Migration of the population of Botswana* James D. Tarver Human settlement in present day Botswana goes back at least two thousand years ago. Beginning in the 17th and 18th centuries, the Tswana speaking people moved Into Botswana in different waves and streams at different times. In fact, many non-Tswana groups were forced to migrate into Botswana in the nineteenth century from Transvaal and other areas seeking refuge from the Difaqane wars and/or Boer incursions. Botswana's population has grown from around 120,000 persons in 1904 to more than 940,000 in 1981, an increase of well over 800,000 during this seventy-seven year period (Table I). Population increases between the censuses have not been regular due, in part, to the incompleteness of enumeration, especially in the early censuses. Nevertheless, in the ten-year period between the last two censuses taken In 1971 and 1981, well over 340,000 persons, or 57.6 percent, were added to the country's population. The population of a c,'untry grows, of course, whenever the number of births surpasses the number of deaths, which is known as natural increase. Population increases also occur when the number of persons moving into (immigrants) exceeds the number of persons moving out of the country (emigrants), or net immigration. Population increases in Botswana due to natural increase cannot be measured precisely, because the vital statistics registration and reporting systems for births and deaths have not yet been fully perfected. Therefore, this report will be devoted specifically to the examination of population changes arising from migration, for those years for which data are available. International Migration International migration statistics are available for the number of immigrants into Botswana for most years from the United Nations Demo~raphic Yearbooks starting in the 1960s and going into the I980s from the Botswana Statistical Bulletin. During the last 15 years somewhere between two to seven thousand persons have immigrated to Botswana each year, but no comparable record is available for emigrants leaving Botswana each year: Nevertheless, the 1981 census did report 41,961 Botswana citizens abroad and It also reported that 10,117 persons from abroad had moved into Botswana during the year preceding the 1981 census.l International movements from Botswana involve ,mainly short-term, I mine workers to South Africa. These Batswana are chiefly between 15 and 34 years of age with low levels of education. Mine labour migration from Botswana comprised around one quarter of the adult males from the 1940s to the mid 1970s, when the peak recruitment of men from Botswana began to decline after 1976.2 In contrast, fewer than five percent of the adult females have been temporarily absent from Botswana during that period. *1 want to thank Ih Hesill 1U111lc8Y8 for his helpful l'_rks end critid_ of this peper. In terms of census definition, only a permanent change in residence is considered migration. Notwithstanding, there is a large volume of temporary migration of the village population to outlying cattleposts for seasonal agricultural activities, but these annual short-term movements do not constitute permanent changes of residence. The Batswana miners recruited for work in the gold mines in South Africa usually sign two-year contracts and mayor may not renew them upon their expiration. Nevertheless, their international movements must be considered temporary, too, for each plans some day to return to join his family in Botswana. Internal Migration The migration figures that will be analysed in this report are for migrants within the country itself, namely internal migrants. These figures are based upon a question asked in the 1981 census about the place of residence of each person a year before. Thus, the migration figures for the one-year period are complete counts from the entire census enumeration and are not estimates. Nevertheless, the volume of migration in the year preceding the census is understated for two reasons. First, all migrants who moved and died and those who returned to their former residences within the one-year period