6—
National Income and Public Finance

The economies of Bophuthatswana and KwaZulu are extensions or subsectors of that of the Republic and are not self-contained. Their economic dependence takes several forms. The lack of opportunities in the homelands compels many workers to seek employment in the white areas of the Republic. The resultant system of transitional migrant and commuter labor provides the bulk of the income by which homeland citizens survive; production within the homelands typically constitutes less than one-quarter of total area income. Wage rates for jobs held by Africans in the Republic are thus a major determinant of homeland economic welfare.

This dependence of the private sector is paralleled by the reliance of the homeland governments upon the financial support of the central government. Lacking a tax base of affluent, productive citizens, and barred from taxing white businesses or mines within their territories, the homeland governments receive most of their revenues directly from the Republic. For agricultural, industrial, and infrastructure development, the homelands depend upon the Republic for investment capital, management, and entrepreneurship. The Bantu Investment and Mining corporations have been given the major part of the responsibility for providing the industrial, commercial, and extractive enterprise necessary to make the homelands economically more self-reliant. These white-dominated agencies have become the major sources of capital and the locus of most decisions about developmental priorities.

Sources of Income and Structure of Output

The pattern of underdevelopment in Bophuthatswana and KwaZulu is in most respects that shared by the world's poorer countries and regions. Income per capita and labor productivity are low; economic growth is sporadic and dependent upon external factors. Rapid population growth causes a deterioration of the man-land and food production-population ratios. The supply of mature labor exceeds the economy's employment capacity and there is a mounting overbalance of persons in nonproductive, dependent age ranges

(especially below age fifteen). Agriculture is the most important economic activity, but is for the most part customary and noncommercial. In Bophuthatswana, where the rainfall is light and there are no dependable rivers, the dominant rural occupation is cattle herding mixed with small-scale farming. In KwaZulu, water is more plentiful and in many areas farming is more important than husbandry. In both homelands, agricultural technology is backward and static, fertilizer and other modern inputs are rarely used, and extension services are limited.

Industrialization in Bophuthatswana and KwaZulu has hardly begun and its distribution is uneven. Unlike many newly independent nations, neither homeland has inherited a colonial industrial, plantation, or mining sector that could constitute the core of an industrial base. The sponsorship and management necessary to generate large-scale factory enterprise have not been forthcoming from either internal or external sources. Facilities to mobilize savings and provide agricultural and industrial credit, such as rural cooperatives, private banks, and specialized development finance corporations, are almost totally absent. Electricity consumption per person is very low, and vast areas of the two homelands remain without any power except that provided by men and animals. Transport facilities are primitive except in those few areas near white cities that are served by trains and buses. There is little piped drinking water and sanitation, and health facilities are rudimentary.

In an underdeveloped economy, the most important measures of economic welfare and growth are the national income accounts that measure annual production and earnings. The best calculations of national income in Bophuthatswana and KwaZulu are those released by the statistical sections of BENBO, the research and planning unit of the Bantu Investment Corporation.^[1] Using the available population, national income, and homeland product statistics, it is possible to derive some approximate indices of per capita output and income.

Several limitations must be considered in assessing these "fairly rough" estimates. Subsistence income is essentially ignored; some farm, domestic, handicraft, and traditional personal service production is not adequately included. Since many workers are absent much of the time, the population

[1] Scattered data are found in: J. A. Lombard and P. J. van der Merwe, "Central Problems of the Economic Development of the Bantu Homelands," Finance and Trade Review, X (1972), 1–46, tables 2 and 10; J. J. Stadler, "Demografiese en Ekonomiese Kenmerke van die Suid-Afrikaanse Bantoetuislande," Agrekon, IV (1970), 22–28; J. A. Lombard, "The Economic Philosophy of Homeland Development," in Rhoodie, South African Dialogue, 168–181; fuller series are found in BENBO's Economic Revues for 1975 of the two homelands (Pretoria, 1976). See also J. A. Lombard, et al., Focus on Key Economic Issues, 9, The Homelands (Pretoria, 1974).

pyramid, or age and sex distribution, of the homelands is skewed toward females and young and old male dependents. Internal production is therefore depressed beneath that of populations of comparable size. Further, it is difficult to determine exactly how much income is earned and remitted by absentee workers: by commuters, migrants, and de jure homeland citizens living in white areas. An undetermined number of people work and earn illegally in white cities, either as commuters or as longer term migrants. There is no attention to the distribution of income, and averages can do little to suggest how many persons are found at the extremes of poverty and wealth.^[2]

The most recent available income statistics are for 1973, with roughly comparable estimates for 1970 and 1960. These are presented in table 6.1 and include the major income aggregates such as internal earnings, absentee labor income, and total income of the de facto and de jure populations. The first column contains estimates of income earned by homeland residents: residents of Bophuthatswana earned R38.0 million and residents of KwaZulu earned R79.7 million in 1973. (An appropriate exchange rate for 1970 and 1973 is R1 = $1.40; for 1976, R1 = $1.15.) By way of contrast, the gross domestic product of the Republic in 1972 was over R15.3 billion, to which both homelands contributed less than 1 percent, far less than their area and population would imply. The total income of each of the homelands is smaller than that of an American town or medium-sized business firm. Computations based on the 1970 income figures and gross domestic product totals (from table 6.2) indicate that about 54 percent of area output in Bophuthatswana and 83 percent in KwaZulu are retained as income by homeland Africans, the remainder presumably going to whites and others as wages and salaries or business income.

