. This is not uncommon. Google Scholar is highly sensitive and inclusive but can lack precision and therefore requires more extensive manual filtering than WoS and BASE (Giustini and Kamel Boulos 2013). 11 Records were retrieved and screened by one coauthor, screening results reviewed by a second coauthor, and each of the remaining full-text articles was inspected and assessed for eligibility by at least two coauthors. 93 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Additionally, and to minimize the chance that we omit relevant studies that are insensitive to our search specification, we cross-verified our results against three related collections of studies, each of which partially overlaps with our review in temporal or thematic scope.12 This process yielded six additional articles, as indicated in Fig.1. The final set of studies covers interventions in fourteen different countries: Uganda (7 studies), Rwanda (5), Tanzania (4), Ethiopia (4), Togo (2), Nigeria (2), Liberia (2), Côte d’Ivoire (1), Ghana (1), Kenya (1), Senegal (1), South Africa (1), Sierra Leone (1), and Mozambique (1). Table2 summarizes key features of the studies that meet our inclusion criteria and are RCTs of skills training interventions targeting employment outcomes in subSahara Africa. We show key sample characteristics, a brief description of the intervention, summarize outcomes and main effects, and indicate heterogeneous effects that the study may report. The included articles cover a wide range of programs and contexts, but with notable areas of concentration: • Regional distribution Our list includes 21 studies that were implemented in East Africa, 10 in West Africa, 2 in Southern Africa, and none in any Central African countries—an obvious gap in the evidence base. • Target populations Most commonly studies focus on youths (19 articles), and in particular youths that are described as disadvantaged in one form or another (10) and youths in urban areas only (8). A second large set of studies works with business owners/entrepreneurs or smalland medium-sized firms (13). Trainings for non-youth workers are rarely studied, as are trainings for individuals that are highly educated already (such as coding programs for university graduates). • Intervention types Technical/vocational training programs (12 studies) or apprenticeship programs (3) are the modal intervention for youths and other workers, as are business trainings for entrepreneurs/firms (12). Life skills/soft skills modules feature in 13 studies, often in tandem with other training components. Other complementary elements such as cash grants (4), financial literacy courses (4), networking/mentorship opportunities (2), digital skills instruction (2), or job placement outside of apprenticeships (2) are relatively less common. Teacher trainings and instructions on collective bargaining/labor law are the focus of one study each. 12 First, Kemper etal. (2022) review a total of 89 experimental and quasi-experimental studies on “vocational training” from 1990–2020, which overlaps in part with our temporal scope. We find that for subSaharan Africa and the relevant years 2019 and 2020, two articles are included in both their and our sets of studies, one study is included only in their review, and seven studies are included only in our search output, given our broader focus on skills trainings. Second, J-PAL (2023) provides an overview of 28 studies on apprenticeships and vocational training programs, six of which fall within our temporal and geographic scope, including three additional studies not yet included here (and again with a larger number of studies that we retrieved not included in the J-PAL overview). Third, we compare our list to relevant studies cited in Carranza and McKenzie (2024) recent overview article on job trainings in developing countries, which yielded one additional article to include, for a total of five additional studies. 94 B.Beber et al. • Study outcomes Income/earnings are the most common metric of program effectiveness, together with participant’s employment status and hours worked. These are followed by investments/savings/assets, entrepreneurship/self-employment, performance/productivity, life satisfaction/physical or psychosocial wellbeing, and aspirations/self-efficacy. Infrequently studied outcomes include job search behavior, consumption, fertility, and unionization levels. At the business level, studies tend to focus on revenue and profits, management practices, and firm survival. Notably few studies consider the number of workers employed by a business. 4 Empirical Evidence 4.1 Primary Results We now turn to a description of effect estimates, first with an overall assessment, followed by a more detailed discussion of the intervention type that appears most promising—high-quality, multipronged programs that are responsive to local demand— and important effect heterogeneities. In each case, we will relate results from this Fig. 1 PRISMA flowchart of search procedure 101 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Table 2 (continued) Reference Country Sample Intervention Outcomes and main effects Heterogeneous effects Chioda etal. (2021) Uganda 4400 high school students at the end of secondary education, from 200 schools across Uganda Hard-skills mini-MBA (ca. 75% hard skills, 25% soft skills), compared to soft skills program (ca. 25% hard skills, 75% soft skills), and control Large increase in earnings for both types of trainings, higher likelihood of enterprise start and survival. Soft skills training linked to improvements in self-efficacy, persuasion, negotiation None reported Crawford etal. (2021) Nigeria