WPS8090 Policy Research Working Paper 8090 Background Paper to the 2018 World Development Report On-the-Job Training Returns, Barriers to Provision, and Policy Implications Priyam Saraf Development Economics Trade and Competitiveness Global Practice Group June 2017 Policy Research Working Paper 8090 Abstract Firms that provide on-the-job training do so when it is literature suggests that for most firms, a comprehensive critical to their productivity—and when productivity is policy approach that resolves external constraints to becom- critical to their survival. This paper begins by confirming ing more productive is likely to have a positive impact on a significant and positive return from on-the-job training the provision of on-the-job training and adoption of com- on wages and productivity, as well as the presence of posi- plementary policies. More direct forms of firm-level support tive externalities from on-the-job training, while discussing to improve management capabilities could also alleviate the methodological considerations at play. The paper then under-provision of on-the-job training. Where societies reviews and validates the presence of market failures such have improved welfare as a goal, public policy measures as information asymmetries within the firm as a result would be needed to complement on-the-job training of low-quality management practices that dampen firm for some specific groups of workers (older, less educated, demand for on-the-job training. Lack of competition in women). In essence, the paper highlights the importance the firm’s external environment appears to undermine of demand-side constraints for firms, rather than sup- adoption of on-the-job training and other complementary ply-side constraints, for the provision of on-the-job training. productivity-enhancing activities within the firm. The This paper—prepared as a background paper to the World Bank’s World Development Report 2018: Learning to Realize the Promise of Education—is a product of the Trade and Competitiveness Global Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at psaraf@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team On-the-Job Training: Returns, Barriers to Provision, and Policy Implications∗ Priyam Saraf+ JEL: I25, J24, J31, M5, O15, O33 Keywords: on-the-job training, skills, human capital, productivity, wages, firms, employer, management, workers, demand, information, competition, constraint  ACKNOWLEDGEMENTS: The author is grateful to Jose Ernesto Lopez Cordova, Alexandria Valerio, Deon Filmer and Marialisa Motta for their guidance and support to this work. Tara Beteille, Vincent Palmade and Leonardo Iacovone provided excellent peer review comments that helped to improve the paper, for which the author is grateful. Valuable inputs were provided at different stages by Alexandria Valerio, Damien Shiels, Daria Taglioni, Denis Medvedev, Gaurav Nayyar, Joern Huenteler, Omar Arias, Rita Almeida, Sanda Liepina and Thomas Farole (in alphabetical order). Asier Mariscal and Lucia Garcia Velazquez provided invaluable research support in collecting data on skills, wages and productivity. The author is grateful to participants at the Skills Workshop in New Delhi on April 28, 2014, for comments on the original proposal, and to seminar participants at the World Bank Human Development Week in Washington DC on May 11th 2017 for helpful inputs. The interpretations of the literature, and any errors associated with it, remain the author’s responsibility.  Economist, World Bank Group email: psaraf@worldbank.org 1 Contents 1. Introduction .......................................................................................................................................... 4 2. Returns to OJT .................................................................................................................................... 10 2.1 At current employer: Effect on productivity, wages and costs......................................................... 10 2.2 At future employer: Evidence of positive externalities .................................................................... 15 3. Barriers to Provision of OJT ............................................................................................................... 19 3.1 Factors within firms: Information asymmetries, low management quality ..................................... 20 3.2 Factors outside firms: Low competition ........................................................................................... 23 4. Policy Considerations and Areas for Future Investigation................................................................ 26 4.1 Demand-side considerations ............................................................................................................ 26 4.1.1 Reforming product market regulation, increasing competition and trade ............................... 26 4.1.2 Reducing uncertainty about returns, increasing information flows, providing certification .... 28 4.1.3 Helping improve management skills for better HR policy formation ........................................ 29 4.1.4 Using more gain-sharing instruments and performance-based compensation for workers .... 30 4.2 Supply-side considerations ............................................................................................................... 30 4.2.1 Targeting selective firms for government subsidies .................................................................. 30 4.2.2 Lower cost of poaching for firms ............................................................................................... 32 4.2.3 Lower coordination costs for a group of firms .......................................................................... 32 4.2.4 Targeting support toward disadvantaged groups...................................................................... 33 4.3 Areas for future investigation ........................................................................................................... 34 5. Conclusion .......................................................................................................................................... 34 References .................................................................................................................................................. 36 Annexes ....................................................................................................................................................... 48 Annex 1: Annotated Bibliography (key literature on various aspects related to OJT) ........................... 48 Annex 2: Survey instrument to measure quality of management in a firm ........................................... 58 2 Boxes Box 1: How Does the Literature Define OJT? ................................................................................................ 5 Box 2: Key Questions Explored in the Literature Review.............................................................................. 6 Box 3: The Methodological Challenge of Comparing Returns across Training Programs........................... 10 Box 4: What Type of Workers Receive Training? ........................................................................................ 13 Box 5: Evidence of the Presence of Industry-specific General Skills........................................................... 18 Box 6: The Limited Role of Credit Constraints ............................................................................................ 25 Figures Figure 1: Effects of OJT on Productivity and Wage Gains ........................................................................... 13 Figure 2: Effects of Worker Mobility on Productivity and Wages .............................................................. 17 Figure 3: Management Practice Scores across Countries ........................................................................... 21 Tables Table 1: Constraints and Market Failures Associated with Providing OJT .................................................... 6 Table 2: Summary of Findings from the Literature on Constraints to Providing OJT ................................... 9 Table 3: Summary of Findings from the Literature on OJT ......................................................................... 35 Abbreviations ALMP Active Labor Market Policies CEO Chief Executive Officer FDI Foreign Direct Investment HRM Human Resource Management IALS International Adult Literature Survey ICT Information and communications technology IRR Internal rate of return MNC Multinational Corporation MNE Multinational Enterprise NCC Non-compete contracts OECD Organisation for Economic Co-operation and Development OJT On-the-job training PPP Public-Private Partnership R&D Research and Development RPL Recognition of prior learning SME Small and medium enterprise TFP Total Factor Productivity TVET Technical and Vocational Education and Training 3 1. Introduction It is widely recognized that on-the-job training (OJT) is linked to an increase in worker wages, firm productivity and, ultimately, long-term economic growth. Workers acquire most of their new and improved skills on the job over the course of their lifetimes. To maintain competitiveness in a context of rapidly changing preferences and technologies, firms need to continually acquire new production and management technologies. This increases demand for better-skilled workers. Even though training is beneficial for both firms and workers, the existence of OJT is far less common than the evidence about its benefits may suggest. Possible explanations have been explored in the microeconomic theoretical literature, including: information gaps and uncertainty about key aspects of OJT, the free-rider problem (positive externalities) and credit constraints regarding the supply of OJT. More recently, literature from other fields, such as management and firm productivity, has pointed to barriers to adoption of managerial practices and new technologies as factors that depress demand for OJT.1 Despite a rich theoretical literature on OJT, very little empirical evidence is available about which constraints and market failures affect firms more. In the past two decades, there has been a rise in the number of rigorous (and innovative) experimental research projects regarding how firms become more productive. This paper leverages these findings together with the traditional literature on training and labor mobility, to improve understanding of the barriers to providing OJT. Several methodological challenges exist to quantifying the links between OJT, wages and productivity. This paper discusses these as well. First, we look at how the literature defines OJT. A review of more than 80 papers in peer-reviewed journals on OJT and skills finds little agreement on the best ways to define and measure skills. But some areas of convergence do surface (box 1). The term skill reveals the modern understanding that what used to be thought of as invariant traits can be altered through experience and intervention. Skills, in general, could be regarded as the ability to carry out a task. More and more, there is growing recognition that not only intelligence quotient (IQ) and cognition as measured by achievement tests (literacy, numeracy), but other types of skills could be important predictors of success in many aspects of life (Heckman 2015). These include socio-emotional, psychological and foundational skills (communications, team work, interpersonal skills, analytical skills, and the ability to learn) that are important for trainability on the job (Heckman, Stixrud and Urzua 2006) and ideally acquired early in life. 1 The theoretical arguments around the market failures for OJT are well documented in the microeconomic literature. In a seminal paper on training, Becker (1962) made a distinction between general and specific training. Under perfect competition, he suggested that firms will not pay for general training of their workers, who can then leave the firm searching for better-paid work that compensates them for the increased productivity they acquired through general training. Hence the worker is the sole recipient of general training benefits and also bears the costs of it. In a series of later papers, Acemoglu and Pischke (1998, 1999, 2000) argue that a substantial amount of training is paid for by firms and is still general in nature. Firms pay for general training in the presence of a compressed wage structure caused by imperfections in the labor market. With a compressed wage structure, training increases the marginal product of labor more than the wage, which creates incentives for the firm to invest in general training. 4 Among these skills, OJT has been referred to as a post-school investment (Heckman, Lochner and Taber 1998). According to them, investments in human capital occur over an individual’s life- cycle, and post-school investments (often in the form of OJT) account for more than half of a person’s life-time human capital. It can happen through a combination of learning-by-doing or formal on-the-job training in the firm. Other authors have mentioned that OJT “plays a crucial role to not only maintain but also improve [the] human capital of the workforce” (Konings and Vanormelingen 2015, p. 485). Some say that “employer-provided training is by far the most important source of further education and training after an individual enters the labor market” (Hannson 2008, p.5). A substantial portion of these human capital investments are financed by firms. Work-based training or OJT could be industry-specific or occupation-specific, general (useful to many firms), or specific (useful to the firm that is conducting the training). However, research on OJT and its impact is still in a preliminary stage; not surprisingly, “economists know less about the incentives and returns to firms of investing in training compared with what they know about the individual's returns of investing in schooling” (Almeida and Carneiro 2009, p. 1). Box 1: How Does the Literature Define OJT? According to the relevant literature, OJT has the following characteristics:  The distinguishing feature of OJT is that the decision about whether to provide it is typically made by the firm with a specific business need in mind.  OJT focuses on the acquisition of skills within the work environment under normal working conditions. It can also include training courses in the workplace or through a training institute for existing workers, master classes, internships and apprenticeships, or even Massive Open Online Courses (MOOCs).  OJT typically includes verbal and written instructions, demonstrations and observations, as well as hands-on practice and imitation. Formal aspects of OJT have been measured in the literature (such as time spent training on the job and the share of employees receiving formal training). But measuring the “unobservable” tacit learning through interactions with co-workers, both within and outside the firm, has proven difficult (also called learning- by-doing). Most studies on OJT measure formal training. Further, most studies do not separate the source of funding for OJT—though in most cases, workers contribute only modestly to the direct cost of OJT, and the firm bears the cost of OJT. The worker contribution is found to be through indirect means—for example, accepting compressed wage structures (see Section 2.1). Lack of credit for the firm also does not appear to significantly constrain the provision of OJT (see Section 3.2). Next, as we review the literature, we develop a framework that looks at demand-side and supply- side constraints to the provision of OJT, whereby each constraint is one that has been suggested by the microeconomic literature on skills and OJT (table 1). In reviewing the literature, we try to determine to what extent these reasons hold empirically.2 At a broader level, the analysis of the constraints in table 1 will help answer the key questions covered in box 2. This is important for policy makers trying different combinations of instruments to increase the incidence of OJT. 2 For example, microeconomics suggests that two kinds of information asymmetry are at play: (i) where firms that stand to gain a lot from OJT do not have correct information about its returns or how much they need it; and (ii) where firms do not know what kind of training is available. Recent literature throws strong light on the first kind of information asymmetry as a real problem that firms across sectors are dealing with; in the second case, there is not enough empirical proof to suggest that it is a binding issue in the provision of OJT. 5 Table 1: Constraints and Market Failures Associated with Providing OJT Information gaps and  Value of OJT not understood by firms (that uncertainty about returns is, accurate information about productivity returns, costs, and wage rises).  Value is most misunderstood by firms that stand to gain the most, and that do not Demand-side realize how poorly run their firms are. constraints Barriers to adoption of  Firm: Low adoption of new managerial managerial practices and new practices or new technologies among firms technologies that reduces their demand for better skills, and hence, for training. This could be due to factors internal or external to the firm.  Worker: Misaligned compensation structures limit worker incentive to adopt and train in new technologies. Information asymmetry  Firms do not know what kind of training is available and how to go about providing it. Free-rider problem  Firms that provide general OJT are not able Supply-side to fully appropriate the rents to it. constraints Credit constraints  Firms do not have adequate resources to finance OJT, even though it may have positive returns. Box 2: Key Questions Explored in the Literature Review Guided by this understanding of constraints/market failures and the available literature, the key questions to be explored in this review are: 1. To what extent does a firm benefit from providing OJT? How do productivity gains for the firm compare to wage gains for the worker? What are the returns after factoring in the cost of training? 2. To what extent do future employers benefit from hiring trained workers? What is the extent of positive externality (and the associated free-rider problem) in the provision of OJT? 3. What are the barriers to provision of OJT at the firm level? Which barriers appear binding in nature? Here we will review the literature about constraints on the demand and supply side (see table 1 for the constraints description). 4. What are the policy implications of the findings from the literature? What are areas for future investigation? The methodology involves reviewing more than 80 peer-reviewed papers that deal empirically with OJT. Given the multidisciplinary nature of this topic (including issues such as education, economic development, labor economics, personnel economics, firm productivity, competitiveness), we do not restrict ourselves to the skills literature. In this regard, we review leading papers from these fields to the extent that we find them useful in understanding the barriers affecting provision levels of OJT. As an example, keywords (and variations on them) such as on- the-job training, work-based training, skills, training, productivity, wages, manager, human 6 resource development, technology adoption and others were used based on the Journal of Economic Literature (JEL) classification system. Within these papers, we prioritize those that we find rigorous in their methodology and treatment of selection issues. We prioritize papers that are rich in spelling out the “why” driving the empirical phenomenon: a justification of the economic mechanisms driving the quantitative result.3 For example, if the finding is that OJT-related wage increases tend to be lower than productivity gains, then a natural question is whether the economic mechanism is one of imperfect competition- general skills or perfect competition-specific skills. Does the paper test for one versus the other? The economic justification would influence the type of policy advice (for instance, more support to sector-specific skills or general training). Most OJT literature deals with developed countries, but we also analyze papers that include developing country examples to validate these findings. There are few rigorous empirical studies that can compare different types of training programs—for example, OJT with active labor market policies (ALMPs) and other types of training; see box 3. We also investigated a few areas where we had hoped to find empirical links to OJT, but ultimately did not. We mention these in the section on “Policy Considerations and Areas for Future Investigation.” Summary of results by key questions Regarding question 1 in box 2, we reviewed recent studies that measure the effect of OJT on wage and productivity separately, and treat for selection issues.4 We also looked at literature that tries to factor in the cost of training (direct costs, forgone earnings and opportunity costs) to determine returns. Carefully designed literature that treats for selection issues suggests that OJT is linked with wage and productivity gains, but with firms benefiting more than workers (Almeida and de Faria 2014; Dearden, Reed, and Van Reenen 2006; and Konings and Vanormelingen 2015). (See Section 2.1 for a detailed review.) For question 2, we studied the literature on labor mobility to better understand whether the future employer experiences productivity gains from hiring previously trained workers (Balsvik 2011; Poole 2013; and Stoyanov and Zubanov 2012). The evidence points to the presence of positive externalities because future employers tend to value and pay a premium for previously trained workers. This also indicates that while much of the OJT is general enough, it is also industry- specific in nature. Where accreditation and recognition of competence are effective, the transferability of training wage premia occurs more often. (See Section 2.2 for a detailed review.) 