Financing Training

In most countries the cost of training is shared by individuals, enterprises and government. There used to be a tendency, especially in developing countries, for it to be assumed that the state should play a dominant role in training provision. In recent years economic recession and shrinking government revenues have forced a reconsideration of this view and growing acknowledgment and appreciation of the role of the private sector. Although the documentation is incomplete, it is now evident that nearly everywhere the private sector is responsible for skill development on a substantial scale. The state may have dominated the provision of high-profile institutional training, but the private sector, through innumerable types of training activities, many of them informal and discernible only to those directly involved, may well be making the greater contribution overall. In recognition of this fact, this chapter will begin by discussing how training is financed by it.

Until the present century, apprenticeship in one form or another was the usual means of entering those occupations which demanded extended initial training. In many countries it still is, either because it has retained popular support as a social institution or because alternative forms of training provision have not developed. It is not always called apprenticeship. Indeed, that term is typically reserved for training for craft occupations. In non-manual occupations, no doubt reflecting status differentiation, other terms are used instead. For example, accountants may serve articles, lawyers pupillage, and doctors internships or housemanships. In some occupations there is no special term at all, but nevertheless the trainee is effectively an apprentice during the training period.

Regardless of the terminology, pay determination has rested on a simple principle which has long been well understood by both practitioners and theorists. Apprentices and other individuals receiving initial training are in a position to appropriate all the benefits of the training if the skills are transferable. If the employer providing the training does not offer the market wage once they have qualified, they can move elsewhere. It follows that, since the training firm cannot reap any part of the benefits of the training, it cannot be expected to bear any part of the costs either, and will shift them to the apprentices by offering a wage lower than that paid to unskilled workers who are not being trained. This is acceptable to the trainees because they can expect to earn more than unqualified workers when they have completed their apprenticeships. The amount that an apprentice might expect to be paid depends on his or her productivity and the cost of training. In the simplest version of the theory, the firm does not incur a net training cost at any stage. It pays a wage exactly equal to the net contribution of the apprentice, that is, output less the direct cost of training (instructor salaries, cost of training facilities, training materials, etc). With the growing productivity of the apprentice, output and net contribution tend to rise over time, and so does the apprentice's wage.

It can well happen that the net contribution of the apprentice to the firm's revenues is initially very low, not even high enough to justify a subsistence wage. Under these conditions rigid adherence to the principle of equating the wage to the net contribution would deter potential trainees from taking up training places, or prevent them altogether if their families are unable to support them or forgo their earning power. This problem is overcome by making apprenticeship a contractual arrangement, with the trainee required to stay with the employer for the complete training period. This allows the employer to redistribute the income of the apprentice, overpaying in the first part of the training and underpaying in the latter part.

In times past, the need for the recovery of training costs by the employer often resulted in slow training. The term of apprenticeship was extended artificially, reducing the rate of training input and giving the employer more time to recover training costs. Such an arrangement is inefficient from the point of view of the trainee and can lead to exploitation. A more common practice now is for the apprentice to bonded to work for the employer at a below-market wage for a number of years after qualification, allowing the employer to incur net costs during training and recoup them later. Bonding is very widespread in the case of public sector employment, but it is not hard to find instances of its use in the private sector. Where bonding is discouraged or prevented by legislation or social norms, it may nevertheless be implemented de facto, the enterprise lending the trainee the cost of the training and cancelling the loan if he or she remains in service for an agreed number of years. In Belgium, France, Spain and the United Kingdom, employers have a legal right to make such provisions in employment contracts.¹

A key contribution by Becker to the theory of training finance is the distinction between general training for transferable skills and specific training for non-transferable ones.² With regard to general training, the analysis follows the traditional theory. In the case of specific training which raises productivity only in the firm that provides it, Becker's analysis shows that the costs and benefits of such investment would be shared between the worker and the employer. Because the worker receives part of the benefits, the risk to the employer that he or she will quit after training is reduced; by the same token, because the employer receives part of the benefits, the risk of layoff is also reduced. Since the benefits are shared, it follows that the costs of training would also be shared to achieve balance in the demand for training and its supply.

Most government-regulated apprenticeship systems are intended to provide broad-based, transferable training. According to the theory, in the absence of government intervention, the cost of the training should mostly be borne by the trainee and one should anticipate that apprentice wages should rise steeply from a low level. Certainly this is consistent with the facts. Usually apprentice wages are set at a fixed proportion of the wages of qualified workers, the proportion increasing according to the year of apprenticeship. In Europe, Asia and Australasia apprentices typically earn 30 to 40 percent of the qualified workers' wage in their first year. By their third year this figure has risen to 75 to 90 percent, although there are exceptions. As one would expect, wages tend to be lowest in those apprenticeship systems which have a reputation for providing the most intensive training. In Germany, for example, a minimum level is fixed nationally by employers' and workers' organizations, but even so, third year apprentices earn only about 40 percent of the qualified workers' wage.³

Perhaps unsurprisingly, the traditional theory is corroborated most transparently by the unregulated apprenticeship systems that have spontaneously grown all over the world. Typically apprentices in these systems receive only nominal pay to cover food and transport expenses and they may have to pay fees, the size of the fee in such cases bearing a clear relationship to the net cost and the value of the training. For example, a Nigerian survey found that the fees paid by apprentice motor mechanics tended to be much greater than those paid by apprentice tailors, reflecting the greater net cost of training mechanics, the higher incomes earned in that occupation and the greater chances for upward social mobility.⁴ An alternative to charging a fee is to extend the duration of the training artificially to allow the master to make a profit, as is described in a study of skill development in small and intermediate enterprises in carpentry and footwear production in Tunisia.⁵

