Prepared by Dr. Frikkie Liebenberg and Prof Johann Kirsten Department of Agricultural Economics, Extension and Rural DevelopmentUniversity of Pretoria
8 October 2013
Introduction
Creating jobs is one of the key challenges facing the South African economy. With agriculture consid-
ered to be a labour intensive industry a lot of hope is put on agriculture to make an important contribution to this challenge. The National Planning Commission for example believes that agriculture has the potential to create one million new jobs by 2030. This target is set in the context of the sector shedding almost one million jobs over the last three decades. At the same time the debate about the numbers of employed peo-ple in agriculture and the wages they receive has been misdirected due to the fact that nobody has a clear idea of the real facts. This is partly caused by the fact that the statistics on farm labour are spread between censuses of agriculture (rather incomplete and infre-quent), the October Household survey and more re-cently the Quarterly Labour Force Survey (QLFS) and also the 2011 population census. In this note we discuss the data, extracted for peo-ple employed in the agricultural sector only, to ensure a comprehensive picture of employment in the South African agriculture.
Statistics on Farm Labour in South Africa
Adjustments to the o� cial data sourcesIn our estimations of total employment in agriculture we extend the working age to include workers aged 65 and above. We have also consistently excluded workers employed in the “informal agricultural” sector and accounted for them separately. Workers in the forestry and � sheries industries are also excluded, and the latest re-weighted datasets of the QLFS are used in an e� ort to fathom the trends in employment in agri-culture since 20081. A key feature of the General Household Survey and the QLFS is that a rolling sample is used – a quarter of the household are replaced with a new sample in each successive round of surveying. Although this presents comparability problems it is claimed to introduce a better re� ection of the dynamics of the labour market.
Farm worker types and occupations: (de� nitions and classi� cations as per data sources)In order to ensure consistency we have adjusted the de� nition of the agricultural sector to only include the crops, horticultural, game and livestock industries, and excluded domestic servants. Based on the an-nual average of the quarterly observations (to address seasonal � uctuations and limit the e� ect of the rolling sample base) employment in agriculture in 2008 was 657 000 from where it decreased by 88 000 to reach 568 000 in 2011. From here it increased by 103 000 to reach an average of 672 000 during the � rst two quar-ters of 2013, see Figure 1. In the QLFS it is possible to identify whom the work-
Figure 1: Employ-ment in Agriculture, March 2008 to June 2013
UP Agricultural Economics Working Paper 2013/#2 | 1
2
er is working for, i.e. working for someone else for pay, an employer (who employ one or more employees), own account workers (not employing anybody) and those workers helping without pay in a household business. Disaggregating the data by geographic dis-tribution it becomes possible to identify commercial agricultural employment separate from informal and tribal (‘subsistence’) workers and employers. Figure 2 compares the composition of the agricul-tural work force according to whom they work for, for the years 2008 and 2012. The relative composition has changed very little over the years. Ninety percent of the workers work for pay – typically 75 percent of these are working in the formal (urban and rural com-bined) agricultural sector. When comparing this with the results reported in the 1993 Agricultural Census some alarming trends appear: • In 2012 there were 34 590 employers and own ac-
count workers in agriculture, down from 48 219 pro-prietors and tenants counted in 19932.
• The QLFS shows that South Africa had 50 332 farm-ers in 2009, but this declined to 34 905 by 2012 – a
Figure 2: Composition of Agricultural Work Force, 2008 and 2012
decrease of 15 427 in just three years.• Working family members in agriculture currently
stand at 3 582 as opposed to 20 428 in 1993. A sur-prising feature is that this class of labour in agricul-ture has almost entirely disappeared in the formal homeland areas and has been absent in urban agri-culture since 2010.
Taken together, the precipitous drop in the num-ber of farmers and the waning involvement of family members in farming spells disaster for the long run sustainability of agriculture.
How many people are employed in agriculture for a wage?Figure 3 shows that in 2008 there were 589 000 work-ers working for a wage in agriculture. This decreased by 78 000 over the subsequent three years to reach 510 000 by 2011. Growing by an average of 9.7 per-cent per year paid workers reached an average of 613 000 during the � rst two quarters of 2013 and rep-resented 99.2 percent of the increase in total employ-ment in agriculture.
1 For the years 2000 to 2007 the earlier de� nition of employment in agriculture includes those involved in farming for the pur-pose of own use. This will be revised as part of longer term research into labour trends in agriculture.
2 In 1993 the total number of farming units were 57 980 and Working Proprietors and Tenants plus Family Members were 68 647 in 1993. In Table 4.1 of the 1993 Agricultural Census Report, Family Members is quoted as 20 428. This leaves 48 219 full- and part-time Farmers and Tenants who farmed; 46 963 of which were white commercial farmers.
