World Bank Employment Policy Primer April 2004 I No. 6 TRAINING LEVIES 33395 Evidence from Evaluations * This note examines one of the most commonly Background and Introduction used techniques for generating resources from There is often broad political support in develop- employers ­ payroll levies. In many countries payroll ing countries for public provision and financing of taxes on enterprises (with reimbursement for training worker training and most countries have policies that, undertaken) have become a significant source of either directly or indirectly, support the training of financing for skills training, both in specialized train- workers. One strategy is to create the right environ- ing institutions and in enterprises. ment for firms to train their workers, rather than leg- One of the key principles underlying these islating compulsory spending on training. Govern- schemes is to try to address the `free rider' problem: ments may promote training by, for example, when training is transferable, the benefits from train- improving access to relevant information, highlight- ing can be reaped by other firms through the poach- ing the skills acquired through training, and encour- ing of workers. Hence firms may decide to underin- aging flexibility so that employers and employees can vest in training. Given that all firms contribute to make use of the acquired skills. financing training (at least theoretically), this free An important element is the financing of training. rider problem becomes less of an issue. An appropriate financing policy ensures both the sta- Another rationale for payroll levis is that as the bility of funding and the level of financing. The claim firms are contributing financially, such interventions on public resources for vocational education and training is much weaker than that for lower levels of encourage them to put greater emphasis on upgrad- education in most countries. Furthermore, in many ing the skills of their workforce and making them developing countries, government budgets constitute more competitive. a vulnerable and unreliable source of financing for Payroll taxes are also attractive to governments training. Thus an important element in the financing because they provide a sheltered source of revenue for of worker training is to increase the contribution of training and a means of mobilizing funds that may beneficiaries ­ both employers and trainees ­ while otherwise be unavailable to the public sector. ensuring that adequate incentives are in place to make training relevant to the needs of the labor market. *This note was prepared by Amit Dar and edited by Tim Whitehead. The World Bank Employment Policy Primer aims to provide a comprehensive, up-to-date resource on labor market policy issues. The series includes two products: short notes, such as this one, with concise summaries of best practice on various topics and longer papers with new research results or assessments of the literature and recent experience. Primer papers and notes are available on the labor markets website at or by contacting the Social Protection Advisory Service at (202) 458-5267 or by email at . I m p a c t E v a l u a t i o n Types of Payroll Levies · Payroll tax exemption: Firms can reduce or elimi- While there are mixed models, two major types of nate their levy obligations by the amount of payroll taxes for financing training exist in developing approved training provided or purchased. Cote `d countries. In the traditional (revenue generating) Ivoire and France are examples of countries scheme, firms are taxed to generate revenues to which implement these schemes. finance training provided by the public sector. In the Administration of these schemes differs by coun- alternative, the levy-grant or levy-rebate, scheme, in- try and type of scheme. In revenue-raising schemes plant training is encouraged by providing firms with the money is usually collected and administered by training incentives. These schemes are described the government body or agency directly responsible below. for training. In reimbursement schemes, governments Revenue-Raising Schemes. In these schemes, rev- and/or employers and workers organizations admin- enues fund national training systems that provide a ister reimbursement mechanisms and levy-grant wide range of pre-employment and in-service train- schemes from national or sectoral funds. These funds ing courses. Training programs in Brazil and Moroc- collect levies and distribute training grants based on co are examples. The emphasis is on public-sector defined criteria. Payroll-tax-exemption schemes usu- training provision rather than the encouragement of ally operate through employers' individual actions firms to undertake training. The coverage of this levy and are supervised by either the Ministry of Labor or also varies, both in terms of sectors and firm size. This the national general revenue agency. levy is restricted in many countries to the industrial sector of the economy, reflecting the limitations of the programs offered by public training