58601 World Bank Employment Policy Primer November 2010 No. 16 ACTIVE LABOR MARKET PROGRAMS FOR YOUTH A Framework to Guide Youth Employment Interventions Y outh are three times more likely to be either has been shown to have predominately posi- unemployed than adults, even in econo- tive impact, as measured by rigorous impact eval- mies with strong economic growth (ILO, uations,2 or has weaker evidence of impact--rig- 2008). This begs the question of what is it about orous evaluations with mixed evidence of impact youth that leads to such high rates of unemploy- or strong positive monitoring data--and is theo- ment? And what can be done to help young people retically sound. Cost-effectiveness information is more efficiently integrate into the labor market? presented when available.3 This Note is a tool to provide policymakers and The Note focuses on programs that are appro- youth-serving organizations with a framework to priate to address constraints faced by youth from better diagnose short- to medium-run constraints disadvantaged backgrounds.4 The sample of pro- facing the stock of unemployed youth and to design grams is supplemented by examples of non-youth evidence based youth employment interventions. While reducing youth unemployment in the long-run is a multi-dimensional approach requir- · Identify the target population and the constraints they face in ing a range of substantial reforms as proposed in finding employment, in the current macro, investment climate, and institutional environments the MILES framework (World Bank, 2007), this Step 1 Note considers only Active Labor Market Programs (ALMP) that are designed to enhance labor market · Using Table 1, select the set of interventions that correspond (re)integration within existing institutional and to the constraints identified in Step 1 macro-economic constraints, i.e. we take as given Step 2 labor market regulations, the investment climate, and the general education system. The Note only · Adjust the design according to the specific needs of the target population and the institutional and administrative capacity of addresses youth employment; strategies to affect the particular country or labor market wages, productivity, underemployment, or job Step 3 quality are not directly discussed. · Evaluate the evidence of the program impact for future Selection Criteria of the Evidence learning and improvement Step 4 This Note presents youth-oriented ALMPs that conform to one of two criteria.1 Each intervention *Prepared by Wendy Cunningham, Maria Laura Sanchez-Puerta, and Alice Wuermli (alphabetical). A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h specific programs that were found to have had economy, will lack all four categories of skills, face above average impacts for youth or other disadvan- severe gender discrimination, and be limited by tages groups (eg. women, ethnic minorities etc.). gender norms, while low-income men in urban Chile may be most constrained by information A `How to' Guide about job opportunities, difficulty in signaling competencies to potential employers, and by inap- The identification of appropriate short- and propriate technical and soft skills. Thus, the short- medium-term responses to support unemployed list of constraints for the reader's specific target youth requires a four-step process: population needs to be identified through a coun- At the end of the process, the reader should have try and labor market analysis. a short list of programs, appropriate for the needs of the target population and the target labor market, Step 2: Select the Interventions and that take into account country-specific factors. that Correspond to the Constraints Step 1: Identify the Target Population Once the employment constraints have been and the Constraints for Finding Jobs identified, we can identify interventions to help the target population overcome those barriers. Table Youth are not a homogenous group. Each 1 summarizes evidence-based program interven- sub-group has its own set of constraints that hin- tions that correspond to each of the constraints der its entry to the labor market. We observe this considered above. Using the list of target group through differential unemployment rates between constraints from Step 1 as a guide, a short list young women and men, youth of different ages, of potential program interventions can be drawn rural and urban dwellers, and so forth. Thus, for from Table 1. this exercise, it is necessary to identify the specific This section presents each constraint and the population(s) of interest. respective knowledge on suitable interventions to This Note proposed five general categories of help youth overcome those constraints. constraints that may limit young people's access to the labor market: Job-Relevant Skills Constraints A wide range of competencies are necessary to Job-relevant skills constraints, including insuf- perform a job well, including ficient basic skills, technical skills mismatch, behavioral skills mismatch, or insufficient basic skills--literacy and numeracy--are the entrepreneurial skills foundation of communication and further Lack of labor demand, both at the macro-level skills development processes. through slow job growth and at the micro- technical skills--trade- or job-specific skills level through employer discrimination range from manual skills to computer literacy. Job search constraints such as information behavioral skills--or noncognitive skills-- about job openings or being able to commu- consist of a range of personality traits such as nicate skills to potential employers motivation, socio-emotional regulation, time Firm start-up constraints including lack of management, and the ability to work with oth- access to financial capital or business networks ers (Heckman, 2008). Social constraints on the supply side, such as entrepreneurial skills--both the creativity to social norms that limit skills development or invent/adopt a new product or process and the labor market entry business skills to market the idea--are essen- tial for both the self-employed and employees. Each sub-population of interest will face a different set of constraints. For example, young Training interventions can yield promising women in rural Rwanda live in a low-growth results to overcome a range of skills deficits. 