The Employment Tax Incentive – a reform proposal

In the context of South Africa’s deep structural unemployment, neither the empirical evidence nor the underlying theory supports the current narrowly targeted wage subsidy of South Africa’s youth employment incentive. We outline recommendations for reform of the incentive as a broad-based employment subsidy, drawing in part on a comparison with the US earned-income tax credit. How should the ETI’s design be adapted, and what are the cost implications?  

Introduction

The National Treasury introduced the Employment Tax Incentive (ETI) in 2014 to encourage employers to hire young work-seekers. A variation on the earlier recommendation of the International Panel on Growth for a “once-and-for-all” fixed wage subsidy to facilitate the school-to-work transition, the ETI also represented a considerably more modest intervention in the labour market than the general wage subsidy then under consideration as part of the comprehensive social security reform process. The social security wage subsidy proposal was to supplement the earnings of all low-wage employees, both as a redistributive measure and to assist in their participation in a statutory payroll-financed retirement and social insurance arrangement.

The evidence from the current ETI arrangements suggests that a narrowly targeted youth incentive is an inadequate intervention in the context of high unemployment. On both redistributive and labour market grounds, there should be a reconsideration of the design of the ETI in favour of a broad-based wage incentive. While education and training reforms and targeted industry-support measures are also key components of an employment-focused growth strategy, South Africa needs to provide economy-wide support to labour-intensive firms.

The current ETI provisions

The ETI operates through employers’ PAYE returns to SARS. Employers can deduct the ETI amount from the total PAYE amount owed to SARS each month. If the ETI exceeds the employees’ tax payable, the excess can be rolled over or reimbursed to the employer. It is in effect a short-term cost-sharing arrangement between government and employers in respect of the wages paid to young employees who earn below a prescribed threshold. It is available to employers other than national, provincial or local government authorities or public entities, subject to a requirement that the incentive does not lead to displacement of employees.

The ETI can be claimed for employees aged 18-29, other than domestic workers or those “connected” to the employer, for a maximum period of two years. The wage paid must meet the requirements of applicable bargaining council agreements, sectoral determinations and the national minimum wage. With effect from August 2020, the age restriction (but not the two-year limit) was removed for qualifying companies within six designated special economic zones. [1]

The ETI amount is calculated in relation to an employee’s monthly wage (or on a pro-rata basis if the employee works part-time or for part of a month). With effect from March 2022, an increased ETI benefit was implemented, though the remuneration thresholds were left unchanged. In April 2025, the remuneration thresholds were raised, but not the maximum benefit. The calculation formulae applicable before and after these changes are set out below:

It is helpful to think of the ETI amount as determined by three thresholds – A: the monthly earnings level at which the maximum ETI amount is triggered (R2,500); B: the earnings level at which the ETI amount begins to decline (R5,500); C: the earnings level at which the ETI falls to zero (R7,500).

The ETI can be claimed only if an employer is compliant with applicable wage determinations and the minimum wage. In general, therefore, it applies only to earnings above the minimum wage (from March 2026: R30.23 an hour, or R4,837 for a 160-hour month). Only within a narrow remuneration band (R4,837 – R5,500 a month) is the maximum ETI amount payable.

Part-time work (less than 160 hours a month) attracts a subsidy proportional to the subsidy that would be paid for full-time work. This is presumably intended to prevent employers from gaining a cash advantage from job-splitting. But it has the effect of reducing the benefit associated with part-time or shift-based work. Considered together with the minimum wage requirement, this means that the earnings threshold A and the flat-rate benefit between thresholds A and B are largely irrelevant. The ETI benefit is in practice determined by a proportion of the maximum calculated by reference to hours of work. Apart from the impracticality of ensuring that employers provide accurate information on hours of work, job splitting brings the benefit of additional work opportunities, which is what the incentive is intended to achieve.  

When the ETI was introduced in the 2013 Budget, it was envisaged as “a graduated tax incentive at the entry-level wage falling to zero when earnings reach the personal income tax threshold.” The PIT threshold for the 2013/14 year was R67,111 (or R5,593 a month). The ETI thresholds introduced in 2014 were consistent with this intent – threshold A was R2,000 a month, threshold B was R4,000 a month and threshold C was R6,000 a month – just above the personal income tax threshold.

Since then, threshold A has increased by R500, while thresholds B and C have increased by R1,500. This represents a substantial erosion in the real level of the ETI remuneration thresholds. The 2014 thresholds, adjusted for CPI inflation, would be as follows in 2025 prices: A: R3,360; B: R6,720; C: R10,080. The maximum ETI amount of R1,000 in 2014 has similarly eroded in value, as its equivalent of R1,680 in 2025 prices has only partially been accommodated by the 2022 increase.

