Journal Issue
Share
Article

South Africa: Selected Issues

Author(s):
International Monetary Fund
Published Date:
March 2000
Share
  • ShareShare
Show Summary Details

II. Labor Legislation in South Africa, 1995-199819

A. Introduction and Summary

53. Between 1995 and 1998 South Africa rewrote almost its entire labor legislation and passed new laws aimed at correcting some of the abuses of the apartheid era. The Labor Relations Act (LRA, 1995) replaced a version of the same name; the Basic Conditions of Employment Act (BCEA, 1997) replaced a version of the same name and the Wage Act; the Skills Development Act (SDA, 1998) replaced the Manpower Training Act (MTA); and the Employment Equity Act (EEA, 1998) was passed. In addition, the Presidential Jobs Summit was convened in 1998, from which was launched a number of initiatives designed to stimulate employment.

54. Given South Africa’s high unemployment rate (37 percent in 1997),20 there have been concerns about the possible impact of these laws on incentives for job creation, quite apart from the positive effects they might have on promoting industrial peace, safeguarding the basic rights of workers, enhancing the skills of the workforce, and redressing anomalies in the workplace caused by the apartheid system.21 The government has committed itself to a review of existing labor legislation to determine whether it contained unintended aspects that were inhibiting job creation.

55. This section summarizes the recently passed labor laws and examines them in terms of their potential impact on job creation, using the recommendations of the OECD Jobs Study as a point of reference.22 The paper also compares labor market legislation in South Africa with that in other selected countries. The section is organized as follows: Subsection B provides some stylized facts about the unemployment problem in South Africa; Subsection C describes the recently passed labor laws, comparing them, where relevant, to the old laws they replaced; Subsection D describes the main recommendations of the OECD Jobs Study; and Subsection E analyzes the laws in terms of the recommendation of the Jobs Study and compares them to laws in other selected countries.

56. The review finds that the recently passed labor laws are not significantly out of step with the recommendations of the OECD Jobs Study, except in the area of enhancing the flexibility of wages. In this regard, the LRA’s promotion of sectoral collective bargaining agreements, which can be extended to nonparticipants, appears to run counter to the recommendations of the Jobs Study. South Africa’s labor legislation is not atypical of that found in many industrialized and emerging market economies, and in many cases the laws passed in the 1995-98 period are in step with the trends taking place in many countries.

B. Unemployment in South Africa

57. The overall unemployment rate was 37 percent in 1997. This figure hides the enormous divergence in unemployment conditions across races, skill levels, sexes, and regions. It also hides the fact that, even with the inclusion of discouraged workers in the labor force, the participation rate (55 percent) is quite low. The unemployment rate varies from a low of 4.3 percent among white men to 54.8 percent for black women (Table 3). Unemployment rates in rural areas are significantly higher than those in urban areas. For black persons, the likelihood of employment does not appear to be affected by the number of years of schooling until they reach the tertiary level, at which point the unemployment rate falls to 10 percent, compared with rates of 50 percent for persons with a secondary education or less.

Table 3.South Africa: Unemployment rates, 1997
South AfricaAfricanColoredIndian/AsianWhite
Total37.446.522.112.86.2
Urban30.841.523.612.86.2
Rural50.453.214.5
Male30.73918.410.24.3
Urban25.234.720.110.24.3
Rural424510.1
Female45.254.826.717.18.9
Urban37.649.227.717.18.6
Rural59.561.920.7
Schooling
No schooling42.444.218.3
Grade 647.651.323.8
Grade 1036.145.024.519.310.8
Grade 1233.650.321.410.96.2
Diploma10.216.92.6
Degree5.610.33.6
sources: Statistics South Africa, 1997 October Household surveyNote "…" indicates that simple size was too small for reliable estimates
sources: Statistics South Africa, 1997 October Household surveyNote "…" indicates that simple size was too small for reliable estimates

58. One of the most striking features of the employment picture in South Africa is the extremely low ratio— 21 percent— of employed persons to total population (compared with, say, Australia, at 67 percent, Brazil, at 44 percent, and Mexico, at 61 percent). This low ratio reflects, in addition to the high unemployment rate, the low participation rate, and the very skewed age distribution of the population toward the young (39 percent of the population is under 15 years old). On the basis of the current age distribution, and even assuming the same low participation rate for the next ten years, the labor force is likely to grow by 30 percent during this period. Thus, for the unemployment rate simply to remain at its current level, employment would have to grow at the same rate as the labor force, or 2.7 percent a year; moreover, to reduce the unemployment rate by 10 percentage points in ten years, employment would have to grow by over 4 percent a year (see Box 2). It is important to note that, even in the latter case, the number of unemployed persons would be reduced by less than 5 percent over the whole period.

59. The daunting challenge presented by this scenario is well appreciated by the government. Beginning with the 1995 Growth, Employment, and Redistribution (GEAR) strategy, and continuing with the Presidential Commission to Investigate Labor Market Policy in 1996 and the Presidential Jobs Summit of 1998, the government has sought to develop and implement a set of consistent policies aimed at significantly increasing the rate of employment growth in the economy.23

60. Against this background, the government has embarked on a review of the recently passed labor laws to ascertain whether they contained any significant unintended disincentives for employment creation. To date, this review has yielded one small modification in the way the BCEA is applied to small businesses (see below). The question arises, however, as to whether any other areas of labor legislation could potentially inhibit employment growth.

