Zambezia (2000), XXVII (ii).EMPLOYMENT, UNEMPLOYMENT AND THEEVOLUTION OF LABOUR POLICY IN ZIMBABWEMKHULULINCUBE*Department of Economics, University of ZimbabweAbstractThis article discusses the development of the labour market in Zimbabwe. Inparticular, we analyse the evolution of labour market policy since thecolonial period. Empirical evidence on the operations of the labour marketare reviewed and the consequences of different labour market policies areanalysed.INTRODUCTIONDevelopments in the Zimbabwean economy have been discussed in severalstudies (Mlambo and Kayizzi-Mugerwa, 1991; Durevall and Mlambo, 1995;and Weeks and Mosley, 1996). One of the conclusions of these studieswas that the economy has not been able to create a sufficient number ofjobs. The capacity to create employment started to diminish in the mid-1970s when sanctions began to bite and the war of liberation intensified.The erratic growth of the economy in the 1980s worsened theunemployment problem. Although public sector employment growth hasbeen relatively high, it was not sustainable. In the 1990s the economyunderwent structural adjustment with market forces replacing controlsand regulations. It was assumed that as public sector employmentcontracted, there would be a net increase in employment due to privatesector employment creation. However, this was not so, as the formalsector as a whole lost more than 40,000 workers between 1991 and 1995.The labour market plays an important part in economic development.It is a crucial channel for the transmission of both internal and externalpolicy shocks. In addition, labour market performance has a crucialbearing on economic growth, income distribution and poverty. Forexample, the labour market was used by colonial regimes to pursuepolicies of racial discrimination. In contrast, the post-independencegovernment used the labour market to try and achieve equitable incomedistribution and alleviate poverty.I wish to thank Professor M. F. C. Bourdillon and two anonymous referees for helpfulcomments.161162 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEThe purpose of this article is to discuss the evolution of labour policyin Zimbabwe. Although the country achieved its independence close totwo decades ago, the influence of the colonial policies in the labourmarket is still important in the current labour market situation. Naturally,our discussion will begin with a review of colonial labour policy. Theother principal aspects of the labour market reviewed here are: thestructure and development of both the formal and informal labour markets,wage developments and the evolution of industrial relations.THE COLONIAL LABOUR MARKETThe colonial labour market was characterised by overt, apartheid style,racial discrimination. Labour policy was guided by the IndustrialConciliation Act of 1934. Africans were not defined as employees andtherefore could not organise or be party to any forum that determinedemployment conditions.1 Many Africans were pushed into unproductiveland, and thus forced to earn a living through wage labour. Wage labouralso became more important with the introduction of poll tax and otherrural levies. The system ensured that industry, mining and commercialfarms had a constant and cheap source of labour. African workers werenot allowed entry into skilled jobs nor were they given any meaningfultraining. It was not until the late 1950s, when manufacturing began togrow rapidly, that the benefits of a stable and permanent labour forcebegan to outweigh those of unreliable casual labour. Internal labourmarkets and African labour unions began to emerge. However, theselabour unions were used as forums for control and not for protection ofworkers' rights QLO, 1978, 6). African unions could not freely bargain forwage adjustments and the right to strike was effectively removed undersuccessive legislative amendments of the Industrial Conciliation Act.2 lncontrast, white workers could be members of unions and could bargainfor their employment conditions. The dualistic nature of the labour marketwas even more evident in the distribution of wages. In 1975 white workersearned on average ten times the wage of an average African employeeThere was hardly any unemployment among white labour.At independence in 1980, African labour was relatively unskilled,with the wage distribution skewed against them, and was organised intoweak and fragmented unions. Unemployment among blacks was close to10%. These were the disparities that the government sought to redressInternational Labour Organisation (1978).The government could veto any intended strike action simply through the President-decree. Between 1950 and 1980, there was not one legal strike bv black worker! ŁRhodesia (Herbst, 1990). ' s '"M.NCUBE 163and consequently, labour market policies after 1980 revolved aroundthese issues.A NOTE ON LABOUR FORCE DATAIt should be noted from the outset that in analysing the Zimbabweanlabour market, data is always a problem. In Zimbabwe there are severaldifferent sources of information related to the labour force, employmentand unemployment. However, no one source has a satisfactory coverageover time and across sectors. The most comprehensive survey is theQuarterly Employment Inquiry (QEI), which provides information overtime for the whole of the formal sector. Another important survey is theCensus of Industrial Production (CIP). It is a yearly survey that gathersinformation about the number of people employed in manufacturing,mining, electricity and water supply, and construction. The other surveysare too limited in either time or coverage to be of much use in evaluatingthe changes in employment. In this section we describe some of theshortcomings of the QEI and CIP and how they are likely to affect the datareported.In the QEI, data on employment and wages expenditures is collectedand summaries are published by the Central Statistical Office (CSO)'sQuarterly Digest of Statistics for the economy divided into 13 sectors.The CIP is a yearly survey that gathers information about the number ofpeople employed in manufacturing, mining, electricity and water supply,and construction. In spite of having a wide coverage both in time andspace, QEI and CIP have several weaknesses. One issue of concern is thecoverage of the enterprises in the surveys. First, questionnaires are sentout to all formal companies, that is, companies registered by the Registrarof Companies. However, all do not respond and so in practice only asample is analysed, which is obviuosly not random. Most likely there is abias towards including large companies in the sample. Before theintroduction of the Economic Structural Adjustment Programme (ESAP)there were few bankruptcies, in particular among big companies, so thebias in the sample was probably not a problem. However, during the1990s several big firms have collapsed, and it is generally believed thatthese have not been replaced promptly by new ones. Hence, it is possiblethat the figures reported on employment are too low for the 1990s.Secondly, since only registered companies are surveyed, employment inthe informal sector is not measured. This is not a great problem as longas the relation between employment in the formal and informal sectors ismore or less constant because then the information on growth rates iscorrect. This has not been the case, however. The informal sector hasgrown rapidly and, according to anecdotal evidence, new firms sometimes164 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEprefer to be in the informal sector even when they are fairly large andtechnologically advanced. Hence, this problem is also likely to create atendency for the surveys to underestimate both the number of peopleemployed and the growth rate in employment.During recent years there has been a strong trend towardscasualisation of labour in Zimbabwe. This means that many workers arehired on short-term contracts. In practice this often implies that a workermight work one week, be unemployed for a couple of weeks, then workfor some weeks, and so on in the same company. How this change inemployment patterns affects the collection of data is not clear. It couldlead to an overestimation of the number of people working full time, orthe opposite, if employers fail to keep track of the hours worked bycasual workers.Another weakness of the QEI and CIP surveys is that children underthe age of 15 are by definition excluded from the labour force, and thusdo not enter the statistics. Since child labour is common in developingcountries, and Zimbabwe is no exception, there can be a substantialdifference between actual employment and what is reported for somesectors. This problem is likely to be most severe in agriculture. Availabledata on child labour, obtained in the 1992 Population Census, indicatesthat close to 39 000 children aged between 10 and 15 years were employed,which is