2 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Table 1: The Menu of Constraints and Interventions Constraints Possible ALMP Interventions Evidence-based interventions Mixed evidence, theoretically sound Insufficient basic · Informationaboutthevalueofeducation · Secondchanceprograms skills Technical skills · Training"plus"/comprehensiveprograms · On-the-jobtraining mismatch · Informationonreturnstotechnicalspecialties Job-relevant skills constraints Behavioral skills · Behavioralskillstraining mismatch Insufficient · Entrepreneurialtraining entrepreneurial skills Slow job-growth · Wageortrainingsubsidies · Publicserviceprograms economy · Labor-intensivepublicworks Lack of labor demand Employer · Affirmativeactionprograms · Subsidiestoemployerswhohiretargetgroups discrimination · Employeementoring Job matching · Employmentservices · Technology-basedinformationsharing Job search constraints Signaling · Skillscertification Competencies · Trainingcenteraccreditation Lack of access to · Comprehensiveentrepreneurshipprograms · Microfinance Firm start-up financial or social constraints capital Social Excluded-group · Targetexcluded-group'sparticipationinprograms · Adjustedprogramcontent/designtoaccountforexcluded-group constraints constraints · Non-traditionalskillstraining specificneeds on the supply (ethnicity, gender, · Safetraining/employmentspacesforspecific side etc) groups Insufficient Basic Skills they had estimated, transition to 9th grade was 4 Functional literacy and numeracy are basic percentage points higher than for boys who did needs for most employment and for further skills not receive the information (Jensen, 2010). development. While literacy rates continue to Second chance programs: Literacy and improve they are still as low as 72% for Sub Saha- numeracy programs, equivalency degrees, and ran Africa, and far lower in specific countries (eg. accelerated learning programs are designed to Benin 39%).5 Women may be particularly con- teach basic skills to youth who did not acquired strained as observed by significant gender dispari- them by the time they left school. Such programs ties in educational attainment in many developing may be as simple as adult literacy programs or countries. Potential interventions include: as complex as a package of cognitive and non- Information about the value of education. cognitive capacities. The programs may--and in Providing students (and perhaps their parents) fact should--provide a post-program certification with facts about the link between school and their to demonstrate to employers that the person has future work lives enables them to make informed acquired these basic skills (Mattero, forthcoming). decisions on their human capital investments. The evidence of impact of second-chance Simple information can be powerful. For example, education programs on employment prospects when 8th grade boys in the Dominican Republic is still developing. Participants in the JOBSTART were told that the average salary for completing (US) program had a higher likelihood of earn- secondary school was four times higher than what ing their General Education Degree (GED), the 3 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h `equivalent' to a high school diploma, as opposed sive, or "Training Plus," programs. These programs to a matched control group of non-participants combine pure technical training with behavioral- (Cave et al., 1993). But the long-term economic skills training, internships, employment services, benefits of earning a GED has been refuted (Heck- and accreditation. They have a strong emphasis man & LaFontaine; 2006). The second chance on demand-driven skills training, ensured by ex component in the Chilean program Chilecalifica is ante agreements with the private sector to provide designed to provide beneficiaries basic education internships to their graduates. Many provide wage and/or technical and vocational training. Unlike a subsidies or a subsidized training period, especially GED, the degree in Chile is indistinguishable from for young mothers. an ordinary school diploma. Preliminary impact A randomized experiment of the US Job Corps evaluation results suggest an increase in wages program found that participants earned 12 percent and greater post-program schooling of Chileca- more than the control group within 4 years of lifica beneficiaries, particularly for women (San- leaving the program, had lower arrest and convic- tiago Consultores, 2009), though the small sample tion rates, and had shorter incarceration duration. size and selectivity issues may positively bias the From society's perspective the program was only results. cost-effective for the group of participants age Monitoring data from other programs for 20­24, where total benefits were double the value girls, such as the Indian Better Life Options pro- of program cost (Schochet et al., 2006). gram that combines non-formal education, fam- A review of six "Jovenes" programs in Latin ily life education, life skills, vocational training, America have shown to increase employment health services and character development, or the by 0­5%, with women and younger participants Ehiopian Biruh Tesfa that creates safe spaces in exhibiting impacts of 6­12% in some countries. which they are taught literacy and other skills, Scarcer evidence suggests that these programs are seem to have increased literacy among participants cost-effective (Ibarraran & Rosas, 2009). (CEDPA, 2001; Elrukar et al., 2010). Program success requires a well-developed Making up for lost learning at the earlier stages national network of private vocational training of development may come at a much greater cost centers, a private sector willing to hire apprentices than the initial investment (Shonkoff & Phillips, (Cunningham et. al. 2008), and a positive macro- 2000). The average net benefit of the JOBSTART economic context coupled with job growth (Ibar- program was negative when considering taxpayer's raran & Rosas, 2009). and societal costs, though positive among women Information on returns to technical special- who were custodial mothers at program entry ties. People who have not worked before do not and among men with pre-program arrests. The have good information about which competen- program benefits may be underestimated since cies the market rewards. Providing this informa- they do not take into account, for example, lower tion to them can be low cost but very effective. fertility and arrest rates, and the intergenerational Preliminary results from the Jua Kali voucher benefits of education. (Cave et al., 1993). program in Kenya, which provided its female beneficiaries with information about wages in Technical Skills Mismatch various occupations, suggest that more than 10% A technical skills mismatch is when workers of those who received the information switched lack the trade- or job-specific skills demanded by to more lucrative (often `male') jobs as compared employers. Such skills range from manual skills to to girls who did not receive the information computer literacy. (Hjort, 2009). Training "plus"/comprehensive programs. On-the-job training may take the form of Institution-based vocational/technical training apprenticeships, internships, or learning-by-doing programs have not proven very successful in devel- while employed. This model has not been evalu- oping countries (Betcherman, 2004; Wodon & ated in developing countries but some developed Minowa, 2001), leading to the rise of Comprehen- country examples are instructive. For example, both 4 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Switzerland and Germany have longstanding post- gram participants have a greater ability to take secondary school vocational training system. While responsibility and work in teams than do their the two systems are structurally very similar, the other employees (entra21, 2009). apprenticeship training in Switzerland