Income earned inside the homelands is strikingly small compared to outside commuter and migrant incomes. Using the data in table 6.1, for the Tswana, internal income, commuter income, and migrant income amount, respectively, to 23, 58, and 19 percent of earnings; for the Zulu, the respective shares are 22, 36, and 41 percent. In the former case, commuter income earned in Pretoria, Rustenburg, and other white areas immediately south of the homeland is the major component, although for KwaZulu migrant income dominates. The average income earned from internal sources in 1973 was about R40 in Bophuthatswana and R35 in KwaZulu, but was only R28 and R26 respectively in 1970. These exceptionally low per capita output

[2] Lombard, van der Merwe, and Stadler mention most of these reservations in their various publications. The net effect of the omissions and qualifications is difficult to assess. The underrepresentation of subsistence income reduces the homeland output figures below what they might be, but, on the other hand, estimates of the portion of absentee worker income that actually returns to the homelands are merely rough guesses and could be too high or too low.

figures reflect the lack of productive employment opportunities, the absence of much of the work force, and, to a lesser extent, underestimation of subsistence production.

These low levels of productivity, and the fact that over three-quarters of Tswana and Zulu income comes from absentee labor, make abundantly evident the dependence of the homelands on the economy of the Republic. Because they are the baselines from which reconstruction of the homeland economies must begin, they are also dramatic measures of the immense gap that must be closed before productivity levels in the homelands can compare to those outside them. The race between population growth and diminishing returns in homeland agriculture (and minimal returns in industry) was lost a generation or more ago, leaving most workers no option but to seek employment in the Republic. In simple terms, the homelands would have to become four to five times more productive merely to reabsorb their de facto work forces at constant levels of living. To close the gap with the living standards in the Republic, or with those in middle-level developing nations, or to reabsorb a fraction of the now permanently absent Tswana or Zulu, would require much greater effort.

These broad aggregate conjectures, which do not allow for the increasing difficulty of sustaining worker productivity as more people are employed in a homeland, make it plain that an exceedingly well-financed and well-designed strategy will be required to overcome existing underdevelopment and provide satisfactory employment for a significant fraction of absentee workers.

The income of Tswana permanently absent from their homeland was R230.1 million in 1973; Zulu living in the Republic earned R368.3 million. If these aggregate incomes are added to the total incomes of the homeland Tswana and Zulu (column 4 of table 6.1), then total Tswana income was R393.5 million and total Zulu income was R727.5 million for 1973. But these global totals have no real meaning for the homelands, except in defining the total income of the de jure population and establishing the broadest definition of the homeland's personal tax base. Little income from permanently absent Tswana and Zulu is sent to the homelands, although some is doubtlessly returned to kin. Likewise migrants probably must spend most of their incomes on housing, food, and necessities, leaving little to send or take back.

Lombard, Stadler, and van der Merwe assume that all commuter income, one-fifth of migrant income, and one-twentieth of the earnings of those de jure citizens permanently living in white areas are remitted to the homelands. The addition of these amounts to income earned in the homeland yields an aggregate total called area national income. This is a measure of the total income generated within and flowing into the homeland economies. In 1973

area national income in Bophuthatswana was R150.1 million, and in KwaZulu it was R258.4 million. Using estimates of 1973 populations, this means that per capital area incomes were R151 and R112, respectively. For 1970 comparable figures were R109 for Bophuthatswana and R79 for KwaZulu. These income estimates lie between the average income of the de facto inhabitants (column 8 of table 6.1) and an average income measure that includes full migrant incomes (column 9 of table 6.1).^[3]

Despite the importance of external sources of income, internal economic development will depend upon the capacity of the homeland governments and Republic developmental institutions such as the Bantu Investment and Mining corporations to accelerate structural changes in homeland output. Generally speaking, although agriculture must grow absolutely and become more commercial, the relative shares of industry, mining, transportation and communication, and the public sector will have to expand relative to the farming sector. BENBO has recently made available data describing the sectoral origins of gross domestic product in Bophuthatswana and KwaZulu. These augment fragmentary information published by Lombard, Stadler, and van der Merwe.^[4] From fiscal 1960/61 to fiscal 1971/72 the two homeland regional economies apparently experienced structural modernization in a moderate, uneven manner (see table 6.2). Of necessity the data depend upon rough estimates, but the estimates are probably no more conjectural than those used for many less-developed areas of the world.

The more important changes in the structure of Bophuthatswana's economic activity are the decline of agriculture, the sharp rise in the contribution of mining, the moderate industrial gains, and the shift away from traditional personal services. Because of the expansion of the value of total output (item 13), all sectors with the exception of agriculture show some absolute growth, but those with higher shares at the end of the period have been growing relatively faster. The stagnation of agriculture and its small share of output reflect the poor prospects for intensive farming in most of the homeland, the failure and small size of government programs, and the inability of indigenous farming and herding to commercialize and become more productive without substantial improvement in the availability of modern inputs, technology and extension advice, credit, infrastructure, and marketing outlets. The government sector (items 7 and 8) grew rapidly in the late 1960s and early 1970s, although this growth was disguised by the overshadowing rise of mining. Likewise, the growth point at Babelegi, the

[3] Cf. Lombard and van der Merwe, "Central Problems," 35. Population growth rates from 1970 to 1973 are taken from 1951 to 1970 averages, ibid. , 33.