about 3% of all children in that age group. If all these wereworking in agriculture, employment would have been about 10% higher inthat sector.Finally, there is the difficulty of finding information on the number ofpeople engaged in domestic services. Up until the mid-1980s this kind ofemployment varied in a way that made sense, but then there has beentoo little change. In 1985 it was 98 000; it rose to 102 000 in 1998, and since1990 the number has been exactly 102 100. Thus, there is no reason tobelieve that 102 100 even closely corresponds to the number of peopleemployed in private domestic services.In reading the subsequent sections it is important that theseshortcomings in data are borne in mind.THE FORMAL LABOUR MARKET AND LABOUR FORCE PARTICIPATIONTable 1 shows the size of the labour force including the unemployedemployed and communal farmers. In 1982 the labour force comprised of2.48 million people, increasing to 4.045 million by 1995. The share of theformal sector labour force to total labour force has declined from 47% jn1982 to 31% in 1995, while the share of the unemployed has trebled (andis almost the size of the formal sector) in the same period.M.NCUBE 165Table 1SIZE AND STRUCTURE OF THE LABOUR FORCETotal Labour-forceEmploymentCommunal FarmersUnemployment1982thousands2 4831 1771038268%10047.441.810.81992thousands3 50212361502764%10035.342.921.81995thousands4 045125016001 195%10030.939.629.5Sources: Government of Zimbabwe (1996; 1988)The distribution of the formal labour force is shown in Table Al inthe appendix. An inspection of this table shows that agriculture andmanufacturing are the major employers of labour. These two sectorscombined contribute more than 40% of total formal employment and 36%of Gross Domestic Product (GDP). The other important contributors areeducation, distribution and hotels, and domestic services. The structureof formal employment seems to have changed over time, however. Theshare of employment in agriculture fell from 35% in 1975 to 24.3% in 1990,only increasing slightly to 26% by 1994. Two potential explanationsunderlie this structural shift: First, the droughts of 1983, 1987, 1990 and1992 might have contributed to the decline in the agricultural share. Itsshare in GDP has also fallen from 17% to 13%. Secondly, minimum wagesmight have forced commercial farmers to become more capital intensive.The manufacturing sector, on the other hand, has continued to increaseits employment level and its GDP share. In 1975 its share in employmentwas 14.5%, reaching a peak of 16.5% in 1990, and slightly declining to 16%in 1994. Retrenchments during ESAP partly explain the decline in 1994.For the same years its GDP share has been 22.8%, 24.7% and 22.5%,respectively.The share of the public sector in total employment has increasedover the years.3 In 1975 it was 9.3%, 19% in 1985, and 23.1% in 1994.4 Thispattern is explained by government policies immediately after 1980. Aspart of its "growth with equity" policy, the new government expanded itsprovision of social services, including health and education. In the labourmarket the new government deliberately assumed the role of "employer-of-last-resort". However, in 1991 under ESAP and specifically as a budgetbalancing measure, policies de-emphasised the leading role of GovernmentPublic sector includes health, education and public administration.The proportion of public sector employment to total formal sector employment which is23.1% is relatively smaller than those of other African countries. For example in Kenya,Tanzania, Uganda and Zambia the proportions are 45.8%, 50.5%, 30.5% and 59,7%respectively.Table 2DISTRIBUTION OF FORMAL SECTOR FEMALE EMPLOYEES (PERCENTAGES)AgricultureMiningManufacturingElectricityConstructionHotelsTransportFinancePublic SectorTotal Females% of formal*198049.50.76.80.10.67.01.82.930.610017.0198235.60.87.60.10.48.11.83.342.310015.7198435.20.78.10.10.46.61.93.04410016.5198635.20.76.30.10.46.51.72.746.410017.7198834.80.76.50.10.66.81.72.646.210017.8199035.30.76.30.20.86.71.72.745.610018.0199236.50.65.80.21.46.71.42.844.710018.0199437.50.67.10.12.46.71.92.840.910019.9Or~OOT!CDoUR POLICYzN<mSource: Calculations based on figures from ILO Statistical Year Books (various issues'); *Proportion of total female employees to the totalformal sector labour-force.M.NCUBE 167in employment creation. The public sector started retrenching part of itslabour force.Table 2 shows the distribution of female employees in the formalsector. Women are largely concentrated in the public sector, followed byagriculture. In 1994, 40.9% and 37.5% of female formal sector employeeswere in agriculture and the public sector, respectively. Electricity andwater, mining, construction, transport and finance are still male-dominatedindustries. Table 2 also shows that female participation in the formallabour market has increased over the years (from 17% in 1980 to 20% in1994). Their share continues to increase in the better-paying public sector,and declining in the low paying agriculture sector.5 This picture seems tohold also in other African countries. In Kenya, for example, 17.8% of theformal sector workers were women in 1983, but their ratio had increasedto 23.1% by 1994. They are also largely concentrated in agriculture andthe public sector (Ikiara and Ndung'u, 1997, 33).Female employment in the formal sector grew at the rate of 2.7%(compared with 1.4% for males) between 1980 and 1994. Explanationsinclude the phenomenal expansion and diffusion of education afterindependence, which saw enrollments for both sexes more than quadruplein the different institutions.6 Secondly, at independence all discriminatorywage and other gender biases in the labour market were abolished.Finally, the introduction of minimum wages in 1980 and paid maternityleave, may have made participation in paid work more attractive thanhousehold work.ECONOMIC SHOCKS AND EMPLOYMENTTable A2 in the appendix shows changes in formal employment from1970-1994. During the Unilateral Declaration of Independence (UDI) period(1970-1979) employment growth averaged 1.7% per annum with publicadministration spearheading it.7 However, if we subdivide this period,interesting patterns emerge. The highest level of employment growth(4.5% per year) was recorded in the period 1970-1974. The manufacturingand construction sub-sectors spearheaded this growth. The two sectorsbenefited from the import-substitution industrialisation strategy, pursuedby the government in the face of sanctions.In the period 1975-79 total employment declined by an average of1.1% per year. Two factors explain this trend. First, there were twoSee Table A3 in the appendix for ratios of wages in different sectors.In the period 1979 to 1994 enrollments increased at primary school by 312%, at secondaryschool by 899%, and in tertiary institutions by 355%.The Rhodesian government unilaterally declared its independence (UDI) from Britain in1965.168 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEnegative external shocks which disrupted the economy; the intensificationof the liberation war and the first oil shock of 1973/4. Second, the dynamicsof the import-substitution strategy had been lost by the mid-1970s as theeasy stages had been exhausted (Davies and Ratts0, 1990, 322). At sub-sectoral level, public administration had the highest growth. Employmentin this sector grew from 50,000 in 1975 to 74,000 in 1979. This wasbecause the UDI government absorbed most of the war-displaced whiteworkers, and recruited additional manpower to service its war machinery.In the post-independence decade (1980-1990) employment grewannually by 1.8%, or 1% during the 1980-1984 phase, and 2.4% in the sixyears that followed. This period was marked by high public sectoremployment growth. Employment in the education sector, for example,increased by an average of 21% per year. This high post-independencepublic sector employment growth is not unique to Zimbabwe. This hashappened elsewhere in Africa (see van der Geest and Wignara, 1996,15;Cohen, 1993; and van Ginneken, 1990). Van der Geest and Wignara, andCohen note that the government of Kenya, for example, acted as "employer-of-last-resort" after independence. The motivation in Kenya was political,and also to strike some ethnic balance. The reforms currently sweepingacross less developed countries have reversed this trend.Other sectors (except agriculture and mining) recorded moderate tohigh employment growth between 1980 and 1990. The decline was severein agriculture as the sector was hit hard by the 1983 drought. 