About 4000 marginalized youths in the northern states of Kaduna, Kano, Katsina, and Jigawa Mafita program, with either apprenticeships with master craft persons or vocational/technical training (including literacy, numeracy, and businessrelated soft skills trainings) at skills development centers Apprenticeships: Increased self-employment, wage employment, business profits, attempts to start business, job search behavior, household assets, consumption. No significant effect on literacy, numeracy, or non-material outcomes. Skills development centers: Similar increases in employment, productivity, entrepreneurship behavior, consumption, household assets. No significant effect on literacy, numeracy. Mixed evidence on non-material outcomes, with higher subjective well-being, attitudes on female empowerment For women, large increases in self-employment and income, and small increases in self-esteem, professional networks (for women in apprenticeships) or subjective well-being (for women in skills development centers) 102 B.Beber et al. Table 2 (continued) Reference Country Sample Intervention Outcomes and main effects Heterogeneous effects Crépon and Premand (2019) Côte d’Ivoire 1842 youths aged 18–24 and 731 firms in seven urban areas Formal apprenticeships with a monthly subsidy of 30,000 FCFA, dual training with on-the-job and theoretical components, and certification Short-term effects on youths: No significant impact on total income, with decrease in labor earnings offset by subsidy. Medium-term effects on youths: Increase in total earnings, higher involvement in complex tasks, increased certification. Firm effects: Limited substitution, so subsidized formal apprentice implies ¾ of a new position None reported Dammert and Nansamba (2023) Liberia About 570 entrepreneurs with small businesses Standard business training, interpersonal skills training, or both Either training led to better customer service, improved business performance, fewer customers lost. No additional gains from combined training. No impact on customer satisfaction None reported Dimitriadis and Koning (2022) Togo About 300 entrepreneurs in Lomé, with businesses having been in operation for at least one year Two-hour social skills training as part of a two-day marketing program Increased monthly profits, value-generating partner interactions, collaborative perceptions, information exchange, relationships formed. Lower ethnic concentration in new relationships formed None reported 103 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Table 2 (continued) Reference Country Sample Intervention Outcomes and main effects Heterogeneous effects Hardy etal. (2019) Ghana 3928 youth program applicants aged 15–30 across 32 districts and all regions Access to apprenticeships with small informal sector firms in trades such as garment making, cosmetology, carpentry, welding, and masonry Modest increases in probability of starting and completing an apprenticeship, in training durations, and in shifts from wage work to self-employment. No significant short-run increase in self-employment profits Effects vary by gender and trade, e.g., women in cosmetology more responsive than men in construction. Higher earnings associated with more experienced or profitable trainers Hirvonen etal. (2023) Ethiopia Beneficiaries of the Productive Safety Net Program (PSNP) Graduation programming that includes financial literacy training, promotion of new income generation activities, nutrition interventions and consumption support Not reported, focuses on effects conditional on drought Program moderated drought effects, protected against nutritional deficits, with savings as mechanism Lafortune etal. (2022) Rwanda Students at 159 high schools Six-week gamified online entrepreneurship training Significant increases in business ownership, participation in school business clubs, employment, earnings, and hours worked Similar effects by gender, stronger effects for public schools 104 B.Beber et al. Table 2 (continued) Reference Country Sample Intervention Outcomes and main effects Heterogeneous effects McIntosh and Zeitlin (2022) Rwanda 1848 underemployed men and women aged 15–30years in the districts of Rwamagana, Muhanga, and Nyamagabe Workforce training program, cash transfers of varying amounts, or a combination of training and cash transfer Training: Increased hours, assets, savings, subjective well-being, and business knowledge. Cost-equivalent cash: Also increased hours, assets, savings, subjective well-being, plus consumption, income, wealth. Minimal additional benefits with higher transfer amounts. Combined program: No compelling evidence of complementarity Limited evidence of heterogeneity, and impacts largely similar across genders, baseline consumption levels, and local labor market conditions Mensmann and Frese (2019) Togo 912 non-agricultural entrepreneurs in Lomé, with fewer than 50 employees Personal initiative training Personal initiative increased post-training but decreased over time. No significant effect on well-being Personal initiative gains over time contingent on “need for cognition” Morris etal. (2023) Tanzania 165 entrepreneurs, recruited through a government agency for entrepreneurship training “Growth mindset” psychological training in addition to technical skills training, compared to technical skills training only Growth mindset training resulted in significantly higher entrepreneurial self-efficacy and greater number of entrepreneurial actions None reported Nkonya etal. (2019) Uganda 1235 individuals from farmer groups across 9 districts Fullor half-dose of Creative Capacity Building (CCB) trainings, focused on harnessing