[4] See chapter 6, note 1. Fuller series are published in BENBO, Bophuthatswana, Economic Revue, 1975 (Pretoria, 1976), 30, and KwaZulu, Economic Revue, 1975 (Pretoria, 1976), 34. See also Lombard, Key Economic Issues , figure 8.

construction of a shopping center at Temba township in the homeland and new shops and beer gardens have been responsible for the growth reflected in the figures for industry, construction, and commerce.

The unique role of mining in Bophuthatswana deserves further comment. The value of the homeland's mineral output rose from R4.2 million in fiscal 1968/69 to R25.8 million in fiscal 1971/72. This increase accounted for 61 percent of the growth in the homeland. If this element is factored out, then the gross domestic product in fiscal 1971/72 would have been only about R38 million instead of R60 million. It is known that few Tswana work in the mines of their homeland, that the mines employ many white workers and are, of course, controlled by white-owned firms, and that the Tswana government does not participate directly in mineral earnings. The total income earned by Africans in Bophuthatswana was probably about R31 million in fiscal 1971/72,^[5] which is only slightly more than half the gross domestic product. Although estimates have been omitted from the appropriate place in the income table (column 5, table 6.1), it is reasonable to suppose that much of the R29 million difference represented wages, salaries, and capital income paid by the white mining corporations. The data thus establish that the most dynamic area of Bophuthatswana's economy has had an extremely limited impact upon African incomes in the homeland. Further, under the present institutional arrangements growth in this sector should not be viewed as contributing in a major way to the development of the homeland.

The economic profile of KwaZulu is more representative of the homelands in general than is that of Bophuthatswana. Agriculture is the foremost activity. Modern sectors — industry and construction, transport and communication, and commerce and finance — have grown absolutely and proportionately, but remain relatively inconsequential. Public administration and education have expanded with rising spending on public services in the homelands. Mining is trivial. The reasons for the loss of over R4 million per year in mining output after 1968 are not clear; it appears either that an operation was terminated or that a boundary change removed a mineral deposit. However, a discovery of a major deposit of high grade anthracite was announced in 1975. Approximately 43 percent of KwaZulu's output in fiscal 1971/72 was generated by the subsistence sector; 57 percent arose in the private and public commercial sectors. This pattern bears out the continuing importance of traditional activities in KwaZulu, the extensive role of the homeland and the Republican governments in providing income and employment, and the inertia of private industry and commercial agriculture.^[6]

[5] R31 million is an interpolation between the figures for income given in column 1, table 6.1. For figures on mineral production, see BENBO, Bophuthatswana, Economic Revue, 1975, 30.

[6] BENBO, KwaZulu, Economic Revue, 1975, 34.

In Bophuthatswana, subsistence production provided only 13 percent of the total; the private sector, dominated by big mining operations, was responsible for 61 percent.^[7]

Welfare and Growth

The chief measure of living standards in a nation or region is per capita income, although it is necessary to recognize that averages mask the realities of the distribution of individual incomes over high, middle, and low ranges. Also average yearly per capita income measures of welfare must be supplemented by using other economic and social indicators in order to obtain a rounded profile of the overall standard of living. Although calculations of homeland output and incomes are still only rough approximations, in 1973 the average per capita income earned inside the two homelands was R40 for the Tswana and R35 for the Zulu, or $56 and $49, respectively (see table 6.1). If these were complete estimates, they would rank among the lowest, if not as the lowest, in the underdeveloped world. But when commuter income is considered, the per capita yearly income of de facto residents rises to R140 ($196) for the Twsana and R92 ($129) for the Zulu. The higher figure for the former homeland is a consequence of its greater dependence on commuter as opposed to migrant workers. If all migrant income is added, then average incomes are R165 ($231) for Tswana and R145 ($203) for Zulu. Incomes for permanently absent de jure citizens are even higher. In response to a question in Parliament, the minister of Bantu administration and development said that the average per capita income of all Africans in South Africa was R166 in 1973.^[8] Although the Tswana are something of an exception, the per capita income of de facto homeland residents is about two-thirds that of all Africans. In other words, there is a sharp difference — partially closed by unmeasured subsistence income — between income in the homelands and African incomes in the Republic.

The area national income concept described in the previous section is an aggregate consisting of all internal income, all commuter receipts, one-fifth of migrant income, and one-twentieth of the earnings of permanently absent workers. In 1973 average area income in Bophuthatswana was R150 ($210) compared to R112 ($157) in KwaZulu. If average area output is arbitrarily doubled to allow for unrecorded subsistence income, and if no subtraction is made for possible overestimation of the remittances of absentee workers, then in 1973 Bophuthatswana had an average per capita income of about

[7] Idem, Bophuthatswana, Economic Revue, 1975, 30.

[8] Survey of Race Relations, 1974, 235. Related information on poverty datum levels and wages is also provided. Most of this is geared to the urban African worker and his family's needs, however, and is only indirectly related to the question of homeland welfare.