1983 wasalso marked by an overall downturn in economic activity partly due tothe world recession Š GDP growth was negative at 4%.The period 1986-1990 was characterised by modest employmentgrowth (2.4% per year). Although growth was positive but lower than inthe previous period in many sectors, the major employers of labour,agriculture and manufacturing, increased their employment levels morerapidly. A boom-slump economic growth pattern characterises thisperiod.8The ESAP period 1991-1995, saw the wholesale restructuring of theeconomy and the overall removal of "rigidities" in the labour market,such as minimum wages and employment protection regulations. As pa^of the general restructuring process, government intervention in thelabour market ended in 1990. Employers took this opportunity an(jadjusted their labour force. Employment in non-agricultural sectors fellfrom 939,800 to 844,000 in the first two years of the reforms. However, thepublic sector had by 1995, retrenched 20,000 workers. The private formal8 See Mlambo and Kayizzi-Mugerwa (1991) for a detailed discussion of the changes ineconomic activity during the 1980s.M.NCUBE 169sector on the other hand, retrenched 25,510 workers.9 Because ofretrenchments, employment growth during the Economic StructuralAdjustment era was lower (see Table A2). What then explains thebehaviour of formal sector employment in Zimbabwe?Economic theory gives us two main approaches to the study ofemployment determination. First, the neoclassical theory which broadlyviews the labour market as capable of reaching a full employmentequilibrium through a flexible wage. Central to neoclassical theory is thatlabour markets clear, a situation that means the absence of involuntaryunemployment. If there is involuntary unemployment real wages adjustdownwards relative to the rental rate of capital. Unemployment is thereforea symptom of high, inflexible wages. If there is continuous unemployment,it would be institutional, caused by government or trade unionintervention. The main policy prescription from this theory is that labourmarkets should be flexible.In the Keynesian approach employment is determined in the productmarket. Deficient demand in the product market leads to unemployment.High prices depress demand for goods and therefore depress demand forlabour. Improving demand conditions in the product market is theprescription for increasing labour demand.In Zimbabwe employment determination has recently been studiedby Kanyenze (1993,1996), Knight (1996) and Ncube (1998). Kanyenze andKnight estimate a model of the formIn It = a + Eg In qt + fS«r In wt + &t In hi + ztwhere /, q, w represent employment, output and the real product wage,respectively, s is the error term. Kanyenze estimates this equation fornine sectors (for the period 1970-90) namely agriculture, mining, finance,construction, manufacturing, electricity, distribution, transport and otherservices. Output was found to be significant in all but two sectors(agriculture and mining). Real product wages were significant inagriculture, mining and finance. The lagged dependent variable wassignificant in seven of the nine sectors and the sizes of the significantcoefficients indicated a leisurely speed of adjustment to equilibrium.Knight's model is applied to the formal sector as a whole, agriculture andmanufacturing, for the period 1979-90. In the formal sector and agriculturereal product wages were significant with elasticities of -0.176 and -0.325,respectively. Output and lagged employment were significant in the formalsector and manufacturing equations. What we learn from these two studies9 Zimbabwe Congress of Trade Unions (1996).170 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEis that aggregate demand factors are important in employmentdetermination.Ncube (1998) models employment determination slightly differently.He estimates a less restrictive translog labour demand function and usespanel data estimating techniques. The empirical analysis is based on across-section of ten manufacturing industries observed during the period1970-1993. A two-way fixed-effect error-component model specification isused. Labour demand is a function of real producer wages, output, interestrates, the lagged dependent variable and rate of technical change. Bothstatic and dynamic versions of the labour demand function are considered.Results of this study suggest that labour demand responds more to realwages, followed by output, and lastly interest rates. On average there istechnical regress (labour-using) in all industries. Total regress in thesample period was found to be 16%. The speed of adjustment is constantacross industries. The mean speed of adjustment was found to be 39%,also, as in the other studies, indicating a relatively slow speed. This is notsurprising considering that the Zimbabwean labour market was underextensive controls in the 1980s.INCIDENCE OF UNEMPLOYMENTDespite unemployment being one of the worst features of Zimbabwe'seconomic performance, statistical data to establish the trends, structureand dynamics of the problem are sketchy. Reliable time-series data arenon-existent. For some indication of the structure of unemployment inZimbabwe we shall use the 1986 labour-force survey data as well as the1992 census data (see Tables 3 and 4). But since the survey and censusdata collection methods differ, the results are not entirely comparable.The census report of 1982 indicates that unemployment was about 10.8%in 1986 it was 7.2%,10 and in 1992 it was 21.8%.HIt is thus not surprising when economists and policy makers estimateunemployment to be currently at least 30% (Government of ZimbabweThe 1986 figure for unemployment is lower than the 1982 figure not because unemploymentdropped, but the definitions of unemployment differ. The 1982 and 1992 figures arederived from the census reports, and unemployment is defined as persons who, as theirmain activity over the previous year, had done no work or were looking for work or wer.available for work. The 1986 figures are from a labour-force survey, and unemploymentis defined as those not working in the previous week but available for. and wanting wo^Unemployment seems to be slightly higher than that of other countries in the region 11991 for example, unemployment in Kenya (urban Š see Kulumiu. l'.)97); and Zambia.!national Š see Seshamani and Kaunga, 1997) is between 17.8"-. and 23.6'X., and 21 s%respectively.M. NCUBE 1711996, 129). The labour market is also characterised by high levels ofunderemployment, especially in the rural areas and within the urbaninformal sector. Government of Zimbabwe (1996) estimates that 300,000people are underemployed in the rural and informal sectors.Table 3INCIDENCE OF UNEMPLOYMENT BY GENDER, AGE AND RESIDENCE1986 (% OF THE LABOUR-FORCE)MaleFemaleAge15-2425-3435-4445+TotalRural areas*2.75.37.14.11.91.23.5Urban areas13.327.446.513.55.517.217.9Total6.57.914.96.94.21.77.2Total (006)111123138592215234INCIDENCE OF UNEMPLOYMENT BY AGE AND EDUCATION 1986 (% OFTHE LABOUR-FORCE)Age15-2425-3435-4445+TotalNoeducation2.22.62.31.31.6Primary10.68.24.83.66.8Secondaryor above19.37.22.21.311.6Total14.96.93.52.37.2Total(000)138592215234Source: Government of Zimbabwe (1989).Notes: 'Includes large scale commercial farms172 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWETABLE 4INCIDENCE OF UNEMPLOYMENT BY GENDER, RESIDENCE, AND AGE1992 (% OF THE LABOUR- FORCE)*NationalNational excluding Communal FarmersBy AreaUrbanRural**By Age Group15-1920-2425-2930-3435-3940-4445-4950-5455-5960+Male22.119.823.855.239.820.112.410.49.69.18.07.15.3Female21.431.317.149.536.118.310.16.64.53.73.02.82.2Total21.828.523.420.852.338.419.412.09.07.77.26.25.74.3INCIDENCE OF UNEMPLOYMENT BY EDUCATION 1992 (V.', OF THELABOUR-FORCE)***By educationNo education 2.4Primary 24Secondary or diploma 25.4Graduate or post graduate 1.9Total 15.9Source: Government of Zimbabwe (1994) and Knight (1996). Notes: *The unemployed areexpressed as a proportion of the working age population. * "Includes larye scale commercialfarming, resettlement and communal areas. ***The overall unemployment rate for a||educational levels is lower than the national figure because the denominator is personsaged 5 and above.From Tables 3 and 4 we can make the following observ.itions:(i) Overall unemployment is higher in urban than in rural areas.'2 |nIn 1992 unemployment was however hiijher in rural aro.is (or IIMI<'S ih.ui (finales.M. NCUBE 1731986 the proportion of the unemployed to the labour force in ruralareas was 3.5% compared with 17.9% in urban areas. Thecorresponding figures for 1992 were 20.8% and 23.4%, respectively.(ii) In urban areas more females are unemployed than males. However,the differentiation is more difficult in rural areas.(iii) Unemployment decreases with age. The youth (aged 15-24) aremore represented among the unemployed than any other group,partly because this group is composed of people who lack experienceand skills. In addition, most of them are school dropouts. Between1980 and 1990, there were about 173,315, Grade 1 to Form 6 dropoutsper year (Siddiqui and Matare, 1993).