establishments. Advantages and Drawbacks of Levy In other countries, the levy also covers the commercial Schemes sector and, in some cases, is all-encompassing. Brazil, Taxes for financing training fall upon the enter- Chile and Morocco are examples of countries that prise and it is believed that the same enterprises bear have implemented such economy-wide schemes. the burden of the levy, thus making it appear to be Levy-Disbursement Schemes. In other countries, the 'fair'. But on whom does the real burden fall ­ do payroll tax is linked to a disbursement scheme, with employers bear the costs or are they passed onto firms receiving grants in proportion to the level of workers in the form of lower wages? training that their employees undergo. The advantage If the firm-sponsored training program is exactly is that payroll-tax proceeds are used to encourage what every worker would have chosen for himself or firms either to set up programs of in-service training herself, then the cost will be borne entirely by the or to purchase training for their workers at a recog- worker (who would willingly accept a wage net of all nized training center. In this way, the firms qualify for payroll tax contributions). If ­ as is more likely ­ the a rebate of up to a specified percent of the tax paid. scheme is not identical to what the worker would These schemes take various forms: choose, the tax burden will, in general be shared by the worker and the employer. Thus, the higher the · Training cost reimbursement/levy-grant schemes: elasticity of demand for labor, the larger the share of Firms are reimbursed based on the cost incurred the tax borne by the worker; and the higher the elas- for certain designated forms of training. Alterna- ticity of supply of labor, the higher the share of the tax tively, grants are paid to firms conditional on meeting criteria once a systematic training borne by the employer. Who actually pays (i.e., bears approach is adopted. Examples of countries with the incidence) therefore depends not on whom the these schemes include Hungary, Kenya, Malaysia, tax is levied but the design of the scheme and the rel- Singapore, and South Africa. 2 I m p a c t E v a l u a t i o n TABLE 1: IN-SERVICE TRAINING INITIATIVES LEVY-GRANT/ REVENUE PAYROLL TAX EXEMP- COUNTRY RATE (%) TRAINING COST REIM- GENERATING TION BURSE. Bahrain 1.0-3.0 Brazil 1.5 Cote d'Ivoire 0.4-1.6 France 1.5 Hungary 1.5 Kenya 1.0 Korea 0.5 Malaysia 1.0 Morocco 1.6 Nigeria 1.25 Singapore 1.0 South Africa 0.5-2.0 Tanzania 2.0 Turkey n.a. U.K. 1.0-2.5 ative market power of the worker vis-à-vis the schemes may assist in developing national training employer. capacities and in financing the training needs of dis- An important issue is how closely training bene- advantaged groups and small employers who may fits match the individual tax payments made. For wish to avail of low-cost public training. However, example, charging a uniform tax across the board such schemes are also fraught with problems. It is dif- (which is the prevalent method in most countries), ficult to sustain employer interest and participation in when costs and type of training vary significantly by these schemes ­ many employers regard their contri- sector, may lead to cross-subsidization of workers in butions simply as a tax. Often when the public system some sectors at the expense of others. A related issue is of low quality, they feel that the levy they are paying is whether training opportunities are distributed is supporting an inefficient supply-led and hence, equitably across workers. If, for example, the tax inappropriate, training system. These schemes are finances the training of high-skilled workers, low- also sometimes used to expand into areas not relevant skilled workers would not benefit and would, in fact, to employer demands. For example, the Colombian subsidize the training of those with higher skills. training authority (SENA) has expanded into areas such as agriculture, construction and training for the Possible Advantages and Disadvantages of Revenue- self-employed ­ services that are only loosely tied to Raising Levy Schemes. These schemes are one of the the needs of firms that pay the levy. most sheltered and reliable sources for financing pub- lic training, especially in countries that face consider- Possible Advantages and Disadvantages of Levy-Dis- able difficulties in mobilizing such resources. These bursement Schemes. These schemes promote self- 3 I m p a c t E v a l u a t i o n financed, employer-based training. By establishing criteria for financing from levy-based funds, the · Levy schemes have led to an increase in quantity of training. In many countries levy schemes have schemes also allow management of the profile and had some impact on increasing the amount of quality of training offered by employers. These advan- training. For example, in Malaysia, HRDF partic- tages are apparent in Singapore's Skills Development ipants were more likely to have trained their Fund where the number of individuals trained tripled employees as compared