3 We provide an annotated list of references (Annex 1) that details key papers reviewed, as well as a complete list of references containing all the papers found using this methodology. 4 While there is considerable research connecting OJT to wage increases (Bartel 1995; Bartel and Sicherman 1999; Booth 1991; Blanchflower and Lynch 1994; Blundell, Dearden and Meghir 1996; Frazis and Loewenstein 2005; Greenhalgh and Stewart 1987; Lillard and Tan 1992; Lynch, 1992; Mincer 1989; Winkelmann 1994), empirical work that connects OJT to firm productivity is more recent and limited (Almeida and de Faria 2014; Black and Lynch 2001; Dearden, Reed, and Van Reenen 2006; Konings and Vanormelingen 2015). For a long time, wages were used as a proxy to determine productivity gains from OJT. However, wage-productivity equality does not hold in imperfect markets. Such studies might then underestimate the gains to firms. 7 For question 3, we reviewed the firm productivity and management literature that emphasizes the importance of information processing, effective human resource policies, and management capability for OJT provision and aligning worker incentives (Atkin and others 2015; Bandiera, Barankay, and Rasul 2011; and Bloom and others 2013). In this regard, we find evidence of information failures within the firm. (See Section 3.1 for a detailed review.) Among the external conditions, we find that product market competition plays a key role in forcing firms to be more efficient and productive by creating the pressure (and incentive) to invest in innovation, exports, management capability, and/or adoption of new technology. These factors have been shown to be complementary to the demand for OJT (Almeida and Aterido 2008; Black and Lynch 2001; Frazis, Gittleman and Joyce 2000; Frazer 2006; Pierre and Scarpetta 2004; and Rosholm, Nielsen, and Dabalan 2007). We did not find strong evidence pointing to credit constraints or lack of information about training programs (or how to go about it) as binding constraints to firm uptake of OJT (Atkin and others 2015; Bloom and others 2012; Carneiro and Heckman and Lee 2004; and de Mel, McKenzie, and Woodruff 2008).5 (See Section 3.2 for a detailed review.) The purpose and contribution of this paper are twofold. First, it provides a review of systematic empirical evidence regarding the links between OJT, wages, productivity, costs in the firm that trains and, to an extent, at the future employer. This is helpful in assessing the extent to which positive externalities are at play, the type of training imparted and who benefits from OJT. Second, the review goes beyond the traditional literature on skills and education, and delves into the literature on firm management (personnel economics) and productivity. As such, it provides a more granular sense of the different types of barriers that the firms may be facing. To the best of our knowledge, this has not been done before. Because firm productivity and survival are necessary conditions for productivity-enhancing activities such as OJT, we need to cross over to that literature to better understand the economic mechanisms at play both within and outside firms. We believe this will enhance the understanding of “why” with regard to OJT provision. In addition, it has direct implications on how governments decide whether to support this important training area. 5 Studies show that access to information about new training in the market is not a binding constraint for sporting goods firms in Pakistan and textiles firms in India (with an intervention that is available on the market, for example, consulting services or sharing of information). Some firms are faster than others in accessing these tools, depending on their management capability and other factors (Atkin and others 2005; Bloom and others 2012). 8 Table 2: Summary of Findings from the Literature on Constraints to Providing OJT Constraint Explanation Findings (from Empirical Literature) Policy Considerations and (from Theoretical Areas for Future Work Literature) Value of OJT is not Literature indicates that the returns to OJT (firm Demand side considerations understood by firms productivity and worker wages) are positive. The Information (accurate information on internal rate of return is positive (see section 2.1). Reforming product market gaps and productivity returns, competition, trade, regulatory uncertainty costs, wage rises); it may environment and other Literature presents validation of information gaps, about be particularly institutions (see section 4.1.1). and finds that firm-level management is often not returns misunderstood by those well-informed about the cost-benefit calculus of new Reducing uncertainty about firms that stand the mosttraining, leading to ineffective HR policies (see gains, increasing information to gain, and that do not sections 2.1 and 3.1). flows, and providing realize how poorly run certification systems (see they are. section 4.1.2). Firm: Low adoption of Evidence of external environmental factors such as Demand- new managerial practices low product market competition linked to low Helping to improve side Barriers to or new technologies adoption of productivity-enhancing activities (OJT, management skills for better constraints adoption of among firms — reducing new technologies, such as information and HR policy formation (see managerial their need for better communications technology [ICT], management section 4.1.3). practices skills and, hence, practices). Within firms, low adoption of managerial and new Using more gain-sharing demand for training. practices in developing countries with a technologies instruments and performance- This could be due to preponderance of family-owned firms depresses based compensation for factors internal, or demand for productivity-enhancing activities such as workers (see section 4.1.4). external, to the firm. OJT (see sections 3.1 and 3.2). Supply side considerations Worker: Misaligned Even if the firm management wants to adopt new compensation structures technology, there is evidence of distorted HR policies Targeting selective firms for limit worker incentives (linked to information gaps) that creates low the provision of government to train in new incentives for workers to participate as they do not subsidies (see section 4.2.1). technologies. share in the gains (see sections 3.1 and 3.2). Lower coordination costs Information Firms do not know what Literature review does not suggest this is a binding among firms (see section asymmetry kind of training is constraint because some firms in the market are able 4.2.2). available on the market, to access this information, while others cannot. and/or how to go about Simple interventions, such as increasing information Targeting support to providing for it. awareness, can help resolve the problem (see section disadvantaged groups (see 3.1). section 4.2.3). Free-rider Firms that provide There is evidence of productivity benefits for future Areas for future work problem general OJT are not able employers profiting from hiring workers trained in More rigorous analytical work Supply-side to fully appropriate the the previous firm. Wage gains for existing workers in on OJT link to Public-Private constraints rents to training. the future employer are higher when the firm hires a Partnerships (PPPs); impact of trained worker (for example, from multinational agglomeration on OJT; and corporations) (see sections 2.1 and 2.2). resolving methodological issues for better OJT Credit Firms do not have For most firms, a lack of convincing evidence about measurement. constraints adequate resources to the presence of credit constraints is a significant finance OJT, even reason why firms under-provide training. This could though it may have be a bigger issue for microenterprises and low- positive returns. income workers (see section 3.2). 9 The next section discusses the findings from the literature with respect to the four guiding questions: OJT returns to the current employer, returns to the future employer of hiring previously trained workers (i.e. the extent of the free rider problem), barriers to adoption (within and outside firms) and policy considerations and suggestions for future research and investigation. 2. Returns to OJT 2.1 At current employer: Effect on productivity, wages and costs To what extent does a firm benefit from providing OJT? How do productivity gains for the firm compare to wage gains for the worker? What are the returns after factoring in the cost of training? At the level of aggregate economic growth, the skills level of the workforce is a first-order determinant of country growth (Mankiw, Romer, and Weil 1990). Adding educational quality measures beyond the quantity (years of education) produces better skills, mostly by improving cognitive skills (Hanushek and Woessmann 2008).6 Indeed, measures of cognitive skills can account for three times the variation in economic growth than models that include only years of schooling.7 The impact is larger for more open economies. Countries have used different kinds of training programs that range from government-subsidized active labor market programs (ALMPs) for disadvantaged young adults to OJT for workers who have completed primary and secondary school. ALMPs tend to prepare young adults for the job, while OJT is typically for workers who are already employed—that is, on the job. Theoretically, there would be nothing better than a systematic comparison across different kinds of training programs (for example, OJT with ALMPs) to determine how best to spend public funds. However, we cannot do this practically without running into serious methodological challenges. Directionally, however, there is evidence of rather modest returns to ALMPs (Betcherman, Dar and Olivas 2004), whereas the returns to OJT have been found to be higher (Konings and Vanormelingen 2015). However, the challenge is that we are not able to attribute causality to program design properly as these programs target different groups of trainees. Their quality varies, and disentangling the priors of trainees from achieved results is hard. See box 3 for a summary of the methodological challenges in comparing returns across training programs, as well as a short review of ALMPs. Box 3: The Methodological Challenge of Comparing Returns across Training Programs Serious methodological issues are involved in comparing types of skill programs (such as ALMPs and OJT), including: (1) Training programs differ vastly in the type and quality of trainees they attract. Disentangling worker selection into programs from causal effects is hard: Are the returns high because of the program or because of the higher-ability students attracted to it (endogeneity)? More disadvantaged workers are likely to enroll 6 There is a significant body of literature regarding the returns to secondary and tertiary education. The paper will refer to it as required, without focusing on it per se. In general, secondary education is an important feeder into OJT, and the two complement one another. 7 Evidence indicates that a standard deviation in cognitive skills increases income by a range of [0, 0.48], being mainly positive in several (mostly African) countries. 10 in government-driven ALMPs, which is not the case for OJT, where high skills beget more skills.8 (2) Training systems vary across and sometimes even within countries. Therefore, it is difficult to successfully isolate, or even compare, system-level effects (local institutions, informality, and public spending on training). For instance, Organisation for Economic Co-operation and Development (OECD) countries have extensive experience with ALMPs, technical and vocational education and training (TVET) and OJT. Developing economies have much less experience with these programs. (3) Program quality could vary widely between ALMPs and OJTs: there are master-trainers within firms, access to the latest machines and equipment, and concrete opportunities to utilize the new skills. The fact that a firm is conducting OJT could also imply less asymmetry of information between the needs of the firms and the training output. However, in an indirect way, by looking at the evolution of the content and setting of training used in ALMPs—a widely studied area in social protection—we can arrive at some sense of what training approaches work better. ALMPs use a mix of approaches, including in-classroom training programs; workplace training; the interaction of the previous two types; and in-classroom and workplace training— plus supplementary services, such as counseling and mentoring, monitoring, job search and placement assistance, and soft and life skills training. A comprehensive meta-assessment of training programs from 90 countries found some patterns of transition from in-classroom only training toward combined measures of in-classroom, workplace training, and other supplementary services (Fares and Puerto 2009). The meta-analysis of 345 studies of training programs indicated that programs that combine different training approaches in ALMPs have a higher probability of positive labor market impacts on employment. In particular, the combination of in-classroom and workplace training increases the likelihood of positive labor market impacts by 30 percentage points over in-classroom training only. When this interaction is combined with other services, the probability of a positive impact increases by 53 percentage points. In another relevant study, Betcherman, Dar and Olivas (2004) take an instructive approach about not viewing ALMPs as a panacea to youth unemployment, given the modest returns to these. They emphasize the following as good features to consider: a comprehensive package of services, programs oriented to demand and linked to workplaces; careful targeting; and early engagement of employers in training design. Both studies lament the lack of rigorous impact evaluations and cost-benefit analyses as a missing piece of the puzzle—and an area for future research. The following subsections focus on reviewing the literature on OJT (given the high returns), to examine wages for workers, productivity for firms, costs of provision, and common measurement challenges (that the more rigorous studies analyzed here have managed to address). 8 ALMPs are typically government programs that intervene in the labor market to help the unemployed find work. Many of these programs grew out of earlier public works projects designed to combat widespread unemployment in the developed world during the inter-war period. ALMPs include: (i) public employment services, such as job centers, help for the unemployed in their job search effort by disseminating information on vacancies, as well as by providing assistance with upgrading interview skills; (ii) training schemes, such as classes and apprenticeships, help the unemployed improve their skills and increase their employability prospects; and (iii) employment subsidies, in the public or private sector, directly create jobs for the unemployed. These are typically short-term measures designed to enable the unemployed to build work experience and prevent skill atrophy. 11 Upside to OJT: Productivity and wage effects What is the upside to OJT, in terms of productivity gains and wage effects? Studies find positive effects of OJT on direct measures of firm productivity and wages. For example, Konings and Vanormelingen (2015) use firm-level panel data of OJT to estimate its impact on productivity and wages in Belgium (figure 1). Treating for endogeneity of input factors and training, they find that the increase in productivity of the trained worker is higher than the increase in wages the trained worker commands because the worker participated in OJT. In numeric terms, the average sales per worker, a proxy for productivity, increased from 1.7 to 3.2 percent, while their wages only increased from 1.0 to 1.7 percent. This occurred in response to a 10 percent increase in the number of workers who benefited from training.9 Another study confirms these results. Dearden, Reed, and Van Reenen (2006) study the impact of work-related training on productivity. Using panel data from the United Kingdom, they find that OJT is associated with significantly higher productivity. In response to a 1 percent increase in training, there is an increase in value-added per hour of about 0.6 percent, as well as an increase in hourly wages of about 0.3 percent. They also provide evidence of training externalities (a topic covered in greater detail in section 2.2), while controlling for firm selection and training. Studies that look at the impact of OJT on productivity in developing countries are more limited due to data constraints. Therefore, to get a directional sense, we reviewed studies that show a link between the incidence of OJT and wage gains (which are studied widely in the developing world). Evidence from two carefully prepared country studies suggests that individuals’ earnings can substantially increase with OJT (Tan and Lopez-Acevedo 2003, for Mexico, and Rosholm, Nielsen and Debalsen 2007, for Kenya and Zambia). Rosholm, Nielsen and Debalsen (2007) find a 20 percent wage increase, and Tan and Lopez-Acevedo (2003) find an impact of 5 to 7 percent due to OJT. These studies also control for endogeneity. These studies also find that the incidence of training is different for different groups by gender, prior skill level and age (box 4). 9 Details on each study (as well as other papers reviewed), including data set descriptions, dependent and independent data sets, and challenges with regard to methodology (if any) are included in Annex 1. Initial data collection work is described in the internal memo (Saraf, Mariscal and Velazquez, 2016). 12 Figure 1: Effects of OJT on Productivity and Wage Gains 6.00 3.2 Note: In the case of Belgium, firm productivity gain ranges Size of the effect in 5.00 4.00 from 1.7 to 3.2 percent in response to an increase of 10 3.00 1.7 percentage points in the fraction of workers who receive % 2.00 1.7 training. The average wage increases by only 1 to 1.7 percent 1.0 1.00 in response to the same independent variable. 0.00 Productivity Gains Wage Gains Source: Konings and Vanormelingen, 2015. 6.00 Note: In the United Kingdom, a 1 percent increase in work- Size of the effect 5.00 4.00 related training is associated with an increase in value-added 3.00 per hour of about 0.6 percent for the firm, as well as an in % 2.00 0.60 increase in hourly wages of about 0.3 percent for the worker. 