(iv) Unemployment is greatest among those with secondary educationor "above" (i.e. in 1986), and the secondary or diploma education (i.e.in 1992) categories. It has also been noted in other studies (Siddiquiand Matare, 1993, 3) that unemployment is higher among secondaryschool leavers than any other group.13 The lowest reported cases ofunemployment are among those with no education, a reflection ofnon-selectivity of this group as far as jobs are concerned. The massiveunemployment of the educated is the consequence of governmentpolicy which focussed on the supply side of the labour market. Primary,secondary and tertiary level enrolments increased by 311%, 898%,and 355%, respectively between 1979 and 1984. Thus school outputoutstripped the jobs created. In 1980 for example, there were 26,000new labour market entrants with primary and secondary education(not continuing) competing for 33,000 non-agricultural job openings.In 1993, there were 225,000 primary and secondary school graduateswho entered the labour market. Instead in that year, non-agriculturalsectors_shed 20,000 jobs (Knight,J996,18).The problem of unemployment in Zimbabwe has been worsened byretrenchments in both the public and private sectors, a result of structuraladjustment. As has been noted above the two sectors retrenched 20,000and 26,000 workers, respectively. Private sector retrenchments cameabout largely as a result of poor economic performance while governmentretrenchments were part of the civil service rationalisation programme.However, the government never envisaged private sector retrenchmentsat the start of ESAP. The government had assumed that reductions incivil service employment would be counteracted by major employmentSecondary (and primary) education expansion has not been accompanied by acorresponding increase in jobs and increases in the tertiary sectors of the educationsystem. Between 1980-1990 secondary enrolment increased from 74,321 to 672,653 whereasenrolment at the University of Zimbabwe increased from 2,240 in 1980 to 9,300 in 1990(Siddiqui and Matare, 1993).174 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEincreases in education, manufacturing, tourism and commercialagriculture. The four sectors were projected to increase their employmentbetween 1990 and 1995 by 12,000, 34,000,13,000 and 20,000, respectively(van Der Geest and Wignaraja, 1996,16). To date little employment growthhas been realised in the formal sector. Many job seekers have beenabsorbed in the informal sector. In the next sub-section we look at theinformal sector alternative.THE INFORMAL SECTOR AND GOVERNMENT POLICYWhen the prospects of employment creation in the formal sector weregood, government attitude towards the informal sector was hostile, withstiff and oppressive regulations. The central government and localauthorities hoped that this sector would one day disappear. The downturnin formal employment growth, high levels of rural-urban migration, andpublic and private sector retrenchments under structural adjustment ledto a change in attitude. Realising the immense employment "sponge"effect of this sector, the public sector relaxed the laws regulating thesector. Because of the changes in the legal environment affecting thesector and retrenchments due to structural adjustment, there has been aphenomenal increase in its size.14 Daniels (1994) estimates the growth ofthe non-formal sector at 14% between 1991 and 1993. As elsewhere inAfrica, it has overtaken the formal sector in employment absorption inrecent years (Ikiara and Ndung'u, 1997, 54, for Kenya).A series of studies have been produced in an attempt to characterisethe size and structure of the informal sector (see for example, Moyo ef.al, 1984; Mead, 1991; Mcpherson, 1991; Peters-Berry, 1993; and Mhone,1995). These studies agree that the informal sector is both large and aviable complement to the formal sector because it is labour intensive,uses simple and inexpensive technology, is less import dependent andcreates employment cheaply compared with the formal sector. Forexample, the cost of creating a single job in the formal sector ranges fromZ$80,000-Z$100,000 whereas in the informal sector it is only about Z$10,OQO(Zimbabwe Congress of Trade Unions, 1996, 74).Mcpherson (1991) estimated that the informal sector had 845,000enterprises each employing an average of less than two people. This isroughly the same size as other informal enterprises in Southern Africa.1514 It is difficult to know how much of the informal sector expansion is due to the adverseemployment effects of ESAP, or direct incentive effects stimulating the sector such asthe relaxation of regulations.15 In Swaziland, Lesotho and South Africa, the average enterprise employs 1.85: 1.60, and2.10 people respectively (McPherson, 1991).M.NCUBE 175There is consensus among the above studies that women constitute thebulk of employees in the informal sector. Mcpherson puts their share at57%. The informal sector is also characterised by underemployment andlow wages. Whereas average annual wages in the formal sector wereabout Z$8,000, in 1990 those in the informal sector averaged only Z$l,000(Zimbabwe Congress of Trade Unions, 1996, 74).DEVELOPMENT AND STRUCTURE OF WAGESIncomes policy: narrowing wage differences and minimum wagesAmong concerns of the government at independence were the skeweddistribution of incomes and high levels of poverty. Many people dependedfor their livelihood on wage income, so the labour market was an importantarea in which government could change the colonial legacy of povertyand income disparities. The government introduced an incomes policy tonarrow income inequalities and also improve the purchasing power ofmany people.Wage gaps were moderated by the introduction of minimum andmaximum wages. The wages of those at the bottom scales adjustedupwards by wider margins while upper scales were adjusted only a little.For example, in 1982, those earning at most Z$l,200 per annum had theirsalaries adjusted upwards by 23.5%, whereas those earning Z$20,000 perannum had a 1% upward adjustment. In 1985 salaries at the top andbottom brackets were adjusted by 2% and 15% respectively.16 Theeffectiveness of such a policy is debatable, however. In the private sector,there was no effective mechanism to detect non-compliance. The policywas however, circumvented through the use of in-kind benefits, fakepromotions or allowances and outright secrecy (Shadur, 1996; Knight,1996; and Herbst, 1990). The system seems to have worked only incentral government. Knight (1996) and the International Monetary Fund(1993) argue that though compressionary income policies were effectivein the public sector, they had disastrous effects in the civil service. Thenegative effect of this policy, according to the IMF, was the apparentdevaluation of skills in this sector. There was inevitably some widening ofthe wage gap between private and public sectors and this could partiallyexplain the migration of manpower from the former to the latter, and thepersistent shortage of skills in the public sector (Kanyenze, 1996, 57).Knight (1996, 15) cites the Public Service Review Commission (1989)document which noted that, "... Wage compressionary policies hadbeen taken too far in the public sector . . . and that scarcity, and highSee Ncube (1997) for salary adjustments in the 1980s.176THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEturnover of professional and technical staff such as accountants, engineers,and financial managers, was due to compressionary policies". Kanyenze,(1996,58) and Ncube (1997,130) also note that falling morale, absenteeismand moonlighting in the public sector were some of the consequences ofthis policy. Over time, the compression of wages might have contributedto widening of wage gaps between the Zimbabwean workers and otherworkers in the region, e.g. Botswana and south Africa (Knight, 1996, 22).For example, median salaries of Zimbabwe chartered accountants were43% of their South African counterparts' median salaries; programmersearned 41%, while employed medical doctors, engineers and universitylecturers earned sightly above half. With such differences in emolumentsthere was considerable brain drain to the south.1'Minimum wages were different for different sectors, as shown inTable 5 below. In real terms minimum wages reached their peak in 1982,thanks to the boom of 1980-1982, and the good rains. After 1982 realminimum wages were allowed to fall as pressure from the donorcommunity and industry to abandon legislated wages was mounting. Theargument from industry and donors was that minimum wages werediscouraging employment creation? When the minimum wages policywas abandoned in 1989, agricultural workers were earning 16% abovetheir 1980 levels, whereas those in mining and industry were 9% belowtheir 1980 levels.Table 5MINIMUM REAL WAGES: (Z$ PER MONTH) (JULY 1980=100)Date oforderJanuary 1980January 1981January 1982September 1983July 1984July 1985July 1986July 1988Source: Knight (1996)AgricultureandDomestic10096.8141.2107.3116.7125.9121.7116.3Note: In 1987 there was a wage freeze.Mining100117.5127.19292.3103.69790.7IndustryandCommerce100117.5127.196.196.2103.69790.7LowIncomeConsumerprice index100103.3118170.9185.7198.5232.8286.717 For example in 1992 alone, 200 Zimbabwean doctors moved to South Africa (Kntoht1996).M. NCUBE 177Classical theory argues that the presence of minimum wages is apossible constraint to labour demand. The textbook model of minimumwages contends that if minimum wages are set above the marketequilibrium level, employment is reduced. The workers likely to be affectedby minimum wages are those at the lower end of the wage spectrum. InZimbabwe the employment reducing effect of minimum wages has beeninvestigated by Hawkins, et. al. (1988). Hawkins et. al. surveyed 168enterprises in Zimbabwe and found that minimum wages had a negativeemployment effect in the agriculture sector. The survey revealed that77% of agricultural employers acknowledged that labour costs deterredthem from employing more labour, compared to only 39% in non-agricultural sectors.We test this hypothesis by estimating a simple econometric model(as opposed to Hawkins et al survey method) where labour demand 0) isa function of output (y), minimum wages (mw) and a time trend (t) torepresent technological progress. Minimum wages are defined as a ratioof minimum wages to average wages in each sector. A panel dataframework is used whereby i indexes sector and t represents time.18 Wewrite this equation asIn lu = do + % In Yit + &w In MWu + &tt + enThis equation is estimated over the period 1980-1988, the time whenminimum wages were in force. The i index represents the followingsectors: agriculture, mining industry and commerce. The estimatedequation is:In La = 2.6340 + 0.0601 Inrit -0.0324 InMKit+0.0421f(1.2310) (4.7670) (3.8230) (1.9840)The results indicate that output, minimum wages and technical change(time trend) were significant factors in explaining the employmentsituation in Zimbabwe. The minimum wage coefficient is negative andstatistically significant, suggesting that minimum wages were a significantconstraint to labour demand when all sectors are combined. This resultlends support to Hawkins et al (1988)'s attitudinal study.Considering that many workers were at the bottom end of the wagespectrum, the other question that naturally follows is: were minimumwages effective in improving the standard of living of many people?There are diverging views about how minimum wages affect the livingWe use a panel data framework to save degrees of freedom.178 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEstandards of people. Advocates of minimum wages hold that theyredistribute resources in a welfare-enhancing way, and can thus reducepoverty, improve productivity, and foster growth. Opponents, on theother hand, contend that minimum wage interventions result in amisallocation of labour and lead to depressed wages in the very sectorswhere most of the poor are found, with the effect of wasting resourcesand reducing the growth rate. This has always been a difficult area toinvestigate empirically. Previous studies have looked at whether minimumwages actually pushed average earnings up, which would have indicatedthat at least incomes improved. In Zimbabwe two studies have investigatedthis issueOne such study is Kanyenze (1996) which investigated whetherminimum wages pushed average wages away from their historic trend.The results from this study suggest that minimum wages actually pushedthe wage structure beyond their historic trend. This "push" was largeduring the 1980-1982 period Š the period when there was a sharp increasein minimum wages. It was more pronounced in sectors with an abundanceof unskilled workers (e.g. agriculture and domestic service) and less so inareas where there were a lot of skilled workers (e.g. finance, insuranceand real estate).The second study is that of Knight (1996). Knight estimated thefollowing earnings model for the period 1980(3) to 1988(4),In e=&o +6/ In mwi+&2 In p,Where e is the average wage in sector /, mw is the minimum wage andp is the low consumer price index. The coefficient of interest £1 wassignificant and positive in mining, agriculture/domestic, and industry/commerce. Respectively, a \% increase in minimum wages raised averagewages by 0.58%, 0.25% and 0.35% Š an indication that minimum wagesdid push average wages up in these sectors.Government intervention in the determination of wages as well asother employment conditions in almost all the sectors was eased in 1989and 1990 in line with reforms in other markets.19 Since then, wage andemployment conditions in the private sector are determined throughcollective bargaining, although it is still subject to indirect governmentinfluence. At the beginning of each year the government advises industryand labour on the performance of the economy and that excessive wagedeals could impact negatively on labour demand and macroeconomicinstability. In collective bargaining, parties are supposed to take note of19 The exception here is the domestic and agriculture workers where the government stillsets minimum wages.M. NCUBE 179this "warning". In the civil service the determination of wages andemployment conditions is still the employer's prerogative.Earnings determinationEarnings determination in Zimbabwe has been analysed by Velenchik(1994). She uses data from the Regional Programme on EnterpriseDevelopment (RPED) survey and estimates an earnings function with thelog of hourly earnings as the dependent variable.20 The results of thissurvey suggest that earnings rise with experience up to 38 years of age.Returns to primary education and secondary are found to be 0.9% and23.9%, respectively. Occupational variables, race and sex variables aresignificant determinants of earnings. The race coefficient suggests thatwhites earned 23,2% more than blacks when control is made for humancapital variables and 11,1% when occupational variables are included.This large difference could be explained by differences in skills or outrightdiscrimination. The gender coefficients indicate that males earn 28%more than women, after controls for human capital and occupationalcharacteristics. Also, this difference could partially be explained by genderdiscrimination.Velenchik (1996) also investigated the effect of firm size on wages.Her estimates point out to the powerful effect of firm size on earnings.The major finding in this context was that there is labour marketsegmentation. There is a labour market for small firms (1-10 employees)and one for big firms (ten plus employees). Small firms are characterisedby low wages, low wage dispersion and low returns to human capitalvariables. The premium on secondary education (compared to primary)is 31% for small firms and 103% for large firms, while the premium onuniversity education is 311% for small firms and 896% for larger firms.Velenchik's results show that small firms are characterised by employeeswith inferior education, and who are less likely to acquire human capital.Larger firms in her study recruit better quality labour and have internallabour markets with proper career development ladders.21INCOME AND WAGE GAPSRural-urban income gapsThe Zimbabwean labour market is characterised by differences in wagesbetween different groups. Labour markets in developing countries can beThe RPED data is based on a panel survey of about 200 manufacturing firms and theirworkers conducted in Zimbabwe between 1993 and 1995. See Gunning (1994) for a fulldescription of this data.Similar results have been found in Kenya (see Kulundu, 1997).180 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEsafely categorised into rural and urban. The incomes in the two marketsdiffer in their magnitude and composition. Rural households are prone torisks and often respond to this by diversification, both within agricultureand between agriculture and non-agriculture (Bigsten and Horton, 1997).In Table 6 we report the composition and magnitude of rural and urbanhousehold incomes.In rural areas as a whole, the main source of income is wageemployment, followed by remittances and lastly the "other" category.Female headed households in rural areas have higher incomes than maleheaded ones. In female headed households remittances are the majorsource of rural incomes (40%) compared to about 12% for male headedones. Employment is the major source of urban incomes (81%) followedby property incomes (12%). Only male headed urban households makeremittances while urban female headed households receive remittances.In total, rural households earn almost 39% of the urban households'incomes, while urban female and male headed households earn 48% and37% more than their rural counterparts, respectively.Table 6RURAL AND URBAN HOUSEHOLD INCOMES (1990/91, Z$ PER ANNUM)All householdsFemale headedMale headedAll