R190 ($266); KwaZulu had roughly R147 ($206). Even these maximal, generously biased numbers are not large enough to lift the average standard of living for homeland Tswana and Zulu beyond the margin of poverty and into the world's middle income range of $300 to $500 yearly per person. The level of income in Bophuthatswana and KwaZulu is above that of the lowest stratum of international poverty — that of Bangladesh, India, Ethiopia, Haiti, and Indonesia — and below that of middle-range developing nations such as Mexico, Iran, Zambia, or Taiwan.^[9] The extraordinarily low levels of average income earned in the homelands indicate how enormous is the task of raising productivity to global middle-range standards. Considered as potentially independent states, Bophuthatswana and KwaZulu are in the anomalous position of having income levels that approach the intermediate range, but output levels that are as low as any in the world.

The standard of living in an area is not adequately represented by per capita income alone. Other economic and social indicators must be considered to obtain a complete picture of welfare. The homeland territories are sometimes called "rural slums" or "rural ghettos," and most casual visitors and professional observers have commented unfavorably upon living conditions in them. Their reactions are doubtless heightened by the juxtaposition of homeland poverty and white opulence. White average yearly income is thirteen times as great as African average income in South Africa.^[10] but homeland Africans are poorer than most and the middle and upper strata of white society lead a suburban existence that is affluent and physically comfortable by American or European standards. The gradient in living styles, readily visible by travelling the few miles between the white suburbs of Durban, Pretoria, and Johannesburg and the adjacent Tswana and Zulu homeland exurbs, is as steep as one can find anywhere in the world.

Social indicators confirm the disparity between homeland economic levels and those outside. Health and nutrition measures reveal differentials that are the product of decades of inattention and unconcern. In South Africa as a whole in 1973 there were 10.0 hospital beds per 1,000 whites, 5.6 per 1,000 urban blacks, and 3.5 per 1,000 homeland blacks. Of about 15,000 medical practitioners and specialists in South Africa, 300 were African, and only 45 whites and 9 Africans practiced in the homelands.^[11] (Another statement, obviously based on a second definition of homeland practice, placed

[9] This impression was derived not only from abstract numbers, but from the subjective method of driving through and observing living areas in many of the main blocks of Bophuthatswana and KwaZulu, on and off the main roads.

[10] SPROCAS Economics Commission, Power, Privilege and Property (Johannesburg, 1972), appendices B and C, 115–117.

[11] Cora E. Erasmus, Director of Strategic Planning, Department of Health, as quoted in Muriel Horrell, The African Homelands of South Africa (Johannesburg, 1973), 162.

the numbers at 471 white and 70 African doctors.)^[12] The average number of beds per 1,000 people seems to have declined in Bophuthatswana from 7.0 in 1972 to 6.5 in 1973 and 6.2 in 1974 as a result of a diminishing number of beds available in government hospitals. For KwaZulu, there appear to have been 3.6 beds per 1,000 members of the population in 1972 and 4.1 per 1,000 in 1973. In 1973 there were reported to be 74 occupied medical posts and 1,757 occupied nursing posts in Bophuthatswana; comparable data for KwaZulu are 191 and 4,256.^[13] People living in the homelands either receive no Western medical treatment when they are ill, relying instead on home remedies or indigenous medical practice, or they must travel some distance, perhaps with sick infants and children, to towns. There are three white dentists practicing in the homelands and only one African dentist, trained overseas, in all of South Africa.^[14] Infant mortality rates in the homelands are very high, but precise records are not kept. For all Africans, infant mortality has been estimated at 140 per 1,000 live births, compared to 22 per 1,000 births for whites and 40 per 1,000 births for Asians.^[15] African male life expectancy is about fifty-two years, contrasted with sixty and sixty-five years for Asians and whites, respectively.^[16]

Malnutrition is common, particularly among persons who are not living within family groups: men away from home, children who have been left with relatives while their parents work in white areas, and women and children trying to survive upon irregular or nonexistent remittances from migrant fathers. Homeland residents suffer from a number of diseases directly or indirectly related to an inadequate, unbalanced diet. High rates of kwashiorkor, marasmus, pellagra, and vitamin deficiency illnesses have been reported.^[17] In addition, tuberculosis is endemic; eye diseases are a major

[12] Survey of Race Relations, 1974, 383, quoting an answer of the deputy minister of Bantu administration and development to a question in the Legislative Assembly.

[13] BENBO, KwaZulu, Economic Revue, 62; idem, Bophuthatswana Economic Revue, 54.

[14] Survey of Race Relations, 1974, 384, quoting the minister of Bantu education.

[15] J. L. Sadie, "The Costs of Population Growth in South Africa," South African Journal of Economics, XL (1972), 112.

[16] Idem., Projections of the South African Population (Johannesburg, 1974), appendix tables.

[17] SPROCAS, Power, Privilege and Poverty, 24–25; M. L. Neser, "Can We Eradicate Malnutrition in South Africa?" South African Medical Journal, XXXIX (1965), 1161; P. J. Pretorius and H. Novis, "Nutritional Marasmus in Bantu Infants in the Pretoria Area," ibid., 237–238, 501–505; F. W. Quass, "The Nutrition of Preschool and Primary School Children," ibid., 1137; M. E. Edginton, J. Hodkinson, and H. C. Seftel, "Disease Patterns in a South African Rural Bantu Population," ibid., XLVI (1972), 974; J. V. O. Reid, "Malnutrition," in Peter Randall (ed.), Some Implications of Inequality (Johannesburg, 1971), 38; J. G. A. Davel, "The Incidence of Malnutrition among Bantu Children," South African Medical Journal, XXXIX (1965), 1148; Alexander R. P. Walker, "Biological and Disease Patterns in South African Interracial Populations as Modified by Rise in Privilege," ibid., XLVI (1972), 1128. See also above, 20–21.

affliction.^[18] African children are on average underweight and shorter than normal.^[19] Although it is theoretically possible for people to feed themselves adequately even with very low levels of income, the patterns of production and consumption that have developed in the homelands are not conducive to good diet. Traditional agriculture and pastoralism have declined, but the loss of customary foodstuffs has not been offset by sufficient quantities of nutritious foods. Low levels of purchasing power, coupled with the lack of nutritional education, the lure of advertising and quick food counters in urban centers, and the shortage of good cafeterias and restaurants serving blacks, all contribute to a very unsatisfactory dietary picture.