to non-participants. Sim- since inception and number of firms benefiting from ilarly, in France under the levy-exemption the fund more than doubled since 1991, with a signif- scheme and in Singapore through the Skills icant portion of training conducted on the job. A Development Fund, the amount of training drawback with these schemes is that incentives may be undertaken by employers increased. A common inadequate to mobilize additional training, resulting feature of schemes in which training has in a deadweight loss ­ subsidizing programs that increased is the presence of an effective system for employers would have provided anyway.. This leads to administering the levy ­ both for levy collection a windfall to firms with well-established training pro- and for administration of grants. On the other grams, while other firms, which should be increasing hand, Korea saw a steep decline in the number of their training, may not do so. trainees once the training levy exemption was put in place in the mid 1970s. Very complicated rules governed training requirements for the levy and Effectiveness of Levy Schemes: Interna- the criteria for approval, hence many employers tional Evidence did not find it worthwhile to apply for it. There are over 30 countries in the world where levy schemes exist or have existed in the recent past. · These schemes are more effective under conditions Generally, tax rates range from 0.5-3.0 percent of pay- of economic growth. This is not surprising - train- roll and stay reasonably stable over time (Table 1). ing has greater impact under conditions of eco- nomic growth. For example, in Singapore, the Countries differ in the sectors covered by the tax ­ success of the Skills Development Fund is related with agriculture and the public sector most often to the fact that the economy and, corresponding- excluded. Smaller employers are also sometimes ly, the demand for training have been growing. excluded from coverage. The government has made a conscious decision Evaluative evidence on levy schemes is limited. In to link skills development policies with the eco- spite of the relative prevalence of these schemes, rig- nomic development strategy. Thus, skill improve- orous evaluations of the effectiveness of levy pro- ment is a continuous and expanding process that grams are uncommon. The evidence presented here is based on economic development needs. How- draws on the evaluations of some of these schemes. ever in countries such as Tanzania and Nigeria where growth has been slow, the levy-grant They include revenue-generating levy schemes such scheme has not significantly increased training as the Brazil SENAI scheme, payroll tax exemption because employer demand has remained schemes such as Cote d'Ivoire's Vocational Training depressed. Development Fund, and levy-grant/ training cost reimbursement schemes such as Hungary's Vocation- · Small employers do not benefit substantially from al Training Fund and Malaysia's Human Resources these schemes. This issue is of great concern to Development Fund (HRDF). developing countries where over 80% of employ- Some general findings on the impact of these ment is in small to mid-sized firms. Small and schemes emerge: mid-sized employers are either non-compliant in paying the levy or, even if they pay the levy, do not claim reimbursements for training their workers. The high rates of non-compliance are 4 I m p a c t E v a l u a t i o n BOX 1: GENERAL PRINCIPLES TO GUIDE THE · Employers buy-in for these schemes is crucial. OPERATION OF PAYROLL-LEVY SCHEMES Employers are also not always well disposed to the levy. For example, the U.K. levy-grant scheme Several points of principle should guide the was viewed by employers as an additional tax that operation of payroll levy schemes. These include: added to industry's costs and reduced overall · Levies should be subject to periodic review competitiveness. In Hungary, employers felt that · Levies should vary across sector/industry to the government exerted too much control over reflect differing skill composition of the labor funds and limited their effectiveness (see Box 2). force and training needs For such a scheme to be successful, governments · The system should be designed and administered should ensure that employers are consulted early in such a manner that encourages employers ­ and are involved in the design, implementation small as well as large- to participate and evaluation. It is important to ensure that · Training authorities should not venture into extra- training that is being demanded by employers is neous activities financed through these schemes. · The range of training services and courses provid- ed should reflect employer needs · Administrative efficiency and transparency is cru- cial. This is a prerequisite for any successful train- · Levies should be used to promote training by ing levy scheme. Characteristics of such a scheme enterprises. include an effective tax system/levy-collection mechanism, efficiency in processing and reim- bursing claims and clear and transparent