Together with the March 2022 increase in the value of the ETI, the National Treasury indicated in the 2022 Budget Review that “improved targeting of the incentive will be considered to support jobs for long-term unemployed work seekers, alongside an expansion of the eligibility criteria for qualifying employees to improve the incentive for small businesses.” These intentions do not appear to have been taken forward.

SARS and National Treasury records indicate that ETI claims amounted to about R4.8 billion in 2019/20, increasing to R7.2 billion and R6.6 billion during the 2020/21 and 2021/22 years in which an expanded benefit was available, and falling back to just under R4.8 billion in 2022/23. The number of individuals for whom claims were submitted, outside of the expanded ETI years, appears to have fallen from a peak of around 700,000 to perhaps 600,000 currently.

Impact and effectiveness of the ETI

Studies of the impact of the ETI have yielded mixed results. Several reviews of the aggregate impact on employment of young workers suggest no significant effect (Ranchhod, V and Finn, A, 2015; Ebrahim, A, et.al., 2017; Muller, S, 2021). However, a number of firm-level studies have found positive effects in that employment increases faster at firms that claim the subsidy, which tend to be larger faster-growing firms (for example, Bhorat, H, et.al., 2020). A recent review of the available evidence (Budlender and Ebrahim, 2022) finds little support for the ETI as an employment intervention and concludes that its future development requires “a fuller and more careful and sceptical engagement with the evidence”.

That the ETI has had little impact on either overall employment of young people or the distribution of income is no surprise. The erosion in the real value of the ETI thresholds, relative to the minimum wage and to young people’s reservation wages, partly explains this outcome and will further limit its impact over the period ahead. But there are several more fundamental reasons.

From a macroeconomic perspective, the aggregate value of the ETI is small. The effective reduction in the personal income tax burden peaked at less than 1.1% in 2016/17 (apart from the expanded ETI years) and is now less than 0.8%. As a share of GDP, the ETI cost is less than 0.1%.

As a redistributive measure, the ETI is similarly weak. Calculated as a share of gross reported income below the tax threshold, it is less than 1.5%. It amounts to less than a quarter of total wages paid on public employment programmes, and one-fortieth (less than 2.5%) of the total value of social grants paid to households.

From the perspective of individual firms, the substitution of existing employees for ETI-qualifying work-seekers (“displacement”) is prohibited. Though enforcement is probably difficult, this means that active recruitment to take advantage of the ETI is inhibited and is mainly driven by the income effect of the incentive. But the income benefit falls away after two years of employment, and in the relevant earnings range above the minimum wage is at best a third of the cost of employment and falls rapidly as the wage rate rises.

The 75% phase-out rate of the incentive, with effect from March 2022, substantially weakens the income benefit to firms employing new recruits who perform well or exhibit capabilities consistent with earnings above the minimum wage. The cost to the employer of raising the wage of a first-year ETI employee from R5,500 a month to R7,500 is (R1,500 + R2,000), an implicit marginal tax of over 40%.

For larger labour-intensive firms, with staff turnover sufficient to target young work-seekers as recruits, the age-limited two-year incentive will be claimable for at most one in ten employees, depending on average age and turnover rates. For small or medium-sized firms, the scope for adjusting the size and composition of the workforce in response to an incentive that attaches to a specific age cohort for a restricted time is limited. For any firm preferring continuity of workforce over rapid turnover, the ETI offers no lasting reduction in costs associated with expanding employment.

The underlying idea behind the ETI is that if firms offer more first-time employment opportunities, younger work-seekers will have better prospects of finding work – and this will stand them in good stead beyond the duration of the incentive. But South Africa’s youth unemployment problem cannot be addressed by a marginal shift in the employment of young people. The 18-29 age cohort is more than a quarter of the total labour force: South Africa’s youth unemployment problem is the sharp end of the overall unemployment problem. This cannot be solved by intervening only in the transition from school to work, it must be addressed on the demand side of the labour market as a whole.

In sum, the available evidence indicates that the ETI has little impact on South Africa’s unemployment rate, as it does little to reduce the cost of labour for labour-intensive firms or activities, or to shift technology in a labour-intensive direction. If fiscal support is to be provided to employment creation in South Africa’s circumstances, a broader labour market intervention must be sought. The case for a broad-based labour subsidy was made in the 1990s by at least one of the international advisors to the South African Labour Market Commission, though its Report limply recommended that “a task force … be convened to consider the pros and cons of employment subsidies.”[2] A subsequent analytic review argued that a wage-subsidy programme would raise employment and GDP and reduce poverty and inequality in South Africa. Links to training activities and targeting of industries with employment growth potential were identified as important complementary measures.