UP Agricultural Economics Working Paper 2013/#2 | 2
3
Figure 3: Working for Pay in Agri-culture, March 2008 to June 2013
Figure 4: Geographic distribution of Farm Workers, 2008 and 2012
Did farm labour really decline since mid-1980s?During the period from 1983 through 1994 the levels of employment of regular workers have � uctuated be-tween 615 000 and 735 000. During the past 6 years workers working for pay in the formal urban and rural areas (taken here to refer to commercial agriculture) have reached a minimum of 385 000 and are currently at 436 000 (71 percent of total agricultural employ-ment), almost the same level as in 2008.
Geographic distribution of farm labourThe geographic distribution of workers is classi� ed as:• urban formal, ordinary town or city area• urban informal, squatter areas in urban area
• tribal, tribal authority area with villages• rural formal, areas outside cities and towns with
farms and agricultural holdings
The majority of farm workers (52 percent) have always been employed in the rural formal areas (commercial farms), followed by about 20 percent of the agricul-tural work force being employed in the tribal areas. Farm workers in the informal urban areas make up less than 8 percent of the agricultural work force. Taken together, the share of primary agriculture of national employment is currently 5 percent of total national employment.
UP Agricultural Economics Working Paper 2013/#2 | 3
UP Agricultural Economics Working Paper 2013/#2 | 4
Changing nature of EmploymentThe QLFS distinguish between permanent and sea-sonal workers on the basis of the nature of the con-tract. This relationship yields four types of workers, ie., permanent workers, limited duration workers (casual workers), an unspeci� ed category (which is read here as seasonal) and a not applicable category which in total add up to the sum of employers and own ac-count workers. Figure 5 shows that:
• Seasonal and unspecifi ed duration workers have declined from 50 percent in 2008 to 47 percent in 2012.
• The total number of permanent workers in agricul-ture has shown a marginal increase of four percent over the same period.
• In the formal rural areas this change has been a little more pronounced. The share of permanent workers has increased by 6.8 percent from 2008 through 2012 replacing seasonal workers.
Why did we see increase in employment in agri-culture in last few quarters? By how much?It is di� cult to ascribe this increase to any single fac-
Figure 5: Farm Workers in Rural Areas According to Contract Type, 2008 and 2012
tor, however, over the 2 years since 2011 agricultural employment in the rural formal areas increased by 9.1 percent per annum, at the same time investment in agricultural machinery increased by 15.5 percent. As we have seen this coincides with a shift to more permanent workers (and a decrease in the number of employers), which hint at a greater reliance on more skilled and more permanent workers.
Farmers are employing more skilled workersThe share of paid workers in agriculture with an edu-cation level higher than primary schooling (grade 7 and more) has increased signi� cantly from what it was in 2008. This fact is relevant for both the com-mercial (formal) and informal agricultural sectors (Figure 6). In the commercial sector it has grown from 43 percent of the total number of paid work-ers in 2008 to an average of 55 percent for the � rst two quarters of 2013. In the informal sector the same metric has changed from 40 percent to 49 percent, an increase of 8.6 percent.
UP Agricultural Economics Working Paper 2013/#2 | 5
Figure 6: Education levels of paid workers: 2008 vs. Average for First Two Quarters of 2013
Formal Agriculture
Informal Agriculture
Are farmers paying more than the rest of the economy?Figure 7 shows the average wage per worker by geo-graphic region for the years 2008 to 2011. In 2008 work-ers in the rural formal areas (commercial farms) received a higher salary than those in other regions. Stated dif-ferently, farmers in the urban formal areas paid about 61 percent of the rates paid by farmers in the rural for-mal areas, with urban informal and tribal areas paying only about 20 percent of the wages paid by commercial farmers in the rural formal areas. By 2011, rural formal workers in agriculture received monthly wages that was double that of 2008. The gap between wages paid in the urban formal areas to that paid in the rural formal areas were essentially wiped out, whilst the informal areas and tribal areas were still pay-ing much less (in the order of 50 percent and less of the formal sector).
How many people practice agricultural activities for survival?The Population Census of 2011 included three ques-tions that tested for peoples’ involvement in agricul-ture. According to this source 2.9 million households were engaged in agricultural activities in that year. We doubt this estimate since it should most likely read “persons”. The commensurate estimate of the QLFS for 2011 is 2.4 million persons (not households) with the highest estimate (2.7 million persons) in the � rst quar-ter and the lowest (2.1 million) in the second quarter. According to Figure 8 this wide � uctuation in the en-gagement in agricultural activities is explained by the varying involvement of a large number of non-eco-nomically active persons in agriculture – on average 52.7 percent for 2011. For the period from 2008 to the second quarter of 2013, an average of 71.3 percent of these persons were either unemployed (11.8 percent), a discouraged job seeker (7.6 percent) or not econom-ically active (NEA) (51.8 percent). This hints that agri-culture is a residual activity, not even part-time – in the household livelihood.