Box 2.Labor Market Projection: Some Simple Illustration Scenarios 1/

POPULATION (In thousands)2/
1997200220072012
TOTAL41,44445,98250,76856,052
0-44,7805,3035,8556,465
5-94,6664,7805,3035,855
10-144,7084,6664,7805,303
15-194,3204,7084,6664,780
20-244,0674,3204,7084,666
25-293,5874,0674,3204,708
30-343,1753,5874,0674,320
35-392,7543,1753,5874,067
40-442,2552,7543,1753,587
45-491,7762,2552,7543,175
50-541.3451,7782,2552,754
55-591,1121,3451,7782,255
60-649151,1121,3451,776
65+1,9732.1332,1762,340
Average annual growth rate2.12.02.0
Working Age Population (15 - 64 year olds)25 30829 10132 65436 089
change3 7933 5533 435
percent change15.012.210.5
A. AVERAGE PARTICIPATION RATE CONSTANTB. ECONOMICALLY ACTIVE POPULATIONS=25-54 YEAR OLDS
1997200220072012199720Q220072012
Average participation rate 3/55.3655.3655.3655.36Average participation rate 3/58.860.561.762.7
Economically active population13,90216,11018,07719,979Economically active population14,89317,61520.15822,611
change2,2081,9671,902change2,7222,5422,453
percent change15.912.210.5percent change18.314.412.2
Scenario 1: Unemployment rate constantScenario 1; Unemployment rate constant
Employed8,70010,08211,31312,503Employed8,70010,29011,77513,209
change1,3821,2311,190change1,5901,4851,433
percent change15.912.210.5percent change18.314.412.2
average annual percent change w.r.t, base year 5/-1.53.02.72.4average annual percent change w.r.t, base year 5/-1.53.43.12.8
As a percent of working age population34.37734.634.634.6As a percent of working age population34.435.436.136.6
Dependency ratio 4/20.99221.922.322.3Dependency ratio 4/21.022.423.223.6
Unemployed5,2026,0285,7647,476Unemployed6,1937,3258,3829,403
change826736712change1.1321,0571,020
Unemployment rate37.437.437.437.4Unemployment rate41.641.641.641.6
Scenario 2: Unemployment rate declines by 0.5 percentage points a yearScenario 2: Unemployment rate declines by 0.5 percentage paints a year
Employed8,70010,48512,21714,002Employed8,70010,58412,44714,339
change1,7851,7321,785change1,8841,8641,892
percent change20.516,514.6percent change21.717.615.2
average annual percent change w.r.t. base year 5/-1.53.83.53.2average annual percent change w.r.t. base year 5/-1.54.03.63.4
As a percent of working age population34.436.037.438.8As a percent of working age population34.436.438.139.7
Dependency ratio 4/21.022.824.125.0Dependency ratio 4/21.023.024.525.6
Unemployed5,2025,6255,8605,977Unemployed6,1937.0317,7108,272
change423235117change838679562
Unemployment rate37.434.932.429.9Unemployment rate41.639.938.336.6
Scenario 3: Unemployment rate declines by 1 percentage point a yearScenario 3: Unemployment rate declines by 1 percentage point a year
Employed8,70010,88813,12115,500Employed8,70010,87713,11915,470
change2,1882,2332,379change2,1772,2422,350
percent change25.120.518.1percent change25,020.617.9
average annual percent change w.r.t. base year 5/-1.54.64.23.9average annual percent change w.r.t. base year 5/-1.54.64.23.9
As a percent of working age population34.437.440.242.9As a percent of working age population34,437.440.242.9
Dependency ratio 4/21.023.725.827.7Dependency ratio 4/21.023.725.827.6
Unemployed5,2025,2234,9574,479Unemployed6,1936,7367,0387,141
change21-266-478change545301103
Unemployment rate37.432.427.422.4Unemployment rate41.638.334.931.6

Labor market projections are based on two different assumptions about the growth of the economically active population (or labor force). In the first set of projections (A) it is assumed that the participation rate remains constant, which, together with the working age population, determine the labor force. In the second set of projections (B) it is assumed that the economically active population is equivalent to the persons in the 25-54 age group. The reason for using these two different assumptions was to examine whether the low participation rate witnessed for South Africa was due to the high proportion of young people, who although in the working age population, were out of the labor force because they were still in school. However, as the results show, the projections are robust to the different assumptions.

The age distribution of the population is estimated by (i) calculating total population on the basis of assumed aggregate population growth rates (equal to Statistics South Africa’s estimates of actual population growth over the past five years); estimating the first age cohort (0 - 4) by applying the aggregate population growth rate to the 0 - 4 cohort of the previous period; assuming a natural aging of each cohort except for the 65+ group; and calculating the 65+ age cohort as a residual. Note that the working age population, which comprises of the 15- 64 year olds, it is independent of the assumption of aggregate population growth under this methodology.

Economically active population divided by the working age population.

Employed persons divided by the population.

1997 number represents the 1994-1998 average annual growth rate of employment in the formal sector.

Labor market projections are based on two different assumptions about the growth of the economically active population (or labor force). In the first set of projections (A) it is assumed that the participation rate remains constant, which, together with the working age population, determine the labor force. In the second set of projections (B) it is assumed that the economically active population is equivalent to the persons in the 25-54 age group. The reason for using these two different assumptions was to examine whether the low participation rate witnessed for South Africa was due to the high proportion of young people, who although in the working age population, were out of the labor force because they were still in school. However, as the results show, the projections are robust to the different assumptions.

The age distribution of the population is estimated by (i) calculating total population on the basis of assumed aggregate population growth rates (equal to Statistics South Africa’s estimates of actual population growth over the past five years); estimating the first age cohort (0 - 4) by applying the aggregate population growth rate to the 0 - 4 cohort of the previous period; assuming a natural aging of each cohort except for the 65+ group; and calculating the 65+ age cohort as a residual. Note that the working age population, which comprises of the 15- 64 year olds, it is independent of the assumption of aggregate population growth under this methodology.

Economically active population divided by the working age population.

Employed persons divided by the population.

1997 number represents the 1994-1998 average annual growth rate of employment in the formal sector.