indigenous knowledge in technology innovation processes Training increased economic output and improved tool repair capacity, but no robust positive effect on household income and no effect on household assets None reported 105 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Table 2 (continued) Reference Country Sample Intervention Outcomes and main effects Heterogeneous effects Rosas etal. (2022) Sierra Leone More than 3000 urban youths aged 15–35, with at least some secondary education Technical/business skills training, in combination with cash injections Improved employment and entrepreneurship opportunities, and protected household consumption and investments against the backdrop of the Ebola crisis Positive entrepreneurship, labor impacts with high initial non-cognitive skills. Improved skills, consumption, but not savings with low non-cognitive skills Tarekegne etal. (2023) Ethiopia 440 smallholder farmers Three-week training program focusing either on a highor a low-intensity version of the Comprehensive Competence‐Based Training (CCBT) High-level training led to greater improvement in both competence development and agricultural performance, including higher yield gains None reported Wheeler etal. (2022) South Africa 1638 young, disadvantaged individuals seeking work, recruited from existing job readiness programs in large South African cities Four hours of LinkedIn training during their job readiness program Significant increase in endof-program employment rate, with effect persisting at least a year Employment effect smaller for those who already had LinkedIn accounts, larger for those with low initial communication skills 106 B.Beber et al. current review back to insights from previous studies and meta-analyses.13 Our core contributions vis-à-vis these other reviews are twofold: First, in terms of coverage, we summarize a set of recent RCTs that have not previously been reviewed systematically. Second, in terms of substantive insight, we find reason to be relatively more optimistic about the potential of skills trainings to contribute to SDG8 objectives, with one possible explanation for this being differences and progress in program implementation, at least for trainings that fall within the scope of our review. Effectiveness in recent skills training RCTs Overall, the studies summarized here appear notably more optimistic on average than prior meta-analyses have been. For eleven of the evaluations of technical/vocational training or apprenticeship programs for youths and other workers that are included here, we are able to see if they had a positive effect overall either on a measure of employment (broadly speaking, any measure of the extent or intensity of income-generating activities) or on a measure of income or both. Six of the eleven studies identify positive effects on both of these dimensions.14 We see positive effects on one or the other dimension in four studies (Bandiera etal., 2020; Hardy etal., 2019; McIntosh & Zeitlin, 2022; Nkonya etal., 2019) and just one study without a positive effect on either of these particular outcome types (Alcid etal., 2023). For business and entrepreneurship trainings, the picture that emerges is similarly encouraging. We can observe an overall effect on a measure of entrepreneurial activity or performance in ten studies, and we see seven of them report at least one relevant significant positive effect,15 while three do not (Bardasi etal., 2021; Bjorvatn etal., 2020; Mensmann & Frese, 2019). This promising summary of recent evaluations stands in contrast to a bleaker set of older results. Overall, those reviews found that the majority of stand-alone training programs appear to have weak or no average effects on employment and job quality outcomes (Betcherman etal., 2004; Card etal., 2011; Grimm & Paffhausen, 2015; J-PAL, 2017; McKenzie & Woodruff, 2014). Kluve et al., (2017, 2019) observed positive effects on employment rates or earnings in only about one in three youth employment programs. Blattman and Ralston (2015) concluded that perhaps development initiatives should shift from a focus on trainings to a focus on 13 Previous reviews overlap but are not exactly congruent with each other or with our review in terms of their geographic scope, target populations, types of interventions, outcomes, and included study designs. For example, Kluve etal., (2017, 2019) cover 113 counterfactual impact evaluations of a wide range of employment programs, including skills training and entrepreneurship programs, for youths in both highand low-income countries, with a focus on earnings and employment as principal outcomes. Cho and Honorati (2014) include 37 experimental or quasi-experimental studies of entrepreneurship programs, including technical skills trainings relevant for self-employment, for individuals in developing countries, with income/profits, labor market activities, and financial behavior and attitudes as outcomes. Grimm and Paffhausen (2015) consider 53 evaluation studies of entrepreneurship trainings and other interventions targeted at micro-entrepreneurs in LMICs, with employment as the principal outcome of interest. While common themes emerge, findings from any one review also reflect its scope. 14 In this group, we place Crépon and Premand (2019), Alfonsi etal. (2020), Bandiera etal. (2023), Atkin etal. (2021), Crawford etal. (2021), and Rosas etal. (2022). 15 Here we see Araar etal. (2019), Alzua etal. (2020), Chioda etal. (2021), Blalock etal. (2022), Dimitriadis and Koning (2022), Lafortune etal. (2022), Dammert and Nansamba (2023). 