householdsFemale headedMale headedTotal1290132012723 3212 7383 414Employ-mentAgricul-tureRural areas406 385145 428557 360Urban areas2 992 2451324 7643 256 163Property443452442Remit-tances290524155ŠOther209223200-359198-447Source: World Bank (1995), Zimbabwe: Country Report (Washington DC, World Bank).As in other studies (Mazumdar, 1994; van Der Geest, 1996) we canalso use the ratio of manufacturing to agriculture wages to proxy urban-rural wage gaps. The ratios are shown in Table a3 in the appendix. Theratios show that manufacturing to agriculture wage gaps have widened inthe 1990s after narrowing in the 1980s. Compared with other Sub-SaharanAfrican countries, the rural-urban wage gaps are wider (and also wideningwith time) for Zimbabwe. Van der Geest provides estimates of the ratiosfor 1985, for some African countries: Botswana (3.9), Ghana (1.8), KenyaM. NCUBE 181(3.04), and Malawi (3.4). The ratios for 1991 were as follows: Botswana(2.06), Ghana (0.87), Kenya (2.71), and Malawi (3.12). For Zimbabwe theratios are 4.2 and 4.6 for 1985 and 1991, respectively.Sectoral wage gapsThe development of sectoral wage gaps can also be investigated bylooking at average ratios of relatively better paying sectors to low payingones. Table a3 in the appendix shows that finance, electricity andmanufacturing are better paying sectors, while agriculture and domesticsare at the lower end. The ratio of the better paying sectors (finance,electricity and manufacturing), to low paying sectors (agriculture anddomestics) for the period 1970-1994, are shown in Table a3 in the appendix.The ratio of the better paying sectors and agriculture was wide in the1970s, narrowed between 1980-1990 and widened during the structuraladjustment phase (1991-1995). The ratio of average wages inmanufacturing, electricity and finance to wages in the domestic sector,were in 1970, on average 5.8, and in 1994, it was 17.7.In inspecting Table a3, we see that the gap between public sectorworkers and the private sector (represented by manufacturing) iswidening. Between 1970 and 1983 public servants' wages were abovethose of the private sector and after 1983 the trend was reversed. Thepublic servants on average earned 141%, 108%, 86% and 70% of theaverage manufacturing employee wage in 1970, 1980, 1990 and 1994respectively.22 The public sector has expanded too fast, and relativelyhigh wages have not been preserved.Occupational, gender and racial wage differencesData on gender, racial and occupational wage differences in Zimbabwe issketchy. The only source of information for such gaps is Velenchik(1996a).23 In Table 7 we show gender, race and occupational gaps for1993 and 1994.Females earn lower wages than their male counterparts. Europeansand Asians on average earn about six and three times the wage of Africans,respectively.24 There are also wide wage gaps between professionals andmanagement on one hand, and low level employees such as labourersand support staff, on the other. The question is then what explains thesedifferences. One possible explanation is that Africans or females haveinferior human capital endowment compared with whites or males,22 Kanyenze (1993) using salaries of civil servants and private sector workers in a numberof occupations shows that there are wide wage gaps between these two sectors.23 This study is based on the Regional Programme of Enterprise Development (REPD).24 Differences in human capital endowment or discrimination could explain such differences.182 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWErespectively. Another explanation is that Africans or females arediscriminated against in the labour market. In Zimbabwe no empiricalwork has been done on discrimination in the labour market. But elsewherein Sub-Saharan Africa the problem of discrimination exists. For example,Oaxaca (1991,50) using the Knight and Sabot data of the 1980s, finds thatin Tanzanian manufacturing, 17% of the differences in wages are explainedby factors other than observed individual characteristics.TABLE 7MEAN WEEKLY WAGE BY GENDER, RACE AND OCCUPATION (Z$)GenderMenWomenRaceAfricanEuropeanAsianOccupationManagementProfessionalOffice workersCommercial and salesSupervisors/foremenEquipment maintenanceSkilled labourMachine operatorsLabourersService/support staff1993281.52241.73236.591 298.63651.471 092.2806.26287.97282.43360.36304.12168.7132.6699.14120.91994368.23297.09304.971 805.11 080.791511.611 525.45396.54388.34467.39364.94211.18163.83127.2143.53Source: Velenchik (1996a).Trends in real wagesIn this sub-section we look at the development of real wages. In Figure 1below we trace the developments of average real consumption wages forall sectors (total formal sector) and three sub-sectors: agriculture,manufacturing, and the public sector. Consumption real wages are nominalwages deflated by the consumer price index. This reflects the purchasingpower of workers' wages. Between 1975 and 1979 real consumer wagesincreased steadily for all the sectors and between 1980 and 1982 theM.NCUBE183Figure laREAL WAGES IN AGRICULTUREReal WagesReal WagesFigure lbREAL WAGES IN THE PUBLIC SECTORPublic Sector includes health, education and public administration184 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEFigure lcREAL WAGES IN THE MANUFACTURING SECTOR1000080006000Real Wages40002000Figure IdREAL WAGES IN THE INFORMAL SECTORReal W agci2000M. NCUBE 185increase in real wages in all the sub-sectors and the formal sector as awhole, was more dramatic. The aggressive minimum wages, economicboom and good harvest that characterised the first two years ofindependence partly explains this rapid growth. Real consumer wages inthe manufacturing sector fluctuated around a stationary path during theperiod 1983-1990. In the 1990s, manufacturing real wages declined severelyand by 1994, they were below those of 1975. In agriculture consumerwages declined after 1982, stabilised between 1987 and 1990, then declinedbetween 1990 and 1992, only recovering slightly in 1993 and fallingthereafter.In the 1990s, high inflation triggered by ESAP eroded wages in theformal sector as a whole. The average purchasing power of all categoriesof workers in 1995 was lower than in 1975, a clear indication of fallingliving standards.This fall in formal sector real wages in the 1990s was part of the ESAPstrategy. The World Bank (1989) argues that a fall in real wages especiallyunder adjustment "... Is a brutal but necessary adjustment to reflectlabour unemployment caused by a growing labour-force that hasoutstripped job creation and the need to become internationallycompetitive." According to the World Bank, industry profit margins, whichare likely to be translated into investment, are eroded by excessivewages. However, in Zimbabwe the benefits in terms of employmentgeneration of falling real wages, expected from ESAP did not materialise.Weeks (1996), Zimbabwe Congress of Trade Unions (1996) and Kanyenze(1997) argue that wage rigidities were not a possible cause ofunemployment, as wages in Zimbabwe have been flexible.INDUSTRIAL RELATIONSPrivate sector industrial relationsThe conduct and performance of the labour market is best understoodwhen put in the broader industrial relations framework. In 1980 thelabour market inherited industrial relations characterised by entryrestrictions to skilled jobs, training, and apprenticeship, and restrictionsto membership of unions to many African workers. Collective bargaining,which was limited to only a few unionised workers (mainly white) wasconducted by government-appointed industrial councils. For the non-unionised workers government appointed industrial boards (which werenot collective bargaining forums) but whose mandate was to recommendwage adjustments to government.Government intervention in the labour market continued in the periodbetween independence (1980) and the introduction of structural186 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEadjustment in 1990. The government, besides being an active participantin the wage determination process, was also instrumental in the formationof trade unions. The independence of trade unions was limited through ahost of repressive and paternalistic regulations. As Shadur (1994) pointsout, during this period the government actually took the role of labour'sadvocate. Government intervened through the Employment Act of 1980.The Act, besides listing the rights of employees, gave the Minister ofLabour unlimited powers in industrial relations. Its distinctive featurewas that employers were required to seek ministerial approval beforeany retrenchment or dismissal of an employee for any reason. Theregulations were designed to offset the possible negative employmenteffects of minimum wages.25 The process of seeking ministerial approvalwas cumbersome and costly. Often the minister denied requests toretrench (Velenchik, 1996b). Employment protection rules made itimpossible