applica- related to the low administrative capacity of gov- tion procedures that promote employer compli- ernments and ineffective levy-collection mecha- ance. The Singapore Skills Development Fund, nisms that are not able to target smaller employ- the French levy-exemption scheme and the ers (e.g. Cote d' Ivoire, Tanzania). Bureaucratic Malaysian levy-reimbursement scheme are exam- and administrative difficulties impose high ples of well administered and transparent opportunity costs, especially on smaller firms, and dissuade them from submitting claims (e.g. in Nigeria). BOX 2: LACK OF EMPLOYER INVOLVEMENT IN HUNGARY'S TRAINING FUND · There is cross-subsidization across skill categories. In a similar vein to the point above, evidence Three hundred and forty top managers in Hungary shows that these schemes do not affect the distri- were asked to assess the degree of influence of var- ious actors on the allocation of the training fund on bution of training across different skill categories a scale of 0 to 100. They judged that government ­ most of the training goes to the more educated and ministries have the strongest influence (score and higher-skilled workers. Hence, less skilled of 87), followed, in descending order, by local gov- workers are subsidizing the training of the more ernments (41), schools and training institutions skilled (e.g. in Malaysia). (37), employers' associations (32), and trade unions (21). In their opinion, the order of precedence Development of Levy Schemes: Some should be reversed, with employers followed by schools exerting the strongest influence on how Guiding Principles funds are allocated and government playing a less Based on the evidence, there are some general dominant role. As an example of excessive govern- principles that should be kept in mind for countries ment influence, employers cited the government's decision to use the training fund to finance public that are planning to implement some sort of levy training institutions that they believe should have schemes for training. These principles are similar to been financed from the government budget. those suggested in the Bank's policy paper on VET (Box 1). 5 I m p a c t E v a l u a t i o n BOX 3: THE MALAYSIAN HUMAN RESOURCES DEVELOPMENT FUND (HRDF) The HRDF, a levy reimbursement scheme, was established in 1992. The Human Resource Development Council, with rep- resentatives from the private sector and from responsible government agencies administer the scheme. Employers pay a payroll contribution of one percent and are eligible to claim a portion of allowable training expenditures up to the limit of their total levy for any given year. The reimbursement rates vary by sector and type of training. The HRDF is considered to be one of the most well-run schemes administratively and is extremely efficient in reimbursing claims and making procedures with which employers can easily comply. Evaluation of HRDF A survey conducted in 1994 of 1450 firms eligible to participate in the HRDF found that 402 firms (28 percent) were not registered with the HRDF. Of those registered, another 35 percent reported that they did not claim reimbursements under HRDF. Small firms (with 50 to 100 workers) are more likely to be non-compliant (49 percent) than large firms (8 percent). These data also revealed wide variations across sub-sectors in the take-up of the HRDF scheme. The high- est take-up rates of the scheme were in general machinery, electric machinery, and ceramics and glass. Sub-sectors with low take-up rates were food, beverages and tobacco, and textiles and apparel. Comparing the experiences of firms registered with the HRDF with those which chose not to register provides some answers about the effectiveness of HRDF. The results show that HRDF increased training modestly - 49 percent of the registered companies said that they had increased training and 39 percent firms said that their training had remained the same. In contrast, of the eligible firms not registered with the HRDF, 27 percent said that their training had increased while 47 percent firms said that their training had remained unchanged. Regression analysis showed that while HRDF did not have any impact on increasing training among small firms, it did have a role in increasing train- ing among medium and large firms. Who are these non-claimant firms and why are they not training? Only 6% of these non-claimants do no training. The majority of firms not claiming (54 percent) are those that only provide informal on-the-job training. The remaining 40 percent report that they provide formal training, yet do not claim reimbursements for expenditures. Analysis showed that firms least likely to claim from HRDF are small firms and firms providing only informal training. Important factors that employers cite as inhibiting their training: the limited resources available for training, the use of mature tech- nology with low skill requirements, the adequacy of skills provided by schools, and the availability of skilled workers who can be hired from other firms. schemes. However, even