107 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… capital-intensive programs, whether in the form of cash transfers or small business loans. Explanations for recent developments A decade ago, McKenzie and Woodruff (2014) cautioned that well-powered studies (with large enough samples and a sufficient number of participants) remained relatively rare at the time, especially in lowincome contexts, so some of the shift toward detectably positive results could simply be the result of researcher-side improvements. But as Carranza and McKenzie (2024) recently pointed out, policy-side efforts have also been underway to improve training programs, “in part due to the evidence from the first waves of rigorous training evaluations”. These efforts have largely taken two forms. First, funders have worked to improve implementation quality and effective program intensity, for example through sustained monitoring of service providers and beneficiary progress and appropriately incentivized contracts with service providers that ensure dependable program delivery.16 These kinds of quality parameters are difficult to observe, but one helpful proxy for implementation intensity may be whether programs are designed to be multipronged, involving providers and beneficiaries in more than one type of programmatic interaction. And when we again look at the list of studies included in our review, we see that five of the six technical training or apprenticeship programs with positive effects on both a measure of employment and a measure of income are indeed multi-component interventions, combining technical trainings with cash grants or subsidies (Crépon & Premand, 2019; Rosas etal., 2022), soft skills coaching (Crawford etal., 2021), or job placement (Atkin etal., 2021), and so on.17 In fact, the gloomy outlook of previous reviews concerned stand-along trainings in particular. Already then skills training had been successfully combined with placement services (Chakravarty etal., 2016) or formalization support (Benhassine etal., 2018), and some best-practice guides were recommending that training interventions should be multipronged if possible (Datta etal., 2018). The second policy adjustment observed by Carranza and McKenzie (2024) and others is a turn toward “demand-driven” programs. This could mean the involvement of private sector firms at the curriculum development stage to ensure that training content is contextually suitable, or the addition of a program in a particular sector if a market study identifies an unmet demand for skilled labor in this area. Again, multipronged interventions may be the safest bet to achieve some degree of demand responsiveness, simply because some training element may turn out to be valued in the labor market—the buckshot logic of effective multicomponent skills trainings. 17 We do not necessarily see that multicomponent treatments within the same program outperform results in a training-only condition (Bandiera etal., 2023; McIntosh and Zeitlin 2022), although it could be that a multicomponent design as such is an indicator of implementation (and implementer) intensity across conditions. 16 Program intensity also depends on training durations (Escudero etal., 2019; McKenzie and Woodruff 2014), although duration effects may vary across contexts and program objectives. For example, Hirshleifer et al. (2014) suggest that for vocational training courses in Turkey longer programs have less impact on employment than shorter ones, perhaps because longer courses leave less time for job searches. 108 B.Beber et al. One point of confusion is that demand responsiveness generally entails the involvement of private sector firms and an interest in their needs but does not necessarily imply that market demand is in fact optimally met by having private sector firms engaged in program implementation. This may help explain why results for apprenticeship programs tend to be relatively mixed. Among the studies included in our review, Crépon and Premand (2019) show lasting positive effects of a dual apprenticeship program in Côte d’Ivoire, Crawford etal. (2021) find that those assigned to apprenticeships in a study in Nigeria are no better off than those trained—more cost-effectively—in skills development centers, and Hardy et al. (2019) sees a shift from wage work to self-employment among those assigned to apprenticeships, but with newly self-employed beneficiaries on average unable to realize enough profits to offset their lost wage income. Alfonsi etal. (2020) offer a particularly helpful illustration of how private sector involvement does not equate responsiveness to market demands. They assign youths to either vocational training or firm-provided training and find that both are effective vehicles for skills transfers. But only those assigned to vocational training saw sustained wage increases because they received certified skills. Here, firm-provided program implementation ran counter to actual market demand.18 4.2 Effect Heterogeneities Frequently, studies report effect heterogeneities. We broadly highlight three factors: Macro-level economic conditions; poverty and disadvantage at the beneficiary level; and gender. One reason to focus on these three factors is that they may explain some portion of the discrepancy between pessimistic earlier reviews and the more hopeful results we have presented here. Broadly speaking, skills trainings tend to be more effective—and have historically been more effective—in less developed contexts, with poorer participants and female beneficiaries, a profile that the recent studies included in this review are more likely to fit than studies that informed earlier reviews, many of which were implemented in developed countries. While