1.00 0.30 0.00 Productivity Gains Wage Gains Source: Dearden, Reed, and Van Reenen, 2006. Box 4: What Type of Workers Receive Training? Training is not equally distributed among employees. Indeed, certain types of workers usually receive more training than others. More education and a higher threshold of skills from before begets more training. In this regard, various studies (Barron, Black and Loewenstein 1989; Hansson 2009; Leuven and Oosterbeek 2004; and Lillard and Tan 1992) show that less-educated and older workers receive less training, and that worker tenure increases the likelihood of training for the same job. Almeida and Carneiro (2009) consistently show that the workforce in firms that provide training is more educated than the workforce in firms that do not. The proportion of workers with a Bachelor’s or other college degree(s) is 6 percent and 3 percent in high and low training firms, respectively, versus 1.3 percent in non-training firms. Balsvik (2011) also finds that firms that provide training to their workers tend to have a workforce with more years of formal schooling. Training incidence is also lower among female workers. Blundell, Dearden and Meghir (1996) find that men have a much higher probability than women of undertaking employer-provided training and work- related training leading to a formal vocational qualification. Evertsson (2004) finds that women are significantly less likely to take part in formal OJT than men based on a Swedish pooled cross-sectional data set. Her research concludes that among those who receive training, women are more likely to participate in industry-specific training, while men participate in general training for lifelong skills. There is no conclusive evidence that returns to training on productivity vary significantly with worker characteristics such as gender, educational or skill levels. Wage premiums to those participating in training while working for an employer are relatively homogeneous across age and educational attainment. However, some factors might be influencing employers to have an unequal distribution of training investments for these groups: discrimination, higher worker turnover for specific groups, prior skill levels, distortionary public policy, selection into job characteristics in the firm or possibly other factors (Evertsson 2004; Hansson 2009; Leuven and Oosterbeek 2004; Lynch 1991; Melero 2004; and Veum 1995). The impact of training on different groups can manifest differently: in the case of relatively young and/or higher-educated workers, training provides transferable skills and increases mobility. For older and lower- 13 educated workers, training allows them to attain and maintain the competencies required to bring productivity in line with market wages, thereby sustaining the employment prospects of these groups. It is clear from the literature that certain groups are likely to receive less OJT, thereby potentially amplifying skill gaps and equity concerns. This would imply an area for government intervention to stimulate acquisition of basic/secondary education, socio-emotional skills and others forms of training for these groups. Given the modest returns to government-driven training programs such as ALMPs, the design of these programs would need to factor into the latest evidence regarding what works and what does not (for evidence on ALMPs, see box 3; for policy considerations, see section 4.2.3). Cost of providing OJT If training were free or cheap, it would suffice to look at its returns using productivity and wage indicators. But the direct costs of training can be quite high. Data on costs are essential for computing meaningful estimates of efficiency. However, for the most part, finding data that are detailed enough and that contain cost information is rare. Important work has been done in this area by Almeida and Carneiro (2009), who analyzed a range of large manufacturing firms in Portugal in the 1990s—with detailed information on investments in training and costs. They estimated training’s internal rate of return on firms and workers. They classified costs into forgone earnings, opportunity costs and direct training costs.10 According to Almeida and Carneiro (2009), about three-quarters of the cost is associated with the direct costs of training and only a quarter with forgone production. For the group of firms that provide training, the internal rate of return on OJT is 8.6 percent, with the return diminishing over time. For the entire sample, the internal rate of return of an additional hour of training is zero, indicating a profitable investment. In general, these firms enjoy similar returns of investment on training and physical capital, and yet, a small fraction of hours is dedicated to training (less than 1 percent). This could be reflecting the difficulty of appropriating the benefits of training. They control for selection issues. However, external validity is limited as the sample is one of large firms (>100 employees), limiting application of results to smaller and medium-size enterprises in developing countries, where similar empirical work would be required. Some other papers come to a similar conclusion regarding the cost of training. Bartel (1995) used detailed data from a manufacturing firm in the United States, and found a 13 percent internal rate of return. Frazis and Loewenstein (2005) characterized costs and obtained a 60 percent opportunity cost of training and a 40 percent direct cost. Their internal rate of return was found to be 50 percent. Methodological challenges associated with measuring returns to OJT The literature on OJT faces a few methodological challenges that merit discussion. First, most data on productivity and costs are for developed countries. Hence caution is required while doing cross- country comparisons and applying results across contexts. Second, the definition of what is considered training could vary by survey. Often, one cannot separate the formal aspects from the 10 Forgone earnings are paid by the worker, direct costs are mostly paid by the firm, and opportunity costs can be associated with both the worker (time dedicated to training that could instead be dedicated to other activities) and the firm (the worker being trained could have been producing instead). The authors only had information about the direct cost of training. They then calculated the forgone productivity costs of training using the marginal product of labor and the time allocated to training. However, they could not disentangle the effect on both firm and worker rates of return, so they estimated a joint rate of return. 14 learning-by-doing. Information about the length, type and timing of programs may not be available, or may be self-reported. In some cases, perceptions of training might vary.11 In most studies the sample includes large and formal firms due to data limitations. Hence it is important to remember what OJT studies measure (and what they exclude). Finally, when measuring the impact of training, it is important to treat for selection bias as the results may capture the superior attributes of the worker rather than the actual effect of training. At the firm level, the main difficulty in measuring impacts comes from the fact that firms offering OJT tend to already have a set of better observable (and unobservable) characteristics. The studies included in the review (for example, Almeida and Carneiro 2009; Dearden, Reed and Van Reenen 2006; and Konings and Vanormelingen 2015) treat for selection issues and indicate that even after this, increasing OJT could have productivity gains. Beyond quantitative challenges, another issue with the literature on OJT is that few studies delve into the economic mechanisms explaining the causal relationship between OJT and productivity. To improve our understanding of the economic mechanisms, we will review the literature on barriers to provision of OJT in section 3.1 and section 3.2 to explore these issues more. In general, returns to firm productivity and worker wages are both positive (factoring in the cost of OJT) with regard to OJT provision. Firm productivity gains are higher than wage gains, indicating an imperfect pass-through of gains to workers and the presence of a high general skills content to the OJT (see Section 2.2. for more information on type of skills). 2.2 At future employer: Evidence of positive externalities To what extent do future employers benefit from hiring trained workers? What is the extent of positive externality (and the associated free rider problem) in the provision of OJT? To explain the gap between the productivity and wage premium for trained employees, theory suggests a situation of general training in imperfect markets (Acemoglu and Pischke 1998, 1999, 2000; Becker 1962). Recent empirical evidence appears to validate this. The evidence points to two specific results about firm provision of OJT: (i) productivity gains more than wage gains indicating worker willingness to bear some cost for the training as the firm cannot fully appropriate the benefits of OJT in imperfect markets (we saw this in the previous section 2.1); and (ii) future employers willing to pay a wage premium for previously trained workers, indicating the presence of transferrable (general) skills and positive externalities (we will discuss in this section 2.2). Therefore, much of OJT is general training relevant to the current firm and others in the market.12 The literature reviewed validates the presence of general skills content in the training imparted. Booth and Bryan (2005) find that workplace training is associated with significantly higher wages at current and future firms, with a larger impact at future firms, using a household panel survey in the United Kingdom. Accredited OJT is more strongly associated with higher wages at both current 11 Black and Lynch (1995) mention that “… a supervisor who is assigned to work side-by-side with a new employee may view the time she spends with a new hire as training, whereas the employee views the fact that the supervisor is always hanging around as monitoring, not training”. 12 If the returns to future employers were not statistically significant, one could have argued for the wedge between wages and productivity to be created by perfect competition and firm-specific training. However, this does not seem to be the case given the rise of productivity at future employers as a result of hiring previously trained workers. 15 and future employers than is non-accredited training—and that only accredited training is transferable between employers. However, the literature did not provide much by means of explaining how firms were paying (individually or via cooperation) for these accredited general skills. Findings from this research are consistent with Loewenstein and Spletzer (1998) and Acemoglu and Pischke (1999, 2000). In another related study involving labor movement between multinational companies (MNCs) and non-MNCs in Brazil, Poole (2013) shows that when former MNC-employed workers move to domestic firms, incumbent workers of the hiring domestic firms enjoy wage gains. This starts to point to the positive externalities associated with more trained workers. At the average wage of a typical domestic worker, a 10 percentage point increase in the share of former multinational workers increases wages by $23 a year. For the average domestic establishment with 60 workers in Brazil, the total implied wage effect from an enhanced foreign presence is roughly $1,400. In a similar study in Ghana, Gorg and Strobl (2005) use firm-level data for a sample of manufacturing plants to find that firms whose owners worked in MNCs in the same industry prior to joining are more productive than other firms. In OECD countries, Stoyanov and Zubanov (2012) show positive productivity effects of hiring workers that come from more productive firms (using Danish data). They find that the effect is 0.35 percent per year for an average firm one year after hiring. The accumulated gain in the fifth year is 1.64 percent. This result is similar to Balsvik (2011), who using matched employer-employee data for Norway, finds that workers with MNC experience contribute 20 percent more to the productivity of non-MNCs than workers without such experience, even after controlling for unobservable characteristics of the workers.13 Through fixed effects, these studies are able to control for the firm-specific advantage of MNCs, such as better technology (so that the gains are not attributed to MNCs simply being bigger or better firms). Regarding the role of human capital developed in the previous firm as a factor in increasing wages, Balsvik (2011) mentions that through on-the-job experience (or training), workers in MNCs may be able to embody valuable knowledge and bring this knowledge to a new employer. Stoyanov and Zybanov (2012) find that while the presence of human capital accentuates the productivity gains, it does not fully explain it. This leads them to conclude that the presence of knowledge spillovers or positive externalities are related to labor mobility. Poole (2013) finds that among different types of workers, highly skilled workers from MNCs produce a larger effect.14 13 Stoyanov and Zubanov also show that of the total output gain, two-thirds goes to the new firm, and 25 percent to the incumbent workers. For Norway, the wage premium for movers from MNEs to non-MNEs is 5 percent compared to workers (with similar characteristics) who stay in non-MNCs. The productivity gain premium is 20 percent in the new employer, as compared to hiring non MNC workers. See Saraf, Mariscal and Velazquez (2016). 14 The literature on R&D spillovers and labor mobility uses a human capital framework: since at least a part of the knowledge acquired in the firm will move with the worker in the case of mobility, workers who receive access to formal training/knowledge should be willing to pay for this by accepting a pay cut (Pakes and Nitzan 1983). 16 Figure 2: Effects of Worker Mobility on Productivity and Wages 2.5 8 100 Note: Danish data shows positive firm productivity productivity gain… 1.64 80 25 effects of hiring workers from more productive firms. The Share of the 2 Share of the effect 60 1.5 effect is 0.35 percent per year for an average firm one 40 66 year after hiring. The accumulated gain is 1.64 percent in in % 1 20 0 4 years. Of the total output gain, two-thirds goes to the 0.5 0.35 Share of the new firm, 25 percent to incumbent workers, and 8 percent 0 productivity gain to the workers who bring the spillover. Productivity gains Source: Stoyanov and Zubanov, 2012 25 Note: Norwegian data from Linked Employer-Employee Share of the effect in % shows that the wage premium for movers from MNEs to 20 non-MNEs is almost 5 percent, compared to workers who 15 stay in non-MNEs. The productivity gain premium is as 10 20 much as 20 percent in the new employer as compared to hiring workers without MNE-experience. 5 5 Source: Balsvik, 2011. 0 Productivity gains Wage premium 1400 1400 Note: Data from Brazil shows a 10-percentage point Average Increase in US increase in the share of former multinational workers’ 900 wages by $23 per year, or roughly $1,400 per year for the Dollars average domestic firm in Brazil (for firms with 400 approximately sixty workers). 23 -100 Source: Poole, 2013. Increase per worker Average firm increase These studies demonstrate that future employers are able to benefit from hiring previously trained workers, indicating the presence of general skills in the training (useful across multiple firms). In addition, they demonstrate the positive externalities that exist due to the provision of OJT. The future employer benefits by not paying the full cost of this advantage. Thus, the benefit to the firm that provides training is less than the benefit to society. This includes the future firm and workers in the future firm who enjoy spillover advantages of having a trained person in the workforce. In an unregulated market, with positive externalities, less is produced than the socially optimal level. How general are the general skills imparted through OJT? Firms do not seem to provide the type of (very important) general skills, such as behavioral, psychological or socio-emotional skills, perhaps because these are better provided during the formal education years of the worker, or because firms may not have the methodology to train in these skills (Gonzales-Velosa, Rosas, and Flores 2016). In fact, some studies indicate the presence of industry-specific general skills in OJT. In other studies, the productivity effect of hiring workers from more productive firms within the same sector was found to be twice as high as for workers from other sectors. This could imply that the knowledge brought by new workers is, in large part, industry-specific, but also general enough 17 to be applied in different firms (Gorg and Strobl 2005; Stoyanov and Zubanov 2012). The skills imparted through OJT are more likely to be general, but industry-specific (see box 5). Box 5: Evidence of the Presence of Industry-specific General Skills The literature review highlights the presence of general skills content in training imparted, as well as productivity gains in the future employers who hire previously trained or more productive workers (Balsvik 2011; Booth and Bryan 2005; Gorg and Strobl 2005; Poole 2013; Stoyanov and Zubanov 2012). The question then becomes: How transferable are these general skills? For example, are these skills applicable across sectors, for example, socio-emotional or behavioral skills? Studies suggest that the type of general skills imparted through OJT are unlikely to be skills such as socio- emotional or behavioral, even though they are highly valued by most employers. Socio-emotional or behavioral skills are the most highly valued by employers and the most difficult to find in the labor market (Busso and others 2012). However, only a fraction of firms give priority to behavioral skills in their OJT, according to a recent study on Latin American countries (Gonzales-Velosa, Rosas, and Flores 2016). Even firms that have the greatest difficulties fulfilling their requirements for socio-emotional skills tended to offer more training in industry-specific skills. The tendency to give less training in behavioral skills may result from a firm’s lack of capacity (for example, methodologies and technologies) to teach general skills. It may also reflect the disincentives that arise due to a lack of appropriation given that most of the OJT in this region is financed by the employer. In fact, there seems to be evidence of the presence of industry-specific general skills in OJT. In two studies, the productivity effect of hiring workers from more productive firms within the same sector was found to be twice as high as for workers from other sectors. This implies that the knowledge brought by new workers is, in large part, industry-specific, but also general enough to be applied in different firms (Gorg and Strobl 2005; Stoyanov and Zubanov 2012). An example of such industry-specific general skills could be machining skills that are valuable across automotive, tooling, electronics and other kinds of manufacturing sectors. It also helps to remember that the distinction between general and specific skills is not binary because industry-specific skills could also have general elements, and general skills could be specific to some sectors. A useful lens to use is whether the skills imparted by one employer are useful to another firm, and whether the future employer is willing to pay for such skills. However, provision of skills may not be fully left to OJT. In times where automation stands to make several existing jobs in manufacturing obsolete, emphasis needs to be placed on the general lifelong skills such as socio-emotional and behavioral training that may be under-provided by the OJT market, but would be critical to help workers adapt to new technologies. In this area, there is an argument for some public support. There are three ways in which firms appear to cope with the presence of positive externalities and the absence of corresponding regulation: increasing labor market rigidities that raises the cost of switching for the worker, voluntary cooperation or under-provision. From 2000-08, many firms used non-compete contracts (NCCs) to increase the cost of switching employers for workers. Indeed, the number of non-compete cases has been on a rise in the United States. In 2009 poaching 18 became such a serious issue in Silicon Valley that firms worked on a cooperative basis not to approach each other’s employees.15 At other times, firms under-provide OJT, consistent with what is shown in the research and development (R&D) literature: firms reduce their R&D expenditures due to the risks of R&D workers moving to other employers. In this context, they also increase patenting (Kaiser, Kongsted and Ronde 2008; and Kim and Marschke 2005). In the section on “Policy Considerations and Areas for Future Investigation,” we revisit the evidence to suggest potential instruments that governments can consider using. In summary, much of OJT is general enough, but at the same time, industry-specific. Where accreditation and recognition of competence is effective, the transferability of training wage premia occurs more often. Some kind of policy effort to internalize the positive externalities and bring provision levels closer to the social optimum might be useful. Under-provision could be a more serious issue in developing countries where property rights are still a work in progress, enforceability of contracts low and cooperation harder due to high informality. Policy support is also perhaps required in the provision of general lifelong skills such as socio-emotional traits that are likely to be under-provided by firms. 3. Barriers to Provision of OJT What can we learn from the literature on constraints and barriers —such as information gaps among firm management, low demand for complementary technologies, low product market competition and credit constraints—that might influence provision of OJT? Which ones appear binding in nature? Complementarity of skills and technology is well acknowledged (Acemoglu 1998; Berman, Bound, and Griliches 1994; Bugamelli and Pagano 2004; Doms, Dunne, and Troske 1997; Disney, Haskel and Heden 2003). This holds for developing countries, as shown in the recent work by the World Bank in South Asia, where new technology adoption was positively and significantly linked with the share of more educated employees in the firm (Lopez-Acevdeo, Medvedev, and Palmade 2016). Adoption of new technologies tends to require training workers who could learn to operate it. Complementarity between skills and management practice has also been shown recently (Bloom, Eifert, Mahajan, McKenzie and Roberts 2013). Management capability is viewed as a form of soft technology (or intangible capital) that helps a firm utilize its factor inputs, such as trained labor, more efficiently. In turn, training helps to improve the quality of managers, creating a virtuous cycle of sorts. Regarding the context of OJT, we are particularly interested in understanding how firms develop effective policies (such as for training and compensation). The role of management (owners, managers) is of great importance. Management accesses information in the market, makes critical decisions about human resources and decides how much training to provide, to whom, and how to compensate them. We want to understand how these mechanisms play out within firms, and what 15 Although with negative results, as they got sued for anticompetitive practices and had to settle in 2014 (more on this issue in the section on policy considerations). 