If the structure of the economy yields fewer jobs than are required, and if there are good reasons for not allowing wages to fall to market-clearing levels, then consideration must be given to a subsidy or incentive design that impacts broadly on the demand for labour. There are many variations on this in international practice, but far the largest programme of support internationally for low-wage earners is the earned income tax credit (EITC) administered as part of the US federal tax system.

A comparative perspective: the US Earned Income Tax Credit

Like the ETI, the EITC has its theoretical roots in the idea of a “negative tax” on earned income. If vertical equity in the tax structure is given effect by progressively lowering earnings in the hands of the employee relative to the wage or salary cost to the employer, above some threshold, the negative tax aims progressively to increase income in the employee’s hands at earnings levels below this threshold. This is essentially a redistributive function, though it also impacts on both labour-market participation and employment.

Both the ETI, and the basic EITC, are structured to rise with earnings up to some maximum value, and phase down to zero at or around the individual tax threshold. The ETI is phased in to its maximum comparatively rapidly and is now phased out even more rapidly. The EITC is phased in slowly and is phased out more slowly than it is phased in. The EITC is administered as a reimbursable credit claimed by and payable to the individual (or couple if filing jointly), whereas the ETI is administered as a credit to the employer.

The federal EITC costs about USD72 billion a year, or about 3.3% of individual income tax revenue and 0.3% of GDP. Together with the child tax credit, it is one of the largest and most effective poverty reduction programmes in the US. This impact is achieved in part because the value of the EITC is substantially higher for claimants with dependent children than for individuals with no children. The EITC is therefore broadly similar in its intent and distributional outcomes to the family allowances and tax-benefit systems of many other OECD economies, though not all are administered by the tax authorities. Reforms to these arrangements in recent years have strengthened links between social security benefits and participation in work. [3]

In 2023, the maximum EITC benefits ranged from $600 (a year) for individuals without children to $3,995 for taxpayers with one child and $7,430 for three or more children. In effect, the EITC serves as a substantial supplement to earnings both for low-income workers with only part-time income from work during the tax year and for earners with family dependents who earn up to about four times the US minimum wage (currently $7.25 an hour, or about $14,000 a year).

Although it is primarily a redistributive measure designed to improve income security of families with children, studies of its labour market effects confirm that it increases employment, particularly of single women with children (Schenzenbach and Strain, 2020). By increasing both incomes and labour market participation of low-income families, the EITC is widely regarded as the most important fiscal channel through which income inequality and poverty are reduced in the US.

A broad-based wage subsidy

The Treasury’s 2007 Social Security and Retirement Reform proposal for a wage subsidy was elaborated in some detail in a subsequent discussion paper, which included cost estimates and the results of three modelling estimates of the potential impact on employment, poverty and inequality. The analysis indicated that a wage subsidy with a maximum value of R5,000 a year (approximately R12,900 in 2024 prices) and phased out at around the tax threshold would cost between R16 billion and R20 billion a year (R42 – R52 billion in 2025 prices), mainly to subsidise social security contributions. The analysis suggested that employment would be raised by 200,000 to 300,000, while also impacting positively on earnings (and retirement security) of the lowest-paid workers.  

The more important impact of a reform along these lines is in its long-term impact on the relative cost of labour-intensive activities and firms. It is through a permanent downward shift in the relative cost of unskilled labour that a sustained shift towards greater labour absorption, and poverty reduction through employment, can be achieved.

In the absence of fiscal support for low-wage employment, the degree of downward pressure on wages that would be needed to achieve near-full employment would be both politically and socially unfeasible. Effective labour market activation programmes and employment services targeted at work-seekers are needed, but a basic redistributive measure to support low-wage employees is also a critical enabling condition for both more rapid employment creation and sustainable poverty reduction.

When the Treasury’s social security reform proposals were tabled in 2007, sectoral determinations and bargaining council agreements covered much of the formal economy, but there was no minimum wage. In the presence of a national minimum wage, which has already been increased several times in inflation-adjusted terms, it is that much more important that a broad-based fiscal support measure should be implemented to support low-wage industries and activities.

Additional to its employment-enhancing and redistributive benefits, a third consideration comes into play in the presence of enforceable basic labour standards. It is well-documented in comparative development experience that rising real minimum wages, particularly when coupled with mandatory social security participation, lead to “informalization,” unevenly applied regulations and an increasing gulf between established high-paying formal businesses and an unregulated informal economy. If this tendency is to be countered, fiscal support for low-wage employment is a helpful element in the policy mix. It is an “incentive-compatible” approach to the modernization of labour standards and broadening of social insurance and retirement funding participation.