is profitable during the training period, while German firms face Insufficient Entrepreneurial Skills significant net costs (Dionisius et al., 2008). The dif- In addition to basic, technical, and behavioral ference is attributed to the higher value of produc- skills, entrepreneurs need creativity and managerial tive work given to apprentices in Switzerland. skills that enable them to "sort out good ideas from An intervention to consider given its strong bad ones, find the resources and means to create a theoretical base, and the large demand for appren- proto-type, and take the idea through its growth ticeships, is skilling-up Mastercraftsmen who lead phases" (World Bank, 2010b). Wadhwa et al. (2009) the massive, yet informal apprenticeship system in find that lack of business management skills and many African countries (eg. Frazer, 2006; Valen- knowledge of how to start a firm were a significant chik, 1995). By adopting more modern equipment obstacle to entrepreneurship in the US. and methods and setting official learning objec- Entrepreneurial training. A FINCA lending tives, apprentices will acquire more valuable skills program serving poor women with little formal (Peeters, et. al. 2009). education, in which business training was added in a randomized experiment, found that entrepre- Behavioral Skills Mismatch neurship skills training increased new entrepre- Employers across the world give increasingly neurs' business knowledge, improved their busi- higher priority to behavioral skills than to techni- ness practices, and led to greater business success cal competencies (Heckman et al., 2006; Blom & (Karlan et al., 2006). Most programs recognize Hobbs, 2008; World Bank, 2010a). They note the that entrepreneurial skills training is necessary but absence of such skills among new hires. But young posit that it needs to be complemented with other people are rarely explicitly taught these skills, thus business support. For more detail on comprehen- limiting their job prospects. sive entrepreneurship programs see the section Behavioral skills training. Behavioral skills below on `Firm start-up constraints'. for the labor market may be taught in school sys- tems, in training programs, or on-the-job. Raising Lack of Labor Demand Healthy Children, a program in the US in which Low demand for labor can stem from a wide elementary school teachers are taught proactive range of factors. It may be due to economy-wide classroom management, cooperative learning factors such as a difficult investment environment, methods, strategies to enhance student motiva- natural disasters, war, a sudden change in the tion, student involvement, and interpersonal and global economy, or patterns of trade, that may hin- problem-solving skills, has shown to reduce anti- der job growth. Or, it may be due to micro-factors, social behavior, and lead to stronger commitment such as employer discrimination. to school and higher academic performance of participants compared to the control group (Cata- Slow job-growth economy lano et al., 2003). More recently, these skills have In many countries, labor supply exceeds job been directly taught as a program in itself, linked creation. For example, the labor force in the to other training programs or linked to job train- Middle East and North Africa is projected to ing.6 For example, comprehensive training pro- grow by more than 4 million people per year grams, such as Entra 21 implemented in 18 Latin between now and 2030, requiring the creation of American countries, include a behavioral (life) more than 50 million new jobs by 2015 (World skills component.7 Bank, 2005). While governments cannot directly While there are no rigorous impact studies create jobs, there are short-run programs they showing that soft skills training alone increases can introduce while spurring longer-term private employability, employers report that Entra21 pro- sector development. 5 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h Wage or training subsidies are intended to civic engagement, more positive attitudes towards encourage hiring by lowering the cost of new employment, and a higher likelihood of public workers for firms that face employment inhibiting service careers, but no significant increase in edu- budget constraints. This is a particularly relevant cational attainment as compared to a control group scheme for youth, whose marginal productivity (Frumkin et al., 2009). Youth Service Canada was may be below market wages. said to have had both positive impacts on post- Wage subsidies in Poland and the Czech program employment as well as further education.8 Republic had a positive impact on employment, Labor-intensive public works programs are especially among women and the less educated an increasingly popular mechanism for address- (Betcherman et al., 2007). An impact evaluation ing youth unemployment and are generally of a randomized experiment of the Proempleo aimed at providing a cash transfer to people in program in Argentina showed significant increases exchange for construction or rehabilitation of in the probability of being employed, largely due public infrastructure (Grosh et al., 2008). While to the impact on women and young participants Trabajar in Argentina showed sizeable income (Galasso et al., 2002). gains for younger participants (Jalan and Rav- While these interventions seemed to be cost- allion, 1999), Gilligan et al. (2008) found lit- effective in Poland and Argentina, we may be tle impact in Ethiopia, perhaps because actual underestimating the costs. Deadweight loss and transfer levels were very low. The larger literature `substitution' effects (employers substituting work- finds inconclusive evidence on the impact of ers with subsidized ones) may not always be ade- public works programs at increasing employ- quately accounted for and employment beyond the ment beyond the program duration (Betcher- subsidized period is questionable (Betcherman et man et al., 2007). However, once participants' al., 2004). opportunity costs are fully considered, alter- A less tested model is to subsidize training by native policies may be more cost-effective at providing students with an `income' while they reaching the short-term employment or poverty are increasing their human capital. For instance, alleviation objective (Ravallion, 1999; Ravallion Training Plus programs typically pay a subsidy & Datt, 1995; Datt & Ravallion, 1994). or stipend. Evidence on the impact of cash trans- To increase the impact of public works on fers conditional on secondary school attendance post-program employment of youth, a new model (CCT), such as Oportunidades in Mexico which is being tested in several countries. These programs increased secondary school attendance by approxi- add mandatory behavioral skills, financial literacy, mately 10­30% (Fiszbein & Schady, 2009), suggests or job search training to the public works project. that this might be a model to explore for second Evaluation results for Kenya and Sierra Leone are chance education programs or training programs. expected in 2013 (World Bank, 2010c, World Bank, Public service