Curiously, in spite of its economic importance, informal apprenticeship has largely been neglected in official policy-making. One factor, undoubtedly, has been a failure to recognize its contribution to skill development. Despite the fact that it may be the dominant means of skill development for most crafts, many planners in developing countries appear to be barely aware of its existence. It is no accident that most studies of informal apprenticeship have been undertaken by non-government bodies or individuals, often foreigners. Even when apprenticeship is well-documented, planners typically take it for granted that institutionalized training must be more cost-effective, or at least more effective, whereas it may be neither. Without doubt high priority in the training field in many developing countries should be given to a comprehensive assessment of this underworld, comprising a quantitative survey of its scope and an evaluation of the degree to which it substitutes or complements other forms of skill development.

Although enterprise-based training is the dominant source of skill development in most countries, privately-financed institutional training is also significant in scale in some. The rationale for purchasing such training is a straightforward extension of the rationale for purchasing privately-financed education: the trainee (or his or her family) incurs costs in the form of fees paid to the institution and income forgone during training, and in exchange receives benefits in the form of higher earnings on completing the training. The decision to purchase the training is rational if the rate of return on the investment exceeds the available return on alternative investments.

The simplest type of privately-financed institutional training is the training provided by proprietary schools, training institutions or back-street training centres which are privately-owned. The increasing concern with husbanding public financial resources in developing countries suggests that policy-makers should look positively on the potential contribution of proprietary schools to skill development. Nevertheless, proprietary training is very poorly documented in most countries. The reasons for this neglect are similar to those responsible for the lack of documentation of informal apprenticeship.

First, policy-makers and planners, and even researchers, are frequently unaware of the scale of this type of training. Quite typical in this respect is a Brazilian study whose objective was the evaluation of the relative importance of SENAI apprenticeship and short courses for training lathe-operators, it being supposed that apprenticeship would be the more important mode. In fact it turned out that apprenticeship was the minority route and that the majority of those who had acquired their skills through short courses were not trained by SENAI at all, but in proprietary schools.⁶

A second reason for the neglect of proprietary training by policy-makers and planners is a prejudice against it caused by the profit motive. They are concerned that the schools may overpromote their courses and generate unrealistic expectations with respect to the jobs that will follow from the training. They suspect that the schools may provide inferior quality training, economizing on inputs. In addition to these substantive criticisms, there is often a general feeling that the profit motive leads to the subordination of broader educational and training objectives. However, when these complaints are well-founded, remedial action is generally preferable to discouragement or prohibition. Remedial action may include requirements that proprietary schools publish records of the success of their graduates in obtaining employment, or perhaps, as a last resort because it can lead to corruption, regulation. It is worth noting in this context that the expectations aroused by public training programs are not necessarily any more reliable. It is unusual for public training establishments to undertake placement or tracer studies and to publicize the results.

In addition to reducing the pressure on public sector training provision, proprietary schools can be more cost-effective than their public counterparts. The lengths and contents of proprietary school courses are often more closely tailored to the actual requirements of each occupation in the particular market served by them. One of the explanations advanced for the co-existence of proprietary schools with competing, free, community colleges in the United States is that the training is more intensive and graduates enter the labour force more quickly.⁷

In some countries it is evident that proprietary schools have the additional merit of being more responsive than their public sector counterparts to changes in the demand for skills. They are typically much more assiduous than public establishments in maintaining close links with local employers, not just for the short-run objective of placing current trainees, but also for adapting curricula so that they can attract clientele and maintain their placement rates. If the demand disappears, the courses are dropped and the school may even close.⁸ By contrast, public training establishments tend to suffer from inertia. In many countries, this is attributable to the heavy centralization of authority. But even where public training establishments enjoy substantial autonomy, their management may not be held accountable for the labour market performance of their graduates and hence they may lack incentives to respond promptly to local feedback.⁹ And as for closing those that have outlasted their usefulness, the problem of writing off facilities and redeploying permanent staff often makes it an impossibility.

In many countries public sector training institutions have been established in response to a perceived need for the skills provided and a failure of the private sector to meet this need, the failure being caused by a lack of know-how or entrepreneurial capacity, a lack of capital, risk aversion or regulation. In developing countries such institutions are frequently established with the aid of international donor agencies which provide technical expertise and funding.

The fact that such institutions are publicly owned does not automatically mean that they should be publicly financed. In principle they could finance themselves like proprietary schools by charging user fees which cover their costs. In practice this is exceptional, especially in developing countries, but partial or even nil cost-recovery can only be justified if there is a case for subsidizing the training. Otherwise a failure to charge full-cost fees will lead to a misallocation of resources and to an unnecessary burden on the public purse. The misallocation of resources results from the price of training not reflecting its cost, with the consequence that social demand is artificially stimulated. A typical outcome would be unemployment of the graduates or their employment in jobs which make little or no use of their training. The unnecessary burden on the public budget is partly caused by this excess demand, and partly by the crowding out of private provision of the training.

In many countries the government actively promotes training provision, taking the view that it would be desirable for more resources to be devoted to skill development. Typically financial incentives are established, for example in the form of a levy-grant scheme, and a national training authority is established to administer it. The authority may also be assigned the task of coordinating, even regulating, training provision.

However, the rationale for this apparatus is by no means self-evident, at least in a market economy. Enterprises are expected to decide upon their products and processes, their investment in plant and equipment, their marketing and promotion, and their sourcing of inputs in response to market forces. Why should the training of their workforces not also be left to the same market forces? Training increases productivity, but there are diminishing returns and beyond some point the benefit does not justify the cost. What are the grounds for assuming that the optimal point has not already been reached through the operation of market forces?