19 barriers exist in their path. In this context, then, we review the recent empirical management capability and organizational (personnel) economics fields. The literature has also identified external conditions positively linked to OJT provision at the firm level. Favorable external conditions include advancing competition, improving regulatory environment, influencing firm conditions by providing incentives to adopt new technologies, and improving management and providing OJT. We review this evidence to understand what type of conditions would facilitate boosting the demand for OJT.16 3.1 Factors within firms: Information asymmetries, low management quality How do firms decide on human resource policies? What steps are involved in accessing market information about new practices and training? Who accesses this information, and who does the cost-benefit analysis of such training? How accurate are their expectations? How adequate are their cognitive skills for information processing? How quickly do they act on it? These are the types of questions addressed in this section. There is a small but growing literature with robust field evidence about the link between management capability and firm productivity. We draw on this literature to provide suggestive evidence regarding the barriers to effective policy-making within firms—which in turn impacts decisions regarding the provision of OJT (Atkin and others 2015; Bandiera, Barankay, and Rasul 2011; Bloom and others 2013). More often than not, management teams - within firms - access information about training and decide whether to provide it. They also tend to decide based on their internal cost-benefit analysis. The literature reveals that, across countries, substantial variations exist in management practices across organizations and sectors (Bloom and Van Reenen 2007; Bloom and others, 2012).17 In developing countries like Brazil, China and India, management scores are lower than in the United Kingdom or United States, reflecting lower management capabilities in general. Among management practices, those pertaining to human resource management are the most relevant for our purposes: policies on evaluating worker performance, designing incentives for worker performance, training provision and so on. 16 In a sense, the provision of OJT could be likened to the adoption of other productivity-enhancing activities such as information and communications technology (ICT) or R&D. In the case R&D and OJT, the ability of a firm to appropriate rents may be more limited than in the case of ICT. 17 They have been using double-blind survey techniques and randomized sampling methods to gather information on management practices for more than 10,000 organizations across 20 countries. See Annex 2 for the survey instrument. 20 Figure 3: Management Practice Scores across Countries Source: Bloom and Van Reenen 2007; Bloom and others 2012 The literature indicates that despite the presence of large gains, management teams within firms do not adopt new practices readily. In a study investigating the reasons for this, Bloom and others (2013) show that adopting new practices—such as factory operations, quality control, inventory, HRM and sales management—raised productivity by 17 percent in the first year for their sample of Indian textile firms. This was achieved through improved quality and efficiency, as well as reduced inventory. Within three years, this led to the opening of more production plants. Their methodology included randomizing free consulting to provide these practices.18 Despite high (and proven) returns, low management capabilities and (resultant) poor human resource policies at the firm level depressed demand for these practices. These mechanisms are likely at play for decisions which are at the discretion of firm management—such as adoption of OJT, R&D or ICT. Based on firm management and organizational economics literature, there is also evidence of information failures within the firm that lead to low demand (and supply) of OJT within firms. Bloom and others (2013) find that firm-level management is often not well-informed about the cost-benefit calculus of new training, Indeed, firm owners have been misinformed about the cost- benefit of new practices. For example, in the case of the textile industry in India, for commonly known business training where information is widely available, owners held the mistaken view that they did not need training. Through deep interviews, the authors found that such owners were comparing themselves with other (low-quality) local firms and judging that their quality was good 18 During the second phase of Bloom’s experiment, the consulting firm worked closely with plant management staff for four months on the job. During this time, they put the procedures into place, fine-tuned them, stabilized them so that employees could readily carry them out. For example, one of the practices was holding daily meetings for management to review production and quality data. The consultant attended these meetings for the first few weeks to help the managers run them, provided feedback on how to run future meetings, and adjusted their design. The third phase involved the measurement of results. 21 based on these comparators. The same firms, when compared to international benchmarks, were lagging. Almeida, Behrman and Robalino (2012) explain that skill provision failures might arise because individual expectations could be based on partial observations of labor market outcomes in the local economy, as well as the experience and advice of peers, friends and family that can be unrepresentative of the average experience. That appears to be the case here. Regarding the issue of information failures, Atkin and others (2015) suggest that if firms believed that adopting OJT would lead to non-trivial costs (due to incorrect information about cost-benefit), that is, they had low prior knowledge about the utility of a new practice, they may be (rationally) unwilling to pay even quite small transaction costs. This could then lead to inefficient outcomes. We ask, then, what prevents the management within firms from having more informed views about the cost-benefit of new practices even when it is available in the market? This question brings us to the composition of the “management” within a firm. In this case as well as in many developing country contexts, several of these (small and medium) firms are family-owned businesses. Firm owners do not fully trust outsiders or non-family members. As a result, these family-owned firms suffer from a limited number of professional managers (in proportion to workers). A lack of delegation, limited managerial time, and limited autonomy are linked to poor decision-making on most fronts: in finance, purchasing, operations, and/or human resources. In Bloom’s work in India, non-family members were given only lower-level managerial positions with authority only over basic day-to-day activities. The span of control constraints also meant that a limited number of managers had time to process new information, do the necessary quality checks, develop an accurate cost-benefit perspective, and arrive at a more informed view about fairly obvious productivity-enhancing practices. In a study on competitiveness of auto-parts firms in India, the on-duty line manager at the plant level was found not to have authority to ‘pull the line’ at the first sight of a defect and needed to gain permission from the SME owner (Saraf 2016). It also raises questions about the extent to which such family-owned firms invest in the skills of the manager to solve the more unusual problems (Caliendo and others 2015). With respect to incentives to adopt training, a recent experimental study from the field of organizational economics shows how poorly informed human resource policies could affect worker incentives to adopt new training (even when the firm management wanted to introduce it). This work is a cautionary tale about poorly designed worker compensation structures that reduce their motivation to adopt a new form of training (Atkin and others 2015). The paper studies technology adoption in a cluster of soccer-ball producers in Sialkot, Pakistan. The authors invented a cutting-edge technology that reduced waste of the raw material (benefiting the firm’s profits), and gave the technology to a random subset of producers. However, after 15 months, take-up remained very low. Not only was the technology not disseminated, it was not even adopted. The suggested explanation has to do with the perverse incentives impacting the workers in charge of cutting activities. Most workers were paid a piece rate, and the new die would slow the cutters in the learning period—but also potentially in the longer run. As a consequence, if cutters are paid at a pure piece rate, their effective wage falls. Therefore, unless owners modify the payment scheme, the benefits of the new technology accrue to the owners and the costs are borne by the plant workers. Realizing this, the workers resist adoption. In this case, information asymmetries 22 seemed to work in both directions. First, management’s misperception about the continued benefit of using piece-rates, and second, employees misinforming the owners about the value of the technology to prevent its uptake. Evidence from the OECD validates this, and suggests that firms that adopt more performance-based schemes are more likely to offer job training (for example, Gittleman, Horrigan and Joyce 1998). Regarding the final issue of information asymmetry about the types of programs available in the market, two studies (Bloom and others 2013; Atkin and others 2015) find that it exists, but is not a critical problem. For textile firms, for less commonly known practices, once this information was provided to firm owners through consulting firms, adoption rates improved drastically. In the Pakistan study, the only firm that successfully adopted the new technology was found to pay a performance bonus (a better gain-sharing arrangement), indicating that such knowledge was available in the market. However, few owners adopted the new practice. These cases indicate that compared to information asymmetry about types of training available, it is in fact, information barriers within the firm at the level of managers/owners, that are perhaps more binding in nature. 3.2 Factors outside firms: Low competition Studies show that the environment in which the firm operates matters in terms of OJT adoption. External factors (such as competition) can act as levers to pull internal factors (such as management capability and adoption of complementary technologies). Firms facing greater competition in their product markets are inclined to raise the minimum productivity threshold to operate profitably and reduce inefficiencies. They do this through both investing in productivity- enhancing activities (such as OJT and ICT), and reducing costs, which in turn helps them capture greater market shares of existing ones (Almeida and Aterido 2008; Lopez-Acevdeo and others 2016). Competition also helps to reallocate resources from the less productive to the more productive firms, increasing the incentives for all firms, to invest in the within-firm productivity levers discussed in section 3.1. The literature also indicates that trade and foreign direct investment (FDI) can affect productivity through providing greater opportunities for exporting and learning by doing, increasing access to new markets, more competition from imports, and selection and reallocation effects (Melitz 2003; Pavcnik 2002; Bloom, Draca and Van Reenen 2016). Complementary studies have examined the role of human capital in the process of adopting new technologies—which becomes even more important as markets open up and standards become more demanding (Bartel and Lichtenberg 1987; Berman, Bound, and Griliches 1994; Bugamelli and Pagano 2004; Doms, Dunne, and Troske 1997; Disney, Haskel and Heden 2003). Let us focus on the case of exporters as an example. Bustos (2011) found that exporters tend to be more skill intensive in both production and non-production occupations in Argentina. She also showed that technology and skill deepening are positively associated at the firm level. Brambilla, Lederman and Porto (2012) observed that the nature of destination markets matters for the type of skills provided. In more sophisticated destination markets, exporting firms need to meet higher standards, while also being cost competitive. These results, in the form of an exporter productivity premium, are confirmed in developing countries by Blalock and Gertler (2004), Clerides, Lachand Tybout (1998), and De Loecker (2007). Export is linked to being able to make higher quality products, which requires better human capital (Fieler, Eslava and Xu 2014; Khandelwal 2010). 23 In the auto-parts sector in India, a handful of auto suppliers who were exposed to greater competition since the 1990s achieved higher productivity and became global suppliers, while car- makers who were protected remained sub-scale and unproductive to a large extent (Saraf 2016). Opening up markets through reduction of import tariffs from 60 to 12.5 percent since the 1990s enabled auto-parts suppliers to cross the fixed cost of exporting through accessing cheaper intermediate inputs, learning from working with discerning buyers in OECD and other markets, and forming technical alliances to learn new technologies. Although there were subsidized public training programs, these global suppliers invested in their own OJT programs to ensure a supply of talented line managers, business managers, and technical floor-level workers to maintain the high productivity levels required to participate in export markets. More product market competition is found to be linked to increasing aggregate management capability—another channel that affects provision of OJT, as previously noted. More competitive product markets would push the badly managed firms out and “raise the stakes” for existing firms that could capture more market share and allocate better resources toward them (Bloom and others 2009).19 In such markets, managers would be more fearful of losing their jobs and firm owners would be more likely to select competent Chief Executive Officers (CEOs) (ideally non-family member).20 Another way more competition and trade could influence internal firm levers is by helping firms grow or creating the conditions for this. Large firms tend to invest more in skills, R&D and other technologies. Indeed, studies also show a positive correlation between firm size and OJT provision in developed and developing countries (Almeida and Aterido 2008; Black and Lynch 2001; Frazer 2006; Frazis, Gittleman and Joyce 2000; Pierre and Scarpetta 2004; and Rosholm and others 2007). Some of these studies also find that firms that initiate more innovative activities (for example, more investment in R&D) tend to train more. Pierre and Scarpetta (2004) and Almeida and Aterido (2008) find that larger and more innovative firms tend to train more (based on data from the World Bank Enterprise Data Surveys for 66 countries). There is evidence of a positive and significant relationship between OJT and firm characteristics (size, innovative activities, and technology adoption) for the Latin America and Caribbean region (Gonzales-Velosa, Rosas, and Flores 2016). Overall, the argument for competition and trade increasing the incidence of OJT is not very different from the reasons why firms exposed to competition increase their demand for other productivity-enhancing activities, such as ICT or R&D.21 However, as in the case of these activities, it remains unclear whether the directional link is from size to productivity, or vice versa. Do firms grow larger first, obtain economies of scale, and then demand new technologies and OJT? Or, does it work the other way around? Do firms that are able to adopt productivity- 19 In their surveys, Bloom and others (2009) asked managers how many competitors they faced, and found the average management score was significantly higher when firms reported facing more competitors. Using other measures of competition not reported by managers, like the import penetration rates (measured by imports as a share of domestic production) or Lerner indices of competition, yields a similar general result that management quality tends to increase with competitive intensity. 20 Meritocratic selection of a CEO (ideally a non-family member) is linked to better management outcomes. However, as family-owned firms typically have less debt, they are less vulnerable to product market pressures, and continue to survive—although with economic losses. 21 See Bayo-Moriones and Lera Lopes 2007; Dasgupta and others 1999; Giunta and Trivieri 2007; Haller and Siedschlag 2011; Hollenstein 2004; Keller, 2004; Kowtha and Choon 2001; Kretschmer, Miravete, and Pernías 2012; Narula and Zanfei, 2005; and Lucchetti and Sterlacchini 2004. 24 enhancing activities cross the fixed cost to export markets and become bigger? Some clues to potential answers are offered by the seminal work by Melitz (2003), who put forth an explanation connecting firm size to a firm’s total factor productivity. Melitz proposed a theory for selection into exporting: only the most productive firms overcome the sunk costs of exporting and foreign direct investment (FDI), start to export, and grow as a result.22 This remains an area for future investigation. Finally, the effect of the regulatory environment, such as labor market reforms, can vary depending on the exact kind of regulation. Reforms that increase the number of temporary workers — and simultaneously protect existing employees through unions—are linked to lower OJT. While stricter hiring regulations and rigidity in the working hours are associated with a higher investment in OJT, stricter firing costs tend to be associated with reduced investment in job training by employers (Almeida and Aterido 2008). There is a lot of variation in results depending on the type of labor regulation. Constraints in the credit markets do not appear as binding constraints (box 6). Box 6: The Limited Role of Credit Constraints The literature review did not provide much evidence regarding the role of credit constraints as binding barriers to the provision of OJT. Although in most developing countries, capital markets could be imperfect, these imperfections do not appear to have a significant bearing on the under-provision problem. Carneiro and Heckman (2004) suggest a low importance of credit constraints in explaining investments in education. Hannson (2008), using the International Adult Literature Survey (IALS), finds that 25 percent of workers reported wanting to take further training. However, only 4-7 percent claimed that it would be too expensive as the reason for not participating. In evidence from Ghana, Gorg, Strobl and Walsh (2007) note that lower wage bounds — like credit constraints, efficiency wages or minimum wages — do not prevent OJT. In the case of microenterprises, credit constraints could play a role. In an experiment from Sri Lanka, De Mel, McKenzie, and Woodruff (2008) find high returns — higher than existing interest rates — to drops of cash of US$100 or US$200 (or of capital of roughly similar magnitudes). This suggests that the microenterprises might be operating under credit constraints. However, for microenterprises, skills have not arisen as the top constraint in analyses, according to the World Bank Enterprise Surveys. Low-income workers might be more affected by credit constraints, suggesting some role for government support. However, for most firms, the marginal importance of credit constraints for OJT provision is likely to be a consequence of employers financing most of the training, and workers being accustomed to receiving training without paying any direct costs (although they contribute indirectly through compressed wages). Studies have pointed to other challenges involving information gaps about returns, ineffective HR policies, or the barriers to demand for training as the main constraints. 22 Empirical data seems to support this (Bernard, Eaton, Jensen and Kortum 2003; Mayer, Melitz and Ottaviano 2014). In a recent World Bank study regarding competitiveness in South Asia (World Bank 2016), researchers found that across sectors, local firms were managing to cross the productivity barrier by learning through working with lead firms in the domestic market, acquiring technologies and/or becoming more efficient. Thereafter, as the markets were deregulated and new markets opened (for example, in the auto-parts sector as well as apparel), they were able to meet the stricter requirements of export markets. In addition, they were also able to compete with incoming imports. In time, such firms grew in size. 25 4. Policy Considerations and Areas for Future Investigation What are the policy implications of these findings? What are areas for future investigation? This paper confirms that OJT could be a powerful instrument in helping firms become more productive and keeping abreast of technological change. At the same time, workers benefit through wage gains at current and future employers. Thus, the first step for policy makers and practitioners would be to familiarize themselves with the empirical data on returns to OJT, as has been captured in this report. OJT programs have shown a high rate of return where studied carefully. The evidence also shows that future employers benefit from hiring previously trained workers, indicating that the social benefit is greater than the private benefit—thereby making a case for OJT as a public good of sorts. This paper clarifies the extent to which oft-mentioned theoretical constraints are real constraints and backed by data. This is critical to any evidence-based policy- making effort. To improve how the public sector supports and engages with OJT, it would need to communicate, partner with and improve its understanding of firm-level challenges (for example, how information failures operate within firms, how human resource policies are formulated, and how to help firms upgrade skills), as well as work more closely with private firms. This would require bolstering the capacity of public institutions that work with the private sector to include skilled personnel who can effectively help the private sector. Finally, regarding specific policy considerations, it is important to highlight that the guiding principle for any policy in this area should be to correct for market failures that prevent firms and workers from fully internalizing the costs and benefits of OJT — rather than crowding them out.23 In the sections that follow, we will discuss considerations for policy makers, utilizing the framework of demand- and supply-side constraints (table 1). Conclusions follow with a discussion of areas for future research/investigation, as well as remaining open questions. 