programs. Youth service, for- 2010d). mal or informal, provides an opportunity for youth to "play an active role in community and Employer Discrimination national development while learning new skills, Even if jobs exist, employers may have biases increasing their employability, and contributing to about hiring youth. For example, an employer their overall personal development" (Cunningham study in Sierra Leone revealed that employers per- et al., 2008). Services may include providing basic ceive youth age 18 to 24 as being less reliable, less health services in public health clinics, building trustworthy, less hard working, less cooperative, sustainable housing, literacy tutoring, protecting and less skilled than adults (Peeters et al., 2009). the environment, and building small-scale infra- In addition to biases against age, hiring prefer- structure, for example. ences may be made along gender, racial, ethnic or A quasi-experimental evaluation of the public religious lines. service program Americore (US) showed that par- Affirmative action programs. Laws that ticipants had a greater incidence of post-program reward employers for hiring commonly discrimi- 6 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s nated against groups or punish those who are result of difficulties in communicating capacities suspected of bias not related to productivity, to a potential employer (signaling). have effectively helped disadvantaged groups in developed countries with little cost to employers Job Matching (Holzer & Neumark, 2000a; Holzer & Neumark, Employers mainly use informal networks to 2000b; Price, 2002). The impact depends largely find new workers, such as family, friends or cur- on enforcement. There is not yet evidence of suc- rent employees (Cunningham et al., 2008; Peeters cessful implementation in developing countries, et al., 2009). These sources provide good infor- though several countries, such as Brazil and South mation about new employees, which is especially Africa, are experimenting with these programs. important in labor markets with high firing costs. Subsidies to employers who hire target Youth, who are new to the labor market, gener- groups. Employers may not hire youth based on ally lack these networks thereby limiting their job the assumption that youth are less productive options. since they have less work experience. Subsidies to Employment services. Job intermediation employers may encourage hiring of young workers centers provide information about job openings. since they compensate the employer's (perceived) In the most basic centers, potential employers post risk of low-productivity until the worker can dem- job openings and potential employees reply direct- onstrate her real productivity (Isbell & Smith, ly to employers of interest. At the other end of the 1991). While not a youth program, a wage voucher spectrum are full service centers, such as the UK program in Dayton, USA, in the early 1980s in Job Centers Plus, which offer a range of services, the presence of rising unemployment rates, was including career counseling, training or education unsuccessful and seemed to have a stigmatizing program placement assistance, job matching, labor effect (Burtless, 1985). In contrast, evidence from exchanges, and other related services.10 the Australian Special Youth Employment Training The New Deal for Young People in the UK, Program indicates lasting impact on employment a mandatory program for unemployment insur- prospects, both through retention of initially sub- ance recipients, provides a combination of wage sidized jobs and enhanced employability (Richard- subsidies to employers and job search assistance. son, 1998). Blundell et al. (2004) find an increase of finding Employee mentoring. Support to new labor a job of 5% on top of a preprogram baseline of market entrants by more experienced, and success- 26% for male participants, of which estimated 1% ful, members of a discriminated group may facili- stemmed purely from job search assistance. They tate entry into a seemingly constrained labor mar- are particularly effective when targeted at disad- ket. Price (2002) finds qualitative evidence in US vantaged workers with little access to the informal highway construction affirmative action programs search channels and when they are privately run of the effectiveness of support groups and one- but publicly funded through an incentive system to-one guidance on successfully integrating and requiring that the private training providers who retaining women in non-traditional trades. Such receive grants have secured a job for the trainee mechanisms are increasingly being used by NGOs, (see World Bank, 2009a). for example, to increase women's employment in Technology-based information sharing. more conservative Islamic legal environments.9 Instead of physical job offices, use of internet, radio, or cell phone text messaging may pro- Job-Search Constraints vide an opportunity to offer employment services In wage economies, young people and poten- across larger areas. SoukTel provides a platform tial employers may have difficulty finding each for firms and potential employees to connect via other due to information failures. This may stem text messages in places where young people are from lack of information about the existence of far more likely to have a cell phone (approx. 85%) potential employees and employers (job match- than internet access (approx 35%), such as Soma- ing). Or, even if they find each other, it may be a lia and Palestine. There is no evidence to date 7 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h involving a control group. However, SoukTel's tem in many Africa countries, certifying skills--as growth experience and financial self-sustainabil- opposed to a completed course--may be particu- ity make it an attractive option. larly useful. Programs to support the job-search and Training center accreditation. Youth may matching process should not discriminate based spend significant resources on job training, but on the legal status of the firm. Excluding the in many countries much of the training occurs in informal sector will drastically reduce the number unregistered, small scale, and often informal insti- of jobs advertised. For example, private training tutions (van Adams, 2008). A national system of providers in the Life Skills Education for Employ- training-provider accreditation may provide infor- ment and Entrepreneurship (World Bank, 2009b) mation to young people to make better decisions program in Indonesia train, certify, and match about which institutions to attend and informa- unemployed youth with domestic and overseas tion to potential employers about the types of skills jobs in either the formal or informal sectors (see that an institution's graduates should possess. The also World Bank, 2009a). impact of training center accreditation on employ- ment has not been measured. Signaling Competencies Workers may have the right skills, but it may Firm start-up constraints be difficult to communicate these skills to potential Youth often look to self-employment as an