C. Labor Legislation, 1995-98

The Labor Relations Act

61. The LRA (1995) does not seem to differ significantly in its core principles from its predecessor.24 It promotes and facilitates collective bargaining; regulates the right to strike and the recourse to lockout; outlines procedures for the resolution of labor disputes through statutory conciliation, mediation, and arbitration (for which purpose the Commission for Conciliation, Mediation and Arbitration has been established); defines what are fair/unfair dismissals and establishes the procedures that must be followed to address this issue, as well as the rights and responsibilities of the parties involved; and bestows on the Labor Court ultimate jurisdiction over labor-related disputes.

62. The LRA guarantees the right of all workers to participate in the formation of a trade union and/or to become a member of a trade union. At the workplace level, the LRA allows for a representative trade union (i.e., a union or unions representing the majority of workers in a place of employment) to conclude collective agreements including agency shop agreements and closed-shop agreements.25

63. A key objective of the LRA is to promote collective bargaining at the sectoral level. In this regard, the bargaining council may be viewed in some sense as the cornerstone institution of the labor market.26 Bargaining councils are forums made up of trade unions and employer organizations covering a particular economic sector and geographical area, with power shared equally between the two sides. The powers and functions of the bargaining council, which are also outlined in the LRA, include concluding and enforcing collective bargaining agreements on wages and working conditions; determining those matters that may not be grounds for a strike or lockout; performing dispute resolution functions; promoting and establishing training schemes; and establishing and administering pension, provident, medical aid, unemployment, and training funds. The constitution of every bargaining council must provide for the representation of small and medium enterprises and contain detailed procedures and rules on dispute resolution procedures and on exemptions from collective bargaining agreements. The LRA outlines the requirements for the registration of these councils.

64. For the most part, agreements concluded by the bargaining council on a given issue (e.g., working conditions or strike procedures) supercede any regulation on that issue contained in the LRA or the BCEA. This reflected a desire on the part of the lawmakers to make workplace rules and conditions reflect, as far as possible, voluntary agreements between employers and workers, as opposed to government-mandated regulations. However, it may be that the very “market-based” labor relations desired by the LRA have been compromised somewhat by virtue of the law’s granting too much power to the bargaining councils, especially in determining workplace practices and conditions of those workers and employers not represented in the bargaining councils.

65. The LRA specifies that bargaining council agreements must be extended to nonparticipants in the same registered scope of the council (the ergo omnes principle) by the Minister of Labor, upon request of the bargaining council, if all of the following conditions are satisfied: (i) one or more trade unions in the bargaining council representing the majority of employees covered by the council vote in favor of extension; (ii) one or more employer organizations in the bargaining council employing a majority of employees covered by the council vote for the extension; (iii) the employees covered by the bargaining council account for the majority of the workers in the registered scope of the council; and (iv) the bargaining council agreement establishes an independent body to grant exemptions to nonparticipants and provides the criteria upon which such exemptions would be granted (which must be fair and promote the primary objectives of the LRA). The Minister of Labor may also extend the bargaining council agreement, even if the above conditions are not met, provided that the parties to the council are sufficiently representative within the scope of the council and the minister believes that failure to extend the agreement may undermine collective bargaining at the sectoral level.27

66. Regarding employment security rules, the LRA does not prescribe formal restrictions on temporary employment or maximum probationary periods. The LRA defines an automatically unfair dismissal as any dismissal for (i) participating in or supporting a strike or protest action that complies with the provisions of the LRA; (ii) a woman’s pregnancy; or (iii) unfair discrimination on the basis of race, age, sex, sexual orientation, etc.28 Such dismissals, if proved by the plaintiff in the Labor Court, could leave the employer liable for damages of up to 24 months of the employee’s salary. For other cases, such as dismissals related to the employee’s conduct or capacity, or the employer’s economic operations (in which the employer is unable to prove that the dismissal was justified), damages can be as high as 12 months of the employee’s salary.

67. For dismissals on account of the employer’s operational restructuring, the employer is required to consult with the trade union representing the affected workers, and an attempt must be made to reach an agreement on measures to avoid or minimize the number of dismissals, the criteria for selecting those to be dismissed, and severance pay. In the absence of an agreement, the LRA does not require any arbitration or further procedures on the matter, except to require that the employer provide written statements to the trade union on the various reasons for disagreeing with its proposals, and to require that the employer’s criteria for selecting those to be dismissed are fair and objective. The LRA does not prescribe any rules on minimum severance payments; however, this is taken up in the BCEA (see below).

68. Employees have the right to strike and employers the right to lockout if the issue in dispute has first been referred to a bargaining council or to the Commission for Conciliation, Mediation and Arbitration, and if a certificate stating that the dispute remains unresolved has been issued. However, neither party can strike or lockout if a collective agreement prohibits it in respect of the issue in dispute; if an agreement requires the issue in dispute to be referred to arbitration; or if the parties are engaged in an essential or maintenance service.

The Basic Conditions of Employment Act

69. As its name suggests, the Basic Conditions of Employment Act (1997) prescribes basic minimum protection for workers. Compared with its predecessor, it provides more generous worker benefits. It establishes a work week of 45 hours (with a stated goal of eventually reducing it to 40 hours), compared with 46 hours previously; a maximum of 10 hours overtime per week, remunerated at1½times the regular wage (compared with 1⅓ times previously); and double pay for work on Sundays and public holidays. Workers in employment for a full year are entitled to 21 days annual leave at full remuneration (compared with 14 days previously); two weeks’ (fully paid) sick leave29; four consecutive months maternity leave; and three days (fully paid) family responsibility leave (nonexistent in previous law).

70. Regarding the termination of an employment contract, an employer or employee must give one week’s notice if the employee has been employed for less than four weeks (compared with one day’s notice under the previous law), two weeks’ notice if employed for between four weeks and a year (compared with one or two weeks’ notice previously); and four weeks’ notice if employed for more than a year.31 Severance pay, in the event of dismissal because of economic restructuring, must be at least one week’s salary for each year of continuous employment.