4.1 Demand-side considerations24 Regarding demand-side considerations, the following question is posed: What policy directions would help increase the uptake of training/complementary activities leading to an attendant increase in the demand among firms for OJT? The demand for OJT is more likely to improve in environments of high product market competition (see section 4.1.1), more information within the firm (see section 4.1.2) and improved firm management capability (see section 4.1.3). For workers, presence of aligned incentives in the form of gain sharing compensation could help increase uptake of OJT opportunities (see section 4.1.4). 4.1.1 Reforming product market regulation, increasing competition and trade The literature suggests an important role for (product market) competition in encouraging the adoption of new practices and technologies. Competitive environments act as a channel to increase 23 The case of OJT is one of positive production externalities — a rare form of externalities. To alleviate this market failure, policies that decrease the marginal cost to the firm of providing OJT would bring levels closer to the social optimal. In order to achieve a level close to the social optimal, the marginal benefit to the firm of providing OJT would need to increase (see section 4.1 on demand-side considerations), and/or, the marginal cost would need to decrease (see section 4.2 on supply-side considerations). 24 These would increase the marginal benefit of providing OJT to the firm, or to all firms in the market. 26 the demand for human capital. There are many types of product market reforms. Although it is outside the scope of this paper to go into the details of all instruments, the key ones are noted here. The principle driving product market reforms is to increase competition in the domestic market, as well as to link markets across domestic and international firms. This not only increases the opportunities for local firms (through the trade route), but also allows them to interact with other firms. In doing so, such firms can learn new technologies and adopt better practices. These could be achieved through: (i) reforming restrictive trade policies that could protect unproductive local firms and restrict the expansion of the more productive firms; (ii) aligning domestic and international standards so that local firms have the incentive to learn new technologies to serve the demands of more sophisticated markets overseas; (iii) strengthening trade logistics to reduce the cost of doing business locally and internationally; (iv) supporting firm-level capacity building through incentives for R&D, as well as improvements in management practices and production processes; and (v) reforming distortionary regulations that create an incentive to remain small e.g. labor regulations for SMEs. These would also enable local firms to access opportunities in international value chains, thereby providing an impetus to improve productivity.25 The example of an auto part firm from India that became a global supplier in the wake of the liberalization of import tariffs (from 60 to 12 percent in 10-15 years) is instructive. In a recent sectoral study from the World Bank, Saraf (2016) traces the evolution of the automotive sector in South Asia. The Chairman of the leading auto-parts company is reported mentioning: “We train workers and lose them to others. But we still train because the ones who stay are crucial for our productivity. Unskilled workers are cheaper, but costs match up when their mistakes are financially accounted for. We cannot afford that type of delay and quality when we need to serve clients in Germany and the US.” The channels through which product market reform would help in increasing the uptake of OJT are likely to be two-fold. First, at the market level, it would amplify the number and size of opportunities, and increase the demand for productive activities that would require better human capital to realize the benefits.26 Firms that do not invest in improving their human capital would 25 A useful area for future research would be to test the sensitivity of OJT to various aspects of product market reforms. Some common product market reform areas include: reducing average rates of protection and harmonizing tariff schedules across intermediate and final goods; improving standards and product market regulations; facilitating imports for exporters (for example, through better functioning duty-drawback schemes); and strengthening trade logistics to reduce customs clearance, tax burden and transit times. Many services-related regulations might go beyond the scope of “product-market” reform, for example, e-commerce rules. Depending on the country and the sector context, further analysis would be required to ascertain the appropriate instrument. World Bank work regarding competitiveness in South Asia (Hallward-Driemeier 2007) suggests that an inefficient business environment would lead to low and uncertain returns on investment, dragging down productivity and reducing the incentive to adopt technical improvements on the factory floor. As seen from the literature, this would, in turn, reduce demand for complementary activities such as OJT. 26 It is possible that even when the size of opportunity increases in the light of reforms, firms may choose production processes and technologies that might be capital-intensive — avoiding the need for more investment in human capital. This seems unlikely in developing countries because labor inputs tend to be relatively cheaper. Further, firms that are profit-maximizing entities will not want to under-utilize a cheaper input. Second, training might increase the 27 be penalized more, thereby raising the marginal benefit to training for firms that do provide training. Such firms can then capture a greater market share (a new share in new markets, as well as capturing shares away from competitors in existing markets). To some extent, this would tackle the issue of free-riders because in competitive markets, the marginal cost of not training for “all firms” would increase during the same time period. The future employer would need to invest in training at the same time as the current employer, and both would have similar risks of poaching. Second, within the firm, managers would have to make greater efforts to conduct due diligence regarding information about new practices and technologies to remain competitive. This would help to alleviate the information failure issues. Further, it would help firms create the right in- house incentives and policies to motivate workers to participate in training. Firms that export, meet international standards, innovate and adopt more advanced production technologies (such as MNCs) tend to have better management capability (Bloom and Van Reenen 2007) and more training incidence (in the Latin America and Caribbean region, see Gonzalez- Velosa and others 2016). In effect, such reforms would increase the ability of the product market to allocate resources efficiently. It would do so by enabling more productive firms to grow and the badly managed ones to exit. The literature about firm productivity also shows a link between productive activities and firm size: firms that are large tend to participate in exports, innovation, R&D and FDI. Large firms also tend to do more OJT.27 Thus, policies that support reforming licensing, size restrictions, labor laws and strengthening property rights might be steps in the right direction for developing countries. 4.1.2 Reducing uncertainty about returns, increasing information flows, providing certification The literature suggested the presence of information failures among firm management that reduce the adoption of beneficial OJT. As noted, product market reforms could provide an incentive for more due diligence on the part of firm managers and owners with regard to information processing and better HR policies. At the same time, governments could provide further support by using low- cost means to help firms interact and learn from each other. One such part of increasing information flows in the market concerns greater coordination among firms. Government-facilitated demonstration projects, networking sessions, and greater informational awareness efforts regarding evidence about productivity, wages and cost effects cost of inputs through wage hikes over time. However, the literature indicates that even in developed markets, firm productivity increases more than the wage gain. 27 Size and productivity: which comes first? Causality is difficult to establish, as firms that are already more productive are more likely to enter export markets than less productive firms (Amiti and Konings 2007; Atkin and others 2014; Melitz 2003; Clerides and others 1998; Goldberg and others 2010; and Wagner 2007.) Lopez-Acevedo, Medvedev, and Palmade (2016) note that “competition plays a key role in forcing inefficient, unproductive, or unprofitable firms to either improve or exit and transfer their resources to more efficient, productive, or profitable firms, thus boosting economy-wide productivity. More formally, there are two mutually reinforcing mechanisms — spurred on by competition in product and factor markets — that increase productivity. First, greater competition, from either domestic or international sources, pushes firms to become more efficient at doing what they do: for example, by learning from international exposure, investing in innovation, improving business practices, adopting better technology including ICT, and improving the input mix. This is the within-firm component of productivity growth. Second, competition also induces inefficient firms to transfer resources to more efficient ones or exit altogether, boosting economy-wide productivity — this is the between-firm component of productivity growth.” 28 could help firms to better understand the cost-benefit calculus of OJT. These could be cost- effective ways for the government to engage with the private sector, while also including academic institutions, industry leaders and business associations. Policies that support the formation and expansion of information platforms between large firms and local suppliers about lead firm requirements/standards on the one hand, and product offerings of local firms on the other hand, might help reduce the information asymmetry in the market. In so doing, it could increase the demand for skilled labor among small and medium enterprises (SMEs). In many cases, the information platform could serve to reduce information gaps regarding certifications of local suppliers by providing these data online. For public programs that help firm upgrading, it could have as a filtering criterion the accreditation of the training provider (including the firm). For workers, the literature clarifies that accredited OJT is more strongly associated with wage gains at both current and future employers. This would also suggest the importance of signaling and policy interventions for providing greater certification and formal recognition of prior learning (RPL). However, reduced asymmetric information between workers and firms could increase turnover, thereby discouraging firms from providing OJT. Thus, it is likely that the effects on OJT provision levels are mixed: the worker demand for OJT would increase due to better signaling, but the firm supply would fall due to greater worker turnover. However, Acemoglu and Pischke (2000) show that by guaranteeing workers a higher wage, certification provides an incentive for workers to exert more effort in the skill acquisition process. If so, then this might alleviate some of the supply contraction on the firm side. 4.1.3 Helping improve management skills for better HR policy formation The literature shows this is one area of skills that has an impact on firm productivity and worker incentives to adopt new skills. Increasingly aware of the importance of investments in managerial capabilities, governments across the world are experimenting with programs that boost these capacities at the firm level. In Latin America and Africa, for example, authorities are piloting interventions which provide SMEs with access to individualized consulting services (similar to the approach by Bloom and others 2013), as well as more novel approaches of providing group consulting services. Such services can be delivered at a lower cost and can help leverage group- learning dynamics (similar to agricultural extension services). Another innovative approach to fill the management capability gap in smaller firms could be to provide incentives for diaspora knowledge transfer and exchange. Building on evidence that shows significant productivity and wage gains to hiring trained workers from MNCs, some countries have targeted programs and policy frameworks to encourage movement of diaspora members who have obtained superior management training from firms outside the country back to the home country (Aguinas and Newland 2012; Brinkerhoff 2006; and Meyer and others 1997). In such programs, the home country may facilitate diaspora members’ personal efforts to transfer skills to the homeland, or work to connect home-country firms to advanced institutions in countries of destination in which diaspora members have ties. In some cases, country-of-destination governments, acting alone or with regional organizations, may partner with home country governments to provide technical assistance. Although some of these initiatives have had a positive 29 impact, particularly in OECD countries where they were conceived, making such programs viable in the long run for developing countries is still a challenge.28 4.1.4 Using more gain-sharing instruments and performance-based compensation for workers Due to heterogeneous parties and agency issues, workers may not have incentives to cooperate with the firm in adopting new technologies and training. At times, the worker could transmit incorrect information about the impact of new technologies to management to restrict its uptake, causing further inefficiencies. To avoid this problem, the greater use of gain-sharing contracts and performance-based pay for workers are suggested. Studies show that firms using such instruments are more readily able to adopt new technologies. Reforms that increase competition in these markets would prompt better compensation policies through two routes. First, firms would want to invest in better technologies, creating a rise in demand for training, a complementary activity. Second, in these markets, the firm management would need to be more diligent about designing attractive HR and compensation policies for fear of losing good workers, thereby alleviating issues of misaligned incentives among workers. 4.2 Supply-side considerations29 In general, government interventions - such as subsidizing training or providing direct financial support - to increase the supply of OJT need to be considered with caution as they could crowd out what private sector firms would have already provided. As seen in many countries, this could then reduce the quality of overall OJT provision.30 Government support towards firms to encourage more OJT should minimize crowding out existing private provision through appropriate targeting and due-diligence of the firms (see section 4.2.1). Through ensuring enforceability of contracts and lowering coordination costs among firms, governments could encourage more supply of OJT (see sections 4.2.2 and 4.2.3). The review indicates that demand constraints are more critical to address as the way to increase OJT provision. However, it also finds that there are cases where a private sector led OJT approach would lead to distributional issues and tend to underprovide certain types of skills that are important for the workers (and for society) in the long run but less for the firm in the short run. As such, when societies have improved welfare as a goal, interventions could be required to restore equality of opportunity rather than reinforce existing inequalities (see section 4.2.4). 4.2.1 Targeting selective firms for government subsidies Government financial support should only go where training would otherwise not occur, or to internalize a positive externality by reducing the marginal cost of OJT provision for firms. In this context, co-financing schemes are a better tool for the alignment of incentives with workers and training providers (in this case, firms), as well as a better use of public monies. Very rarely should the government pay the full cost of training given the presence of other parties that benefit. The 28 Some examples of diaspora programs that encourage skill transfer are: the Return of Qualified Afghans (RQA) program, the Moroccan Migrations and Développement Economique dans la région de l’Oriental (MIDEO), Colombia’s Red Caldas de Colciencias, Senegal’s Senexpert programme, and ChileGlobal in Chile. 29 Supply-side considerations would decrease the marginal cost of OJT to the firm, or to all firms in the market. 30 Since training is difficult to monitor or quantify, there are serious risks of government failure, such as low-quality training by an external third party (for example, an ALMP or private TVET institutes). An external party may find it more difficult to motivate both the firm and the worker to behave efficiently. 30 employer-worker combination should bear the remaining cost. This would also require better targeting mechanisms for public support to training. In order to encourage an activity that has positive externalities for society, policies should use appropriate targeting methodology to help those firms that have a higher propensity to provide training, but need an additional push to cross the fixed cost point of training. What the specific targeting indicators would be would depend on the country and the sector context. However, the literature confirms a few common features that targeting methodologies could consider including: firms engaged in productive activities (exports, innovation, R&D, selling high-quality products, ICT adoption) tend to offer more OJT. Other firm-level metrics could also be considered, such as the presence of professional managers in family-owned firms (better management); performance- based pay (alignment of worker incentives); and the number of technical alliances (new technologies). The targeting methodology should include an assessment of the sensitivity of these characteristics (export-orientation, involvement in R&D) to the propensity to provide OJT, the growth trajectory of the firm (over the short and medium run), and a sense of what the firm is already investing to prevent crowding out. How to target and with what kind of training is an emerging area of research in private sector development. In a recent randomized control trial (Anderson, Chandy & Zia 2016), businesses that received training increased profits significantly compared to those in the control group. Marketing training was significantly more effective for businesses with lower prior business exposure, whereas finance training was significantly more effective for larger, more established businesses. This indicates that programs targeting training at firms to improve management skills must look at the starting point of the company.31 For firms at the size spectrum ends, that is, for the very large and micro firms, government support is likely to crowd out other training, or simply be ineffective. For instance, large firms linked to export markets are known to train with or without government support. If government offers its support through subsidies, it would crowd out this segment of the private sector, creating a deadweight loss for society. For firms in the middle – SMEs – with the above-mentioned characteristics, there is a higher likelihood of success. However, for firms on the margin, government intervention may have a net positive effect by encouraging greater training that would not otherwise have occurred without its support. The question of policy instrument could be an important one. Whether a tax deduction, a levy/grant scheme, or a grant, these instruments are prone to abuse and require careful design and monitoring by the implementing agencies. For example, in the absence of appropriate measures and monitoring, the employer may use government support toward other organizational needs, subsidize wages of employees, or train those who would otherwise have received training. Thus, a correct evaluation of the impact of different instruments requires information regarding the responsiveness of wages and training prices to changes in payroll taxes and other instruments. Each policy choice also has different administrative costs, as well as a bureaucratic responsibility 31 Companies in two treatment groups were given marketing training and finance training respectively. The authors found that businesses assigned to marketing training were more likely to implement practices connected to top- line business growth (e.g. market research, marketing tactics, and sales tactics). In contrast, finance training tended to shift a firm owner’s focus towards greater efficiency in the business through more finance/accounting activities that economized on costs to boost profits. 