employers. The most commonly used signals of a alternative to tight labor markets. In addition to person's employability are whether he/she has held entrepreneurial skills constraints, as discussed a job, or has a school or training institution certifi- in the above section on "job-relevant skills con- cate. The former is a particular challenge for young straints," limitations of financial capital and social people who are just entering the labor market. And capital (business relationships) may be important. the latter is a challenge for poor youth, who have Unfortunately, there does not seem to be any evi- higher school dropout rates and may not have the dence identifying which constraints most affect resources to obtain a certificate. young people's ability to start a business. Skills certification. Whether a person acquires Comprehensive entrepreneurship pro- skills through formal or informal mechanisms, the grams. The sparse evidence to date suggests that fact that he or she has certain abilities can often be comprehensive programs that provide a com- measured against a generally accepted standard. bination of business management and leader- One might argue that skills certification is superior ship skills, financial literacy, micro-credit and/ to certification of completion of a level of schooling or insurance schemes, and the opportunity to or training, as it measures what the person actually network with other new and established entre- learned and is not concerned about how or when preneurs, successfully support new firm owners. such skills were acquired (Bouder et. al., 2008). For example, Peru's Formación Empresarial de la To date, there is little evidence about the effec- Juventud, and Calificación de Jóvenes Creadores de tiveness of skills certification in increasing young Microempresas, as well as TechnoServe in Central people's access to employment. However, labor America suggest that key factors of success are markets in OECD countries all have established access to credit combined with intense business some sort of skills certification scheme. Many mentoring. Klinger and Schuendeln (2007) find developing countries have started creating cer- a higher probability of starting a business and tification systems. For example, Chile has been expanding an existing business when receiving experimenting with certification of training pro- training by 9­11% (statistically insignificant) and grams within Chilecalifica; the program has a posi- 23­26% respectively of TechnoServe participants. tive impact, but the evaluation may overestimate Beneficiaries of Formación Empresarial de la the program's success due to sample selection bias. Juventud have a 40 percentage point higher prob- In markets with extensive informal skills acqui- ability of having a successful business (Jaramillo sition, such as the informal apprenticeship sys- & Parodi, 2003). 8 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Cost-effectiveness is questionable, as dead- tions, but they may be applicable to other groups weight loss and displacement costs may be substan- facing similar constraints due to social norms. tial (Betcherman et al., 2004). In addition, take-up Most of the suggestions made in this section relate rates of entrepreneurship programs are often low to design features and have yet to be evaluated with and failure rates of new firms are high--regardless regards to their effectiveness at increasing youth of whether they received assistance or not (Betch- employment rates. Nevertheless, they are prom- erman et. al., 2004). Nevertheless, these types of ising and have been proven effective to increase programs are essential in stimulating innovation participation of other excluded groups. and job growth (World Bank, 2010b) and, thus, Target young women's (effective) participa- merit further experimentation and learning via tion in programs. Attempting to make programs program impact evaluations. `gender neutral', for instance by implementing Microfinance. Financial capital constraints quotas, will not necessarily lead to gender parity. can be significant to new entrepreneurs, making For example, anecdotal evidence from a skills and microfinance an indispensible component of an microfinance program in Uganda that required 1/3 entrepreneurship program.11 As a standalone inter- of each team be female in order to qualify suggests vention, however, the evidence of positive impact is that while women were registered and on-site, they very mixed, and micro-lending to youth is still in its may not have actively participated in the skills infancy. The heterogeneity of program designs, cli- training activities (Blattman et al, 2010). Results ents, and country, regional and local contexts leads from more rigorous testing of this hypothesis is to a whole range of outcomes and impacts (for a expected by 2012. Women may need to be explic- review of studies see Goldberg, 2005).12 itly targeted, sought out, and actively encouraged to participate (Katz, 2008). Social Constraints Non-traditional skills training. Women tend Local customs and social norms may act as to be concentrated in traditionally female occupa- a deterrent for certain groups of people to par- tions which have been shown to pay less, even after ticipate in employment programs or the labor controlling for measurable characteristics of work- market, even though employers may be willing to ers and occupations (Blau & Kahn, 2000). Training hire and clients may be willing to buy from them. women in modern or `traditionally male' skills For instance, women's income earning opportuni- can increase their employment options (see for ties may be limited by competing homecare and example Jua Kali under `Information on returns home production responsibilities, men's decision- to technical specialties'). Pre-training courses can making power over women's time use and skills prepare women for training in male dominated development possibilities, or self-discrimination trades, while also providing skills to navigate com- that limits women's perceptions about the types peting in men's trades (Katz, 2008). of activities that are suitable (see for example Safe training/employment spaces for girls. Bryceson, 2002). Also, occupational segregation Women's mobility outside of the house may be along racial, ethnic, caste, or religious lines is not limited due to religious beliefs (eg. purdah) or uncommon (Price, 2002; Moyo & Kawewe, 2002). concerns about gender based violence. In the What Appaduray (2004) identifies as a lack of short-term, gender segregated transportation and "capacity to aspire, conceived as a cultural capacity, spaces in which girls can learn and work could especially among the poor" or other discriminated convince husbands and parents to let their wives against groups, may lead to self-exclusion from and girls participate. Studies have also shown that certain programs, occupations, or jobs. such female exclusive environments have their While the interventions outlined in Table 1 own dynamic and can have a significant impact may be appropriate for excluded groups, design on girls' self-esteem and confidence, with posi- features may need to be tweaked in order to reach tive implications for labor market success (see for some excluded youth. The evidence cited in this example CEDPA. 