to adjust the workforce levels in accordance with economicdownturns. Some employers however, circumvented these constraintsthrough employing casual labour. The laying off of casual workers did notrequire ministerial approval (Velenchik, 1996b).Two studies have looked at the effect of these regulations on labourdemand. According to Hawkins et. al. (1988, 18) the "inability to dismissworkers" was ranked second among the factors that discouragedemployment of more workers; 42% non-agricultural and 81% agriculturalemployers considered regulations as a prohibitive factor.The other study which has looked at the effect of these restrictionson labour demand is that of Fallon and Lucas (1993). This study estimatesa labour demand function over the period 1967-1985 in 29 manufacturingsub-sectors. The dependent variable is the number of employees and theindependent variables are the logarithms of real wages and output, andtwo lags of each. Included on the right-hand side also is the first lag of thedependent variable and a dummy which takes the value of one when jobregulations were in force and zero otherwise. The coefficient foremployment lagged once in all but a couple of the industries was positiveand significant, suggesting that even before the imposition of job securityregulations, employers were reluctant to adjust their employment levelsrapidly under different market conditions. Probably speedy adjustmentmay already have been costly. The coefficient on the job security dummywas negative in 24 of the 29 industries and in 20 of the 24, it was significantin at least 75% level. But this dummy coincides with the period minimumwages were in force. However, by including a wage variable in the model,25 As has been noted above, Fallon and Fallon (19lJ3) analysed the employment effects ofjob security regulations and found that they actually reduced labour demand.M. NCUBE 187the effect of minimum wages is controlled for. The other variable thatmight account for the negative and significant effect of the dummy couldbe that it is capturing the exodus of white workers in the immediate post-independence period. This is countered by including a net Europeanmigration variable in the regression, and still, its inclusion does notchange the magnitude and sign of the job security dummy. The conclusionthen from this study is that job security regulations reduced employmentin many manufacturing industries. In addition, their results indicate thatemployees in large establishments were affected most by new regulationsand that there was greater difficulty or reluctance to reduce employmentof Africans than Europeans. In some countries employers have respondedby substituting permanent workers by casual workers who are exemptfrom regulations. India is a good example (see Fallon and Lucas, 1993).Although there was an increase in casual labour in Zimbabwe, thegovernment curtailed this by stipulating a double minimum wage forcasual employees.In 1985 came the most comprehensive labour law, the Labour RelationsAct, which repealed the Industrial Conciliation Act of 1934 and itssubsequent amendment in 1959, and brought together the provisions ofthe Employment Act (1980) and the Minimum Wage Act (1980). The mainprovisions of the new Act were: the definition and protection of interestsand fundamental rights of workers; definition of unfair labour practices;the appointment and description of functions of workers' committees;and formation, registration, certification, and functions of trade unions.The Act safeguarded basic labour standards, such as minimum wagesand maximum working hours, and more important, it encouraged collectivebargaining.However, besides ushering in a new era in industrial relations, theAct had some inherent contradictions. Industrial action by workers wasprovided for in the Act, but technically it was taken away from theemployees in three ways: first, most of the workers were categorised asin the essential services and therefore not legally allowed to engage inindustrial action.26 Secondly, the process of getting permission for any26 The definition of essential service included workers in "any service relating to thegeneration, supply or distribution of electricity, any fire brigade or fire service; anysewerage, rubbish disposal or other sanitation service; any health, hospital andambulance service; any service relating to supply or distribution of food or fuel; anyservice relating to supply or distribution of water, mining, including any service requiredfor the working of a mine; any communications service; any transport service andanything relating to the repair and maintenance, or to the driving, loading and unloadingof a vehicle for use in any transport service; any service relating to any road, railway,bridge, ferry or airfield, and any other service that the minister may, after consultationwith appropriate Trade Union and employers' organisation, declare by notice in theGazette to be an essential service"(Labour Relations Act, 1985). This definition coveredmost formal sector workers.188 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEindustrial action was cumbersome. Workers wanting to engage in industrialaction had to give notice of the pending action to the government. Thegovernment often took a long time to respond. Thirdly, as the Act alsoprovided for collective bargaining through employment councils, thelabour minister was invested with powers to amend or annul collectivebargaining agreements if he deemed them not in the interest of theeconomy (all collective agreements were submitted to the minister forapproval).With structural adjustment, regulations were relaxed and the LabourRelations Act (1985) was replaced by the Labour Relations AmendmentAct (1992). Employment security, previously granted to labour, was takenaway in the spirit of promoting flexibility in the labour market. Hiring,layoffs and dismissal decisions were given back to the employer. Underthe new labour regulations, collective bargaining in unionised industriesis handled by employment councils composed of representatives ofemployers, workers and government, whose agreements have the forceof law. In non-unionised ones, collective bargaining is conducted throughemployment boards, whose members are appointed by the Minister ofLabour to represent employers and employees in a given industry, andwhose agreements are non-binding recommendations to the minister.Regulations regarding the activities of trade unions and industrial action,however, have remained intact.The civil service industrial relationsThe public service labour market is governed by the Public Service Act of1931 and to date, the essentials of this Act have not changed. Strikes bypublic servants are illegal in Zimbabwe as in many other developingcountries.27 Wage determination in the civil service is the prerogative ofthe central government. Public servants are grouped into associationsthat have no legal status of collective bargaining. The governmentappointed" Public Service Commission (PSC) recommends wageadjustments and it "may consult with any association recognised by theminister" regarding employment conditions. Associations can recommendchanges to the PSC, but effective decisions on wages and employmentconditions are debated and made by Cabinet.282' In Nigeria, for example, strikes in essential services are illegal and oil public servants aredefined as in essential services. In Tanzania and India public servants also have no rightto strike (Van Ginneken, 1990).See Ncube (1997) for a detailed discussion of labour relations in the civil service.M.NCUBE 189CONCLUSIONThis article provided an outline of the main features of the Zimbabweanlabour market. This overview shows that employment creation in theformal sector has not matched the number of job-seekers. The ranks ofthe unemployed have continued to swell each year since independence.Today more than 200,000 school leavers enter the labour market, whilethe formal sector job market is expanding by a couple of thousands. Theunemployment rate has increased from 10.8% in 1982 to 30% in 1995.The first decade of independence was also characterised bygovernment intervention in the determination of wages. Through itsincomes policy the government set minimum and maximum wages. In1990, intervention in the private sector labour market was replaced bycollective bargaining while the government still unilaterally determinescivil service wages.Racial and gender income gaps still exist in Zimbabwe, althoughunfortunately, we cannot determine whether this is due to differences inhuman capital endowment or discrimination. Real consumer wages in theformal sector have fallen over the years, especially during structuraladjustment. This has negatively affected the standards of living quitesubstantially. Elsewhere in literature (Kanyenze, 1996) it has beendemonstrated that real producer wages have also fallen rapidly duringESAP. However, this fall in producer wages during ESAP has not beenaccompanied by any meaningful increase in employment.ReferencesBlGSTEN, A. AND HORTON, S. (Forthcoming), "Labour markets in Sub-SaharanAfrica", in Poverty in Sub-Saharan Africa (AERC, Nairobi).COHEN, J. M. (1993), "Importance of public service reform: The case ofKenya", Journal of Modern African Studies, 31, (iii).DANIELS, L. (March 1994), "Changes in the small-scale enterprise sectorfrom 1991 to 1993: Results of the second nation wide survey inZimbabwe", Gemini Technical Report, 71.DAVIES, R. AND RATTS0, J. (1990), "Macroeconomics Conditions For MediumTerm Development in Zimbabwe: From Liberation to Liberalisation"(Mimeo, Department of Economics, University of Zimbabwe, Harare).DuREVALL, D. AND MLAMBO, K. (1995)," Zimbabwe: Inflation and StabilisationPolicies in Zimbabwe" (Macroeconomic Studies, No. 63, Stockholm,Swedish International Development Authority).FALLON, P. R. AND LUCAS, R. (1993), "Job security regulations and thedynamic demand for industrial labour in India and Zimbabwe", Journalof Development Economics, Vol. 40.GOVERNMENT OF ZIMBABWE (1985), Labour Relations Act (Harare).