these schemes encounter ginalized groups. While this may be a socially jus- high rates of non-compliance ­ over one-third of tifiable objective in itself, it may not be economi- employers in Malaysia do not comply with HRDF cally optimal. regulations because they find them quite burden- some (Box 3). · Fund design should ensure that non-governmental providers are not crowded out. In some countries, · Levy schemes should finance in-service and not pre- especially where the levy collected is being chan- employment training. Firms are more likely to neled back to fund public training institutions, share in the cost of firm-specific training but less there has been crowding out of non-governmen- likely to share in the cost of general training. To tal training providers (e.g. Kenya and Tanzania). the extent pre-employment training is general in It is essential that the fund should be designed in nature, the burden of paying for that training such a manner that it allows both government should not fall upon firms. However, some coun- and non-government providers to compete on a tries have used a portion of the resources levied level playing field. Not only will this ensure that from firms to fund training authorities and pub- non-government providers are not crowded out, lic training centers that provide pre-employment but through ensuring greater competition, it will training for unemployed youth and other mar- also increase the effectiveness and efficiency of 6 I m p a c t E v a l u a t i o n Furthermore, evidence suggests that the levies can public provision of training. The government has be subsidies to large employers while a tax on smaller an important role in ensuring quality control employers. Given that most of the employment is among training providers, both government as concentrated in small and medium-sized enterprises well as non-government. Governments need to in the low-income developing countries, governments ensure that providers of good quality and rele- should not dampen their growth by imposing a fur- vant training are the ones that are allowed to access the fund. ther tax on them. · Governments should play a significant role in eval- uating the effectiveness of these schemes. As has Annotated Bibliography been pointed out, these schemes have rarely been Adams, A.V. and D. Fretwell (eds.) (1997). International Conference on Employment and Training Funds. Eco- evaluated in a rigorous manner. While over 30 nomic Development Institute (World Bank) and the countries have implemented these schemes, only Turkish Employment Organization. Provides an overview a handful of schemes have been evaluated rigor- and evaluation of levy schemes in different countries. ously. Governments have largely neglected their Dar, A., S. Canagarajah and P. Murphy (2001). Training Levies: Rationale and Evidence from Evaluations. Dis- roles as providers of information about the cussion Paper. This note is based on this paper. impact and effectiveness of these schemes. Effec- Dougherty, C. and J.P. Tan (1991). Financing Training: tiveness of these schemes will be enhanced when Issues and Options. World Bank Working Paper Series they are combined with rigorous evaluations to No. 716. Examines the rationale for different types of financing of training schemes. ensure that the training is actually implemented Gasskov, V (ed.). (1994). Alternative Schemes of Financing as intended, the quality of training is high, and Training. International Labor Office. Examines evidence employers are benefiting from the scheme. With- on effectiveness of levy schemes. out rigorous and systematic monitoring and eval- Gill, I., F. Fluitman, and A. Dar (2000). Vocational Education and Training Reform: Matching Skills to Markets and uation, there is significant opportunity for mis- Budgets. A joint World Bank-ILO publication. Provides use in these schemes. While the cost of evidence on VET reform in a cross-section of low-income monitoring and evaluation is likely to be high, an and middle-income developing countries and OECD countries. expansion of this role may be one of the most World Bank (1991). Vocational and Technical Education effective ways for governments to foster the and Training. A World Bank Policy Paper. The World development of relevant and cost-effective train- Bank's policy paper on vocational training. ing systems. Conclusions This note discussed the rationale behind levy schemes and looked at some of the evidence on the evaluation of these schemes. It is quite clear that implementing an efficient and equitable levy scheme may be more suitable for middle-income countries rather than most low-income developing countries. While these schemes have usually led to an increase in the quantity of training undertaken, some of the pre- requisites for such a system are not yet fully in place in most of these countries. Their economies are growing relatively slowly. Effective administrative mechanisms may not be in place to collect and disburse the levy in most of these countries. 7 I m p a c t E v a l u a t i o n 8