71. As mentioned above, a collective agreement concluded by a bargaining council may alter, replace, or exclude any provision in the BCEA except for those governing the daily and weekly hours of work, maternity and sick leave, and child employment provisions.32 The Minister of Labor also has this power, which he may exercise by way of a ministerial determination, either at his own discretion or following a petition by an employer or employer organization, and after receiving advice from the Employment Conditions Commission.33 The Minister of Labor issued such a determination in November 1999 regarding the applicability of certain aspects of the BCEA to small businesses. For businesses with less than ten employees, the maximum number of overtime hours was raised to 15 hours per week; the minimum overtime remuneration rate was lowered to “time and a third”; and their eligibility for three days of family responsibility leave was rescinded. These changes, however, did not apply to small businesses that were already covered by a bargaining council agreement or any other ministerial sectoral determination.

The Employment Equity Act

72. The purpose of the EEA is to promote equal opportunity and fair treatment in employment through the elimination of unfair discrimination, and to implement affirmative action measures to redress the disadvantages experienced by designated groups (blacks, women, and people with disabilities) in order to ensure their equitable representation in all occupational categories. The law prohibits unfair discrimination, with the burden of proof on the employer to show that it is fair,34 it also prohibits ad hoc medical HIV/AIDS or psychological testing.

73 The law requires designated employers (those with more than 50 employees or with an annual turnover above a given threshold)35 to prepare and implement employment equity plans containing affirmative action measures, after consultation with employees and/or their trade unions. These plans, which are to be implemented over one to five years, must specify numerical goals and a timetable to achieve equitable representation of suitably qualified persons36 from designated groups within each occupational category. The employer is required to submit a report to the Director-General of the Ministry of Labor within 12 months of the passing of the act, outlining the initial development and consultation regarding the employment equity plan; thereafter, an annual (for employers employing more than 150 persons) or biannual update on the progress in the implementation of the plan is required.

The Skills Development Act

74. The SDA (1998), which replaced the Manpower Training Act (MTA), seeks to encourage employers to provide opportunities for employees to acquire new skills, employ job seekers who find it difficult to gain employment, enable new entrants to the labor market to gain work experience, and improve the employment prospects, through training, of persons previously disadvantaged by unfair discrimination. Financing for the SDA will come from the Skills Development Levy, a payroll tax of 0.5 percent that will begin in April 2000 and increase to 1 percent in April 2001.

75. The SDA envisages the creation of a National Skills Authority (analogous to the National Training Authority in the MTA) to advise the Minister of Labor on the design of a national skills development policy; and the creation of sectoral training boards, composed of representatives of organized labor and business, and government departments.37 Bargaining councils having jurisdiction in the particular sector may also participate, if applicable. The training boards are responsible for developing and implementing a skills plan consistent with the national skills development strategy by establishing and promoting apprenticeships or learnerships, approving workplace skills plans, allocating training grants to employers, education/training providers, and workers, and collecting and disbursing the skills development levies in its sector.

76. Learnerships are to consist of a structured learning component, as well as practical work experience of a specified nature and duration, leading to a qualification recognized by the South African Qualifications Authority. If a learnership agreement is concluded between a learner and employer with whom the learner is already employed, the learner’s contract of employment is not affected. If the learner was not in the employment of the employer with whom he signed a learnership agreement, the learner and employer must enter a contract of employment. This contract would be subject to any terms and conditions of a ministerial determination, as outlined in the BCEA (see above), which would seek to ensure that any proposed condition of employment would not have a negative effect on the employment of learners or the achievement of the purposes of the SDA. The Minister of Labor has yet to make a determination regarding learnerships, including on the rate of remuneration for learners.

77. It is not yet clear what the scope of the learnership program will be, although the objectives of the law suggest that it will broader than its predecessor. Under the MTA, the apprenticeship system was largely geared to training artisans and attracting persons who already could be classified as having some skills. Similarly, although these apprentices may have been paid at a rate equal to (or some fraction of) the lowest artisan wage, this wage was certainly higher than the wage of the unskilled workers.

78. Twenty percent of the Skills Development Levy paid by an employer will automatically go to the Skills Development Fund. Eighty percent of the levy will be transferred to the SETA that has jurisdiction over the employer, or, if a SETA does not exist, the money will go to the Skills Development Fund. The money in the fund may be used only for projects identified in the national skills development strategy as national priorities or as determined by the Director-General of the Ministry of Labor.

D. The Recommendations of the OECD Jobs Study

79. The main conclusion of the OECD Jobs Study in 1994 was that the root cause of the steadily rising unemployment rates in much of the OECD since the mid-1970s was the failure of member countries to adapt quickly and creatively to the transformation of the world economy that was being driven by the forces of globalization and rapid technological change. The Jobs Study made a number of detailed recommendations constituting an integrated policy mix of labor market reforms, macroeconomic policies designed to stimulate noninflationary growth, and policies to foster technical progress, increased competition, and entrepreneurship.38

80. The strategy’s emphasis on a balanced mix of policies was deliberate and reflected an appreciation for the need to tackle the unemployment problem through actions on many fronts.39 Nevertheless, the policy recommendations recognized the key role played by the operation of the labor market and labor market institutions in causing the high and persistent rates of unemployment. The recommendations that directly affect the labor market are summed up in the following six points; (i) make wage and labor costs more flexible by removing restrictions that prevent wages from reflecting local conditions and individual skill levels, in particular of younger workers; (ii) reform employment security provisions that inhibit the expansion of employment in the private sector; (iii) increase the flexibility of working time voluntarily sought by workers and employers; (iv) reform unemployment and related benefit systems—and their interaction with the tax system—such that society’s fundamental equity goals are achieved in ways that impinge far less on the efficient functioning of labor markets; (v) strengthen the emphasis on active labor market policies and reinforce their effectiveness; and (vi) improve labor force skills and competencies through wide-ranging changes in education and training systems.