31 for monitoring and compliance. While much data exists for developed/OECD countries, it is a policy area requiring further experimentation and data collection in developing countries. To summarize, in order to calculate the optimal size of government support, policy makers would need to consider the type of firm, whether its skill efforts are systemic, how close it is to the frontier, its absorptive capacities (including management quality), and the extent of externalities. 4.2.2 Lower cost of poaching for firms One question to consider regarding costs is: In which ways could companies coordinate to reduce the cost of poaching? In this regard, it would be instructive to learn from examples that have not worked before moving on to some new approaches. When four key Silicon Valley companies (Apple, Adobe Systems, Google, and Intel) signed a non-poaching agreement (2005-2009) to prevent employee attrition, it resulted in a lawsuit by the employees. In 2009, the companies decided to settle. The rationale of the companies was two-fold: first, to avoid having to rehire and retrain employees given the high levels of labor mobility in the technology sector; and second, to protect intellectual property to which the employees had become privy. However, this provoked a heated debate about anti-competitive practices and forced the four companies to settle. These tactics are also unlikely to work in developing country contexts where enforceability of contracts is often weak. Other options to consider could be instruments, such as Non-Compete Contracts (NCCs), payback clauses or back-loading bonuses to increase the worker cost of switching, as well as the attrition for the firm.32 With NCCs, the employee is not allowed to join a competitor for a defined period of time after quitting the current firm. This reduces labor mobility. In effect, it allows the company to reduce the benefits that the future employer could use. In the case of back-loading contracts, the employee receives a bonus at the end of a period during which he or she agrees to stay with the company. It is up to the employee to decide whether he/she prefers to stay for that length of time [to obtain the bonus], or whether the outside offer is more attractive. All of these instruments increase the cost of switching for the employee, and reduce the (probability adjusted) marginal cost for the firm. With payback clauses, the employee has to pay back the cost of training if he/she leaves before a certain time period, as contractually agreed upon prior to the training. The common basis among these instruments that help to increase the switching cost among employees and reduce the cost of poaching is the assurance of rule of law and enforceability of contracts. This is an area where developing countries would need to pay more attention and communicate policies that would ensure a stronger rule of law among private sector firms. 4.2.3 Lower coordination costs for a group of firms Governments could try to reduce coordination costs among firms through the provision of common/sharing facilities or industry-/cluster-based training centers. In the past, the initiative, land, and some financial support came from the state and federal governments. However, the curriculum design, methodology, teacher recruitment, and other management matters were decided by the private sector or non-profit organization.33. The private or non-profit entity assumes the role 32 Downloaded on December 11, 2016 from: http://knowledge.wharton.upenn.edu/article/silicon-valleys- poaching-case-growing-debate-employee-mobility/ 33 See the example of the Penang Skill Development Centre (a public-private partnership [PPP] skill center), Dundar and others (2014). 32 of an ‘aggregator’, and the government acts as a facilitator. The government supports capital expenditures, such as equipment for training. After a few years of operation, the training center can then assume responsibility for recurring costs (materials, trainers, and so on). In some cases, the private sector firms can share the cost of introducing accredited providers and content. In other similar models, grant funding from governments could help industrial clusters cover basic start-up costs by establishing common/shared facilities — with the firms donating space and paying the cost of training (now arguably lower per person due to economies of scale). Thus, instead of financing the direct cost of training, government resources support a group of firms, helping them to meet the fixed costs of cooperating to provide skills that are commonly required. In theory, these arrangements could work because they do not help a single firm, but rather a group of firms to assess common skill gaps, produce skill upgrading plans, create apprenticeship programs, and obtain skills for a fee. In practice, the impact and efficiency of such programs is not yet fully known. Rigorous impact evaluations are required to measure the outcomes of such programs, which are becoming a popular way of increasing coordination among firms and addressing positive externality issues. 4.2.4 Targeting support toward disadvantaged groups Better targeting is also required for workers to alleviate distribution issues. The literature notes that highly skilled workers are more likely to receive training, whereas certain groups of workers (including women and older people) receive less training. The government approach could vary based on the type of skills in question, as well as the seriousness of the distribution issue. For more general skills and groups that may be at a disadvantage but remain capable of accessing credit markets (that is, at a non-serious disadvantage), policies that support an easing of regulations around access to finance for education and training could encourage greater participation. This could take the form of training and education loans with low interest rates or credit default guarantees. With a credit default guarantee, instead of directly paying for skills, the government allows for a guarantee or form of insurance to the training provider (in case the trainee defaults on the loan). This is achieved by the issuer of the bonds insuring the buyer’s potential losses as part of the agreement. The efficiency of these investments would need to be tested, taking into account the welfare gains. In cases of serious distribution issues, the government could direct greater resources toward general training for disadvantaged groups. In this regard, a short-term policy intervention to provide targeted government support on the grounds of equity would be best. However, governments should use more efficient training methods than ALMPs. The literature review shows that the following programs maximize the chances for success: (i) a comprehensive program combining traditional in-classroom training with workplace training; and (ii) supplementary services, such as counseling and mentoring, monitoring, job search and placement assistance, as well as soft and life skills training. This means that training programs for targeted worker groups should aim to create an opportunity for all of these elements to come together, if they are to have a chance of genuinely helping them. For instance, classroom learning provides general skills that increase labor mobility across a worker’s lifetime. However, it may not provide practical knowledge of technical skills in specific occupations which may be required to cross initial barriers to entry and survival in the firm. This 33 is where the workplace training component refines practical skills of the trainees and exposes them to industry standards, technical equipment, and gives them a sense of operating in a real workplace. It is important for trainees to fit in well into the new environment. This will also help to increase worker retention. The trainee needs access to a system for counseling, mentoring and job matching services to find the right opportunities. Over the longer term, more investment and better learning outcomes in basic and secondary education, as well as achievement of lifelong skills for such groups, would be needed to correct the problem at its origin in the future. 4.3 Areas for future investigation There are a few interesting analytical questions for future investigation. However, robust literature was found to be limited. For example: What is the relationship between economic geography/agglomeration and provision of training? What is the relationship between different public-private partnership (PPP) models of training and the impact on a firm’s provision level of OJT? What are the empirical reasons for the lack of private training markets for general OJT? These could be stimulating questions for future investigation. In general, much more work is also required to identify the conditions (that is, the characteristics of the firms, type of training, and types of workers) in which OJT is offered, as well as the economic mechanisms through which OJT contributes to firm productivity. This would improve the targeting of policy interventions. Finally, more attention is needed with regard to techniques to resolve methodological concerns associated with OJT measurement. 5. Conclusion While creating gains for workers and society, firms that provide on-the-job training do so when it is critical to their productivity—and when productivity is critical to their survival. This paper begins by confirming a significant and positive return from on-the-job training on wages and productivity, as well as the presence of positive externalities from on-the-job training, while discussing the methodological considerations at play. The paper then reviews and validates the presence of market failures such as information asymmetries within the firm as a result of low- quality management practices that dampen firm demand for on-the-job training. Lack of competition in the firm’s external environment appears to undermine adoption of on-the-job training and other complementary productivity-enhancing activities within the firm. The literature suggests that for most firms, a comprehensive policy approach that resolves external constraints to becoming more productive is likely to have a positive impact on the provision of on- the-job training and adoption of complementary policies. More direct forms of firm-level support to improve management capabilities could also alleviate under-provision of on-the-job training. Where societies have improved welfare as a goal, public policy measures would be needed to complement on-the-job training for some specific groups of workers (older, less educated, women). In essence, the paper highlights the importance of demand-side constraints for firms, rather than supply-side constraints, for the provision of on-the-job training. 34 Table 3: Summary of Findings from the Literature on OJT Constraint Explanation Findings (from Empirical Literature) Policy Considerations and (from Theoretical Areas for Future Work Literature) Value of OJT is not Literature indicates that the returns to OJT (firm Demand side considerations: understood by firms (accurate productivity and worker wages) are positive. Information information on productivity Internal rate of return is positive (see Section 2.1). Reforming product market gaps and returns, costs, wage rises); it competition, trade, uncertainty may be particularly regulatory environment and Literature presents validation of information gaps, about misunderstood by those firms other institutions (see and finds that firm-level management is often not returns that stand the most to gain, Section 4.1.1). well-informed about the cost-benefit of new and that do not realize how training, leading to ineffective HR policies (see Reducing uncertainty about poorly run their firms are. Section 2.1 and 3.1). gains, increasing Firm: Low adoption of new Evidence of external environmental factors such as information flows, providing managerial practices or new low product market competition linked to low certification systems (see Demand- Barriers to technologies among firms — adoption of productivity-enhancing activities (OJT, Section 4.1.2). side adoption of reducing their need for betternew technologies, such as information and constraints managerial skills, and hence, demand for communications technology [ICT], management Helping to improve practices training. practices). Within firms, low adoption of management skills for better and new managerial practices in developing countries with HR policy formation (see technologies Worker: Misaligned preponderance of family-owned firms depresses Section 4.1.3). compensation structures limit demand for productivity-enhancing activities such worker incentives to train in as OJT (see sections 3.1 and 3.2). Using more gain-sharing new technologies. instruments and Even if the firm management wants to adopt new performance-based technology, there is evidence of distorted HR compensation for workers policies (linked to information gaps) that creates (see Section 4.1.4). low incentives for workers to participate as they do Supply side considerations: not share in the gains (see sections 3.1 and 3.2). Targeting selective firms for Information Firms do not know what kind Literature review does not suggest this is a binding the provision of government asymmetry of training is available in the constraint because some firms in the market are subsidies (see Section market, and/or how to go able to access this information, while others 4.2.1). about providing for it. cannot. Simple interventions, such as increasing information awareness, can help resolve the Lower coordination costs problem (see section 3.1). among firms (see Section 4.2.2). Free-rider Firms that provide general There is evidence of productivity benefits for problem OJT are not able to fully future employers profiting from hiring workers Targeting support to Supply-side appropriate the rents to trained in the previous firm. Wage gains for disadvantaged groups (see constraints training. existing workers in the future employer are higher Section 4.2.3). when the firm hires a trained worker (for example, Areas for future work: from MNCs) (see section 2.1 and 2.2). More rigorous analytical Credit Firms do not have adequate For most firms, a lack of convincing evidence work on OJT link to PPPs; constraints resources to finance OJT, about the presence of credit constraints is a impact of agglomeration on even though it may have significant reason why firms under-provide OJT; and resolving positive returns. training. This could be a bigger issue for methodological issues for microenterprises and low-income workers (see better OJT measurement. section 3.2). 35 References Acemoglu, D., & Pischke, J.-S. (1998). Why Do Firms Train? Theory and Evidence. The Quarterly Journal of Economics, 113(1), 79–119. Acemoglu, D., & Pischke, J.-S. (1999). The structure of wages and investment in general training. Journal of Political Economy, 107(3), 539–572. Acemoglu, D., & Pischke, J.-S. (2000). Certification of training and training outcomes. European Economic Review, 44(4-6), 917–927. Retrieved from http://econpapers.repec.org/RePEc:eee:eecrev:v:44:y:2000:i:4-6:p:917-927 Acemoglu, D., Kremer, M., & Mian, A. R. (2008). Incentives in Markets, Firms and Governments. Journal of Law, Economics, and Organization, 24(2), 273–306. Acemoglu, D. (1998). Why Do New Technologies Complement Skills? Directed Technical Change and Wage Inequality. The Quarterly Journal of Economics, 113 (4), 1055-1089. Acevedo, G. L., & Tan, H. W. (Eds.). (2011). Impact evaluation of small and medium enterprise programs in Latin America and the Caribbean. World Bank Publications. Aghion, P., & Tirole, J. (1997). Formal and Real Authority in Organizations. Journal of Political Economy, 110(3), 567–591. Retrieved from http://econpapers.repec.org/RePEc:ucp:jpolec:v:105:y:1997:i:1:p:1-29 Agunias, D. R., & Newland, K. (2012). Developing a road map for engaging diasporas in development: A handbook for policymakers and practitioners in home and host countries. IOM/MPI. Akerlof, G. A., & Kranton, R. E. (2005). Identify and the Economics of Organizations. Journal of Economic Perspectives, 19(1), 9–32. Albinger, H. S., & Freeman, S. J. (2000). Corporate Social Performance and Attractiveness as an Employer to Different Job Seeking Populations. Journal of Business Ethics, 28(3), 243– 253. N- Alfaro, L., & Charlton, A. (2009). Intra-industry foreign direct investment. American Economic Review, 99(5), 2096–2119. http://doi.org/10.1257/aer.99.5.2096 Almeida, R. K., & Aterido, R. (2008). The incentives to invest in job training: do strict labor codes influence this decision. The World Bank Social Protection Discussion Paper, (0832). Almeida, R. K., & de Faria, M. L. (2014). The Wage Returns to On-the-Job Training: Evidence from Matched Employer-Employee Data. IZA Discussion Papers. Retrieved from http://ideas.repec.org/p/iza/izadps/dp8314.html Almeida, R., & Carneiro, P. (2009). The return to firm investments in human capital. Labour Economics, 16(1), 97–106. Retrieved from http://econpapers.repec.org/RePEc:eee:labeco:v:16:y:2009:i:1:p:97-106 Almeida, R., Behrman, J., & Robalino, D. (2012). The Right Skills for the Job? (R. Almeida, J. Behrman, & D. Robalino, Eds.). The World Bank. http://doi.org/10.1596/978-0-8213- 8714-6 36 Ambec, S., & Barla, P. (2006). Can Environmental Regulations be Good for Business? an Assessment of the Porter Hypothesis. Review of Energy Studies, 14(2), 42–62. Amiti, M., & Konings, J. (2007). Trade liberalization, intermediate inputs, and productivity: Evidence from Indonesia. The American Economic Review, 97(5), 1611-1638. Anderson, S., Rajesh, C., & Zia, B. (2016). Pathways to Profits: Identifying Separate Channels of Firm Growth through Business Training. World Bank Policy Research Working Paper no. 7774 Andreoni, J. (1989). Giving with Impure Altruism: Applications to Charity and Ricardian Equivalence. Journal of Political Economy, 97(6), 1447–58. Retrieved from http://econpapers.repec.org/RePEc:ucp:jpolec:v:97:y:1989:i:6:p:1447-58 Andreoni, J. (1990). Impure Altruism and Donations to Public Goods: A Theory of Warm-Glow Giving? Economic Journal, 100(401), 464–77. Retrieved from http://econpapers.repec.org/RePEc:ecj:econjl:v:100:y:1990:i:401:p:464-77 Aronsson, T., & Johansson-Stenman, O. (2008). When the Joneses’ consumption hurts: Optimal public good provision and nonlinear income taxation. Journal of Public Economics, 92(5- 6), 986–997. Retrieved from http://econpapers.repec.org/RePEc:eee:pubeco:v:92:y:2008:i:5-6:p:986-997 Aterido, R., & Hallward-Driemeier, M. (2007). Investment climate and employment growth: The impact of access to finance, corruption and regulations across firms. Atkin, D., Chaudhry, A., Chaudry, S., Khandelwal, A. K., & Verhoogen, E. (2015). Organizational barriers to technology adoption: Evidence from soccer-ball producers in Pakistan (No. w21417). National Bureau of Economic Research. Backhaus, K. B., Stone, B. A., & Heiner, K. (2002). Exploring the Relationship Between Corporate Social Performance and Employer Attractiveness. Business and Society, 41(3), 292–318. Bagnoli, M., & Lipman, B. (1989). Provision of Public Goods: Fully Implementing the Core through Private Contributions. Review of Economic Studies, 56(4), 583–601. Retrieved from http://econpapers.repec.org/RePEc:oup:restud:v:56:y:1989:i:4:p:583-601. Balsvik, R. (2011). Is labor mobility a channel for spillovers from multinationals? Evidence from Norwegian manufacturing. The Review of Economics and Statistics, 93(1), 285-297. Bandiera, O., Barankay, I., & Rasul, I. (2011). Field experiments with firms. The Journal of Economic Perspectives, 25(3), 63-82. Banerji, A. C. W. F. A. K. E. P. H. R. D. T. J.