2001; Elrukar et al., 2010; Gili- section is drawn from gender oriented interven- brand et al., 1999). 9 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h Adjustment of program content and design courses provided by employment offices can be to account for time use constraints and other complemented by child care services when target- gender specific needs. Most women's time use ing unemployed mothers. patterns do not allow for full time work dur- If program design cannot be adjusted to ing daytime hours (Benería, 2003). Alternative account for these realities, the program should arrangements, such as part-time work, job-share, be dropped from the short list. For example, if non-contiguous work hours, or non-standard the target group is youth living in remote areas work hours may provide market access opportu- where there is no paid employment, wage subsidies nities. In addition, Community or center-based will not be an appropriate strategy to overcome childcare services and other social support systems slow job-growth. This intervention should thus be can free up women's time. At the most basic, these dropped from the short-list. are community-based arrangements where women At the end of Step 3, the short-list of program may trade childcare duties with each other. options will include interventions that: are appro- priate to the challenge facing the target population Step 3: Adjust the Design According in the relevant labor market, reflect the world's to Country and Target-Group Factors knowledge based on the limited evidence base to date, and have been adjusted for the specificities By the end of Step 2, the reader should have a of the context and the target group in the labor short-list of evidence-based programs that address market of interest. This list is ready for discussion the constraints of the target population. Given the with stakeholders. economic, social, institutional, and administrative diversity within and across countries and the spe- Step 4: Evaluate the Program Impact cific needs of the target group, all the interventions on the list of promising options will not necessarily The ultimate test of whether the correct set of be feasible in a specific country context. Two next programs was selected is to evaluate program impact steps are necessary to refine the list. and calculate benefit-cost ratios. Interventions that The first is to learn more about the necessary may have proven effective elsewhere may have great- economic, social, institutional, and administrative er or lesser effectiveness once program design has conditions for each program's success and assess been adjusted (Step 3) and the program operates whether those conditions exist in the target coun- under the specific circumstances of the country, try or labor market. And if these necessary condi- labor market, and target group. tions are absent, whether the program design can To ensure that a new program is a good use be adjusted such that the program is feasible. For of resources and of young people's time, a pro- example, the Jovenes programs in Latin America gram evaluation should be built into the program and the Caribbean were able to take advantage of design. This requires specialized researchers and an existing large public training sector. In countries budget. However, there are numerous resources on that lack this institutional capacity, comprehensive how to conduct rigorous evaluations (Heckman et training programs to address skills deficits will have al., 1999; Ravallion, 2008; Baker, 2000; Duflo et al., to find alternative ways to provide both the physical 2006; Ravallion, 1999a; sources cited in footnote infrastructure as well as the manpower to deliver 213), and increasingly there are independent fund- the training. The studies cited in the presentation ing sources to allay evaluation costs. of each program type provide information about Findings from program evaluation have pro- program design. vided the basis to adjust program design to better Second, the list should be further refined to meet objectives. They have been used to inform only include programs that are appropriate for the potential program participants and stakeholders target group or for which the program design can (including funders) of the effectiveness of a par- be adjusted such that it meets the needs while not ticular program design. And they are increasingly creating perverse incentives. For example, training the basis for new policy, as presented in this Note. 10 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Only through experimentation and learning can tion and training: Volume 2. Fourth report on we more fully populate Table 1 and better support vocational training research in Europe: Back- the short- and medium-term employment oppor- ground report. Cedefop (ed) Luxembourg: tunities for today's young people.14 EUR-OP Bryceson, D. F. (2002). The Scramble in Africa: References Reorienting Rural Livelihoods. World Develop- ment, Vol. 30, No. 5., pp. 725­739. Appaduray, A. (2004). The Capacity to Aspire: Cul- Burtless, G. (1985). Are Targeted Wage Subsidies ture and the Terms of Recognition. In Rao, V. Harmful? Evidence from a Wage Voucher & M. Walton (eds.), Culture and Public Action. Experiment. Industrial and Labor Relations Stanford, CA: Stanford University Press. Review, Vol. 39, No. 1, pp. 105­114. Baker, J. (2000). Evaluating the Impact of Develop- Catalano, R., Mazza, J., Harachi, T., Abbott, R., ment Projects on Poverty: A Handbook for Prac- Haggerty, K. & Fleming, Ch. (2003). Raising titioners. Washington, DC: The World Bank. Healthy Children through Enhancing Social Benería, L. (2003). Gender Development and Glo- Development in Elementary School: Results balization: Economics as if all People Mattered. after 1.5 Years. Journal of School Psychology, New York: Routledge. Vol. 41, No. 2, pp. 143­164. Betcherman, G., Godfrey, M., Puerto, S., Rother, F. Cave, G., Bos, H., Doolittle, F. & Toussaint, C. & Stavreska A. (2007). A Review of Interven- (1993). Jobstart: A final Report on a Program tions to Support Young Workers: Findings of the for Schooldropouts. Manpower Demonstration Youth Employment Inventory. SP Discussion Research Corporation. Paper No. 0715. The World Bank. CEDPA (2001). Adolescent Girls in India ­ Choose a Betcherman, G., Olivas, K. & Dar, A. (2004). Impacts Better Future: An Impact Assessment. of Active Labor Market Programs: New Evidence Santiago Consultores (2009). Evaluación en Pro- from Evaluations with Particular Attention to fundidad Chilecalifica. Chile. Developing and Transition Countries. Social Cunningham, W., Cohan, L., Naudeau, S., & Protection Unit, Human Development Net- McGinnis, L. (2008). Supporting Youth at Risk: work. The World Bank. A Policy Toolkit for Middle-Income Countries. Blattman, Ch., Fiala, N. & Martinez, S. (2010). Washington, DC: The World Bank. Delivering Opportunities to Youth in Northern