81. In the following section, South African labor laws are analyzed in terms of the above basic policy guidelines, with no presumption made that deviations from them constitute the main factors underlying South Africa’s unemployment problem. Rather, the purpose is to show how South Africa’s labor laws stand up in the light of policy recommendations made to a group of countries also attempting to overcome the problem of high and persistent unemployment and which, in some cases, possess similar legal and institutional frameworks underlying their labor markets.

E. South African Labor Laws and the OECD Jobs Study Recommendations

Increasing Flexibility in Wages and Labor Costs

82. In this area, the Jobs Study emphasized the importance of aggregate wage flexibility (i.e., ensuring that real wages were sensitive to changes in the unemployment rate) and relative wage flexibility (i.e., ensuring that wages reflected underlying productivity). The study noted the negative effects on employment in certain countries of minimum wage legislation, wage indexation policies, and centralized collective bargaining agreements, to the extent that they inhibited the market determination of wages.

83. As a means of supporting faster employment growth, the Jobs Study advocated a relaxation of policies that impeded the emergence of market-determined wage differentials. Even though it acknowledged that its recommended policies could create low incomes that might put households below the poverty line, the Jobs Study argued that, in the context of the serious unemployment problem, more weight could be given to establishing market-clearing wages while simultaneously pursuing equity objectives through other instruments. Moreover, the greatest concerns about wage inequality and in-work poverty generally related to adults with family responsibilities. Young persons and other first-time job seekers could gain valuable work experience and skills in low-wage jobs and then eventually move up the job ladder; low-wage jobs, it was suggested, were bad only if people became trapped in them.

84. In South Africa, there is no legislated minimum wage, although, as mentioned above, the Minister of Labor is authorized to make determinations on conditions of work (including minimum wages) in cases where workers are not covered by collective bargaining agreements. The wage rates prescribed in these determinations generally tend to be significantly lower than the minimum wage rates for comparable occupations negotiated in bargaining council agreements.40 There is no official wage indexation process in South Africa. Nevertheless, real wages in South Africa have shown a remarkable insensitivity to labor market conditions, as a result of which adjustment has taken place almost entirely on the employment side (Figure 6).

Figure 6.South Africa: Wages, Employment and Economic Activity, 1988-98

Sources: South Africa Reserve Bank; Andrew Levy and Associates; and IMF staff estimates

85. Moreover, with its promotion of sectoral collective bargaining, the LRA seems to go against the Jobs Study recommendation that wages be more reflective of productivity differences across firms and across workers; this tendency is even exacerbated by the automatic extension of collective bargaining contracts to nonparticipants. While there is nothing in the LRA that precludes more firm-level bargaining, the law serves to reinforce the system of sector-level agreements that has been slow to voluntarily move to a more flexible wage formation system. Although there are a few examples where productivity frameworks have been incorporated into wage agreements,41 bargaining council agreements generally prohibit any firm-level modification of the agreement; firms must formally apply to the council for approval to be exempted from a given provision in an agreement.42 One measure that could possibly improve relative wage flexibility would be to grant the Minister of Labor more discretion in not extending bargaining council agreements to nonparticipants (as was the case in the old LRA), a proposal that was put forward in the GEAR in 1995 (and made by the Jobs Study to OECD member countries), but which has not yet been implemented.43

86. Centralized collective bargaining, with the extension of agreements to nonparticipants, is not uncommon in either the industrialized world, where it occurs in France and Germany, or in the emerging market economies, where it exists in Argentina and Brazil. However, it is noteworthy that, during the 1990s, the trend has been unmistakably in the direction of introducing more flexibility in the wage formation process: in Norway, the government is encouraging social partners to pay more attention to local conditions when negotiating wage bargains; in Australia, the centralized wage system has been largely replaced by firm-level wage bargaining based on productivity improvements; and in the Netherlands and Spain, new legislation permits firms and employees to agree to exclude themselves from sectorally negotiated minimum wages.

Reforming Employment Security Provisions

87. In recommending reforms in this area, the Jobs Study was particularly concerned that overly strict employment protection legislation in certain countries was reducing the incentive of firms to hire workers because the laws made it too difficult or expensive for firms to terminate workers who were no longer required. However, it acknowledged that a delicate balance needed to be struck because employment protection legislation could provide important benefits; specifically, longer-term contracts could encourage employers and workers to invest in on-the-job training that would otherwise likely be hindered by high labor turnover.

88. As mentioned above, the LRA defines what constitutes just or unjust firing; the damages for which the employer might be liable; and the procedures required to terminate workers on account of economic restructuring. While the rate of severance pay is substantially lower than in other countries, a negative aspect of South African labor laws from the standpoint of overly strict employment security is the absence of a cap on severance pay, which tends to raise the prospective termination costs and may create a disincentive for keeping long-term employees.

89. South Africa’s employment protection legislation is not overly strict when compared with other countries (Table 4). In fact, a major concern on the part of the union movement regarding the LRA is that it requires only that firms consult unions in the case of layoffs, instead of requiring that the size and terms of any layoff be the product of employer-union negotiation. In contrast, in France, which has been cited by the OECD for its onerous employment protection legislation, dismissals must still be approved by the Ministry of Labor; in Indonesia, also, a long process is required for collective dismissals.

Table 4.South Africa: Layoffs and Severance Procedures in Selected Countries
NotificationSeverance PaySpecial Procedure for Collective

Dismissals?
x=monthly salary; N=no. years employed
Firing with just causeFiring without

just cause
Dismissal due to

economic

conditions
Argentina1 month0l*x*N.