The greatest inequality that homeland and other Africans confront in South Africa is not that of income or health. White South Africans benefit from the array of public services associated with a modern state, although they do lag behind residents of Western welfare states in the availability of many social services. From municipal governments, whites receive clean, paved streets, piped water, sewage disposal systems, and power. The provincial and national governments provide old age, disability, and war pensions, maintenance grants for needy families, and institutions to care for orphans, the mentally ill, and the aged. Africans receive most of these services, but at segregated and markedly lower levels of financing, care, and availability. Homeland towns and rural locations offer few urban amenities and services.

In addition to all of these economic and welfare inequalities are the disorganization of family lives caused by the migratory labor system, the many devices of petty and grand apartheid with their attendant indignities and inequities, and the constraints upon political expression and personal mobility. Poor housing, inadequate police protection, and the use of the police and the courts to enforce legal segregation further diminish African living standards.

The Republic's efforts to deal with these welfare problems are tied to the expansion of the functions of the homeland governments and increases in their funding. The central government has gradually taken over health care from provincial and private authorities. In 1965 the Department of Bantu Administration and Development assumed responsibility for capital expenditures for hospitals, major items of medical equipment, and maintenance of existing medical facilities. In 1970 control over health services and hospitals was assumed, and in 1973 a gradual takeover of mission hospitals was begun.

[18] Survey of Race Relations, 1972, 1973, and 1974, chapters on "Health."

[19] P. M. Leary and J. E. S. Lewis, "Some Observations on the State of Nutrition of Infants and Toddlers in Sekhukhuniland," South African Medical Journal, XLVI (1972), 1157; B. S. Richardson, "Mortality Rates in Preschool Children," ibid., XLII (1963), 964; F. M. Leary and D. Obst, "The Use of Percentile Charts in the Nutritional Assessment of Children from Primitive Communities," ibid., XLII (1969), 1165–1168; Rand Daily Mail, 11 Nov. 1969.

New facilities are being added.^[20] Spending on health services for homeland Tswana, including government, mission, and provincial hospitals, and clinics, rose from R5.2 million in fiscal 1970/71 to R10.6 million in fiscal 1974/75. Similarly expenditures for the Zulu approximately doubled during the same period from R11 to R21 million.^[21]

Presumably, as health departments are created, responsibility for hospitals and clinics will gradually be transferred to the homeland governments from the agencies of the Republic. One proposal calls for a number of linked clinics and hospitals to provide health care to homeland districts. Some emphasis is being placed on nutritional education in the schools, and there is a plan to provide skimmed milk powder to local authorities in the homelands. Training of doctors and other medical personnel has for some time been undertaken at the Natal Medical School, at black colleges, and in technical schools. A new medical school, to be opened at GaRankuwa near Pretoria in 1978, will become the major training school for Africans. The Bantu Trust and the Bantu Investment Corporation are devoting substantial sums to creating townships and providing them with housing, businesses, theaters, and sports facilities.

Because national output, income, and population data on the homelands, as well as information on related measures of development and welfare, are sparse and incomplete, there is no way of precisely determining the rate at which the homeland economies have grown in per capita terms. Given the evidence available, and relying upon the impressionistic estimates of those in and out of government who are familiar with the homelands, the average rate of real output expansion per capita in the homelands was in the neighborhood of 1 percent per year between 1960 and 1975.^[22] The homelands vary among themselves in performance, and there were considerable fluctuations from year to year attributable to the weather. Recession in the South African

[20] Horrell, African Homelands, 160–161.

[21] BENBO, Bophuthatswana, Economic Revue, 54; idem, KwaZulu, Economic Revue, 62.

[22] This conclusion is based on a compound rate of output growth of around 7.0 to 8.0 percent (see table 6.3), less inflation of about 3.0 to 3.5 percent and population growth of about 3.0 percent. The inflation rate is taken as roughly equivalent to the change in consumer prices in South Africa since there is no price deflator for homeland output. It is unlikely that the average change in per capita real output was much greater than 1.0 percent, and it may well have been nil, at least in most areas. There is a good deal of discretion involved in formulating the accounts, and information about crop output, a major component of homeland output, is highly unreliable. It requires a certain amount of optimism to believe that output in the public sector, commerce and trading, industry, and (outside Bophuthatswana) mining has risen in real terms by the 4.5 percent per year necessary to offset population growth of about 3.0 to 3.5 percent. On the other hand, those familiar with the homelands, in and out of the government, are reluctant to say that nothing at all has happened in the past fifteen years, and especially since 1970.

economy, which reduces purchases from the homelands and also restrains the amount of public funds that the Republic is willing to commit to homeland finance, can also depress the actual output of these small, peripheral regions temporarily below any growth trend line. During the early 1970s and afterwards, there appears to have been some acceleration in growth, attributable to increased public sector spending, the building of homeland growth points, construction of infrastructure and housing, and, in Bophuthatswana, mining.