are not counted as migrants. Second, obviously there was some under-enumeration in the 1981 census, the precise amount being unknown. Say, for instance, that four percent of the total population was missed in the actual census; in other words, the census count was 96 percent complete. Consequently, the estimated total migrants would be obtained by multiplying the enumerated number by 106 percent. Thus, ih such a hypothetical illustration the total number of in-and-out-migrants, corrected for under-enumeration would be 83,712 rather than the actual enumerated figure of 78,974 (Table 2). Information obtained from questions in the 1981 census make it possible to examine the streams of migration from one census district to another in that one-year period, to compute in-, out-, and net migration rates for census districts, and to relate migration to age, sex, education, fertility, occupation, industry, and other related variables. In the 'year preceding the 1981 census, a total of 77,348 persons moved from one residence to another within the same census district, 78,974 moved from one district to another within Botswana, and 10,117 persons from abroad moved into the country (Table 2). Thus between August 1980 and 1981, 8.2 percent were intra-district movers and 8.4 percent were inter-district movers. Rural-Urban Migration Some people suspect that migration in Botswana involves a large scale exodus of people from the rural areas who are very prone to shift residences from one district to another far-removed district. Actually, migration was relatively high in the urban census districts and comparatively low in the rural census districts in 1980-81. First, in the six urban districts about 18 percent of the population moved to another district, whereas in the 15 rural census districts only slightly more than 6 percent moved to another district (Table 3.) Second, in-and-out- migration rates ranged from over 15 to nearly 50 percent in the six urban districts, while the comparable rates for the rural census districts were less than 10 percent. Consequently, the urban w~re much more migratory than the rural residents, regardless of the measure or comparison used. 81 A second major characteristic of the 1980-81 population movements in Botswana is that the rural rather than the urban census districts had net gams resultmg from in-and-out-migration. Specifically, in terms of rural-urban population shifts among inter-district movers in 1981, the rural census districts gained about 10,000 new residents as a result of net migration. Should immigrants from abroad to Botswana during that one-year period be included, the rural areas experienced net populatIOn gains of over 15,000 persons as a result of migration. Looking at the rural-urban migrants in another perspective, over three-quarters of all 1980-81 migrants lived In rural areas at the end of the one-year period (Table 4). These numbers mclude all migrants, both within and between census districts, both rural and urban. There are other distinguishing characteristics about rural and urban migrants, Very importantly, population movements in the six urban census districts were stnctly mto or out of each district, as less than one percent of all movers were mtra-dlstnct movers (within district movers). In contrast, within district movers actually outnumbered both the in-and-out-migrants from the most highly rural census districts such as Ngamiland, Kweneng, and Ghanzi. Therefore, local, short-distance movement tended to be much more common among migrants living in rural than in urban areas. Rural and urban census districts tend to share somewhat the same type of exchange patterns among their inter-district migrants: there are dominant flows out of each district and counter flows back into each respective district. For example, migrants who leave urban census districts tend to move to adjoining rural census districts or else to the nearest non-contiguous district. In turn, there are counter flows from both the outlying, adjacent and closest non-contiguous rural census districts back into the nearest urban census districts. Thus, the largest movements