Minimum: 2*x.

Maximum: limit on

N.
½12;*x*N.

Minimum: 2*x

Maximum: limit on N.
Brazil1 month. During this month employee entitled to 2hours per day (paid) to look for workAccumulated balance plus accrued interest in FGTS (Fundo de Garantia por Tempo de Servicio). At time of hiring, employer opens a bank account for employee into which is placed 8 percent of the employee’s wage.Includes dismissals for economic reasons. In addition to FGTS balance, employer pays a “fine” equal to 40 percent of FGTS balance.
Chile1 month01.2*x*N.

Maximum: N=11.

For voluntary

Resignation,

employer pays

½12;*x*N, if N>=7.
l*x*N
Mexico0- 1 month3*xIncludes dismissals for economic reasons. No

maximum limit

2/3*x*N
Negotiation with union. In case of no agreement, approval by Conciliation and Arbitration Board required
France1 - 8 weeks00.1*x*N0.1* x*N“Social plan” required, which must be approved by labor Market authorities
Spain1 month1.1*x*N.

Maximum: 42*x
2/3*x*N

Maximum: 12*x
In 1994, prior administrative authorization for dismissals for economic reasons was abolished
Indonesia0x*N.

maximum: 4*x.

Workers with more than five years tenure are entitled to an additional bonus which can reach 5 month’s salary for workers with more than 25 years tenure.
Negotiation between employer and worker representatives required. Mediation by Ministry of Manpower Services if no agreement is reached, and an extensive appeals process is provided for.
Korea0-1 month0l*x*Nl*x*N

(In 1998, legal permission granted for dismissals due to economic conditions)
Duty to inform and consult with union… Must notify Minister of Labor
Australia1 - 4 weeks000Employers required to consult with unions. Notification of
New ZealandNo specific period required by law; case law requires reasonable notice000public authorities required Duty to inform and consult with union only if required by contract. No notification of public authorities required
South Africa1 - 4 weeks00¼14;*x*NEmployers required to consult with unions
Sources: Cox Edwards(1993); Cox Edwards(1996); Gill and Neri(1998); Guash(1999); Mizala(1998); and of OCED Economic Surveys for various countries.
Sources: Cox Edwards(1993); Cox Edwards(1996); Gill and Neri(1998); Guash(1999); Mizala(1998); and of OCED Economic Surveys for various countries.

Flexible Working Time Arrangements

90. The Jobs Study recommended that introducing more flexible working-time arrangements in some countries could enhance job creation and employment prospects. It noted that the labor requirement of firms varied over time and across sectors, both as a function of seasonal and cyclical fluctuations in demand and differences in production processes. As such, a single set of rules for all firms or sectors constrained flexibility for both employers and workers, leading to reduced output and employment.

91. As described above, the BCEA contains regulations on conditions of work, including maximum hours of daily, weekly, and overtime work, holiday and Sunday work, overtime compensation, and leave entitlements. However, in certain ways, the system is flexible, presenting no formal constraints on part-time employment, allowing employers and employees to average hours over a four-month period or substitute additional time off for overtime, and even overriding many of these regulations in collective bargaining agreements. While one could argue that South Africa is not out of step with recent trends in the rest of the world, they are nevertheless behind Europe, which has been more aggressive in adopting flexible working-time arrangements. For example, in France, rules on maximum weekly hours have been relaxed by a reform that allows firms and their employees to annualize working hours (i.e., average them over a 12-month period), thereby reducing overtime costs. In Spain, the 40-hour work week was annualized, minimum overtime rates abolished, and rules allowing firms to offer additional time off in lieu of overtime introduced. Also, in a number of countries, including France, Spain, Argentina, and Brazil, recent reforms have been introduced to facilitate part-time and/or temporary work.

92. Although not specifically related to working-time flexibility, there have been some concerns expressed by the business sector that the EEA could hinder flexibility in the workplace if it were to constrain the ability of employers to organize their workforce as they desired.44 The EEA declares that an employer is not required to implement any policy that would establish an absolute barrier to the prospective or continued employment or advancement of persons not from designated groups. It thus remains to be seen to what extent operational flexibility will be compromised as a result of having to comply with the affirmative action requirements of the law.

Reform of Tax and Benefit Systems

93. Tax and benefit systems not only have an impact on the demand for labor by affecting the nonwage costs of employment and, hence, the demand for labor, but they also have an important impact on the supply of labor by affecting the incentive to take a job or to improve one’s skills. The Jobs Study noted that, in a number of countries, payroll taxes, including social security contribution rates, were contributing to low employment levels.45 It advocated that governments shift part of the base of financing social programs away from payroll taxes to other sources of revenue, especially in those countries where payroll taxes were high.

94. Under the Skills Development Levy Act, a payroll tax of 0.5 percent (later rising to 1 percent) will be levied on firms in order to finance worker training programs. While the importance of skills development is unquestioned, alternative sources of funding might be preferable, since they would avoid the potential disincentives to employment from the payroll tax. Nevertheless, compared with other countries, South Africa’s payroll taxes axe small, totaling less than 3 percent (including the Skills Development Levy).46 In contrast, Argentina, Brazil, France, Germany, and Spain all have payroll taxes greater than 30 percent, with the median for OECD countries higher than 20 percent (Table 5). However, the trend of recent reforms is to lighten the payroll tax burden, as European countries have either reduced employer social security contributions across the board, as in Portugal, or for the long-term unemployed (e.g., in Belgium, France, and Spain), or for youth and low-wage workers (e.g., in Belgium, France, and Ireland).