It is the hallmark of a good trainer that this question will never have entered his or her head. However, if a proper case is to be made for government intervention, it should be based on one or more of three grounds: market failure, externalities and equity. Market failure occurs when firms do not attempt to maximize their profits. It is also caused by distortions attributable to minimum wages and wage rigidity, imperfect capital markets, risks and uncertainty, and rigidities associated with in-service training. Externalities exist when the benefits of training to society exceed those that accrue to private firms and individuals, justifying a greater investment in training than that undertaken by firms and individuals in their narrow self-interest. Equity arguments can be used to justify the subsidization of training for groups who would otherwise be unfairly treated or who are disadvantaged in economic or other respects.

The list of possible arguments under these three headings is long and has been surveyed elsewhere.¹⁰ In practice, there are only three or four situations where a compelling case can be made, and then only in specific circumstances. Under market failure, there may be cases for subsidizing transferable in-service training and apprenticeship training. Under equity, for subsidizing pre-employment training and training of the disadvantaged. It is unusual to find any persuasive arguments under the externalities heading.

In those countries where secondary and higher education are subsidized from public funds, there is a powerful equity argument for providing parallel subsidies for pre-employment training. In many countries the government pays part or all of the direct cost of secondary and higher education. In some, students in higher education receive a stipend in addition. Whatever their merits on other grounds, such subsidies can constitute a major source of inequity, for they benefit precisely those individuals who have the best prospects for employment, income and status on entering the labour force. Furthermore, they help to transmit inequality from one generation to the next, for the students who survive longest in the educational system tend to come from those families which are most able to provide supplementary resources for education and to forgo the potential earnings of their children while in school.

In such a situation, the issue is not so much whether the subsidization of pre-employment training is justified, but the scale of the subsidies to be provided. The logic of the equity argument suggests that the total subsidy per trainee should match the total subsidy received by an individual remaining in formal education. Given the variety of pre-employment training programmes in terms of duration and cost per month, this in turn implies different degrees of subsidization for different programmes, with trainees being charged partly-subsidized fees for the more costly ones. In practice, however, there is a tendency for institutional pre-employment training to be subsidized without discriminating among courses, and this can introduce a new element of inequity because those who receive high-cost training leading to relatively well paid jobs benefit disproportionately.

Efficiency considerations are also relevant in determining the scale of the subsidization of pre-employment training. The proportion of young people benefiting from training programs is invariably small, in developed and developing countries alike, because the proportion of entry-level jobs requiring extended training is much lower than seems to be generally assumed. A U.S. study using data from the National Longitudinal Survey of the High School Class of 1972 found that only 5 percent of the labour force entered occupations requiring six months or more of training.¹¹ The figure might be higher elsewhere, but it is inconceivable that it could be more than a very small proportion. Nevertheless the social demand for training may be very high. The subsidization artificially increases the private rate of return and the authorities can find themselves under pressure to allow enrolments unrelated to the needs of the labour market. The obvious answer is to replace subsidies with student loans, both for general education and pre-employment training, and this has met with some success, but it is politically not easy to arrange and the administrative demands limit applicability to the more industrialized countries.

Subsidization of training for the disadvantaged or retrenched (an equity argument)

The equity argument for subsidizing the training of those who are disadvantaged or who have been retrenched is widely accepted. Often it can be reinforced by the externality argument that failure to take special measures may lead to the creation of a socially-divisive underclass detached from the labour force. Persuasive though they may be, these arguments are blunted by the fact that the cost-effectiveness of such training is often low. Those who are illiterate or barely literate tend to have a low absorptive capacity and low motivation, even for the adult literacy programs which might have greatest long-term impact.¹²

The subsidization of transferable in-service training (a market failure argument)

Much in-service training is provided by firms voluntarily for firm-specific purposes. Being firm-specific, it does not raise the value of the employee to competitors and so the firm does not need to increase the employee's pay, or can afford to offer an increase which is lower than the gain in productivity. In this way it recoups its investment. In the case of transferable training, the basic theory holds that if labour is mobile the firm should shift the cost of the training to the employee by reducing wages during the training period. While this may be feasible in the case of apprentices, it is usually unthinkable to ask a regular employee to accept a cut in wages during a period of in-service training. Equally, it is generally impracticable for the firm to get an individual to pay the cost directly. If the firm attempts to recoup an investment in transferable training by not increasing pay in line with productivity, it risks losing not only the new investment in general training but also any previous investment in firm-specific training.

As a consequence, in-service training in transferable skills may be underprovided and there is a case for subsidizing its cost through training grants. However, the determination of the most appropriate source of funding for such grants is not easy. In some countries, general tax revenues are used, a common arrangement being the double deduction of training expenses for tax purposes. In such a scheme a firm paying 20% tax on its net revenues would effectively have 40% of its training costs paid from public funds. Alternatively, firms may be granted a credit equal to the cost of training which can be set against tax liabilities. In either case the fact that the subsidy can be identified as coming from tax that would otherwise have been paid by the firm in question in no way alters the fact that it is coming from general tax revenues, but it does create an incentive for the firm to undertake more training that would otherwise have been the case. However the use of general tax revenues is inequitable since there is no reason why ordinary tax-payers should be expected to pay for gifts of transferable training to a small proportion of the work-force.