from the urban districts of Jwaneng, Selebi-Phikwe, and Orapa were to three respective adjacent rural census districts and the largest movements from Gaborone, Lobatse, and Francistown were to the nearest non-contiguous rural census districts. In all dominant streams the movements were from urban census districts to nearby rural census districts. In turn, the contiguous and nearby non-contiguous rural districts contributed the largest single number of migrants to the nearest urban census district. Finally, somewhat the same types of exchange patterns for inter-district migrants apply to rural census districts, there being dominant flows and counter flows between nearby rural districts. In all, 80 percent of the migrants from urban census dlstncts. moved to rural districts and about 55 percent of all migrants from rural census districts moved to urban districts. Socio-Economic Characteristics of Migrants Age is the one variable probably most often related to migration. Those who moved In the one-year period 1980-81 were no exception. The migration rates of Botswana males rose to a peak of about 32 percent at 20-24 years of age, then declined regularly thereaf.ter (Table 5). Females followed somewhat the same pattern as males, but reaching the highest migration rate at 15-19 years of age then declining thereafter. Studies have generally shown that females move at so~ewhat younger ages than males. In the 1981 census, questions were also asked about the education of each person, the number of children born to each woman, the occupation and industry in which each was workmg as well as about other social and economic matters. For example, every person 12 years of age and over was asked about the amount of schooling he had com- pleted. Table 6 shows that the likelihood of a person twelve years and older moving is directly related to the amount of formal education he had finished. Proportionally, nearly twice as many with a higher education as with no education moved in the year preceding the census. Moreover, males were slightly more likely to migrate than females in each educational category. The proportionate number of females who moved in the year preceding the 1981 census declined consistently as the number of their children increased (Table 7). A fourth of all women with no children moved whereas only about one out of every seven with six or more children did so. Finally, the proportionate number of women moving declined consistently as age increased beyond 15 years of age. In the year preceding the 1981 census, employees had the highest migration rate of all persons in the different economic activity status groups (Table 8). Those actively looking for work had the second highest rates and self-employed persons were the least mobile. In terms of occupational attachment, professional and construction workers were the most mobile and sales persons the least mobile of a11occupational groups in Botswana in 1980-81 (Table 9). And, in terms of the industrial affiliation of workers, those engaged in tourism, construction, and education had the highest migration rates whereas those employed in manufacturing and trade had the lowest migration rates (Table 10). Summary and Conclusions This study has shown that the population was rather mobile in the one year preceding the 1981 census, as nearly 80,000 persons moved within districts and another 80,000 moved between districts. Including the ten thousand persons from abroad, there were over 168,000 migrants in all, about 18 percent of the total population. Although migration into and out of urban areas was relatively high, most Botswana migrants went to rural census districts, as three-fourths of the migrants lived in rural areas at the end of the period. Furthermore, migration in a rural census district was more Iilcely to involve intra-district rather th?n inter-district movement. From an analysis of the Botswana migration data for the one-year preceding the census, it is possible to draw the following conclusions about the socio-economic characteristics of the migrants: First, migration rates are directly associated with age of females up to 