Table 6.3 summarizes the available statistics on growth in homeland output. There are three general patterns. First, as data for fiscal 1967/68 and fiscal 1968/69 indicate, growth has not been smooth, largely due to agricultural instability. Second, some acceleration in growth rates occurred at the end of the 1960s. Last, the statistics for all the homelands, and for Bophuthatswana and KwaZulu in particular, do show growth rates, uncorrected for inflation and population growth, in the range of 6 to 9 percent, or of about 12 percent for Bophuthatswana if its enclave mining operations are included. On balance, even with the mining sector omitted, Bophuthatswana has done somewhat better than KwaZulu. After allowing for higher rates of inflation, about 5 to 7 percent, there were still sharp jumps in output

in both homelands between fiscal 1970/71 and fiscal 1971/72. In the mid-1970s continued public sector spending, especially on construction, education, and industrial development, has probably elevated real per capita output growth rates somewhere into the 7 to 10 percent range. Of course, whether this growth can be sustained, whether it has occurred in the sectors that the Tswana or Zulu might wish, and the extent to which the benefits of this growth have been distributed across the homeland populations remain questions to be considered.

The rise in real average yearly income levels in the homelands has probably exceeded long-run output growth of 1 percent per year, as the real incomes of commuter, migrant, and permanently absent workers have been rising in recent years.^[23] Considerable controversy exists about the most appropriate measure of change in African real incomes in the Republic and of the gains of other groups; thus it is difficult to choose a single estimate. Because there is no accurate information about the remittance habits of absentee workers, it is also very difficult to say what the impact of rising black wages in the Republic has been on the homelands. Commuter income gains have directly influenced homeland incomes, but migrant and permanently absent workers may or may not transmit wage gains to the homelands.

The three income measures reported in Table 6.1 suggest that there has been modest growth in homeland average incomes, much of which is due to increased earnings of absentee labor. Average yearly per capita income from internal sources rose from R29.8 to R40.2, or 34 percent, in Bophuthatswana between 1960 and 1973, and from R28.4 to R34.7, or 22 percent, in KwaZulu during the same period. With a 3 percent annual rate of inflation, these apparent gains actually represent modest declines in real per capita income. In contrast, there was material improvement in real incomes when outside earnings are included, again using estimates: the average incomes of residents and migrants combined appear to have risen by about 4 percent in real terms after 1960, and the incomes of de facto inhabitants may have grown slightly faster. These estimates, based on the homeland income accounts compiled for BENBO, are somewhat higher than those made for the whole black population of South Africa by other authorities. Steenekamp estimates that the average growth of real income of all South African blacks was about 3.18 percent between 1964 and 1970. On the other hand, there is some evidence that black wages and incomes rose faster than this—and

[23] Incomes and wages of the various population groups are constantly monitored and estimated by various organizations and reported in the annual editions of A Survey of Race Relations in South Africa published by the South African Institute of Race Relations. See Survey of Race Relations, 1973 , 197ff, 206; also ibid., 1974 , 231ff.

somewhat faster than white wages and incomes—in the early 1970s.^[24] The very rapid inflation of the mid-1970s (14 percent from mid-1974 through mid-1975), which particularly affected food and essential consumer items, probably eroded some of these total and relative black income gains.

The following conclusions appear justified on the basis of the imperfect data available. By all measures, average yearly per capita incomes of homeland Tswana are significantly above those of homeland Zulu. The income gains of commuters and migrants have been responsible for raising average levels of living in Bophuthatswana and KwaZulu by a maximum of 3 to 4 percent per year. Because of the stagnation of the internal economy, incomes in the two homelands have become ever more dependent during the last fifteen years upon earnings that "trickle down" from the Republic's migrant labor system. Since these remittances would at best allow homeland incomes to keep pace with those of other blacks, it is unlikely that homeland residents have improved their position relative to other blacks. There has probably been some slippage. The remittance streams to the homelands, even when supplemented by rising public sector spending, have been insufficient to ensure that these laggard regions grow rapidly enough to keep pace with the growth of the central economy.

The Absentee Labor System and the Income Multiplier

Against this backdrop of internal underdevelopment and relative regional stagnation, it is evident that the genesis of income-creating forces lies outside the homelands in the private and public sectors of the Republic. The overwhelming importance of these external links distinguishes the homelands from the remainder of the world's less developed nations, excepting Lesotho. Few other independent or potentially independent countries lie in such close proximity to a rapidly modernizing industrial economy. Where other small economic units have at least subsistence levels of output, and rely upon exports of goods and services to finance imports, the homelands have sub-subsistence output levels and rely upon exports of labor. Moreover, the homelands (except the Transkei) are fragmented and more open to external economic influences than other small states. Nowhere else is the degree of monetary, fiscal, and employment dependence apt to remain as great after political independence.