skills trainings have arguably gotten better about being implementation-intense and demand-responsive, we suspect this shift in who is being targeted is one reason for the distance between our summary of results and those of earlier systematic reviews. Regional or country-specific economic conditions Historically reviews have shown that program effects vary with the development level of a country. Despite the generally moderate impacts, Kluve et al. (2017) find systematic evidence of larger training effects on employment and earnings outcomes in lowand middleincome countries than in high-income countries. In advanced economies, demand for skilled labor is high and cohorts are overall relatively well-educated to start with, so a limited-duration skills intervention targeting disadvantaged workers might not be enough to help them catch up with their peers. 18 We see mixed evidence concerning apprenticeships in previous literature too. For example, Alzúa etal. (2016) report positive effects from a firm-linked training program in Argentina, while Cho etal. (2013) find no labor market impacts of an apprenticeship training program in Malawi. 109 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Both within and between developing countries, the performance of training programs should vary with labor demand conditions and the extent to which programs’ target skills match local market demand. Indeed, Ibarrarán etal. (2019) document greater returns to training investments in regions with higher demand for skilled labor in the case of a training program in the Dominican Republic. More generally, Escudero etal. (2019) find more pronounced positive effects of training programs on labor market outcomes in times of economic expansion in a meta-analysis of 296 impact evaluations of active labor market programs in Latin America and the Caribbean. An expanding, innovating economy often sees demand for skilled workers rise (Vivarelli, 2014).19 We do not see any obvious pattern of this type emerging from our systematic review, but the number of RCTs is still small and effectively covers just a small and non-random subset of sub-Saharan Africa’s many labor markets. Beneficiary-level disadvantage and poverty Following a similar logic, prior reviews have suggested that program activities targeted at particularly disadvantaged, poor beneficiaries tend to produce the largest effects. Training interventions are often narrow in scope and duration, so they might not move the needle much for better-off individuals, but they can make a difference for those starting out with more limited skill sets. Abebe etal. (2017), for example, detect the most substantial effects of an intervention including a job application workshop and transport subsidies in Ethiopia among the most disadvantaged participants. In their meta-analysis of interventions in Latin America and the Caribbean, Escudero etal. (2019) find that training programs increase employment rates and earnings in particular when they explicitly target poor individuals. Kluve etal., (2017, 2019) also find that disadvantaged, low-income individuals benefit most in their meta-analysis of youth employment interventions. We also see these patterns in our systematic review (Beaman etal., 2021) but not nearly as commonly as one might expect. This could be a function of only a surprisingly small share of studies reporting heterogeneous effects of this kind to begin with. Gender Heterogeneity by gender is by far the most commonly estimated and reported type of heterogeneous effect, in earlier studies and also in the set of RCTs in our systematic review. Earlier reviews indicated that, if anything and on average, women benefit slightly more from skills training than men. Escudero etal. (2019) find this divergence in their meta-analytic study of programs in Latin America and the Caribbean. Blattman and Ralston (2015) note that youth technical and vocational trainings seem to positively affect women but not men, although business skill programs appear to have similar effects for both. And Kluve etal. (2017) observe that effect sizes with respect to employment and earnings may be larger for young women than young men, but caution that this should not be taken to imply that programs targeting women will generally hold greater promise. There is in fact tremendous variation in how gender interacts with interventions across studies, and the extent to which beneficiaries of one gender outperform others may be heavily dependent on context. In many earlier studies, women appear to be 19 Hirshleifer et al. (2014) note that the links between contextual economic indicators—in their case regional unemployment rates—and demand for skilled labor can be ambiguous. 110 B.Beber et al. better served by a given intervention. For example, Chakravarty etal. (2016) show that women gained employment to a greater extent than men following a youth skills training and job placement program in Nepal, and Attanasio etal. (2011) find large effects on employment and earnings for women only after a training program in Colombia.20 But in other studies, effects are observed for male but not female participants. Ibarrarán etal. (2019) find an effect of youth trainings on formal employment for men only in the Dominican Republic. So do Alzúa etal. (2016) for a vocational training program including skills training and private sector internships in Argentina as well as Hirshleifer etal. (2014) in the context of short-term vocational skills programs in Turkey, with the latter also reporting an effect for men only on hours worked. In Yemen, McKenzie etal. (2016) too find an effect of a short skills training and