Cunningham, W., McGinnis, L., García Verdú, R., Uganda: Short Term Impacts of NUSAF YOP on Tesliuc, C., & Verner, D. (2008). Youth at Risk Non-Migrant Youth. Preliminary Results. Pre- in Latin America and the Caribbean: Under- sentation, 27th May, 2010. standing the Causes, Realizing the Potential. Blau, F. & Kahn, L. (2002). Gender Differences in Washington, DC: The World Bank. Pay. The Journal of Economic Perspectives, Vol. Datt, G. & Ravallion, M. (1994). Transfer Benefits 14, No. 4, pp. 75­99. from Public-Works Employment: Evidence for Blom, A. & Hobbs, C. (2008). School and Work Rural India. The Economic Journal, Vol 104, in the Eastern Caribbean: Does the Education No. 427, pp. 1346­1369. System Adequately Prepare Youth for the Global Dionisius, R., Muehlemann, S., Pfeifer, H. et al. Economy? Washington, DC: The World Bank. (2008). Cost and Benefit of Apprenticeship Blundell, R., Meghir, C., Costa Dias, M., & Van Training: A Comparison of Germany and Swit- Reenen, J. (2004). Evaluating the Employment zerland. IZA DP 3465. Impact of a Mandatory Job Search Program. Duflo, E., Glennerster, R. & Kremer, M. (2006). Using Journal of the European Economic Association, Randomization in Development Economics Vol. 2, No. 4, pp. 569­606. Research: A Toolkit. Found on July 2nd 2010 at Bouder, A., Dauty, F., Kirsch, J.-L. & Lemistre, Ph. http://www.povertyactionlab.org/sites/default/ (2008). Readability of qualifications: a question files/documents/Using%20Randomization%20 as old as Europe. Modernising Vocational educa- in%20Development%20Economics.pdf. 11 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h Elrukar, A., Semunegus, B. & Mekonnen, G. (2010). Goldberg, N. (2005). Measuring the Impact of `Biruh Tesfa' Program Brief. Population Coun- Microfinance: Taking Stock of What we Know. cil, Ethiopia. Grameen Foundation USA Publication Series. EMpower (2010). It's Her Business: A Handbook Grosh, M., del Ninno, C., Tesliuc, E & Ouerghi, for Preparing Young, At-Risk Women to Become A. (2008). For Protection and Promotion: The Entrepreneurs. Found on May 28th at http:// Design and Implementation of Effective Safety itsherbusiness.org/?page_id=72. Nets. Washington DC: The World Bank. Entra21 (2009). Final Report on the entra21 Pro- Heckman, J. (2008). Schools, Skills, and Synapses. gram Phase I 2001-2007. International Youth UCD Geary Institute Discussion Paper Series. Foundation. Heckman, J., Lalonde, R. & Smith, J. (1999). The Field, E., Jayachandran, S., & Pande, R. (2010). Economics and Econometrics of Active Labor Do Traditional Institutions Constraint Female Market Programs. Handbook of Labor Econom- Entrepreneurship? A Field Experiment on ics, Vol. 3. Edited by A. Ashenfelter & D. Card. Business Training in India. American Economic Heckman, J. & LaFontaine, P. (2006). Bias-correct- Review: Papers & Proceedings. Vol. 100, Nr. 2, ed Estimates of GED Returns. Journal of Labor pp. 125­129. Economics, Vol. 24, No. 3. Filmer, D., King, E., Pritchett, L. (1998). Gender Heckman, J., Stixrud, J. & Urzua, S. (2006). The Disparity in South Asia: Comparisons within Effects of Cognitive and Noncognitive Abilities Countries. World Bank Working Paper, 1867. on Labor Market Outcomes and Social Behav- Fiszbein, A. & Schady, N. (eds.) (2009). Conditional ior. Journal of Labor Economics, Vol. 24, No. 3. Cash Transfer: Reducing Present and Future Hjort, M. (2009). Presentation at `Impact Evalua- Poverty. Washington, DC: The World Bank. tion Workshop'. Bonn, Germany, April 2009. Frazer, G. (2006). Learning the Master's Trade: Holzer, H. & Neumark, D. (2000a). What Does Apprenticeship and Human Capital in Ghana. Affirmative Action Do?. Industrial and Labor Journal of Development Economics, Vol. 81, No. Relations Review, Vol. 53, No. 2, pp. 240­270. 2, pp. 259­298. Holzer, H.& Neumark, D. (2000b). Assessing Affir- Frumkin, P., Jastrzab, J., Vaaler, M. et al. (2009). mative Action. Journal of Economic Literature, Inside National Service: AmeriCorps' Impact Vol. 38, No. 3. on Participants. Journal of Policy Analysis and Ibarraran, P. & Rosas, D. (2009). Evaluating the Management, Vol. 28, No. 3, pp. 394­416. Impact of Job Training Programs in Latin Galasso, E., Ravallion, M., & Silvia, A. (2002). America: Evidence from IDB Funded Oper- Assisting the Transition from Workfare to Work: ations. Journal of Development Effectiveness, A Randomized Experiment. Policy Research Vol.. 1, No. 2, pp. 195­216. Working Paper. World Bank. ILO (2008). Global Employment Trends for Youth. Gillibrand, E., Robinson, P. et al. 1999. Girls` Par- October s2008, International Labour Office. ticipation in Physics in Single-Sex Classes in Geneva: ILO. Mixed Schools in Relation to Confidence and Isbell, S. & Smith, L. (1991). Wage Subsidies as Achievement. International Journal of Science an Anti-discrimination Policy. Population Education, Vol. 21, No. 4, pp. 349­362. Research and Policy Review, Vol. 10, pp. 257­ Gilligan, D. O., Hoddinott, J. & Taffesse, A. S. 271. (2008). The Impact of Ethiopia's Productive Jalan, J. & Ravallion, M. (1999). Income Gains to the Safety Net Program and its Linkages. IFPRI Poor from Workfare: Estimates for Argentina`s Discussion Paper 00839. Trabajar Program. Policy Research Working Goetz, A.M. & Gupta, R.S. (1996). Who Takes Paper. World Bank. the Credit? Gender, Power, and Control over Jaramillo, M. & Parodi, S. (2003). Jóvenes Emprend- Load Use in Rural Credit Programs in Ban- edores. Instituto Apoyo. Lima, Peru. gladesh. World Development, Vol. 24, No. 1, Jensen, R. (forthcoming, 2010). The Perceived pp. 45­63. Returns to Education and the Demand for 12 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Schooling. Quarterly Journal of Economics, Vol. to Evaluation. Policy Research Working Paper. 125, No. 2. Washington, DC: The World Bank. Karlan, D., Harigaya, T., & Nadel, S. (2006). Evalu- Richardson, J. (1998). Do wage subsidies enhance ating Microfinance Program Innovation with employability? Evidence from Australian Youth. Randomized Controlled Trials: Examples from Discussion Paper no. 387, April 1998, Centre Business Training and Group versus Individual for Economic Performance, London School of Liability. 