In other countries, the funding comes from a levy imposed on enterprises, usually 1%-2% of the wage-bill, the argument being that firms which do not provide training themselves but merely "poach" trained workers from other firms will therefore be forced to compensate the latter. The logic, however, is severely flawed. The beneficiary in this situation is the trainee, who has received a free gift of transferable training and consequently commands a higher wage. The loser is the training firm. The poaching firm itself is neither a gainer nor a loser. It earns the odium of the training firm by precipitating its loss, but it has to pay the full market wage to secure the services of the trainee.

If there is a rationale for a levy, it has to be of the level playing-field variety: firms that do not provide in-service training have a commercial advantage over firms that do because they avoid this source of loss. According to this argument, it is only just that this advantage should be removed by imposing a levy on all firms so that they share the losses between them. If the argument is accepted, there still remains the issue of the scope of the levy. In some countries, all employers have to pay the levy. This is manifestly unfair to those sectors of the economy which have little or no need for workers with in-service training. In some countries, the levy is restricted to the manufacturing sector, but the same issue arises, since some industries have far greater need of in-service training than others. In the U.K. and some other countries, industry-specific schemes have been tried. Even within a single industry, however, there is usually no uniformity in the skill-needs of firms. In particular, small firms tend to have smaller in-service training needs than large ones because they employ proportionally fewer technical and managerial workers than large ones and the beneficiaries of in-service training are concentrated in the managerial, technical and craft occupations. The outcome is that small firms pay the levy but are not able to make use of the grant, with the consequence that they subsidize their larger brethren.

The analysis is complicated by the fact that the ultimate incidence of a payroll levy is unclear, since it may be partly shifted to employees by causing a permanent reduction in wages.¹³ If this is the case, the cost of the training gift is also shared among all employees, irrespective of whether they profit from the training or not.

There is no simple solution to the issue of the most appropriate source of funding. The unfairness of any definition of scope varies from sector to sector and country to country. One measure which may alleviate unfairness is to make small firms exempt from the levy and, by the same token, ineligible for grant.

The use of a levy as the source of finance is attractive to policy-makers because it usually has the effect of broadening the tax base, adding a new source of revenue without seriously undermining existing ones. The fact that it is usually a very stable source of funds sheltered from the vicissitudes affecting general government receipts can make it irresistible, irrespective of its theoretical merits.

In some countries with coordinated market economies the problem has been avoided by making labour markets non-competitive, thereby allowing enterprises to recoup investments in transferable training in the same way that they do in the case of specific training. Examples are Germany, where poaching is discouraged by tacit agreements among enterprises, sustained by the influence of such institutions as chambers of commerce, and the large-enterprise sector of Japan, where employment has traditionally been for life.

Whether training is general or specific, it must be financed, at least partly, by the worker or apprentice, in the form of a reduced initial wage. If this is lower than the minimum wage established by legislation, firms will be discouraged from providing the training. Restrictions on wage flexibility may also distort the incentives to invest in training. Such restrictions, often the result of collective bargaining activities, tend to compress before-and-after-training wage differentials and interfere with the sharing of the costs and benefits of investment in training between workers and employers. Trade unions may seek to limit wage differentiation, ostensibly to improve equity in pay among workers, but more importantly, to reinforce the union's bargaining power.¹⁴ By increasing trainees' wages as a fraction of those of experienced workers, unions create incentives for employers to hire fewer trainees, reducing the threat that a younger and cheaper labour force would pose to incumbent workers.

The most obvious solution to the problem is the removal of the institutional obstacles that restrict wage flexibility. In particular, apprentices may be assigned a legal status separate from that of the regular labour force and be exempted from minimum wage legislation. For other workers, a direct solution may be difficult to implement. Minimum wages are often desired for various social and political reasons, and their possible adverse effect on training may not outweigh the importance of these other goals. Providing subsidies to cover part of trainees' wages is a common intervention in such cases.

Table 8.2 presents a broad categorization of training promotion schemes by type of subsidization.

The direct subsidization of institutional training is commonly used to promote pre-employment training, the institutional training component of apprenticeships and the training or retraining of those who are disadvantaged or have lost their jobs, subsidization using general tax revenues being justified by the equity arguments outlined in Section 8.2.

Institutional in-service training may also be directly subsidized, justification in this case being provided by the market failure argument. However such schemes invariably have some controversial aspects. First, there is the difficulty in determining an equitable source of funding. Second, such schemes fail to have an impact on training needs for which institutional delivery is unsuitable. Third, they tend to promote institutional delivery of training even when it is less cost-effective than employer-sited training. Fourth, they often impose heavy administrative burdens both on the national authorities managing them and on the firms in scope to them.

In a few countries institutional training is indirectly subsidized by giving individuals training grants which can be used at a choice of approved training institutes. In some, tax credit schemes, in which individuals can deduct part or all of training fees from personal tax liabilities, have a similar effect. In either case the main objective is to provide subsidized training for specific target groups while maintaining a market-driven framework for training institutes. For private training institutes or self-financing public ones, the chief effect of such schemes is to expand the market; complying with the criteria for approval may also cause some changes, not necessarily cost-effective, in training delivery. For public and other training institutes previously in receipt of direct grants, the change requires a major change of culture and can be traumatic.