Reform of the ETI: design considerations

If the ETI is to serve as an effective demand-side fiscal incentive, while contributing to poverty reduction and supporting compliance with the national minimum wage, how should its design be adapted?

The ETI should be available to employers outside of government for all low-wage employees, subject to compliance with applicable labour standards.  

The “grossing up” and “grossing down” provisions of the current ETI arrangement should be dropped in favour of a benefit calculated irrespective of hours worked in a month, or alternatively a de minimus requirement of, say, 40 hours a month.  

The maximum age of ETI qualifying employees should be set at the age thresholds applicable to the old-age grant and complementary retirement funding arrangements.  

A standard ETI benefit should apply to all qualifying employees, irrespective of age and years of service.  

The ETI should be phased down more slowly than the current phase-out rate.  

Recognising that the primary employment impact of a broad-based ETI would be in labour-intensive low-wage sectors or zones, these should also be the target of complementary industrial and development policies and support programmes. Though sectoral targeting of the subsidy should be avoided, growth opportunities should be sought in agriculture and food processing, light manufacturing, residential construction, restaurants and other service activities, for example – supported through local development planning, training and related supportive policies.

Costs and financing of a broad-based ETI

Cost estimates of a broad-based demand-side employment incentive designed also as a poverty reduction measure are set out here. Three alternative ETI models are assessed, calculated for the distribution of taxable income reported to SARS in 2017/18, raised to 2024 prices. Option 1 is similar in design to the current ETI second-year benefit with a maximum value of R750 a month, except that it is phased out more slowly, down to zero at R8,500 a month, which is approximately the current tax threshold. Option 2 adopts the same benefit design parameters as option 1, but without the proportional downward adjustments of benefits for employees working less than 160 hours a month. Option 3 phases up more slowly to a maximum of R1,000 a month (25% of R4,000) and has a single-point peak, before phasing down to zero at R8,500 a month.

The estimated total cost at current levels of formal employment of the broadest of these alternatives, option 3, is around R29 billion. This represents just under 3.5% of 2025/26 PIT revenue, and 0.35% of GDP. Higher employment stimulated by the ETI would lead to somewhat higher fiscal costs.

Over time, the ETI will support growth of labour-intensive activities and will partially fund itself through the resulting revenue gains. But in its initial phases, an expanded ETI will need to be accommodated through tax measures or complementary spending reductions. Consideration might also be given to a phased introduction, perhaps targeted at employment-intensive sectors or special development zones.

A two-percentage point increase in the UIF contribution (one percent each by employers and employees) would be sufficient to finance an ETI of R25-R30 billion, considering that the UIF currently runs a surplus on its statutory obligations. Repeal of the Skills Development Act and diversion of its payroll-based tax to the ETI would release R26 billion in largely ineffective sectoral training and related activities, though some of this might need to go to compensating TVET college funding. A one-percent increase in the VAT rate would yield around R33 billion. A one-percent increase in the base rate of PIT (from 18% to 19%) would yield around R30 billion. Over time, progress in employment creation would reduce the need for the social relief of distress grant, which currently costs R34 billion a year.

Conclusion

Implementing a broad-based ETI even on more limited terms than outlined above, would represent a sizeable shift towards a more redistributive and employment-oriented fiscal structure in South Africa, operating through greater cost-competitiveness of large numbers of labour-intensive enterprises. It would be a substantial contribution to more inclusive growth, implemented through tax administration capabilities that are largely already in place. It is an incentive-compatible tax-benefit reform, in that it encourages registration and formal participation of small enterprises that employ low-wage or irregular employees. The administrative feasibility of this approach, by comparison with more discretionary and institutionally demanding spending measures, is perhaps its strongest argument.

A broad-based wage subsidy is not a sufficient response to South Africa’s extreme unemployment conditions. Improvements in basic education and improved access to vocational education and training are critical. A re-designed employment incentive should be linked to social security reform and to appropriate industry-targeting measures. But the administrative capabilities to implement a broad-based employment incentive have now been built and should surely be put to work.

 

A more complete version of this paper is published as SALDRU Working Paper 310, November 2024. Funding for an earlier draft was provided by the Centre for Development and Enterprise.  

 

Footnotes

[1] For two four-month periods in 2020 and 2021, in the context of Covid-19 lockdown restrictions an “expanded ETI” was available in which the two-year restriction and age limits were lifted, and the calculated incentive amounts were raised.

[2] Presidential Commission to Investigate Labour Market Policy, 1996, p. 135.

[3]Comparative;reviews of countries’ tax and benefit systems are published regularly by the OECD. See www.oecd.org/social/benefits-and-wages.

/sites/default/files/articles/Donaldson%2C%20A_Employment%20Tax_March26.pdf