Table 5.South Africa: Payroll tax and Unemployment Insurance Benefits in Selected Countries
Payroll Tax RateUnemployment Insurance System
ArgentinaOld Age, Disability, Death:

27 percent (employee: 11, employer: 16)

Sickness, Maternity:

9 percent (employee: 3, employer: 6)

Unemployment Insurance:

2.5 percent (employee: 1, employer 1.5)

Family Allowance:

6 percent (employer: 6) Total: 43.5 percent
Duration of benefit

(B) based on months of contributions

(C) to national employment fund, financed by 1.5% surcharge on firm’s

wage bill and 3% surcharge on payroll of temporary employment agencies.

Condition for eligibility: Termination without just cause.
ContributionCompensationReplacement rate
12 < C < 24

24 < C < 36

C>=36
B =4

B =8

B=12
(percent of net salary):

1st 4months: 50

2rd 4 months: 42.5

3rd 4 months: 35
BrazilOld Age, Disability, Death:

28-31 percent (employee: 8-11, employer 20)

Other:

2-6 percent (employer: 2-6 percent) Total: 30-37 percent
Conditions for eligibility: Means tested. Termination without just cause;

employment for past six months, or legally self-employed for 15 months;

no pension or income sufficient for family subsistence.

Benefit period:

up to 5 months-Replacement rate: 50 percent of avg. salary of past 3 months,

but not lower than minimum wage, not greater than 3 times min. wage.
ChileOld Age, Disability, Death:

13 - 20.7 percent (employee: 13 - 20.7)

Sickness and Maternity:

7 percent (employee: 7 percent)

Other:

1 percent (employer: I percent) Total: 21-28.7 percent
Condition for eligibility:

registered for employment. Benefit period: 1 year

Benefit:

17,338 pesos (US$37) per month for first 90 days, falling to 8,700 pesos (US$19)

per month for last 6 months. Subsidy not widely utilized because of its small amount.
MexicoOld Age, Disability, Death:

14.9 - 20.7 percent (employee: 1.8-2.1,

employer: 12.8-18.9)

Sickness and Maternity:

11.9 percent (employee: 3.1, employer: 8.8)

Other:

1.3 percent (employer: 1.3) Total: 28.1 -33.9 percent
None
FranceOld Age, Disability, Death:

14.8 percent (employee; 6.6, employer; 8.2)

Sickness and Maternity:

19.6 percent (employee; 6.8, employer: 12.8)

Unemployment Insurance:

6.2 percent (employee: 2.2, employer 4.0)

Other:

8.7 percent (employer: 8.7)

Total: 49.3 percent
Benefit: Replacement rate of 57.4 percent for a period

varying according to age and length of covered

employment. An additional period of reduced benefits

may follow, which also varies according to age and

work history. The benefit is reduced every 4 months.
SpainOld Age, Disability, Death:

28.3 percent (employee; 4,7, employer: 23.6)

Unemployment Insurance;

7.8 percent (employee: 1.6, employer: 6.2)

Other:

2.0 percent (employer: 2)

Total: 38.1 percent
Eligibility condition: 12 months of contributions during

past 6 years. Registered for employment.

Benefit: 70 percent replacement Tate for first 6 months,

60 percent thereafter.

Benefit period: 2-year maximum (with 6 years of

contributions)
IndonesiaOld Age, Disability, Death:

6 percent (employee: 2, employer 4)

Sickness and Maternity

3 percent (employer: 3 percent)

Other:

0.5 percent (employer: 0.5)

Total: 9.5 percent
KoreaOld Age, Disability, Death:

9 percent (employee: 4.5, employer: 4.5)

Sickness and Maternity:

2-8 percent (employee: 1 - 4, employer: 1 -4)

Other:

1.7 percent (employer: 1.7 percent)

Total: 12.7- 18.7 percent
AustraliaOld Age, Disability, Death:

7 percent (employer: 7)
Conditions for eligibility: Means-tested. Must be

actively seeking employment; cannot refuse reasonable

job offer. Financed from genera] revenues.

Benefit: A$146 - A$I75 (USS93 - US$111) per week.

Payable as long as insured remains qualified.
New ZealandConditions for eligibility: Means tested. Registered for

employment. Financed through general revenues.

Benefit: NZ$147.89 (USS106) per week.
South AfricaUnemployment Insurance:

2 percent (employee: 1, employer: 1)

Other:

0.2 - 0.4 percent
Eligibility condition 13 weeks of contribution during

past 12 months. Registered for employment.

Benefit: Replacement rate of 45 percent

Benefit period: 26 weeks.

95. The Jobs Study specifically cited overly generous unemployment insurance (UI) benefits as important contributors to the high level of unemployment in certain countries. The UI system in South Africa is designed to deal with frictional or short-term unemployment. Eligible workers and their employers each pay one percent of wages to the Unemployment Insurance Fund (UIF), Benefits (set at a replacement rate of 45 percent of the last wage received) accrue only to unemployed contributors.47 One week’s benefit is paid for every 6 weeks worked, up to a maximum of 26 weeks’ benefit. The replacement rate mid the maximum period of benefit payments sure low by international standards (see Table 5), and thus it is unlikely that the receipt of UIF benefits discourages job search on the part of unemployed beneficiaries in South Africa.

Active Labor Market Policies and Education and Training Programs

96. The Jobs Study recommended that more active labor market policies (ALMPs) be implemented than passive income support programs, and noted that, if properly targeted and managed, ALMPs could raise the employability of the unemployed. ALMPs improve access to the labor market and jobs; develop job-related skills; and may lead to wage moderation by strengthening the ability of “outsiders” to compete more effectively for jobs.48

97. The SDA provides for the setting up of public employment services to facilitate job searches, arrange matches between employers and job seekers, and provide counseling for job seekers. The Presidential Jobs Summit has already initiated similar programs. While it is fair to say that the SDA, the Presidential Jobs Summit, and the EEA (through the employment equity plans that are required to prepare previously disadvantaged workers to fill jobs that they hitherto have not occupied) have made important commitments in the area of ALMPs, careful attention will need to be paid in the implementation of such policies. The Jobs Study warned that international experience had shown that these policies often suffered from poor targeting and other design problems, as well as ineffective delivery and monitoring mechanisms.