In many countries the promotion of in-service training has been extended to include employer-based training as well as institutional training, firms being reimbursed part or all of the cost of training provided directly by them or purchased from third parties. As in the case of institutional training, the usual sources of funding are general tax revenues via tax deduction schemes or a levy on employers, the latter arrangement being described as a levy/grant scheme. In some cases enterprises are allowed to deduct all or part of the cost of training from their liability to levy, in which case the arrangement is described as a levy/exemption scheme. The main advantages are that it enables employers to tailor their training programmes more closely to the needs of their employees and, in particular, allows cost-effectiveness to determine the choice between institutional and employer-sited training. However, the cost of administering such a scheme is even greater than that of levy-financed institutional training. Typically the administrative body tries to contain the cost of monitoring training expenditure by imposing rigid eligibility criteria, usually restricting eligibility to relatively formal employer-sited training, but this limits the flexibility of training decisions and can lead to distortions.

In most countries the institutional training component of apprenticeship is provided free from public funds on equity grounds. The same equity argument can be used to justify the subsidization of the employer-based component of apprenticeship, grants either being made directly to apprentices to supplement their low wages or indirectly, via grants to employers who are thereby put in a position to offer higher apprentice wages.

The use of general tax revenues to subsidize school-based pre-employment training is almost universal.¹⁵ The cost of providing training for the retrenched or disadvantaged is also typically met from general revenues, in line with the equity argument; in some, mainly European, countries it is financed by social insurance contributions. (Sometimes the contributions are nominally paid by the employer, sometimes by the employee, but the distinction is cosmetic and in practice the contributions constitute a special payroll levy).

There is tremendous diversity in the financing of other types of training but one can make out a stylized history for the evolution of national training policy on the following lines:

Stage 2 makes the smallest demands on the administrative capabilities of the state and employers and remains appropriate for lower-income developing countries where organizational structures are weak. It enables training provision to be established on a significant scale where there was previously a vacuum and it provides an opportunity for employers or their representative bodies to be drawn into training policy-making through the creation of advisory groups and task forces. At one time this was the dominant pattern in Latin America, SENAI, the Brazilian industrial training authority having led the way, but the larger countries in that region have now moved on. The main weakness of the arrangement is that it places responsibility for training provision in the hands of a centralized agency which is seldom subject to effective controls. Isolated from the discipline of market forces, there is a tendency for training to be unresponsive to changes in the needs of the labour market and for it to be biased towards institutional delivery. Often unit costs are high because resources are dissipated on specialized programmes with low enrolments and high development costs (typically they are not even evaluated). The levy may constitute an oversheltered source of funding, encouraging extravagance or leading to the diversion of funds to non-training purposes.¹⁶ In some countries the introduction of a levy/grant scheme has had a sclerotic effect on training provision, the bureaucratization of the training system affecting not only the national training authority but also the enterprises in scope to it.¹⁷

In stage 3 the bias towards the institutional mode is redressed, at least as far as initial training is concerned. In some countries no incentives are offered to either employers or apprentices, the national apprenticeship scheme serving solely to provide a legal framework for apprenticeship contracts. That is now uncommon, for most modern apprenticeship schemes make provision for day-release or block-release theoretical training at a training institute, and this is typically financed from public funds, as in Germany and the United Kingdom.¹⁸ Some systems provide subsidization of trainee wages, a measure which is a particularly popular measure in developing countries. For example, the National Apprenticeship and Industrial Training Authority in Sri Lanka assumes responsibility for 100 percent of the cost of apprentice allowances, off-site instruction and tests; however the scale of the regulated apprenticeship scheme appears to be very small, with only 10,000 enrolled in 1990, less than one percent of the relevant age group and about 10% of enrolment in technical and vocational schools.¹⁹ In India, the authorities pay 50 percent, up to a certain limit, of the stipends of apprentices in enterprises employing fewer than 500 workers. In the Netherlands, firms employing apprentices may be reimbursed for a percentage of their wages, as in the road transport industry, where the rate is 20%, or may be awarded a fixed grant, as in the metalworking industry. In Colombia, enterprises can claim up to 130% of apprentice salaries as tax-deductible costs. In Chile, under a scheme established in 1988, firms can deduct 60% of apprentice wages from their liability to profits tax. In Australia and Singapore, firms taking on apprentices are paid fixed grants. In Korea, under a scheme introduced in 1992, one of the ways in which firms can secure levy-exemption is by taking in government trainees for a year of practical training.²⁰

However, although formal apprenticeship schemes are easy enough to design on paper, they are much more complicated to administer than public institutional training and have tended to make a significant impact only in higher-income developing countries and industrialized countries.

Stage 4 brings the arrangements closer to market forces. Where the national training authority still provides training directly on a significant scale, it is likely to have developed mechanisms for making its training institutes more sensitive to changing employers' needs: effective links will be created with the private sector and an entrepreneurial spirit introduced, courses being tailored in terms of content, delivery mode and length to real needs rather than administrative convenience and serious efforts being made to recover some costs or even render some courses self-financing. Where provision is made for reimbursement of the cost of employer-provided training, the bias towards the institutional mode is reduced. However the need for meeting approval criteria can still lead to a bias towards inappropriately-formal training and other distortions.²¹ Efforts to monitor compliance and training quality, generally unsuccessful, impose high administrative costs on both the national training authority²² and for the firms in scope to it. Indeed the cost of obtaining approval for training plans, keeping training records, documenting training expenditure and preparing applications for reimbursement typically leads to participation being limited to larger firms, especially if the section of the national training authority responsible for the scheme is under-resourced and unable to process applications promptly and cooperatively. In extreme cases, the national training authority may deliberately discourage levy-exemption in order to maintain the flow of funds to its own training institutes.²³