-P. (2010). Stepping up skills for more jobs and higher productivity, 1–36. Retrieved from http://documents.worldbank.org/curated/en/2010/06/12515032/stepping-up-skills-more- jobs-higher-productivity Baron, D. P. (2001). Private Politics, Corporate Social Responsibility, and Integrated Strategy. Journal of Economics & Management Strategy, 10(1), 7–45. Retrieved from http://econpapers.repec.org/RePEc:bla:jemstr:v:10:y:2001:i:1:p:7-45 37 Barrett, A., & O'Connell, P. J. (2001). Does training generally work? The returns to in-company training. Industrial & labor relations review, 54(3), 647-662. Barron, J. M., Black, D. A., & Loewenstein, M. A. (1989). Job Matching and On-the-Job Training. Journal of Labor Economics, 7(1), 1–19. Retrieved from http://ideas.repec.org/a/ucp/jlabec/v7y1989i1p1-19.html Bartel, A. P. (1995). Training, Wage Growth, and Job Performance: Evidence from a Company Database. Journal of Labor Economics, 13(3), 401–25. Retrieved from http://econpapers.repec.org/RePEc:ucp:jlabec:v:13:y:1995:i:3:p:401-25 Bartel, A. P., & Lichtenberg, F. R. (1987). The comparative advantage of educated workers in implementing new technology. The Review of Economics and statistics, 1-11. Bartel, A. P., & Sicherman, N. (1999). Technological Change and the Labor Market. NBER Reporter, 16. Bayo-Moriones, A., & Lera-López, F. (2007). A firm-level analysis of determinants of ICT adoption in Spain. Technovation, 27(6), 352-366. Becker, G. S. (1962). Investment in human capital: A theoretical analysis. Journal of Political Economy, 70, 9–49. Benabou, R. J., & Tirole, J. (2003). Intrinsic and Extrinsic Motivation. Review of Economic Studies, 70(3), 489–520. Retrieved from http://econpapers.repec.org/RePEc:oup:restud:v:70:y:2003:i:3:p:489-520 Bergstrom, T., Blume, L., & Varian, H. (1986). On the Private Provision of Public Goods. Journal of Public Economics, 29, 25–49. Berman, E., Bound, J., & Griliches, Z. (1994). Changes in the demand for skilled labor within US manufacturing: Evidence from the annual survey of manufacturers. The Quarterly Journal of Economics, 367-397. Bernard, A. B., Eaton, J., Jensen, J. B., & Kortum, S. (2003). Plants and productivity in international trade. The American Economic Review, 93(4), 1268-1290. Besley, T. J., & Ghatak, M. (2003). Incentives, Choice, and Accountability in the Provision of Public Services. Oxford Review of Economic Policy, 19(2), 235–249. Retrieved from http://econpapers.repec.org/RePEc:oup:oxford:v:19:y:2003:i:2:p:235-249 Besley, T. J., & Ghatak, M. (2007). Retailing public goods: The economics of corporate social responsibility. Journal of Public Economics, 91(9), 1645–1663. Retrieved from http://econpapers.repec.org/RePEc:eee:pubeco:v:91:y:2007:i:9:p:1645-1663 Betcherman, G., Dar, A., & Olivas, K. (2004). Impacts of active labor market programs: New evidence from evaluations with particular attention to developing and transition countries. Social Protection, World Bank. Biewen, M., Fitzenberger, B., Osikominu, A., & Waller, M. (2007). Which Program for Whom? Evidence on the Comparative Effectiveness of Public Sponsored Training Programs in Germany. ZEW - Zentrum für Europäische Wirtschaftsforschung / Center for European Economic Research. Retrieved from http://econpapers.repec.org/RePEc:zbw:zewdip:5868 38 Biggs, T., Shah, M., & Srivastava, P. (1995). Training and productivity in African manufacturing enterprises. World Bank Policy Research Working Paper Series, 1–48. Retrieved from http://documents.worldbank.org/curated/en/1995/08/439656/training-productivity- african-manufacturing-enterprises Bishop, J. (1994). The Impact of Previous Training on Productivity and Wages. In Training and the Private Sector – International Comparisons (Ed.), Training and the Private Sector – International Comparisons. Chicago: University Chicago Press. Black, S. E., & Lynch, L. M. (2001). How to compete: the impact of workplace practices and information technology on productivity. Review of Economics and statistics, 83(3), 434- 445. Blakemore, A. E., & Hoffman, D. L. (1988). Seniority Rules and Productivity: An Empirical Test. Earnings in Britain. London: Institute for Fiscal Studies. Blalock, G., & Gertler, P. J. (2004). Learning from exporting revisited in a less developed setting. Journal of Development Economics, 75(2), 397-416. Blanchflower, D. G., & Lynch, L. M. (1994). Training at work: a comparison of US and British youths. In Training and the private sector: international comparisons (pp. 233-260). University of Chicago Press. Blank, R. M. (2000). When Can Public Policy Makers Rely on Private Markets? The Effective Provision of Social Services. Economic Journal, 110(462), C34–49. Retrieved from http://econpapers.repec.org/RePEc:ecj:econjl:v:110:y:2000:i:462:p:c34-49 Bloom, N., & Van Reenen, J. (2007). Measuring and explaining management practices across firms and countries. The Quarterly Journal of Economics, 122(4), 1351-1408. Bloom, N., Genakos, C., Sadun, R., & Van Reenen, J. (2012). Management practices across firms and countries. The Academy of Management Perspectives, 26(1), 12-33. 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., Draca, M., & Van Reenen, J. (2016). Trade induced technical change? The impact of Chinese imports on innovation, IT and productivity. The Review of Economic Studies, 83(1), 87-117. Bloom, N., Kretschmer, T., & Van Reenan, J. (2009). Work-life balance, management practices and productivity. In International differences in the business practices and productivity of firms (pp. 15-54). University of Chicago Press. Blundell, R., Dearden, L., & Meghir, C. (1996). The Determinants of Work-Related Training in Britain. London. Blundell, R., Dearden, L., Meghir, C., & Sianesi, B. (1999). Human Capital Investment: The Returns from Education and Training to the Individual, the Firm and the Economy. Fiscal Studies, 20(1), 1–23. Booth, A. L. (1991). Job‐related formal training: who receives it and what is it worth? Oxford bulletin of economics and statistics, 53(3), 281-294. 39 Booth, A. L., & Bryan, M. L. (2005). Testing some predictions of human capital theory: New training evidence from Britain. Review of Economics and statistics, 87(2), 391-394. Brambilla, I., Lederman, D., & Porto, G. (2012). Exports, export destinations, and skills. The American Economic Review, 102(7), 3406-3438. Brinkerhoff, J. M. (2012). Creating an enabling environment for diasporas’ participation in homeland development. International Migration, 50(1), 75-95. Brinkerhoff, J. M. (2006). Diasporas, skills transfer, and remittances: Evolving perceptions and potential. Converting migration drains into gains: Harnessing the resources of overseas professionals, 1-32. Bugamelli, M., & Pagano, P. (2004). Barriers to Investment in ICT. Applied Economics, 36(20), 2275-2286. Busso, M., Bassi, M., Urzúa, S., & Vargas, J. (2012). Disconnected: Skills, education, and employment in Latin America. Inter-American Development Bank. Bustos, P. (2011). Trade liberalization, exports, and technology upgrading: Evidence on the impact of MERCOSUR on Argentinian firms. The American economic review, 101(1), 304-340. Caliendo, L., Mion, G., Opromolla, L. D., & Rossi-Hansberg, E. (2015). Productivity and organization in Portuguese firms (No. w21811). National Bureau of Economic Research. Carmichael, L. H. (1983). Firm Specific Human Capital and Promotion Ladders. Bell Journal of Economics, 14. Retrieved from http://econpapers.repec.org/RePEc:qed:wpaper:452 Carneiro, P., Heckman, J., & Lee, S. (2004). Comparative advantage and schooling. Working Paper, University College London, London, UK. Chung, T.-P. (2000). The Returns to Education and Training: Evidence from the Malaysian Family Life Surveys. School of Economics, University of Kent. Retrieved from http://econpapers.repec.org/RePEc:ukc:ukcedp:0007 Clerides, S. K., Lach, S., & Tybout, J. R. (1998). Is learning by exporting important? Micro- dynamic evidence from Colombia, Mexico, and Morocco. Quarterly journal of Economics, 903-947. Conti, G. (2005). Training, productivity and wages in Italy. Labour Economics, 12(4), 557–576. Retrieved from http://econpapers.repec.org/RePEc:eee:labeco:v:12:y:2005:i:4:p:557-576 Cornes, R., & Sandler, T. (1984). Easy Riders, Joint Production, and Public Goods. Economic Journal, 94(375), 580–98. Retrieved from http://econpapers.repec.org/RePEc:ecj:econjl:v:94:y:1984:i:375:p:580-98 Cornes, R., & Sandler, T. (1994). The comparative static properties of the impure public good model. Journal of Public Economics, 54(3), 403–421. Retrieved from http://econpapers.repec.org/RePEc:eee:pubeco:v:54:y:1994:i:3:p:403-421 40 Dasgupta, S., Sarkis, J., & Talluri, S. (1999). Influence of information technology investment on firm productivity: a cross-sectional study. Logistics Information Management, 12(1/2), 120-129. De Loecker, J. (2007). Do exports generate higher productivity? Evidence from Slovenia. Journal of international economics, 73(1), 69-98. De Mel, S., McKenzie, D., & Woodruff, C. (2008). Returns to capital in microenterprises: evidence from a field experiment. The Quarterly Journal of Economics, 1329-1372. Dearden, L., Reed, H., & Van Reenen, J. (2006). The Impact of Training on Productivity and Wages: Evidence from British Panel Data. Oxford Bulletin of Economics and Statistics, 68(4), 397–421. http://doi.org/10.1111/j.1468-0084.2006.00170.x Dewatripont, M., Jewitt, I., & Tirole, J. (1999). The Economics of Career Concerns, Part I: Comparing Information Structures. Review of Economic Studies, 66, 183–198. Disney, R., Haskel, J., & Heden, Y. (2003). Restructuring and productivity growth in UK manufacturing. The Economic Journal, 113(489), 666-694. Doms, M., Dunne, T., & Troske, K. R. (1997). Workers, wages, and technology. The Quarterly Journal of Economics, 253-290. Dundar, H., Millot, B, Savchenko, Y., Aturupane, H., & Piyasari, T. (2014). Building the Skills for Economic Growth and Competitiveness in Sri Lanka. World Bank, Washington DC. Engel, E. M. R. A., Fischer, R. D., & Galetovic, A. (1997). Highway Franchising: Pitfalls and Opportunities. American Economic Review, 87(2), 68–72. Retrieved from http://econpapers.repec.org/RePEc:aea:aecrev:v:87:y:1997:i:2:p:68-72 Evertsson, M. (2004). Formal on-the-job training: A gender-typed experience and wage-related advantage? European Sociological Review, 79-94. Fares, J., & Puerto, O. S. (2009). Towards comprehensive training. World Bank. Fieler, A. C., Eslava, M., & Xu, D. (2014). Trade, Skills, and Quality Upgrading: A Theory with Evidence from Colombia (No. w19992). National Bureau of Economic Research. Fitzenberger, B., & Volter, R. (2007). Long-run effects of training programs for the unemployed in East Germany. Labour Economics, 14(4), 730–755. Retrieved from http://econpapers.repec.org/RePEc:eee:labeco:v:14:y:2007:i:4:p:730-755 Frazer, G. (2006). Learning the master’s trade: Apprenticeship and human capital in Ghana. Journal of Development Economics, 81(2), 259–298. Retrieved from http://econpapers.repec.org/RePEc:eee:deveco:v:81:y:2006:i:2:p:259-298 Frazis, H., & Loewenstein, M. A. (2005). Reexamining the Returns to Training: Functional Form, Magnitude, and Interpretation. Journal of Human Resources, 40(2). Retrieved from http://econpapers.repec.org/RePEc:uwp:jhriss:v:40:y:2005:i:2:p453-476 Frazis, H., Gittleman, M., & Joyce, M. (2000). Correlates of training: An analysis using both employer and employee characteristics. Industrial & Labor Relations Review, 53(3), 443- 462. 41 Friedman, M. (1970, September 13). The Social Responsability of Business is to Increase its Profits. The New York Times Magazine, pp. 222–126. Frye, M. B., Nelling, E., & Webb, E. (2006). Executive Compensation in Socially Responsible Firms. Corporate Governance: An International Review, 14(5), 446–455. Retrieved from http://econpapers.repec.org/RePEc:bla:corgov:v:14:y:2006:i:5:p:446-455 Gittleman, M., Horrigan, M., & Joyce, M. (1998). “Flexible” workplace practices: evidence from a nationally representative survey. Industrial & Labor Relations Review, 52(1), 99-115. Giunta, A., & Trivieri, F. (2007). Understanding the determinants of information technology adoption: evidence from Italian manufacturing firms. Applied Economics, 39(10), 1325- 1334. Goldberg, P. K., Khandelwal, A. K., Pavcnik, N., & Topalova, P. (2010). Multiproduct firms and product turnover in the developing world: Evidence from India. The Review of Economics and Statistics, 92(4), 1042-1049. Gonzalezs-Velosa, C., Rosas, D., & Flores, R. (2016). On-the-Job Training in Latin America and the Caribbean: Recent Evidence. IDB Publication. Görg, H., & Strobl, E. (2005). Spillovers from foreign firms through worker mobility: An empirical investigation. The Scandinavian journal of economics, 107(4), 693-709. Greenhalgh, C., & Stewart, M. (1987). The effects and determinants of training. Oxford Bulletin of Economics and Statistics, 49(2), 171-190. Hall, R. E., & Lazear, E. (1984). The Excess Sensitivity of Layoffs and Quits to Demand. Journal of Labor Economis, 2(2), 233–257. Retrieved from http://econpapers.repec.org/RePEc:nbr:nberwo:0864 Haller, S. A., & Siedschlag, I. (2011). Determinants of ICT adoption: Evidence from firm-level data. Applied Economics, 43(26), 3775-3788. Hansmann, H. (1996). The Changing Roles of Public, Private, and Nonprofit Enterprise in Education, Health Care, and Other Human Services. National Bureau of Economic Research, Inc. Retrieved from http://econpapers.repec.org/RePEc:nbr:nberch:6565 Hansson, B. (2009). Job-related training and benefits for individuals: A review of evidence and explanations. Hanushek, E. A., & Woessmann, L. (2008). The role of cognitive skills in economic development. Journal of economic literature, 46(3), 607-668. Hashimoto, M. (1981). Firm-Specific Human Capital as a Shared Investment. American Economic Review, 71(3), 475–82. Retrieved from http://econpapers.repec.org/RePEc:aea:aecrev:v:71:y:1981:i:3:p:475-82 Heckman, E. R. (2015). Financial Literacy in the Workplace. In J. Liebowitz (Ed.), Financial Literacy Education: Addressing Student, Business, and Government Needs (pp. 179- 196). Auerbach Publications. 42 Heckman, J. J., LaLonde, R. J., & Smith, J. A. (1999). The economics and econometrics of active labor market programs (Vol. 3, Part A). Elsevier. Retrieved from http://econpapers.repec.org/RePEc:eee:labchp:3-31 Heckman, J. J., Lochner, L., & Taber, C. (1998). Explaining rising wage inequality: Explorations with a dynamic general equilibrium model of labor earnings with heterogeneous agents. Review of economic dynamics, 1(1), 1-58. Heckman, J. J., Stixrud, J., & Urzua, S. (2006). The effects of cognitive and noncognitive abilities on labor market outcomes and social behavior (No. w12006). National Bureau of Economic Research. Hollenstein, H. (2004). Determinants of the adoption of Information and Communication Technologies (ICT): An empirical analysis based on firm-level data for the Swiss business sector. Structural change and economic dynamics, 15(3), 315-342. Holmstrom, B., & Milgrom, P. (1991). Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design. Journal of Law, Economics and Organization, 7, 24– 52. Retrieved from http://econpapers.repec.org/RePEc:oup:jleorg:v:7:y:1991:i:0:p:24-52 Jaffe, Adam, B., Peterson, S. R., Portney, P. R., & Stavins, R. N. (1995). Environmental Regulation and the Competitiveness of U.S. Manufacturing: What Does the Evidence Tell US? Journal of Economic Literature, 33(1), 132–163. Jara, O. (2002). Evaluacion del Sistema de capacitación en Chile: Inversion publica y los beneficios para las personas. (O. Jara, Ed.). Santiago, Chile: Editorial Universitaria. - Book Jensen, M. C. (2002). Value Maximization, Stakeholder Theory, and the Corporate Objective Function. Business Ethics Quarterly, 12(2), 235–256. Johanson, R., & Wanga, G. (2008). Skills Training for the Informal Sector in Tanzania. World Bank. Johnson, J. P. (2002). Open Source Software: Private Provision of a Public Good. Journal of Economics & Management Strategy, 11(4), 637–662. Retrieved from http://econpapers.repec.org/RePEc:bla:jemstr:v:11:y:2002:i:4:p:637-662 Johnston, R. J., & Roheim, C. A. (2006). A Battle of Taste and Environmental Convictions for Ecolabeled Seafood: A Contingent Ranking Experiment. Journal of Agricultural and Resource Economics, 31(02). Retrieved from http://econpapers.repec.org/RePEc:ags:jlaare:8617 Kahyarara, G., & Teal, F. J. (2008). The Returns to Vocational Training and Academic Education: Evidence from Tanzania. World Development, 36(11), 2223–2242. Retrieved from http://econpapers.repec.org/RePEc:eee:wdevel:v:36:y:2008:i:11:p:2223-2242 Kaiser, U., Kongsted, H. C., & Rønde, T. (2008). Labor mobility and patenting activity. University of Copenhagen, Department of Economics, Centre for Applied Microeconometrics (CAM) Working Paper, 7. Katz, E., & Ziderman, A. (1990). Investment in General Training: The Role of Information and Labour Mobility. Economic Journal, 100(403), 1147–58. Retrieved from http://econpapers.repec.org/RePEc:ecj:econjl:v:100:y:1990:i:403:p:1147-58 43 Keller, W. (2004). International technology diffusion. Journal of economic literature, 42(3), 752- 782. Khandelwal, A. (2010). The long and short (of) quality ladders. The Review of Economic Studies, 77(4), 1450-1476. Kim, J., & Marschke, G. (2005). Labor mobility of scientists, technological diffusion, and the firm's patenting decision. RAND Journal of Economics, 298-317. Kitzmueller, M., & Shimshack, J. P. (2012). Economic Perspectives on Corporate Social Responsibility. Journal of Economic Literature, 50(1), 51–84. Retrieved from http://econpapers.repec.org/RePEc:aea:jeclit:v:50:y:2012:i:1:p:51-84 Konings, J., & Vanormelingen, S. (2015). The Impact of Training on Productivity and Wages: Firm-Level Evidence. Review of Economics and Statistics, 97(2), 485–497. http://doi.org/10.1162/REST_a_00460 Kotchen, M. J. (2006). Green Markets and Private Provision of Public Goods. Journal of Political Economy, 114(4), 816–845. Retrieved from http://econpapers.repec.org/RePEc:ucp:jpolec:v:114:y:2006:i:4:p:816-845 Kowtha, N. R., & Choon, T. W. I. (2001). Determinants of website development: a study of electronic commerce in Singapore. Information & management, 39(3), 227-242. Kretschmer, T., Miravete, E. J., & Pernías, J. C. (2012). Competitive pressure and the adoption of complementary innovations. The American Economic Review, 102(4), 1540-1570. Leete, L. (2001). Whither the Nonprofit Wage Differential? Estimates from the 1990 Census. Journal of Labor Economics, 19(1), 136–70. Retrieved from http://econpapers.repec.org/RePEc:ucp:jlabec:v:19:y:2001:i:1:p:136-70 Leuven, E., & Oosterbeek, H. (2004). Evaluating the Effect of Tax Deductions on Training. Journal of Labor Economics, 22(2), 461–488. Retrieved from http://econpapers.repec.org/RePEc:ucp:jlabec:v:22:y:2004:i:2:p:461-488 Leuven, E., & Oosterbeek, H. (2008). An alternative approach to estimate the wage returns to private-sector training. Journal of Applied Econometrics, 23(4), 423–434. Retrieved from http://econpapers.repec.org/RePEc:jae:japmet:v:23:y:2008:i:4:p:423-434 Leuven, E., Oosterbeek, H., & Van Ophem, H. (2004). Explaining international differences in male skill wage differentials by differences in demand and supply of skill. The Economic Journal, 114(495), 466-486. Lillard, L. A., & Tan, H. W. (1992). Private sector training: who gets it and what are its effects?. Research in labor economics, 13, 1-62. Lillard, L. A., & Tan, H. W. (1992). Private Sector Training: who gets it and what are its effects? Research in Labor Economics, 13. Loewenstein, M. A., & Spletzer, J. R. (1999). General and specific training: Evidence and implications. Journal of Human Resources, 710-733. Lopez-Acevedo, G, Medvedev, D, Palmade, V. (2016). South Asia’s Turn: Policies to Boost Competitiveness and Create the Next Export Powerhouse. World Bank, Washington DC. https://openknowledge.worldbank.org/handle/10986/25094 44 Lopez-Acevedo, G., & Tan, H. W. (2011). Impact Evaluation of Small and Medium Enterprise Programs in Latin America and the Caribbean. Retrieved from https://openknowledge.worldbank.org/handle/10986/2298 Loureiro, M. L. (2003). Rethinking new wines: implications of local and environmentally friendly labels. Food Policy, 28(5-6), 547–560. Retrieved from http://econpapers.repec.org/RePEc:eee:jfpoli:v:28:y:2003:i:5-6:p:547-560 Loureiro, M. L., & Lotade, J. (2005). Do fair trade and eco-labels in coffee wake up the consumer conscience? Ecological Economics, 53(1), 129–138. Retrieved from http://econpapers.repec.org/RePEc:eee:ecolec:v:53:y:2005:i:1:p:129-138 Lucchetti, R., & Sterlacchini, A. (2004). The adoption of ICT among SMEs: evidence from an Italian survey. Small Business Economics, 23(2), 151-168. Lynch, L. M. (1991). The role of off-the-job vs. on-the-job training for the mobility of women workers. The American Economic Review, 81(2), 151-156. Lynch, L. M. (1992). Private Sector Training and its Impact on the Earnings of Young Workers. American Economic Review, 82-1. Retrieved from http://www.nber.org/papers/w2872 Macchiavello, R., Rabbani, A., & Woodruff, C. (2015). The Market for Training Services: A Demand Experiment with Bangladeshi Garment Factories. American Economic Review, 105(5), 300–304. Retrieved from http://ideas.repec.org/a/aea/aecrev/v105y2015i5p300- 304.html Mankiw, N. G., Romer, D., & Weil, D. N. (1990). A contribution to the empirics of economic growth (No. w3541). National Bureau of Economic Research. Margolis, J., Elfenbein, H. A., & Walsh, J. P. (2007). Does it pay to be good? A meta-analysis and redirection of research on the relationship between corporate social and financial performance. Martin, J. P., & Grubb, D. B. (2001). What works and for whom: a review of OECD countries’ experiences with active labour market policies. IFAU - Institute for Evaluation of Labour Market and Education Policy. Retrieved from http://econpapers.repec.org/RePEc:hhs:ifauwp:2001_014 Mayer, T., Melitz, M. J., & Ottaviano, G. I. (2014). Market size, competition, and the product mix of exporters. The American Economic Review, 104(2), 495-536. Melero, E. (2004). Evidence on training and career paths: Human capital, information and incentives. Melitz, M. J. (2003). The impact of trade on intra‐industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695-1725. Mincer, J. A. (1989). Job Training: Costs, Returns, and Wage Profiles. National Bureau of Economic Research, Inc. Retrieved from http://econpapers.repec.org/RePEc:nbr:nberwo:3208 Monk, C., Sandefur, J., & Teal, F. J. (2008). Does Doing an Apprenticeship Pay Off? Evidence from Ghana. Centre for the Study of African Economies, University of Oxford. Retrieved from http://econpapers.repec.org/RePEc:csa:wpaper:2008-08 45 Montenegro, C. E., & Patrinos, H. A. (2014). Comparable estimates of returns to schooling around the world. World Bank Policy Research Working Paper Series. Retrieved from http://ideas.repec.org/p/wbk/wbrwps/7020.html More and Better Jobs in South Asia. (2011). World Bank Default Book Series. Retrieved from http://elibrary.worldbank.org/doi/abs/10.1596/978-0-8213-8912-6 Morgan, J. (2000). Financing Public Goods by Means of Lotteries. Review of Economic Studies, 67(4), 761–784. Retrieved from http://econpapers.repec.org/RePEc:oup:restud:v:67:y:2000:i:4:p:761-784. Narula, R., & Zanfei, A. (2005). Globalisation of innovation (pp. 318-345). Oxford: Oxford University Press. OECD. (2004). OECD Employment Outlook 2004. OECD Employment Outlook. OECD Publishing. http://doi.org/10.1787/empl_outlook-2004-en Oosterbeek, H., & Leuven, E. (2001). Firm-Specific Human Capital as a Shared Investment: Comment. American Economic Review, 91(1), 342–347. Retrieved from http://econpapers.repec.org/RePEc:aea:aecrev:v:91:y:2001:i:1:p:342-347 Pavcnik, N. (2002). Trade liberalization, exit, and productivity improvements: Evidence from Chilean plants. The Review of Economic Studies, 69(1), 245-276. Pasurka, C. A. (2008). Perspectives on Pollution Abatement and Competitiveness: Theory, Data, and Analyses. Review of Environmental Economics and Policy, 2(2), 194–218. Retrieved from http://econpapers.repec.org/RePEc:oup:renvpo:v:2:y:2008:i:2:p:194-218 Pierre, G., & Scarpetta, S. (2004). Employment Regulations through the Eyes of Employers: Do they matter and how do firms respond to them? Pigou, A. C. (1912). Wealth and welfare. Macmillan and Company, limited. Poole, J. P. (2013). Knowledge transfers from multinational to domestic firms: Evidence from worker mobility. Review of Economics and Statistics, 95(2), 393-406. Porter, M. E., & van der Linde, C. (1995). Toward a New Conception of the Environment- Competitiveness Relationship. Journal of Economic Perspectives, 9(4), 97–118. Retrieved from http://econpapers.repec.org/RePEc:aea:jecper:v:9:y:1995:i:4:p:97-118 Ridoutt, L., Dutneall, R., Hummel, K., & Selby Smith, C. (2002). Factors influencing the implementation of training and learning in the workplace. Adelaide: National Centre for Vocational Education Research. Rosholm, M., Nielsen, H. S., & Dabalen, A. (2007). Evaluation of training in African enterprises. Journal of Development Economics, 84(1), 310–329. Retrieved from http://econpapers.repec.org/RePEc:eee:deveco:v:84:y:2007:i:1:p:310-329 Samuelson, P. A. (1954). The Pure Theory of Public Expenditure. Review of Economics and Statistics, 36. Saraf, P. (2014). Developing Practical Skills for Firm Productivity: The How To. Competitive Industries Note, World Bank, Washington DC. 46 Saraf, P. (2016). Automotive in South Asia : From Fringe to Global. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/25119 Saraf, P., Mariscal, A. & Velazquez, L.G. (2016). Internal draft – literature review of skills, productivity and wages. World Bank mimeo (unpublished). Schlesinger, M. (1998). Mismeasuring the consequences of Ownership: external influences and the comparative performance of public, for-profit, and private non-profit organizations. In Powell and E. Clemens (Ed.), Private Action and the Public Good (pp. 85–113). New Haven: Yale University Press Stevens, M. (1994). A Theoretical Model of On-the-Job Training with Imperfect Competition. Oxford Economic Papers, 46(4), 537–62. Retrieved from http://econpapers.repec.org/RePEc:oup:oxecpp:v:46:y:1994:i:4:p:537-62 Stoyanov, A., & Zubanov, N. (2012). Productivity spillovers across firms through worker mobility. American Economic Journal: Applied Economics, 4(2), 168-198. Tan, H., & Lopez-Acevedo, G. (2005). Evaluating Training Programs for Small and Medium Enterprises: Lessons from Mexico. World Bank Policy Research Working Paper, (3760). The World Bank. (2016). World Development Report 2016: Digital Dividends. Retrieved from http://www.worldbank.org/en/publication/wdr2016 Tirole, J., & Benabou, R. J. (2006). Incentives and Prosocial Behavior. American Economic Review, 96(5), 1652–1678. Retrieved from http://econpapers.repec.org/RePEc:aea:aecrev:v:96:y:2006:i:5:p:1652-1678 Veum, J. R. (1995). Training wages and the human capital model. [Unpublished] 1995. Presented at the Annual Meeting of the Population Association of America San Francisco California April 6-8 1995.. Wagner, J. (2007). Exports and productivity: A survey of the evidence from firm‐level data. The World Economy, 30(1), 60-82. Weisbrod, B. (1998). The nonprofit Economy. Cambridge University Press. Winkelmann, R. (1994). Training, earnings and mobility in Germany (No. 982). CEPR Discussion Papers. 47 Annexes Annex 1: Annotated Bibliography (key literature on various aspects related to OJT) Effects Country Other Training on Controls Name of Dependent and Time Training Independent productivit Endogeneity Study Variable Period variables y and Training wages Literature looking at impacts of OJT on productivity and wages separately; also Internal Rate of Return Sector dummies. Proportion of 1% male increase in employees, the proportion in proportion YES, Control occupation, of workers for qualification, trained in unobserved tenure, an industry heterogenei Mean proportion in is ty and the change in (Dearden, small firm, associated potential United ln(Sales per L., Reed, H., average log with an endogeneity Kindom// worker) Share of trained & Van capital-labor increase in of training Panel data Mean workers Reenen, J. ratio, avg log value using a 1983-1996 change in (2006)) real value added per variety of ln(Average added per worker of methods Wage) worker, avg log about 0.6% including gross output per and an GMM worker, avg log increase in system hourly wages, wages of estimation. avg hours about worked, avg 0.3%. R&D spend as a proportion of output Effective YES, Control labor input function increases approach to by 1.7% to estimate 3.2% in production (Konings, J., Capital, type of response to UK// Firm functions. & Share of trained worker, an increase level panel Wages, Wage Vanormelin workers at the proportion of of 10 dataset productivity equations at gen, S. firm workers/trained percentage 1997-2006 the firm (2015)) workers. points in level to infer the fraction productivity of workers and wage who premiums of receive training training, 48 while the average wage increases by only 1% to 1.7% Investment, Statistically Change in significant employment, Trainees/Employ positive personnel ees, Training effects on Policies, Days/Employees productivit Enterprise Corporate , Training y are found (Barret, A., surveys, Proportionat Innovation, Expenditure/Pay both for all O’Connell, Ireland// e change in Corporate NO roll, General training J., (2001)) 1993- productivity Restructuring, Training Days/ and for 1996/7 Labor Cost/ Employees, general Employees in Specific Training training, 1993, Number Days/ Employees but not for of employees in specific 1993. Sector of training. activity. Yes. First Portugal Firms location, difference On average (Ministry of Sector of and it is -7% for Employmen activity, value instrumenta Internal rate firms not t)// panel of Training, added, number l variables (Almeida, of return (r). providing 1500 firms Training hours of workers, (following R., Carneiro, training (with more above the capital, average the P. (2009)) and 24% than 100 median (1,489 h) workforce age, approach by for those employees) yearly number Blundell and providing (1995- of hires and Bond training. 1999) fires, (2000)). Tenure, experience at start of job, AFQT,10 years of education, 60% married, part- YES: adjust US// NLSY Gap opportunit time, union, for data set. between log y cost of dummies for heterogenei (Frazis and Harley wage after Hours of training training, initial ty in wage Loewestein, Frazis, two years on the current 40% direct occupation in growth. (2005)) Bureau of and log job cost of the job, Black, Evidence of Labor starting training. Hispanic, heterogenei Statistics wage IRR equal female, enrolled ty in returns to 50%. in school, missing value indicators for AFQT, union, and part-time 49 Sample of 852 small Profit Variables businesses improveme covering in the Cape Treatment nts in the (Anderson, business owner Town area groups receive range of YES: S., Chandy, background, of South Firm profits marketing or 41-61 randomized R., and Zia, business owner Africa finance business percent control trial B., (2016)) exposure, through a training over the business randomized control establishment controlled group trial Literature looking at impact of OJT on worker wages (or using wages as proxy for productivity) Educational attainment, gender, age, tenure, experience, marital status, occupation, union participation, Matched computer, bank employer- Dummy: having Yes, account, 7.7% for employee received formal Matching (Almeida R. internet Malaysia data for Log hourly on-the-job Estimators K. & de transaction, and 4.5% Malaysia wage training since Method Faria, 2014) training at a for (2002) and having joined (Local Linear previous Thailand Thailand the firm Matching) employer, size, (2004) foreign capital, exports, education of the work force, education of the manager, new production technologies, industry, region Organizatio nal changes Industry (80% Vs Value added dummy 50% for Mexico// per worker variables to control Panel data Log (value control for Tan and firms). from CIMO added) Wages, set of industry Lopez- Participation in Adoption of program in Log industry dummy effects. Acevedo CIMO quality Mexico (production variables, Indicator (2005) control (1991- value – variable to systems 1993) materials control for (23% cost) ec. more). downtown Increased investment 50 in worker training Employee Positive development and YES: Control training, vector significant for selection US// Z: years of effect bias in Personnel Wages, job Training education, years on both (Bartel, training records of a performanc received in 12 of company wage (1995)) programs. 2 large e months service, source growth and step company of the change estimation hire and type of in job method. occupation. performanc Relative salary. e scores Yes, Excluded instruments Age, marital (selection Worker status, model): Level Data nationality, having a Dummy: having 20%-30% (Chung, for Log hourly schooling, bank ever participated (for 2000) Malaysia wage dummies for account, in training women) (1976, employers and level of 1988) unpaid family education in workers 1976, and parent occupationa l status Yes, selection Not equation statistically with different Worker instruments: Dummy: having Gender, from zero Level Data father's (Frazer, Log hourly participated in potential for the for Ghana education 2006) wage an experience, whole (1991- and apprenticeship schooling sample but 1999) occupation, 17% for as well as self- total employed household assets. Different Wage Returns to datasets returns, training are from internal higher if it Partially. different IEs employabilit Worker is Some Training, Hansson, B. (Switzerlan y and job- characteristics employer- studies do Employer- (2009) d, EU, UK, security. (age, education, financed. control for financed training USA, External: gender), Returns to endogeneity Netherland labour training do . s, Italy, participation not vary France, rates, with age, 51 unemploym education ent. or gender, but tend to be unequally distributed among firm. Employer Positive Intercept, Opportunit and education, y Pilot statistically gender, union, Partially: set Loewenstei Project significant Log (wage) temporary or of individual n and (EOPP) On-the-job impact of after two seasonal and job Spletzer survey training training on years employment, characteristi (1999) and the wages. firm size, cs. National General industry, and Longitudina nature of oc l Su OJT. Age, ethnicity, experience, gender, occupation, schooling, tenure, union participation Matched and familiar 2.3% for Yes, employer- Dummy: having relations within Kenya and Matching employee participated in the owners of not (Rosholm et Log monthly Estimators data for training in the the firm, statistically al., 2007)) wage Method Kenya and previous 12 ownership, different (Local Linear Zambia months industry, from zero Matching) (1995) location, size, for Zambia financial situation, skill demand, turnover, unionization, training annual expenses Cognitive skills have powerful Gender, effects on Alternative Data on Years of potential individual estimation Log annual Hanushek & school schooling, experience, earnings, methods, earnings, Woessmann attainment quality of living in rural on the and economic , (2008) across schooling. areas, test distribution samples. growth countries Cognitive skills score, of income, control openness. and on variables. economic growth. The current 52 situation in developing countries is much worse than generally pictured on the basis just of school enrollment and attainment as it doesn’t factor in schooling quality. Literature looking at gains in future employer using various methods including MNC spillovers Worker’s identification code,4 annual real wages in reais,5 job tenure in months, type of job separation, age, educational Positive attainment, Matched multination Log Multinational occupational establishme al wage Control for (Poole, J. P. individual spillovers though classification, nt-worker spillovers omitted (2013)) annual skilled worker the tax number database through variables. wages mobility. of the worker’s from Brazil worker establishment, mobility and the industrial classification of the worker’s establishment. Dummy worker works at a foreign-owned establishment. Firms Ghana// Years of which are Gorg and Fim-level Training/ schooling, set of Partially: run by Strobl data from TFP experience in sector and time GLS-AR(1) owners (2005) manufactur foreign firm dummy estimator that ing firms variables worked for 53 multination als, same industry prior to opening up their own firm have higher productivit y growth than other domestic firms. Real wage, Tenure, Experience, Age, Years of YES, schooling, Plant Controlling MNE size, Labor for Norway// experience productivity, differences Annual increases (Balsvik, R. Worker’s Skill share, in Norwegian MNE experience productivit (2011)) productivity Female share, unobservabl statistics y by 20%. Worker/plant e worker 1990-2000 Positive observations. and plant correlation. Year and characteristi industry cs dummies: Plant, worker characteristics. Total output gain from SPs is Age, gender, 0.11% per education, skill year. At Employer- group, least two- employee experience, thirds (Stoyanov data from Output and worker coming – is Yes, use of and Danish Workers from worker from productive retained by instrumenta Zubanov, manufactur growing firms wage firm, human the hiring l variables (2012)) ing for the capital firms, period estimated, whereas 1995-2007 number of job the SPs transitions. themselves receive a trivial 6% at most EU// Panel Set of dummy Both YES: control Training data from Forms of variables: current and for (Booth and incidence, 1 if the output and gender, age future firm unobserved Bryan the individual European worker group, type of benefit, heterogenei (2005)) started training, Community wages contract, part- with the ty; Control 0 otherwise Household time, full time, future firm for industry, 54 Panel, education level, benefiting occupation, 1994-1999 private/ public more than firm-size. sector worker the current Uses firm random effects probit model. Literature connecting OJT to firm management and technology adoption YES, RCT. Instrumenta l variables: (Bloom, N., Productivity instrument Eifert, B., Male family TFP: 11%; - Quality (log the Mahajan, India// members, post Profits: QDI), managemen A., Experiment Management, treatment, plant 330.000 inventory t practice McKenzie, al data, RCT Intervention. manager annually. (log tons), score with D., & large firms related, plant Statistically output (log log(1+weeks Roberts, J. manager tenure significant. picks). since the (2013)) implementa tion phase began) Lack of adoption is a misalignme nt of incentives within firms: Fearing Output/month, reductions employment, (Atkin, D., in their employment Chaudhry, effective (cutters) A., Chaudry, wage, Pakistan/ Rs/Ball, size, S., Technology A new cutting employees Jan-Apr promotional, Yes, RCT Khandelwal, adoption technique resist 2012, (RCT) age of firm, CEO A. K., & adoption experience, Verhoogen, and cutter E. (2015)) misinform experience, owners. cutter tenure. Introducing incentives positive effect on adoption. The results are supportive of the 55 hypothesis that misalignme nt of incentives within firms is an important barrier to technology adoption. Yes, Portugal// One structural Firm-level manageme productivity balance nt layer, estimation. sheet data Value added, quantity Instrumenta set, and a hours, and wage (Caliendo, based l variables firm- Quantity, L., Mion, G., productivit for product- revenue, Opromolla, Log revenue Number of y increases endogenous level data demand shocks L. D., & labor management by about variables, set and markups at Rossi- productivity layers 4%, while methodolog containing time t-1; capital Hansberg, E. revenue- y of information at time t; and (2015)) based Wooldridge on the productivity at productivit (2009) to production time t – 2. y drops by deal with of more than the manufactur 4%. endogeneity ed goods. in input use. Literature on evolution of curriculum in ALMPs Net impact evaluations greater in Exogenous OECD variables: countries training type, followed by specific ECA and characteristics 90 LAC. In- Probability of the program, Partially. countries// classroom (Fares, J. , & of positive economic/workf Some set of 345 Type of training and Puerto, O. labor orce and studies do studies of program workplace (2009)) market institutional control for training training effect country context, endogeneity programs combined characteristics with other of the labor services supply. Dummy increases variables for probability control groups. of positive labor market 56 outcomes, compared to in- classroom training up to 20%. 57 Annex 2: Survey instrument to measure quality of management in a firm Source: Bloom, N., and J. Van Reenen. (2007) “Measuring and Explaining Management Practices Across Firms and Countries,” Quarterly Journal of Economics, 122(4), 1341-1408. The Management Practice Dimensions - Categories Score from 1–5 Categories Score from 1-5 based on: 1) Introduction of modern What aspects of manufacturing have been formally introduced, including manufacturing techniques just-in-time delivery from suppliers, autonomation, flexible manpower, support systems, attitudes, and behavior? 2) Rationale for Were modern manufacturing techniques adopted just because others were introduction of using them, or are they linked to meeting business objectives like modern manufacturing reducing costs and improving quality? techniques 3) Process problem Are process improvements made only when problems arise, or are they documentation actively sought out for continuous improvement as part of a normal business process? 4) Performance tracking Is tracking ad hoc and incomplete, or is performance continually tracked and communicated to all staff? 5) Performance review Is performance reviewed infrequently and only on a success/failure scale, or is performance reviewed continually with an expectation of continuous improvement? 6) Performance dialogue In review/performance conversations, to what extent is the purpose, data, agenda, and follow-up steps (like coaching) clear to all parties? 7) Consequence To what extent does failure to achieve agreed objectives carry management consequences, which can include retraining or reassignment to other jobs? 8) Target balance Are the goals exclusively financial, or is there a balance of financial and nonfinancial targets? 9) Target interconnection Are goals based on accounting value, or are they based on shareholder value in a way that works through business units and ultimately is connected to individual performance expectations? 10) Target time horizon Does top management focus mainly on the short term, or does it visualize short-term targets as a “staircase” toward the main focus on long-term goals? 11) Targets Are goals too easy to achieve, especially for some “sacred cows” areas of the firm, or are goals demanding but attainable for all parts of the firm? 12) Performance clarity Are performance measures ill-defined, poorly understood, and private, or are they well-defined, clearly communicated, and made public? 13) Managing human To what extent are senior managers evaluated and held accountable for capital attracting, retaining, and developing talent throughout the organization? 14) Rewarding high To what extent are people in the firm rewarded equally irrespective of performance performance level, or are rewards related to performance and effort? 58 15) Removing poor Are poor performers rarely removed, or are they retrained and/or moved performers into different roles or out of the company as soon as the weakness is identified? 16) Promoting high Are people promoted mainly on the basis of tenure, or does the firm performers actively identify, develop, and promote its top performers? 17) Attracting human Do competitors offer stronger reasons for talented people to join their capital companies, or does a fi rm provide a wide range of reasons to encourage talented people to join? 18) Retaining human Does the firm do relatively little to retain top talent or do whatever it capital takes to retain top talent when they look likely to leave? 59