UP Agricultural Economics Working Paper 2013/#2 | 6
Figure 7: Average Monthly Wage per Agricultural Worker: Geographic Di� erences, 2008 to 2011
Figure 8: Engagement in Agricultural Activity, March 2008 to June 2013 Note: NEA – Not economically active
UP Agricultural Economics Working Paper 2013/#2 | 7
Agricultural employment in the secondary and tertiary industries of the economyAccording to the QLFS, employment in agriculturally related industries in the secondary and tertiary sec-tors of the economy equals about 90 percent of the employment in primary agriculture. According to this broader de� nition of agriculture total employment were on average 1.153 million during the � rst half of 2013, down from 1.238 million in 2008. Quarterly observations show that this reached a minimum in
March 2011 from where it increased by 1.36 percent per quarter to reach 1.216 million in June 2013; this growth was driven largely by the growth in the prima-ry agricultural sector over this period. If the workers in the agricultural services, food man-ufacturing and trade are included, agricultural and ag-riculturally related employment represent 9 percent of national employment. As a share of all jobs created in the rural formal areas 52 percent work in agricul-tural and agriculturally related industries.
Figure 9: Total Employment in Agriculture, March 2008 to June 2013
UP Agricultural Economics Working Paper 2013/#2 | 8
Where to � nd the real numbers? Data sources and important facts about the data sources. The Abstract of Agricultural Statistics has traditionally sourced its information on regular employment on farms from the Agricultural Census and Survey of Ag-riculture reports of Stats SA. The numbers quoted are the total of regular workers (inclusive of domestic serv-ants) and seasonal workers, although the census and
survey reports do sporadically report statistics for farm owners and family members who regularly are work-ing on the farms. Workers in the forestry and � sheries industries have traditionally been excluded from these numbers, as were those employed in the informal ag-ricultural sector. With the prolonged absence of an Agricultural Cen-sus from 1996 through 2002 (the latter only released
Period
Workers in Agriculture, Hunting, Forestry and FishingAbstract1 Agricultural
Census/ SurveyOHS/LFS
Census LFS Originally Revised Annual AverageFeb. ‘93 1 093 1 093Feb. ‘94 922 922 980Feb. ‘95 891 891 1 295 949Feb. ‘96 915 915 759 974Oct. ‘97 717 966Oct. ‘98 935 967Oct. ‘99 803 979Sep. ‘00 667 779 689Sep. ‘01 666 682 721Sep. ‘02 941 941 811 825 805Sep. ‘03 1 212 832 714 722Sep. ‘04 1 063 626 534 979Sep. ‘05 628 925 628 579 519 780Sep. ‘06 1 088 496 606 595 805Sep. ‘07 774 1 041 771 668 580 846Sep. ‘08 767 815 767 813 832Sep. ‘09 653 850 653 695 731Sep. ‘10 640 866 640 669 672Sep. ‘11 624 822 624 654 649Sep. ‘12 661 661 686 693
Sources: Abstract of Agricultural Statistics (2005, 2010, 2013); Labour Force Survey (2000-2007); Quarterly Labour Force Survey (2008 - 2013).Notes: 1Number of farm employees and domestic servants on farms
Table 1: Employment numbers in agriculture, hunting, forestry and � shing from di� erent sources.
UP Agricultural Economics Working Paper 2013/#2 | 9
by 2005) the Directorate of Agricultural Statistics be-gan to quote the Labour Force Survey (LFS) from 2003 in addition to observations from the Agricultural Cen-suses and Surveys. The � rst Labour Force Survey was conducted on a bi-annual basis in 2000. This source di� ers in terms in its de� nitions of farm labour used in the agricultural census in a number of important re-spects:• The agricultural census includes farm workers of all
ages, the LFS only report employees of a ‘working age’ (ages between 15 and 65) and often included informal employment in agriculture in the totals on employment by industry. During the late 1990s workers aged 65 and above amounted to about 50 000 persons and the informal sector upwards of 300 000. The periodic inclusion of the latter explains the major change in employment in agriculture from 2007 to 2008.
• Whereas the Agricultural Census traditionally (in this country at least) only included the agricultural and game industries, the LFS classify the Forestry and Fisheries industries as part of the agricultural sector. In the second Quarter of 2008 the Forestry and Fish-eries industries accounted for 102 000 workers out of the total reported for 811 000 (14 percent) work-ers in the LFS de� nition of agriculture. This has de-clined to 61 000 (or 8 percent) in the corresponding quarter of 2013.
• The LFS estimates employment using an intricate method of weighting to adjust the raw counts to re� ect the ‘universe’, or total population. Since its inception the published observations have been re-weighted twice to re� ect the total population count obtained from the two population censuses. These changes were never replicated in the num-bers quoted in the Abstract of Agricultural Statistics.