By stage 5a the role of the national training authority has been reduced to that of a catalyst and it may even cease to exist as a central authority, its functions being disseminated among a variety of regional, local and sectoral bodies. The most notable example of this type of evolution among developing countries is provided by the 1976 reform in Chile precipitated by a report criticizing the national training authority, INACAP, for excessive rigidity. Under legislation enacted that year, there was a wholesale shift towards the promotion of employer-based training, with a coordinating body, SENCE, responsible for promoting employer-based training with financial incentives becoming the main instrument of training policy.²⁴ An equally radical example is the abolition of the Manpower Services Commission, the national training authority of the United Kingdom, and its replacement by a decentralized network of over 100 employer-dominated Training and Enterprise Councils (Local Enterprise Councils in Scotland) in the late 1980s. Most European countries have reached this stage, France perhaps being the exemplar. General revenues are used for pre-employment training (on a large scale: 1.5% of GDP). In addition to this there is a "apprenticeship tax" of 0.6% the wage-bill, of which 0.5% can be refunded against apprenticeship costs but is more commonly used for making grants to training institutions or associations, used for paying fees to external institutions, or paying compulsory contributions to chambers of commerce or industry. Any part of the 0.5% not distributed must be paid directly to the government, which reserves the remaining 0.1% for training youth. Further, there is a compulsory 1.5% in-service training tax on all enterprises with 10 or more employees. Employers can use 1.05% on in-house training or contracted-out training; 0.3% is used for youth training; and 0.15% for individual paid-leave training.²⁵

As publicly-subsidized training provision becomes less important, there is less pressure to impose a levy for the sake of generating revenue. As a consequence it becomes possible to provide training incentives through tax deduction schemes instead of levy/grant or levy/exemption. In Brazil, under a scheme initiated in 1975, double the expenditure on approved training programmes could be deducted, with a maximum remission of 10 percent (reduced in 1987 to 8 percent) of the liability to profits tax. A review of the impact of the scheme undertaken in 1985 by the National Manpower Council of the Ministry of Labour showed discouraging results: only 5,000 out of 2 million enterprises in scope participated in the scheme and there were few indications of an increase in training activity. The tax rebates were being used by enterprises that already had ambitious training programmes.²⁶ The scheme was discontinued in 1990, partly to reduce the budget deficit, but also partly because participation had remained at a very low level, only 1 percent. Many firms in Brazil manage to avoid paying tax on a significant scale. For them the practical value of the incentive was outweighed by the cost of participation in terms of complying with bureaucratic procedures. Many firms did not participate because their training needs were low. Most of those that did participate were larger firms that could cope with the paperwork and which tended to be better controlled by the tax authorities. However these were precisely the firms that were already providing training on a significant scale and it is possible that the incentive did not have much additional impact on them.²⁷

As a slight variation, in some countries enterprises may receive tax credits, equal to expenditure on training, which can be used to discharge liabilities to national taxes. In Chile, under a scheme introduced at the time of the establishment of SENCE in 1976, expenditure on training may be deducted from the profits tax liability up to a maximum of 1 percent of the payroll.²⁸ In Argentina, under a scheme established in 1980, tax credits equal to training expenditure are issued up to a maximum of 0.8 percent of the payroll.²⁹

Alternatively, institutional training may be promoted through individual entitlement schemes. Typically such schemes are restricted to certain categories of worker not in employment. Since 1944 military veterans in the United States have been eligible under successive G.I. Bills of Rights for grants facilitating access to up to 45 months of full-time training. The programme has been particularly effective in promoting the education and training of blacks.³⁰ A more comprehensive example will be provided by the Training Credits Scheme being introduced in the United Kingdom, under which by the year 1996 all 16 and 17 year old school leavers will be offered portable training grants. Large-scale schemes of this type are unusual in developing countries, but an example is provided by the series of programmes operated by SENCE in Chile since 1977. One scheme provides up to 300 hours of training and is targeted at those not eligible for employer-based training under the mainstream tax deduction scheme: the unemployed, the self-employed, and those working in very small enterprises, the awarding of grants being the responsibility of the municipal authorities. A separate youth training scheme, partly funded by a loan from the Inter-American Bank, is targeted at young people from low-income families.³¹

Further developments associated with stage 5a are an extension of the scheme to cover training-related expenditure as well as direct expenditure on training, measures to reduce the bureaucratic cost to firms, measures to encourage group training, and measures to target the incentives towards certain categories of worker. Chile provides examples of all of these: the cost of training-related travel, administrative costs associated with training and 10% of expenditure on training needs assessment qualify for reimbursement subject to certain ceilings;³² the bureaucratic costs have been reduced by allowing firms to contract out their training programmes to authorized training institutions (OTEs) which can take care of the paperwork; group training is promoted on a significant scale by organizations representing employers' federations (OTIRs), 50% of tax rebate in 1989 originating from programmes organized by five major OTIRs and the state mining company; and the scheme is biased towards lower paid workers by a provision that only half the training costs of those earning more than ten times the minimum wage may be set against the tax liability.³³

Similar developments are found elsewhere. In Argentina the Construction Chamber has established an OTIR-like support body, the Cesar M. Tolledo Training Centre Association.³⁴ In Singapore, the problem of the administrative burden imposed by a levy/grant system has been addressed by providing subsidies at a flat rate per trainee hour for in-plant training, eliminating the need for detailed costing, and at a fixed percentage of course fees of approved training establishments. To encourage small firms to participate, the grants are increased for first-time applicants. Singapore also provides an example of the use of a scheme to promote restructuring. Both the payroll levy and the main training grants programme are focused on low-paid workers. The levy, currently set at 1%, is imposed on the wages of those who earn S$750 or less per year (due to rise to S$1,000 in 1995). The training grants are targeted at those with secondary school qualifications or below and with monthly earnings below S$1,000.³⁵