subsequent internship placement program on hours worked for men only. In the studies included in our systematic review, heterogeneous effects across genders are relatively frequently reported, and here too women appear to benefit slightly more than men on average (Atkin etal., 2021, Crawford etal., 2021, and women in some trades in Hardy etal., 2019). However, counterexamples exist as well: Araar etal. (2019) shows that male-owned MSEs see greater revenue gains in an intervention featuring improved credit access and skills training for personnel. The reason why women benefit more from training interventions in some contexts could be that they have more to gain, as they usually start out with worse outcomes than men. In other contexts, constraints may still be binding. Cho etal. (2013), for instance, suggest that female participants suffer disproportionately from credit constraints and external shocks (pregnancy, illness or injury in the household, other family obligations), which make sustained investments in their own human capital difficult. Even if an effect is present for both men and women, different mechanisms could be underpinning each. For example, Kugler etal. (2015), a study in Colombia, found that participants in a training course were more likely to be enrolled in tertiary education programs more than a decade later. For women, vocational training appears to have relaxed credit constraints. For men, training improved their fieldspecific understanding of the returns to tertiary education. Further untangling gender-specific mechanisms across contexts remains a task for future research. 20 Note that Chakravarty etal. (2016) observe that effects on employment type, hours worked, and earnings did not differ across gender, and that the employment effect for women could be due to the fact that men started at a much higher level of employment and it was easier for women to make large marginal gains, or could stem from women’s course choices as employment effects were primarily observed in traditionally female-dominated training fields. Attanasio etal. (2011) report no gender difference in program impact on rates of formal employment, but warn that the effect for men could be attributable to attrition bias. 117 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Attanasio, O., Guarín, A., Medina, C., & Meghir, C., (2015). Long term impacts of vouchers for vocational training: Experimental evidence for Colombia. NBER Working Paper Series, Working Paper 21607, 1–38. Attanasio, O., Kugler, A., & Meghir, C. (2011). Subsidizing vocational training for disadvantaged youth in Colombia: Evidence from a randomized trial. American Economic Journal: Applied Economics, 3(3), 188–220. Bandiera, O., Bassi, V., Burgess, R., Sulaiman, M., Vitali, A., & Rasul, I. (2023). The search for good jobs: Evidence from a six-year field experiment in Uganda. National Bureau of Economic Research. Bandiera, O., Buehren, N., Burgess, R., Goldstein, M., Gulesci, S., Rasul, I., & Sulaiman, M. (2020). Women’s empowerment in action: Evidence from a randomized control trial in Africa. American Economic Journal: Applied Economics, 12(1), 210–259. Bardasi, E., Gassier, M., Goldstein, M., & Holla, A. (2021). The profits of wisdom: The impact of a business support program in Tanzania. The World Bank Economic Review, 35(2), 328–347. Beaman, L., Herskowitz, S., Keleher, N., & Magruder, J. (2021). Stay in the game: A randomized controlled trial of a sports and life skills program for vulnerable youth in Liberia. Economic Development and Cultural Change, 70(1), 129–158. Benhassine, N., McKenzie, D., Pouliquen, V., & Santini, M. (2018). Does inducing informal firms to formalize make sense? Experimental evidence from Benin. Journal of Public Economics, 157, 1–14. Bertrand, M., & Crépon, B. (2021). Teaching labor laws: Evidence from a randomized control trial in South Africa. American Economic Journal: Applied Economics, 13(4), 125–149. Betcherman, G., Godfrey, M., Puerto, S., Rother, F., & Stavreska, A. (2007). A review of interventions to support young workers: Findings of the youth employment inventory. The World Bank. Betcherman, G., Olivas, K., & Dar, A. (2004). Impacts of active labor market programs: New evidence from evaluations with particular attention to developing and transition countries. The World Bank. Bjorvatn, K., Cappelen, A. W., Sekei, L. H., Sørensen, E. Ø., & Tungodden, B. (2020). Teaching through television: Experimental evidence on entrepreneurship education in Tanzania. Management Science, 66(6), 2308–2325. Blalock, G., Nesin, B., & Simons, A. M. (2022). Developing sustainable supply chains: Evidence from entrepreneurship training in Ethiopia. Africa Journal of Management, 8(1), 36–58. Blattman, C., Fiala, N., & Martinez, S. (2020). The long-term impacts of grants on poverty: Nine-year evidence from Uganda’s youth opportunities program. American Economic Review: Insights, 2(3), 287–304. Blattman, C., & Ralston, L. (2015). Generating employment in poor and fragile states: Evidence from labor market and entrepreneurship programs. SSRN Electronic Journal. https:// doi. org/ 10. 