98. The Jobs Study had similar warnings for worker training programs, even while according a high priority to the upgrading of skills and competencies. Thus, while the SDA clearly demonstrates a serious commitment to worker training, there are as yet insufficient details on the national skills development plan or on operational issues concerning the SET As and learnerships to make any assessment of the likely impact of the worker training initiative.

Prepared by Trevor Alleyne.

From Statistics South Africa’s 1997 October Household Survey. This rate of unemployment is based on the so-called “expanded definition”, which includes “discouraged workers” (i.e., persons not working and not actively seeking employment, but who are willing to work) and therefore differs from the official unemployment rate of 22 percent that excludes those workers. See Statistics South Africa (1999).

See, for example, the policy statement of Business South Africa at the 1998 Presidential Jobs Summit (Business South Africa, 1998).

The 1994 OECD Jobs Study (OECD, 1994) was the outcome of a major research effort by the OECD, launched in 1992, to analyze and provide recommendations to combat the rising unemployment rates experienced in much of the OECD area since the mid-1970s. The paper does not take the view that the recommendations of the OECD Jobs Study, if followed, constitute a blueprint for the solution of the employment problem in South Africa. However, it assumes that following the recommendations should improve the situation

The Presidential Jobs Summit was convened in October 1998 and consisted of representatives from the government, organized business, and organized labor. Its purpose was to seek specific measures and strategies to increase employment growth. The parties agreed on a number of measures, which are now being implemented, including special employment programs focusing on rural infrastructure development, mentoring and microfinancing programs for the promotion of small enterprises and labor-intensive sectors (particularly tourism), and skills development and training initiatives targeted at youth and women.

Neither the new version nor the old version of the LRA applies to the military or the national intelligence and secret services. However, the new version applies to domestic servants in private households and to government workers, including teachers, which the old version did not.

Agency shop agreements are collective agreements that bind nonparties to the agreement and also require nonparties to contribute an agreed agency fee to the representative union. Closed-shop agreements require all workers covered by the agreement to be members of the trade union once they gain employment. These agreements are permitted provided that they are ratified by a two-thirds majority vote involving all workers.

Bargaining councils were called industrial councils under the old Labor Relations Act.

Under the old act (1988 amendments), the minister had complete discretion. He could, if he deemed it appropriate, extend (or not) a collective agreement to all nonparties in the industry. The minister was also required to satisfy himself that the parties to the agreement were “sufficiently” representative of the employers and employees engaged in the industry. The law provided no guidelines in this regard.

Discrimination may be considered fair if it is based on an inherent requirement of the particular job.

For sick leave, the BCEA actually specifies that an employee is entitled to six weeks’ sick leave over a three-year period of continuous employment, or one day for each 26-day period worked.

Under the old BCEA, if an employee had been employed for more than four weeks (no intermediate period of between four weeks and a year was specified), the employee or employer had to give one week’s notice if the employee was paid weekly, or two weeks’ notice if the employee was paid monthly,

Instead of giving notice, an employer may pay the employee the remuneration the employee would have received during the notice period. This applies to all cases described above.

In the case of annual leave, a bargaining council agreement may reduce an employee’s annual leave to not less than two weeks.

The determination can be made in respect of the basic conditions of employment covering any category of employees or employers, for any period determined by the minister. The minister may make a sectoral determination, establishing basic conditions of employment, including minimum rates of remuneration, for employees in a given economic sector in a specific geographical area, provided that the employees are not already covered by a bargaining council agreement.

Discrimination on the basis of inherent requirements of a job or affirmative action measures consistent with the act, including preferential treatment and numerical goals but excluding quotas, are not unfair.

The threshold varies from R 2 million for businesses in the agriculture sector to R 25 million for businesses in the wholesale trade and commerce sectors.

In the act, these are defined as persons possessing any one or combination of formal qualifications, prior learning, relevant experience, or capacity to acquire the ability to do the job within a reasonable time.

These Sector Education and Training Authorities (SETAs) are analagous to the training boards of the MTA.

The Jobs Study was followed up with detailed recommendations for each individual member country.

For example, in 1998, a ministerial determination for the security sector set minimum wages for “general” workers in urban areas at R 3.98 per hour. This compared with rates of R 5.66 per hour for urban general workers under the Gauteng Building Bargaining Council; R 6.07 per hour for the (national) Motor Industry Bargaining Council; and R 8.33 per hour for the National Bargaining Council of the Leather Industry (Footwear Section Collective Agreement).

In a few cases, bargaining council agreements grant automatic exemptions to small enterprises.

Of course, it would be important for the Minister of Labor to use this discretion to aggressively promote more flexibility in the wage formation process.

. See Andrew Levy and Associates (1999). Firms have complained about increased costs attributable to increased training requirements, higher administration costs, and difficulties in finding suitably qualified persons.

If wages are flexible, high nonwage labor costs are unlikely to have major effects on employment in the long run. However, where wages are inflexible, employment will suffer if nonwage labor costs rise.

Apart from the forthcoming Skills Development Levy, there is currently a 1 percent unemployment contribution on employers and workers, and payroll taxes levied by local authorities range from 0.2 percent to 0.4 percent.

Workers still excluded from contributing to or receiving benefits from the UIF include those earning above R 69,420, contract workers, piece workers, casual workers working less than eight hours a week, domestic workers, and central government employees.

The Jobs Study noted that countries that rejected the widening of wage differentials as a means of increasing employment would probably have to rely that much more heavily on ALMPs, education, and training policies.

Other Resources Citing This Publication