A common criticism of levy/grant schemes at stage 4, or even levy/exemption and tax-deduction schemes at stage 5, is that they do not appear to have a significant impact on training activity. Often participation is mainly restricted to large firms which can afford to devote managerial time to the bureaucracy of securing approval for training expenditure leading to exemption, and often there is reason to believe that such expenditure would have been incurred anyway. In such cases the net result is that large firms suffer needless expenditure on administration and are induced to distort their training to meet the official exemption criteria, while smaller firms find themselves paying an arbitrary tax with no benefit. Even where a scheme is restricted to larger firms, a satisfactory participation rate is not guaranteed: in 1991 only 20% of the firms in scope to the Korean levy/exemption scheme were providing training, despite the fact that the scheme was at that time restricted to enterprises employing 300 or more workers.³⁶ Given these problems, the next stage may be to begin dismantling the system rather than refining it, 5b rather than 5a, a prime example being provided by the United Kingdom.

Proponents of 5a can cite the fact that French employers now devote an average of 3.2% of their payroll to training, a very impressive amount by international standards, and they have been active participants in the development of their complex system.³⁷ Sceptics can point to the fact that French employees tend to remain with their employers for long periods of time and it is this that encourages employers to make training investments in them. This is also the situation in Germany, where employers spend 4.0% of their payroll on training, with hardly any training incentives at all. The case for levy/grant or levy/exemption is further undermined by a tendency for such schemes to encourage training which is not always cost-effective.³⁸

In both versions of stage 5 responsibility for managing existing public training institutes is likely to have been devolved to local authorities, sectoral bodies or even the institutes themselves, with cost-recovery high on the agenda, as the United Kingdom, Sweden and Chile. In the United Kingdom, many public training institutes (the former Skillcentres) have been privatized and most of those that remain within the public domain, the Colleges of Further Education, formerly managed by local authorities, were incorporated as free-standing institutes in 1993. At the same time their exposure to the discipline of market forces has been increased by the diversion of part of their funding to individuals in the form of trainee credits and career development loans to be spent at approved training providers.³⁹ In Sweden, following reforms in 1986 and 1993, the training institutes of the AMU Group (the national training authority) have had to compete with other training providers for public training contracts and the AMU itself in the process of being transformed from a public agency to a corporation, albeit with only one shareholder, the state.⁴⁰ In Chile the INACAP was transferred to the National Employers' Federation in 1981 and competes with over 600 other private and public training institutions.⁴¹ Some countries have established similar systems of decentralized public training by other routes. In the United States, for example, community colleges, whose original mission was to provide publicly-subsidized entry-level training and general education, have expanded their provision of self-financing in-service training to the point where it now rivals or even dominates their previous function.⁴²

One conclusion of this review of training finance is that the case for financial intervention, and the analysis of the incidence of interventions, are complex and that this complexity is often not fully appreciated. Indeed sometimes national policy-making appears to take place in an analytical vacuum and government initiatives to promote training are launched without reference to market failure, externalities, equity or social concerns. Often, it would appear, the justification is simply an opinion that "not enough" training is provided by the private sector and that more would be desirable. The truth of the assertion that the level of training is too low is frequently taken to be self-evident rather than justified by reference to objective criteria.

In particular, the observation that relatively few firms provide structured training is not, in itself, a ground for intervention. Small firms tend not to provide it for several reasons. One is that it may be prohibitively expensive for them because they are not able to take advantage of economies of scale. Another is that they may have less need for skilled workers, their products and processes being less sophisticated than those of larger firms. It is therefore normal to find that they do not provide much training, in the same way as it is normal for them not to spend large amounts of money on research and development or marketing, to bid for government contracts or to export their products. In the ecology of enterprises, it is not reasonable to expect a small firm to behave like a downsized version of a larger one. When they do need skilled workers, it may well be rational for them to bid for them in the labour market, and this is likely to be the optimal solution from the social point of view as well. To condemn them as poachers is both unfair and likely to lead to ill-formulated policy prescriptions.

As a first approximation, apprentices and others undergoing extended training on entering the labour force should be expected to finance their own training by accepting appropriately-low wages or via some other device like bonding. However, if the education of peers remaining in school is subsidized, there is likely to be an equity argument for some financial support from public funds. The equity argument is reinforced by the fact that apprentices tend to be lower achievers than those remaining in general education and thus in any case have inferior earnings prospects. It may therefore be reasonable to provide day-release training free of charge, as is the practice in many countries.

It may also be reasonable to provide a modest element of income support. In principle the low training wage which is part of the traditional apprenticeship bargain should be acceptable prospective apprentices as well as the firms that take them in. However, market imperfections, notably myopia and uncertainty, may undermine the arrangement, justifying subsidization on a modest scale. Financing the whole of apprentice allowances, as is the practice in Sri Lanka, is unjustified since it is likely to result in windfall gains shared by apprentices and training firms. Where minimum wages inhibit apprenticeship, the most effective long-run solution is to exempt apprentices from the legislation or to leave apprentice wage-setting to collective bargaining

In some countries there may also be an efficiency argument for subsiding apprenticeship. It is not uncommon to find heavily-subsidized full-time public institutional training competing with apprenticeship training, and the institutional training is often much less cost-effective, being more expensive and having lower rates of training-related job placement. In such circumstances interventions designed to increase the proportion of young people taking the apprenticeship route, by enhancing its attractiveness, could lead to a net saving of public resources as well as to more effective skill development.