2139/ ssrn. 26222 20 Blimpo, M. P., & Pugatch, T. (2021). Entrepreneurship education and teacher training in Rwanda. Journal of Development Economics, 149, 102583. Bloom, N., Eifert, B., Mahajan, A., McKenzie, D., & Roberts, J. (2013). Does management matter? Evidence from India*. The Quarterly Journal of Economics, 128(1), 1–51. Bloom, N., Mahajan, A., McKenzie, D., & Roberts, J. (2010). Why do firms in developing countries have low productivity? American Economic Review, 100(2), 619–623. Bloom, N., Mahajan, A., McKenzie, D., & Roberts, J. (2020). Do management interventions last? Evidence from India. American Economic Journal: Applied Economics, 12(2), 198–219. Brudevold-Newman, A. P., Honorati, M., Jakiela, P., & Ozier, O. (2017). A firm of one’s own: Experimental evidence on credit constraints and occupational choice. World Bank. Brudevold-Newman, A., & Ubfal, D. J. (2023). Returns to soft skills training in Rwanda. The World Bank. Bruhn, M., Karlan, D., & Schoar, A. (2018). The impact of consulting services on small and medium enterprises: Evidence from a randomized trial in Mexico. Journal of Political Economy, 126(2), 53. Card, D., Ibarrarán, P., Regalia, F., Rosas-Shady, D., & Soares, Y. (2011). The labor market impacts of youth training in the dominican republic. Journal of Labor Economics, 29(2), 267–300. Carnicelli, L., Kosonen, T., & Pirttila, J. (2022). The impacts of union-led training on working conditions: Experimental evidence from Mozambique. Policy Report. https:// doi. org/ 10. 4419/ 96973 254 Carranza, E., & McKenzie, D. (2024). Job training and job search assistance policies in developing countries. Journal of Economic Perspectives, 38(1), 221–244. Chakravarty, S., Lundberg, M. K. A., Nikolov, P. V., & Zenker, J. (2016). The role of training programs for youth employment in Nepal: Impact evaluation report on the employment fund. World Bank. 118 B.Beber et al. Chioda, L., Contreras-Loya, D., Gertler, P., & Carney, D. (2021). Making entrepreneurs: Returns to training youth in hard versus soft business skills. National Bureau of Economic Research. Cho, Y., & Honorati, M. (2014). Entrepreneurship programs in developing countries: A meta regression analysis. Labour Economics, 28, 110–130. Cho, Y., Kalomba, D., Mobarak, A. M., & Orozco Olvera, V. H. (2013). Gender differences in the effects of vocational training: Constraints on women and drop-out behavior. World Bank. Cooke, A., Smith, D., & Booth, A. (2012). Beyond PICO: The SPIDER tool for qualitative evidence synthesis. Qualitative Health Research, 22(10), 1435–1443. Crawford, E., Crost, B., Dube, O., Ferrah, S., Holmlund, M., Khan, N., Mvukiyehe, E., and Umar, N. S., (2021). A way out? Evidence from two trials of the Mafita apprenticeship and community-based skills training programs in northern Nigeria. World Bank, DIME Department, Endline Evaluation Report. Crépon, B., & Premand, P. (2019). Direct and indirect effects of subsidized dual apprenticeships. SSRN Electronic Journal. https:// doi. org/ 10. 2139/ ssrn. 34957 70 Cummings, D., & Bloom, D. (2020). Can Subsidized Employment Programs Help Disadvantaged Job Seekers? A Synthesis of Findings from Evaluations of 13 Programs. Washington, DC: Office of Planning, Research, and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services, No. 2020–23. Dalton, P. S., Rüschenpöhler, J., Uras, B., & Zia, B. (2019). Learning to grow from peers: Experimental evidence from small retailers in Indonesia. World Bank. Dammert, A. C., & Nansamba, A. (2023). Skills training and business outcomes: Experimental evidence from Liberia. World Development, 162, 106120. Datta, N., Assy, A. E., Buba, J., & Watson, S. M. (2018). Integration: A new approach to youth employment programs—general guidelines for project teams. World Bank. Del Carpio, L., & Guadalupe, M. (2022). More women in tech? Evidence from a field experiment addressing social identity. Management Science, 68(5), 3196–3218. Demerouti, E. (2023). Effective employee strategies for remote working: An online self-training intervention. Journal of Vocational Behavior, 142, 103857. Dimitriadis, S., & Koning, R. (2022). Social skills improve business performance: Evidence from a randomized control trial with entrepreneurs in Togo. Management Science, 68(12), 8635–8657. Escudero, V., Kluve, J., LópezMourelo, E., & Pignatti, C. (2019). Active labour market programmes in Latin America and the caribbean: Evidence from a meta-analysis. The Journal of Development Studies, 55(12), 2644–2661. Evans, L., Lund, C., Massazza, A., Weir, H., & Fuhr, D. C. (2022). The impact of employment programs on common mental disorders: A systematic review. International Journal of Social Psychiatry, 68(7), 1315–1323. Giustini, D., & Kamel Boulos, M. N., 2013. Google Scholar is not enough to be used alone for systematic reviews. Online Journal of Public Health Informatics, 5(2). Grimm, M., & Paffhausen, A. L. (2015). Do interventions targeted at micro-entrepreneurs and small and medium-sized firms create jobs? A systematic review of the evidence for low and middle income countries. Labour Economics, 32, 67–85. Gusenbauer, M., & Haddaway, N. R. (2020). Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of Google Scholar, PubMed, and 26 other resources. Research Synthesis Methods, 11(2), 181–217. Hardy, M., Mbiti, I., McCasland, J., & Salcher, I. (2019). The apprenticeship-to-work transition: Experimental evidence from Ghana. World Bank Policy Research Working. Hirshleifer, S., McKenzie, D., Almeida, R., & Ridao-Cano, C. (2014). The impact of vocational training for the unemployed: Experimental evidence from Turkey. The Economic Journal, 126(597), 2115–2146. Hirvonen, K., Gilligan, D. O., Leight, J., Tambet, H., & Villa, V. (2023). Do ultra-poor graduation programs build resilience against droughts? Evidence from rural Ethiopia. International Food Policy Research Institute. Ibarrarán, P., Kluve, J., Ripani, L., & Rosas Shady, D. (2019). Experimental evidence on the long-term effects of a youth training program. ILR Review, 72(1), 185–222. J-PAL. (2017). Skills training programs. Abdul Latif Jameel Poverty Action Lab. J-PAL. (2023). Vocational and skills training programs to improve labor market outcomes. 