In the theory of economic growth, the investment multiplier plays a vital role. Basically, the multiplier relationship recognizes that an addition to

[24] Ibid., 1974, 235–236.

expenditure in the form of investment has a cumulative impact on incomes as it passes from hand to hand: from investor to wage-earner, from wage-earner to shopkeeper, from shopkeeper to employee, and so on.^[25] This notion has prompted an analysis of homeland underdevelopment called the "leakages" model. The model seeks to explain why growth in the South African economy does not automatically spill over into the homeland regional economies. It does not explicitly analyze the institutional, structural, and resource barriers to development, but lays stress upon the income flows that are a product of these conditions. This idea has wide currency in planning and academic circles in South Africa and constitutes part of the current conventional wisdom that underlies the government's analysis of homeland underdevelopment.^[26] The chief fear is that public spending and private investment in the homelands will "leak" back immediately into the white economy, which would itself then experience the multiplied effects on income, consumption, savings, and employment rather than the homelands.

Homeland economies are linked in the macroeconomic sphere to the South African economy as shown in Figure 6.1. One block represents Bophuthatswana, KwaZulu, or any homeland; the other represents the economy of the Republic. The flow chart shows the major connections between the two units, but omits some internal circulations. The major flows are shown as solid lines and weak, underdeveloped linkages are shown as dashed lines. The chief real outflow or export from the African economic regions is labor: commuters, migrants, and permanently absent workers. These persons spend a large part of their earnings for goods in white and Asian stores and for services ("local purchases"). A portion is diverted to taxes, the estimated value of which is supposed to be converted automatically through the developmental agencies of the Republican government into expenditures inside the homelands.

[25] The standard form of the investment multiplier is D Y = k·D I, where an initial change investment, D I, has a multiplied effect, k, on income, Y. If each person withholds 1/5 of his receipts (as "savings") from the income stream, then k = 5. where s is the proportion of an increment to income that will be saved, k = 1/s. Any injection of money into an economy has a similarly magnified result (neglecting the government and trade factors).

[26] The Tomlinson Commission recognized the problem of deflections of the homeland income stream, but apparently did not give the matter formal treatment. See Summary of the Report, 150–151, and elsewhere. Lombard had made extensive analytical use of the national income flow concept. See, for illustrations, Lombard and van der Merwe, "Central Problems," 13; J. A. Lombard, "Political and Administrative Principles of Homeland Development," Bantu, XVIII (Feb. 1971), 28; idem., "Background to Planning the Development of Bantu Homelands in South Africa," presented at a Conference on Accelerated Development in Southern Africa, Johannesburg, 1972. More concrete is G. D. K. Fölscher, "The Economic and Fiscal Relationships of the Transkei vis-à-vis the Rest of the Republic as Determinants of its Economic Development," South African Journal of Economics, XXXV (1967), 203–218.

Figure 6.1
The Flows of Labor, Money, and Goods Between Homelands and the Republic
^a These proportions have been suggested by Lombard and van der Merwe 
and appear reasonable. See, "Central Problems of the Economic Development 
of the Bantu Homelands,"  Finance and Trade Review,  X (1972), II.
^b Regular commuters probably spent one-quarter to one-third of their incomes 
for transport and meals, and most shop in white or Indian stores 
near bus and train depots before returning home.

Commuters, migrants, and residents with some ties to a homeland take home or remit a fraction of their earnings, and these amounts constitute the most important component of homeland income. To these public sector transfers to the homeland governments and citizens (e.g., pensions) must be added. Consumers and businesses in the Republic may make purchases from homeland producers ("consumer expenditures" and "business purchases"), but they are small. These six items comprise the money inflows of the homeland from the Republic. Most funds go to homeland consumers or to a homeland government. Some are directed at homeland businesses, farms, and industries. As the remainder of the pathways make clear, these monetary movements have almost no secondary multiplicative or cumulative effects because they "leak" away immediately to non-African businesses and industries as "consumption expenditures." Only about 2 percent of this income is saved. As the dashed lines show, some consumption and public funds may go to African farms, shops, and industry, but the proportion is very low. Similarly, homeland businesses and industries may generate some monetary income that remains in the homeland as factor incomes or payments to local farms or firms, but there will be substantial leakages from their payrolls back to the Republic. Virtually all of their capital expenditure will be made outside the homeland. Local enterprises may pay some taxes to a homeland government. Most tax receipts come from local citizens and are then disbursed as teachers' or civil servants' salaries or are paid to contractors who meet local payrolls (these flows are not shown on the chart).

The pattern of circulation depicted on this chart has few dynamic effects inside a homeland like Bophuthatswana or KwaZulu. The absentee labor system and the dependence of the homeland populations on outside purchases, to the detriment of local farmers and businessmen, dominate the scene. The flow of goods out of the homeland is paltry compared to the reverse flow of goods from the Republic, leaving the regional economies with sizeable trade deficits that can only be covered by absentee labor earnings and fiscal supplements from the Republic. The circulation of funds does not stimulate local enterprise and employment. The absentee labor mechanism has become a rigidly institutionalized system rooted in the labor needs of the white economy and the money needs of Africans. The legislative and administrative details of the system, its origins, its influence on black and white society, and their interrelationships are well known. The labor pattern and the income flows associated with it have been little affected by the recent developmental efforts of the Republic, nor will greater homeland political independence lead to any sudden shifts.^[27] It is probably true that the more