The equity argument used in the case of apprenticeship can also be used for justifying the provision of subsidies for institutional pre-employment training equal to the subsidies received by those remaining in general education. Beyond that point, however, cost-recovery should be the rule. Either fees should be charged or there should be a system of trainee loans. If neither of these measures is practicable, consideration should be given to replacing full-time institutional training with an apprenticeship scheme.

In countries with competitive labour markets, a market imperfection argument can be used to justify the subsidization of transferable in-service training. The fact that firms providing such training are unable to recover its cost inhibits its provision and constitutes a ground for its subsidization. The main problem is the practical one of subsidizing it in a cost-effective manner. One approach is the direct provision of subsidized public training by a national training authority. However, the performance of national training authorities has on the whole not lived up to expectations, complaints of rigidity and high costs being common. A national training authority is probably the only solution, despite the problems, in lower-income developing countries. It is needed primarily to provide training, given that there is no alternative, but it should be expected to discharge two other important roles. It should provide a focus for establishing a tripartite approach to determining training policy, and it should act as a conduit for channelling a training culture and good training practice to the larger enterprises. In this it will be assisted by technical cooperation from the donor agencies that helped to establish it in the first place.

In other countries, increasing reliance is being placed instead on more market-driven schemes intended to promote employer-based training. Such arrangements turn out to have problems of their own. They can lead to a misallocation of resources, particularly when the intervention involves subsidies which reduce the effective price of training services and encourage training which is not cost-effective, or when employers are induced to choose inefficient modes of training in order to qualify for grants or levy-exemption. Worse, there is seldom any evidence of a significant incremental impact on training activity, for the number of workers trained, as a proportion of the target group, is often low and much of the training credited to such schemes would have taken place anyway. Further, they can be costly to administer and tend to impose heavy bureaucratic costs on firms participating in them.

Perhaps the most successful schemes a the present time are those of Chile and Singapore, both of which have features intended to address the problems that inevitably arise. However the fact remains that in-service training appears to be well-established only in those countries where internal labour markets are important or where there are other socio-institutional restraints on inter-firm labour mobility.

There seems to be a view that since a payroll levy increases wage costs by only 1 or 2 percent, the imposition of one is a marginal measure for which a rigorous justification is not really needed. This is not the case. Given that wages account for at least 50 percent of national income, more in many countries, the resources tapped by a 2 percent levy could amount to 1 percent of GDP. This may not seem high in relation to, say, expenditure on education, which usually accounts for about 4 or 5 percent. However it is huge, given that the beneficiaries typically are numbered in tens of thousands, while school enrolments amount to millions. The disparity is greater still in terms of hours. The courses attended by in-service trainees seldom exceed several hundred training hours, whereas there are over a thousand annual contact hours in the school year.

To put the same point in a different way, the unit cost of a trainee-hour promoted by a financial scheme will typically be vastly higher than the unit cost of a student-hour in primary or secondary education. This is not in itself a cause for concern, since training is nearly always considerably more expensive than general education. The issue is cost-effectiveness: is the increase in productivity caused by the training commensurate with its high cost? And this is by no means self-evident.

In such cases the reallocation of even a fraction of the 1 percent of GDP absorbed by a training promotion scheme to improving the quality of basic education could have a far greater long-run impact on productivity, if only because better-educated workers are more trainable. All the evidence shows that firms are disproportionately willing to invest in the training of their more skilled or educated workers, whether or not the government provides training incentives. One reason for this may be that large firms both tend to employ a greater proportion of such workers and are more likely to mount systematic training programs. But another, and probably much more important reason, is that the value-added of training is greater for such workers. Or, to put it another way, that training and education are complementary.

The conclusions of Pang and Salome (1986, p. 121) concerning the diversion of the proceeds of the payroll tax in Singapore from upgrading low-skilled workers are typical. Their survey revealed that few companies attempted to upgrade the skills of the unskilled workers, but "this was not entirely due to companies' lack of interest in training the unskilled workers. More correctly, it was due to the lack of basic education among the unskilled workers that makes it almost impossible to send them on skills upgrading courses."

In other words, the most effective financial intervention for training may be to upgrade the basic skills of the work force, in terms of literacy, numeracy, cognitive ability, communication and interpersonal skills, etc. This may take the form of a remedial program. For example, in Singapore the BEST (Basic Education for Skills Training) programme was established in 1983, supplemented by the WISE (Worker Improvement through Secondary Education) programme in 1987. Pang and Salome consider that "the implementation of the BEST program by the Government to help the unskilled workers to obtain the basic literacy and numeracy skills, necessary for further training, was vital for their upgrading." But there are problems with remedial programs. Adult workers, particularly older ones, tend to lack the motivation to participate, and participation is often made difficult by disruption caused by shiftworking.

The alternative, which has the additional merits of cheapness, high coverage and equity, is to improve the quality of basic education. From a political point of view, this alternative may be unappealing because basic education does not have a lobby and, unlike direct training initiatives, such an intervention confers little political benefit because its effects are long-term and almost invisible. But it is the standard against which all other initiatives should be measured.

EMP/SKILLS - Skills and Employability Department

Uses of funds	Types of scheme
Direct subsidization of institutional training	Subsidization of pre-employment training Subsidization of the institutional component of apprenticeship schemes Subsidization of training for unemployed or disadvantaged Subsidization of institutional in-service training
Indirect subsidization of institutional training	Individual training grant schemes Individual tax credit schemes
Subsidization of employer-based training	Levy/grant and levy/exemption schemes Tax deduction schemes Apprenticeship support schemes