2006 Microcredit summit. Economics. Katz, E. (2008). Programs Promoting Young Wom- Schochet, P., Burghardt, J. & McConnell, S. (2006). en's Employment: What Works? Found at http:// National Job Corps Study and Longer-term Fol- siteresources.worldbank.org/INTGENDER/ low-up Study: Impact and Benefit-Cost Findings Resources/GenderYouthEmploymentKatz.pdf. Using Survey and Summary Earnings Records Klinger, B & Schuendeln, M. (2007). Can Entrepre- Data. Mathematica Policy Research, Inc. neurial Activity be Taught? Quasi-Experimental Sen, A. (1995). The Political Economy of Targeting. Evidence from Central America. CID Working In Van de Walle, D. & Nead, K. (eds.), Public Paper No. 153, Harvard University. Spending and the Poor. Johns Hopkins Univer- Mattero, M. (forthcoming). Second Chance Educa- sity Press. tion: A Conceptual Framework and Review of Shonkoff, J. & Phillips, D. (2000). From Neurons Programs. World Bank/HDNCY. to Neighborhoods: The Science of Early Child- Moyo, O. & Kawewe, S. (2002). The Dynamics of hood Development. Washington, DC: National a Racialized, Gendered, Ethnicized, and Eco- Academy Press. nomically Stratified Society: Understanding UNESCO (2008). International Literacy Statistics: the Socio-Economic status of Women in Zim- Review of Concepts, Methodology and Current babwe. Feminist Economics, Vol. 8, No. 2, pp. Data. Montreal: UNESCO Institute for Sta- 163­181. tistics. Peeters, P., Cunningham, W., Acharya, G. & Van Valenchik, A. (1995). Apprenticeship Contracts, Adams, A. (2009). Youth Employment in Sierra Small Enterprises, and Credit Markets in Leone: Sustainable Livelihood Opportunities in Ghana. The World Bank Economic Review, Vol. a Post-Conflict Setting. Washington, DC: The 9, No. 3, pp. 451­475. World Bank. Van Adams, A. (2008). Skills Development in the Price, V. (2002). Race, Affirmative Action, and Informal Sector of Sub-Saharan Africa. The Women's Employment in US Highway Con- World Bank. Draft. struction. Feminist Economics, Vol. 8, No. 2, Wadwha, V., Aggarwal, R., Holly, K. & Salkever, A. pp. 87­113. (2009). The Anatomy of an Entrepreneur: Mak- Ravallion, M. & Datt, G. (1995). Is Targeting ing a Successful Entrepreneur. Kauffman: The through a Work Requirement Efficient? Some Foundation of Entrepreneurship. Evidence for Rural India. In Van de Valle, D. & Wodon, Q. & Minowa, M. (2001). Training for the Nead, K., Public Spending and the Poor: Theory Urban Unemployed: A Reevaluation of Mexi- and Evidence. Baltimore: The Johns Hopkins co's Training Program, Probecat. MPRA Paper University Press. 12310. Ravallion, M. (2008). Evaluating Anti-Poverty Pro- World Bank (2010a). Indonesia Skill Report: Trends grams. Handbook of Development Economics, in Skill Demand, Gaps and Supply in Indonesia. Vol. 4, Schultz, T. P. & Strauss, J. (eds). Amster- di Gropello, E. (ed.), Green Cover Report. dam: Elsevier. World Bank (2010b). Stepping up Skills for More Ravallion, M. (1999). Appraising Worfare. The Jobs and Higher Productivity. Washington, DC. World Bank Research Observer, Vol. 14, No. 1, World Bank (2010c). Project Appraisal Document pp. 31­48. on a Proposed Credit in the Amount of SDR 7.5 Ravallion, M. (1999a). The Mystery of the Vanish- Million and Grant in the amount of SDR 6.2 ing Benefits: Ms. Speedy Analyst's Introduction Million to the Government of Sierra Leone for a 13 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h Youth Employment Support Project. Report No. Indonesia. Project Information Document 54419-SL. The World Bank, Washington DC. (PID). World Bank (2010d). Project Appraisal Document World Bank (2007). Miles to Go: A Quest for an on a Proposed Credit in the Amount of SDR Operational Labor Market Paradigm for Devel- 38.7 Million to the Republic of Kenya for a oping Countries. World Bank/HDNSP, draft. Kenya Youth Empowerment Project. Report No. World Bank (2006). World Development Report 53090-KE. The World Bank, Washington DC. 2007: Development and the Next Generation. World Bank, (2009a). Creating Alternative Path- Washington, DC: The International Bank for ways to Employment: An Assessment of the KPP Reconstruction and Development. Program. World Bank (2005). Sector Brief. Social Protection World Bank (2009b). Life Skills Education for in MENA: Regional Issues and Strategic Direc- Employment and Entrepreneurship (LSE3). tions. The World Bank. 14 A F r a m e w o r k t o G u i d e Yo u t h E m p l o y m e n t I n t e r v e n t i o n s Endnotes 3 Only few impact evaluations also analyze cost-effec- tiveness. The information cited on a program's cost- 1 effectiveness is taken from the program's evaluation Only a small set of programs are presented in this and, therefore, is based on the costs and benefits that Note due to the limited availability of rigorous evalu- a specific evaluation considered important. ations of ALMPs for youth in developing countries. In 4 In some countries, particularly in the Middle East and fact, of all the ALMPs worldwide, only a small fraction North Africa Region, a large percentage of unemployed of interventions have undergone rigorous evaluation. youth are university educated (World Bank, 2006). And, most programs that have undergone such evalua- These youth, who have had more access to opportuni- tions are from OECD countries. Only in the past decade ties, are not the target population for this Note. Many have substantial investments been made in evaluat- of their primary constraints are best overcome by ing the effectiveness of interventions in developing larger structural changes outlined in the MILES frame- countries. 2 Rigorous evaluations are those that convincingly work. However, several of the constraints and program options discussed in this note may provide short-term solve the problems of attribution and identification solutions for them, as well. to ensure that the observed outcomes are due to the 5 Numbers taken from World Development Indicators policy or interventions being evaluated. In the case of for 2007. ALMPs the outcomes (difference between pre-program 6 For an example see http://www.iyfnet.org/document. and post-program behaviors) among those participat- cfm/30/968. ing in the program must be compared to the outcomes 7 http://www.ilo.org/youthmakingithappen/PDF/lo/ among an identical group that does not receive the lo/LatinAmerica-Caribbean Entra21 26Nov07. pdf treatment. 8 http://www.hrsdc.gc.ca/eng/cs/sp/hrsdc/edd/ A common concern for impact evaluations of ALMPs reports/1999-000414/page00.shtml is the self-selection process of participants into the 9 See http://www.injaz.org.jo/ program. Unobservable characteristics, such as motiva- 10 http://www.jobcenterplus.org.uk tion or other factors determining a person's participa- 11 See also http://www.cgap.org/p/site/c/pubs/. tion, may also determine their ability and make them 12 For a youth related case study see "Microfinance, different from non-participants. Unless the existence Youth and Conflict: Central Uganda Case Study" found of unobservables can justifiably be ruled out, we can- on June 22nd 2010 at http://pdf.usaid.gov/pdf_docs/ not conclude that the program had an impact. PNADF657.pdf. The simplest way to establish this is to assign treat- 13 See also http://www.povertyactionlab.org/ and ment randomly to some part of the study population http://www.3ieimpact.org/. and to withhold treatment from the others. However, 14 Our thank Diego Angel-Urdinola, Jochen Kluve, others resort to quasi-experimental methods where the David Robalino, and the World Bank's Children & Youth control group is formed by matching non-participants Team for helpful feedback and comments on earlier with participants based on observable characteristics drafts of the Note. that are thought to determine program participation and/or eligibility. For more detail on evaluating Active Labor Market Programs see Heckman, Lalonde, & Smith (1999). 15 A c t i v e L a b o r M a r k e t P r o g r a m s f o r Yo u t h The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development / The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. 16