Š CSO (Various Issues), Quarterly Digest of Statistics (Harare).190 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEŠ (1988), Main Results of The Labour Force Survey 1986-1987 (Harare).Š CSO (1994), Income and Expenditure Survey Report 1990/91 (Harare).Š (Various Issues), Census of Production (Harare).GOVERNMENT OF ZIMBABWE (1996), Zimbabwe Programme For Economic andSocial Transformation (Zimprest), 1996-2000 (Harare).GUNNING, W. (1994), "The Manufacturing Sector in Zimbabwe: Dynamicsand Constraints" (Mimeo, Free University, Amsterdam).HAWKINS, A. M ., MCBURNEY, P. J., SHADUR, M. A. AND CLATANOFF, W. (1988),Formal Sector Employment Demand in Zimbabwe (Harare, Universityof Zimbabwe Publications).HERBST, J. J. (1990), State Politics in Zimbabwe (Berkeley, University ofCalifornia Press).IKIARA, G. K. AND NDUNG'U, N. (1997), The Labour Market During Adjustment:The Case of Kenya (International Labour Organisation, Geneva).IMANI DEVELOPMENT REPORT (1990), "Impediments Confronting the InformalSector Enterprises in Zimbabwe" (Harare, Paper Prepared for theWorld Bank).INTERNATIONAL LABOUR ORGANISATION (1978), Labour Conditions andDiscrimination in Southern Rhodesia (Zimbabwe) (Geneva, ILO), 65.INTERNATIONAL LABOUR ORGANISATION (1995), World Employment Outlook(Geneva, ILO).INTERNATIONAL MONETARY FUND (1993), Zimbabwe: Background Paper (NewYork, IMF).KANYENZE, G. (1993), "The Impact of Economic Stabilisation on the WageStructure in Zimbabwe" ( PhD. Thesis, University of Sussex).Š (1996),"Adjustment, Employment and Labour Market Institutions inZimbabwe" (Paper Presented at a Seminar on Adjustment,Employment and Labour Market Institutions in Zimbabwe, Harare).Š (1997), Adjustment, Employment and Labour Market Institutions inZimbabwe (Geneva, International Labour Organisation).KNIGHT, J. (1996), "Labour Market Policies and Outcomes in Zimbabwe"(Paper Presented at the 16th Arne Ryde Symposium, Lund, Sweden).KULUNDU, MANDA, D. (1997), "Labour Supply, Returns to Education, andthe Effect of Firm Size on Wages: The Case of Kenya" (PhD. Thesis,Gothenburg University).MAZUMDAR, D. (1994), Wages in Africa (Washington DC, World Bank).MCPHERSON, M. A. (1991), "Micro and Small-Scale Enterprises in Zimbabwe:Results of a Country-Wide Survey" (Gemini Technical Report No. 25,Michigan State University).Mead, D., Mukwenha and Reed (1993), "Growth and TransformationsAmong Small Enterprises in Zimbabwe" (Gemini Working Paper,Harare).MHONE, G. C. Z. (1993), The Impact of Structural Adjustment on the UrbanInformal Sector in Zimbabwe (Geneva, ILO).Š (1994), "Can ESAP sustainably transform the non-formal sectors inZimbabwe," SAPEM, No. 7.M. NCUBE 191MLAMBO, K. AND KAYIZZI-MUGERWA S. (June 1991) "The macroeconomics oftransition: Zimbabwe in the 1980s" African Development Review, 3,(0-MOYO, N. P., DAVIES, R. J., MHONE, G. C. Z. AND PAKKIRI, L. (1984), TheInformal Sector in Zimbabwe Š Its Potential For Employment Creation(Lusaka, ILO, Satep).NCUBE, M. (1997), "Public sector pay in Zimbabwe under adjustment" inColclough, D. (Ed.), Public Sector Pay and Adjustment, Lessons FromFive Countries (Routledge).PETERS-BERRIES, C. (1993), Urban Informal Sector in Zimbabwe: FromInsignificance to Employer of Last Resort (Geneva, ILO, Wep), 2-19.SESHAMANI, V. AND KAUNGA, E.C. (1997), Employment and Labour MarketsDuring Adjustment: The Case of Zambia (Geneva, International LabourOrganisation).SHADUR, M. (1994), Labour Relations in a Developing Country: A Case Studyof Zimbabwe (Avebury).SlDDiQUl, F. AND MATARE, S. K. (1993), Training School Leavers For Self-Employment (Harare, International Labour Organisation).VELENCHIK, A. (1994), "The labour market: Wages and earnings", in J. W.Gunning, The Manufacturing Sector in Zimbabwe: Dynamics andConstraints (Amsterdam, Mimeo, Free University).Š (1996a), "Market Power, Firm Performance and Real Wage Growth inZimbabwe Manufacturing" (Regional Program on EnterpriseDevelopment, Washington DC, World Bank).Š (1996b), "Government Intervention, Efficiency Wages, and the EmployerSize Wage Effect in Zimbabwe" (Regional Program on EnterpriseDevelopment, Washington DC, World Bank).VAN DER GEEST, W. AND WIGNARAJA, G. (1996), Adjustment, Employment andLabour Market Institutions in Sub-Saharan Africa in the 1990s: A Survey(Geneva, ILO).VAN GINNEKEN, W. (1991), Government and its Employees: Case Studies ofDeveloping Countries (Averbury).WEEKS, J. AND MOSLEY, P. (1996), "Structural Adjustment and Employment:A Comparative Study of Zambia and Zimbabwe" (Paper Presented atthe 16th Arne Ryde Symposium, Lund, Sweden).WORLD BANK (1995), Zimbabwe: Country Report (Washington DC, WorldBank).Š (1987), Zimbabwe: A Strategy For Sustained Growth (Washington DC,World Bank).Š (1989), Sub-Saharan Africa: From Crisis to Sustainable Growth(Washington DC, World Bank).ZIMBABWE CONGRESS OF TRADE UNIONS (1996), Beyond Economic StructuralAdjustment Programme (ESAP): Framework for a Long TermDevelopment Strategy in Zimbabwe (Harare).192 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWEAPPENDIXTable AlSECTORAL SHARE OF REAL GDP AT FACTOR COST AND EMPLOYMENTSectorAgriculture% share employment% share GDPManufacturing% share employment% share GDPMining% share employment% share GDPConstruction% share employment% share GDPHealth% share employment.% share GDPEducation% share employment.% share GDPFinance% share employment% share GDPElectricity% share employment% shareGDPPub. admin.% share employment% share GDPDistribution% share employment% share GDPTransport% share employment% share GDP197534.616.514.922.86.06.75.74.81.31.93.33.31.26.60.72.64.76.67.413.24.38.1197934.011.814.725.36.07.14.12.91.52.31.53.61.26.30.72.57.59.96.913.74.47.8198032.413.515.824.76.67.84.22.51.52.14.24.91.26.00.72.27.08.47.014.04.56.7198526.215.716.123.05.27.34.31.62.02.58.59.11.55.90.72.08.69.67.46.04.86.0199024.312.116.524.74.36.86.41.32.12.09.18.81.56.30.73.27.89.38.111.34.55.9199425.513.416.022.55.67.19.11.82.82.612.19.12.35.70.92.78.29.511.310.55.66.0M. NCUBE 193Table Al (cont)SectorPvt. domestic% share employment% share GDPOther% share employment% share GDPTotal formalemployment;thousandsReal GDP Z$ mil.1980=100197511.82.34.85.01 050.21 958197911.22.04.34.9984.72 689198010.72.04.35.41 009.93 42419859.41.55.56.01 052.53 92319908.61.16.25.11 192.24 537199410.91.39.78.0934.24 739Source: CSO, Quarterly Digest of Statistics (various issues). Notes: GDP is at factor cost,constant 1980 prices.Table A2ANNUAL CHANGES IN EMPLOYMENT, 1970-1994 (%)Post ESAPUDI Period Independ. Period Period1970- 1970- 1975 1980- 1980- 1985- 1991-Sector 1979 1974 -1979 1990 1984 1990 1994AgricultureManufacturingMiningElectricityConstructionFinanceDistribution HotelsTransportPub. Admin.EducationHealthPvt. DomesticOthersAll sectors1.03.51.02.30.83.20.42.17.11.03.80.53.31.73.67.82.85.410.25.53.24.42.92.45.03.46.74.5-1.7-0.8-0.8-0.9-8.60.9-2.4-0.211.2-0.52.7-2.3-0.2-1.1-1.22.9-1.22.66.13.53.31.92.311.74.92.75.21.8-4.13.0-1.62.12.55.43.53.04.120.56.25.05.51.01.23.5-0.93.09.22.03.11.10.84.43.90.75.02.42.80.90.2-0.74.26.62.6-0.1-3.61.0-0.70.06.11.6Source: Calculations based on Central Statistical Office, Quarterly Digest of Statistics (variousissues).194 THE EVOLUTION OF LABOUR POLICY IN ZIMBABWETable A3RATIOS OF MANUFACTURING, ELECTRICITY, AND FINANCE REALWAGES TO AGRICULTURE AND DOMESTIC WAGES197019751980198519901994M/A5.35.75.54.24.45.3M/D3.33.94.25.35.610.6E/A7.69.08.56.05.89.0E/D4.86.16.57.67.516.3F/A14.515.413.78.49.313.2F/D9.210.510.510.711.926.2M/A: Manufacturing/AgricultureM/D: Manufacturing/DomesticsE/A: Electricity/AgricultureE/D: Electricity/DomesticsF/A: Finance/AgricultureF/D: Finance/DomesticsSource: calculations based on Quarterly Digest of Statistics (various issues).