15-19 and of males up to 20-24 years of age, then the migration rates declined thereafter. Second, migration is directly related to formal education. Third, the migration of women is inversely related to the number of their children and to their own age. Fourth, employees were the most and self-employed persons the least mobile of all economic activity status groups. Fifth, professional and construction workers were the most mobile and sales workers the least mobile of all occupatIonal groups. Finally, persons engaged in tourism, construction, and education were the most mobile of all industrial groups whereas manufactUrIng and trade workers were the least mobile. 83 TABLE 1 Population of Botswana, 1904-1981 Year Population Change During Period Numerical Percent 1904 120,776 1911 124,350 3,574 3.0 1921 152,983 28,633 23.0 1936 265,756 112,773 73.7 1946 296,310 30,554 11.5 1956 309,175 12,865 4.3 1964 549,510 240,335 77.7 1971 596,9941 47,484 8.6 1981 941,027 344,033 57.6 I. Includes an estimated 10,550 nomads and 12,400 in areas not enumerated. Source: Census Administrative/Technical Report and National Statistical Tables, 1981 Population and Housin~ Census, Central Statistics Office, Ministry of Finance and Development Planning, Republic of Botswana, Government Printer, Gaborone, February, 1983, p.l0. TABLE 2 Internal Migration within and Mip;ration into Botswana in the year preceding the 1981 Census, by Census District Inter-District Intra- Migrants Migrants District From Abroad Migrants Census District n-Migrants Out-Migrants Net Migrants Gaborone 12,271 10,226 2,045 247 230 Francistown 7,9/0 5,736 2,234 94 615 Lobatse 3,509 4,107 - 598 67 331 Selebi-Phikwe 6,586 5,713 873 78 487 Orapa 1,569 915 654 2 189 Jwaneng 2,705 867 1,838 6 483 Ngwaketse 5,078 6,295 - 1,217 10,5J!9 1,445 Barolong 1,247 1,101 146 322 278 South East 2,359 2,060 299 832 458 Kweneng 4,210 4,373 - 163 12,699 982 Kgatleng 2,656 3,389 - 733 3,931 535 Central Serowe 5,627 8,397 - 2,770 8,817 530 Central Mahalapye and Tuli Block 5,663 5,132 531 8,977 666 Central Bobonong 3,310 4,058 - 748 4,724 203 Central Boteti - 1,835 2,155 - 320 3,536 49 Central Tutume 4,526 6,842 - 2,316 3,546 481 North East 2,953 3,123 - 170 1,215 486 Ngamiland 1,575 2,128 - 553 10,534 1,096 Chobe 774 482 292 1,083 296 Ghanzi 855 721 134 3,794 92 Kgalagadi 1,696 1,154 542 2,255 185 TOTAL 78,974 78,974 0 77,348 10,117 Source: Table 29, 1981 Population and Housing Census, Census Administrative/ Technical Report and National Statistical Tables, Republic of Botswana, Gaborone, February 1983. 85 TABLE 3 Migration Rates of all Internal Migrants between Census Districts, Botswana, 1980-81 Census District Rate of Migration by Rural and Urban Residence In-Migration Out-Migration Net Migration Rate Rate Rate Urban 23.1 18.4 4.7 Gaborone 20.6 17.1 3.4 Francistown 25.7 18.5 7.2 Lobatse 18.4 21.6 -3.1 Selebi-Phikwe 22.3 19.4 3.0 Orapa 30.0 17.5 12.5 Jwaneng 1+8.6 15.6 33.0 Rural 5.6 6.5 -.9 Ngwaketse 4.9 6.0 -1.2 Barolong &.1 7.1 .9 South East 7.7 6.7 1.0 Kweneng 3.6 3.7 -0.1 Kl1,atlenF, 6.0 7.6 -1.6 Central Serowe 6.1 9.1 -3.0 Central Mahalapye and Tuli Block 6.8 6.2 0.6 Central Bobonong 7.1 8.7 -1.6 Central Boteti 6.9 8.2 -1.2 Central Tutume 6.0 9.1 -3.1 North East 8.1 8.5 -.5 Ngamiland 2.3 3.1 -0.8 Chobe 9.8 6.1 3.7 Ghanzi 4.5 3.8 0.7 Kgalagadi 7.0 4.8 2.3 TOTAL 8.4 8.4 .0 16 TABLE 4 People who moved in the year preceding the 1981 Census by size of present locality of residence Size of Place in 1981 Number of Migrants Percent of Migrants URBAN 41,378 24.6 5,000 and over 20,905 12.4 1,000 - 4,999 18,760 11.1 500 - 999 14,794 8.8 100 - 499 43,620 25.9 Under 100 28,968 17.2 TOTAL 168,425 100.0 Source: TABLE 28, 1981 Population and Housing Census, Census Administrative/ TecflOica! Report and National Statistical Tables, Republic of Botswana, Gaborone, February 1983.. TABLE 5 Proportionate Ntmber of Persons who moved in the year preceding the 1981 Census, by sex Age Sex of Population Group Total Male Female 0-4. 