119 Employment Effects ofSkills Trainings inSub‑Saharan Africa:… Karimli, L., Lecoutere, E., Wells, C. R., & Ismayilova, L. (2021). More assets, more decision-making power? Mediation model in a cluster-randomized controlled trial evaluating the effect of the graduation program on women’s empowerment in Burkina Faso. World Development, 137, 105159. Kemper, J., Stöterau, J., & Ghisletta, A. (2022). The Impact of Vocational Training Interventions on Youth Labor Market Outcomes: A Meta-Analysis. SSRN Electronic Journal. https:// doi. org/ 10. 2139/ ssrn. 42175 80 Kimou, A. J. C., Ahoure, A., Djebbari, H., & Becho, I., (2022). Soft skills training and social inclusion of vulnerable youth in post-conflict Côte d’Ivoire: Evidence from a randomized control trial. Kluve, J., Puerto, S., Robalino, D., Romero, J. M., Rother, F., Stöterau, J., Weidenkaff, F., & Witte, M. (2017). Interventions to improve the labour market outcomes of youth: A systematic review of training, entrepreneurship promotion, employment services and subsidized employment interventions. Campbell Systematic Reviews, 13(1), 1. Kluve, J., Puerto, S., Robalino, D., Romero, J. M., Rother, F., Stöterau, J., Weidenkaff, F., & Witte, M. (2019). Do youth employment programs improve labor market outcomes? A quantitative review. World Development, 114, 237–253. Kugler, A., Kugler, M., Saavedra, J., & Prada, L. O. H. (2015). Long-term educational consequences of vocational training in Colombia: Impacts on young trainees and their relatives. National Bureau of Economic Research. Lafortune, J., Pugatch, T., & Ubfal, J. T. D. (2022). Can interactive online training make high school students more entrepreneurial? Experimental evidence from Rwanda. IZA DP No. 15064. Lafortune, J., Riutort, J., & Tessada, J. (2018). Role models or individual consulting: The impact of personalizing micro-entrepreneurship training. American Economic Journal: Applied Economics, 10(4), 222–245. Maïga, W. H. E., Porgo, M., Zahonogo, P., Amegnaglo, C. J., Coulibaly, D. A., Flynn, J., Seogo, W., Traoré, S., Kelly, J. A., & Chimwaza, G. (2020). A systematic review of employment outcomes from youth skills training programmes in agriculture in lowand middle-income countries. Nature Food, 1(10), 605–619. McIntosh, C., & Zeitlin, A. (2022). Using household grants to benchmark the cost effectiveness of a USAID workforce readiness program. Journal of Development Economics, 157, 102875. McKenzie, D., Assaf, N., & Cusolito, A. P. (2016). The demand for, and impact of, youth internships: evidence from a randomized experiment in Yemen. IZA Journal of Labor & Development. https:// doi. org/ 10. 1186/ s401750160048-8 McKenzie, D., & Woodruff, C. (2014). What are we learning from business training and entrepreneurship evaluations around the developing world? The World Bank Research Observer, 29(1), 48–82. McKenzie, D., & Woodruff, C. (2015). Business practices in small firms in developing countries. National Bureau of Economic Research. Mensmann, M., & Frese, M. (2019). Who stays proactive after entrepreneurship training? Need for cognition, personal initiative maintenance, and well-being. Journal of Organizational Behavior, 40(1), 20–37. Methley, A. M., Campbell, S., Chew-Graham, C., McNally, R., & Cheraghi-Sohi, S. (2014). PICO, PICOS and SPIDER: A comparison study of specificity and sensitivity in three search tools for qualitative systematic reviews. BMC Health Services Research, 14(1), 579. Morris, S., Carlos, C., Kistruck, G. M., Lount, R. B., & Thomas, T. E. (2023). The impact of growth mindset training on entrepreneurial action among necessity entrepreneurs: Evidence from a randomized control trial. Strategic Entrepreneurship Journal, 17(3), 671–692. Nkonya, E., Bashaasha, B., Kato, E., Bagamba, F., & Danet, M. (2019). Impact of creative capacity building of local innovators and communities on income, welfare and attitudes in Uganda: a cluster randomized control trial approach. 3ie Grantee Final Report. Novella, R., Rucci, G., Vazquez, C., & Kaplan, D. S. (2018). Training vouchers and labour market outcomes in Chile. Labour, 32(2), 243–260. Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J.M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., McGuinness, L. A., Stewart, L. A., Thomas, J., Tricco, A. C., Welch, V. A., Whiting, P., & Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. The BMJ, 372, n71. Quinn, S., & Woodruff, C. (2019). Experiments and entrepreneurship in developing countries. Annual Review of Economics, 11(1), 225–248. 120 B.Beber et al. Rosas, N., Acevedo, M. C., & Zaldivar, S. (2022). Starting points matter: Cash plus training effects on youth entrepreneurship, skills, and resilience during an epidemic. World Development, 149, 105698. Tarekegne, C., Wesselink, R., Biemans, H. J. A., & Mulder, M. (2023). The effects of comprehensive competence-based training on competence development and performance improvement of smallholder farmers: An Ethiopian case study. International Journal of Training and Development., 28, 119. Vicente, P. C., & Vilela, I. (2022). Preventing Islamic radicalization: Experimental evidence on antisocial behavior. Journal of Comparative Economics, 50(2), 474–485. Vivarelli, M. (2014). Innovation, employment and skills in advanced and developing countries: A survey of economic literature. Journal of Economic Issues, 48(1), 123–154. Wheeler, L., Garlick, R., Johnson, E., Shaw, P., & Gargano, M. (2022). LinkedIn(to) job opportunities: Experimental evidence from job readiness training. American Economic Journal: Applied Economics, 14(2), 101–125. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.