9.2 9.1 9.4 5-9 14.6 14.0 15.1 10-14 20.2 19.6 20.9 15-19 26.3 25.0 27.3 20-24 27.5 31.7 24.5 25-29 24.7 30.6 20.4 30-34 21.6 27.1 17.2 35-39 19.5 25.1 15.0 40-44 17.9 21.9 14.5 45-49 16.6 19.4 14.1 50-54 15.6 17.9 13.5 55-59 15.0 16.6 13.7 60-64 14.6 15.8 13.5 65+ 12.4 13.3 11.7 , TOTAL 17.9 18.8 17.1 Source: Table 28, 1981 Population and Housing Census Administrative/Technical Report and National Statistical Tables, Republic of Botswana, Gaborone, February, 1983• • TABLE 6 Proportionate Number of Persons twelve years of age and over who moved in the year preceding the 1981 Census, by sex and educational attainment Sex and Educational Attainment Migration Status Total Never at Not Completed Higher School-Not Completed Primary known Primary Total Population Number 569,387 247,066 199,038 82,783 40,500 Number of movers 120,986 43,034 39,843 23,432 14,677 Percent movers 21.2 17.4 20.0 28.3 36.2 Males Number 257,844 122,618 79,584 34,444 21,198 Number of movers 60,424 24,198 17,752 10,590 7,884 Percent movers 23.4 19.7 22.3 30.7 37.2 Females Number 311,543 124,448 119,454 48,339 19,302 Number of movers 60,562 18,836 22,091 12,842 6,793 Percent movers 19.4 15.1 18.5 26.6 35.2 Source: Table 31, 1981 Population and Housing Census, Census Administrative/ Techical Report and National Stati~tical Tables, Republic of Botswana, Gaborone, February, 1983 TABLE 7 Proportionate Number of Women twelve years of age and over who moved in the year preceding the 1981 Census, by age and n umber of children ever born Number Percentage of Females Moving of Children Age of Women Ever born 12-14 15-29 30-44 45-59 60+ Total 0 21.6 29.0 20.8 17.4 14.4 25.3 1 14.3 24.2 19.2 15.1 11.3 22.1 2 9.1 21.5 18.6 15.9 12.6 19.7 3 19.4 17.0 14.0 11.8 17.4 4 17.6 17.0 13.8 10.7 15.9 5 16.3 14.1 13.0 11.8 13.9 6+ 16.8 13.8 13.1 12.1 13.3 TOTAL 21.6 24.5 15.7 13.8 12.1 19.5 Source: Table 32, 1981 Population and Housing Census, Census Administrative/ Technical Report and National Statistical Tables, Republic of Botswana, Gaborone, February 1983. TABLE 8 Proportionate Number of Persons twelve years of age and over, by sex and economic activity status, who moved in 1980-81 Economic Activity Percentage Moving in 1980-81 Status Total Male Female Employee 28.9 29.7 27.0 Self-Employed 17.5 18.4 16.1 Periodic Piece jobs 21.1 20.9 21.6 Family Agriculture 15.8 15.4 16.3 Actively looking for work 25.3 27.4 23.5 Economically inactive 20.3 24.5 18.7 TOTAL 21.3 23.5 19.4 Source: Table 30, 1981 Population and Housing Census, Census Administrative/ Technical Report and National Statistical Tables, Republic of Botswana, Gaborone, February 1983. 91 TABLE 9 Proportionate Number of Workers twelve years of aRe and over, by sex and major occupation, who moved in 1980-81 Major Occupation Sex of Workers Total Male Female Professional 35.6 35.6 35.6 Administration 20.3 19.7 22.3 Clerical 22.0 23.2 20.7 Sales 17.1 17.2 17.0 Service 27.5 28.2 27.0 Agricultural 22.4 22.4 22.1 Production 27.3 29.5 16.4 Construction 31.5 31.5 29.7 Transport 27.3 27.4 19.4 Others 33.0 33.2 29.8 TOTAL 28.1 29.0 26.1 Source: Table 33, 1981 Population and Housing Census, Census Administrative/Tech- nical Report and National Statistical Tables, Republic of Botswana,Gaborone . February 1983. TABLE 10 Proportionate number of workers twelve years of age and over by sex and industry who moved in 1980-81 Industry Sex of Workers Total Male Female Central Government 29.7 31.8 23.6 Local Government 22.5 21.9 24.5 Education 37.1 38.9 36.3 Agriculture 26.0 25.3 31.3 Mining 31.0 31.3 26.0 Manufactur ing 15.1 15.4 13.8 Utilities 22.1 22.5 16.2 Construction 39.3 39.7 31.5 Trade 16.8 16.9 16.7 Transport 28.4 29.4 19.0 Finance 21.0 22.8 18.8 Personal Services 25.9 23.9 27.9 23.8 28.7 18.4 Health/Social Services 41.3 42.6 33.3 Tourism 42.6 41.6 46.8 Not known 28.1 29.0 26.1 TOTAL Source: Table 34, 1981 Population and Housing Census, Cen~us Administrative/Tech- nical Report and National Statistical Tables, Republic of Botswana, Gaborone February 1983. 93 Footnotes 1. Source: Tables 29 and 55, 1981 Population and Housing Census. 2. Isaac Schapera, Migrant labour and Tribal Life: A study of the condition in the Bechuanaland Protectorate. O.U.P., london, 1947, Report of the Population Census 1971, Central Statistics Office, Gaborone.