[27] In general, all small, underdeveloped nations or regions, not isolated by distance or financial and trade barriers from larger, dynamic centers, experience similar problems. See Michael Ward, "Economic Independence for Lesotho?" Journalof Modern African Studies, V (1967), 363ff. The most recent major study of absentee labor in South Africa is Francis Wilson, Migrant Labour in South Africa (Johannesburg, 1972). The most interesting parts of his book are those derived from the author's personal investigations of the system. These impressions provide a sense of immediacy lacking in more abstract writings. Other recent general sources include: M. Ernest Sabbagh, "Some Geographical Characteristics of a Plural Society: Apartheid in South Africa," Geographical Review, LVIII (1968), 1–28; John C. Williams, "Lesotho: Economic Implications of Migrant Labour," South African Journal of Economics, XXXIX (1971), 149–178.

rapid the growth of the Republic's economy is and the greater urban black and migrant participation in those gains are, the more present tendencies will be thrown into relief. With commensurate internal economic development, however, the dependence of the homelands on the Republic's economy will be somewhat changed in form, although not reduced. Greater goods exports from the homelands may be expected, and the composition of absentee labor may conceivably shift from wage-earners resident in the Republic to commuters, following the shift of black labor-intensive industries to border industry sites near the homelands.

Although it generates needed income for many homeland families, the absentee labor system has a number of negative effects on the homelands' potential for development. The male members of the labor force are withdrawn for long periods from participation in local economic activities, creating a concavity in the age distribution of the resident population. The laborers' oscillations between their homes and a variety of work places do not encourage the acquisition and retention of useful skills. There is little possibility of meshing early education, on-the-job training, and work experience to create valuable human resources. The fragmentary urban skills accumulated by absentee workers are not usually applicable to tasks in the home area. Consumption habits and the superficial traits of urban life, rather than economically useful knowledge, are far more likely to be brought back.

If a fraction of the absentee workers were required to remain permanently at home, total output in the homeland would rise, but average income would fall. The existence of relatively high earnings on the outside constitutes a barrier to enlisting manpower in internal development, unless the returns from participating in a local occupation rise in one step to a level competitive with earnings outside. In concrete terms, a worker who can obtain 100 wage units outside his homeland will not be drawn to employment in local agriculture or industry if it yields a net income of, say, only 50 or 75 units. Economic theory suggests that in a free labor market returns at the margin to labor in the homelands should equal earnings from absentee work. But the market is not free and too much labor is available in the homelands already. To raise internal agricultural or industrial productivity to equality with outside earnings in one step, on a large enough scale to reabsorb a

noticeable volume of labor, would require immense expenditures and substantial structural changes in both the homelands and the Republic. For those who are mobile the income advantages of external work constitute an opportunity cost barrier to the success of development projects aimed at providing them alternatives at home. Agricultural extension workers, for instance, feel they are limited to counselling a few men, most of them old, and women who tend gardens and small maize patches while their husbands work in the central economy. Extension workers and others who seek to increase homeland employment are correct in realizing that headway can be achieved only by creating jobs that offer incomes that are competitive, or nearly so, with those obtainable in the Republic.

The Homeland Public Sector

The largest single component of monetized area output in the homelands as a group is the public sector. Agriculture would be the most important economic activity if subsistence and commercial production were added together. But among monetized economic activities the homeland governments overshadow the private sector. Bophuthatswana's basic economic structure is somewhat anomalous because of its poor agricultural base and high mineral output. In fiscal 1971/72 the subsistence sector accounted for only 15 percent of output, compared to 24 percent for the public and 61 percent for the private sectors. In KwaZulu in fiscal 1971/72 about 24 percent of output was generated in the private market sector and 43 percent in the subsistence sector.^[28] The public sector is also responsible for virtually all of the capital formation in the homelands, although the agency system does bring some white investment in the growth points. Most of this spending comes from the Bantu Trust, but the homeland governments are increasingly engaging in expenditures for housing, fencing, wells, roads, schools, utilities, and other fixed capital items. Provincial governments and departments of the central government also make some investments, presumably in housing and infrastructure.

The expenditures of the homelands are contained in their annual budgets. These budgets show the growth in spending by the homelands, and their composition affords some insight into the relative emphasis on different activities. In summary form, the budgets of Bophuthatswana and KwaZulu distinguish internal and external sources of funds and show expenditures according to departments as allocated by the legislative assemblies. These initial allocations and revenue estimates are sometimes changed or supplemented during the financial year, but without distorting the general pattern.

[28] BENBO, Bophuthatswana, Economic Revue, 1975, 30; idem, KwaZulu, Economic Revue, 1975, 34.

The homelands obtain revenue from two sources: 1) taxes, fees, and licences directly under their control, collectively referred to in the estimates as Part I; 2) sums voted by the Parliament of the Republic, called Part II revenues. Until fiscal 1975/76, the Republic's contribution was composed of: (i) A sum based on the cost of services at the time of transfer to the homeland government, less revenue from certain services accruing at the time to the homelands, and less the salaries and allowances of seconded white personnel. This sum is referred to as the "statutory grant." (ii) A grant, determined annually and drawn from the Consolidated Fund, referred to as the "additional grant." In addition, the Department of Bantu Administration and Development pays a supplementary amount representing general overhead expenses, such as the salaries of white personnel on loan. The amount is small, amounting in fiscal 1974/75 to 9.2 percent of total revenue for Bophuthatswana and 6.1 percent for KwaZulu.

In late 1974 Vorster announced a new formula for "providing funds" for homeland governments that would furnish "a more accurate reflection of the degree to which their expenditure was financed from sources which were due to them." These suggestions in part met demands by homeland leaders that their fiscal footing be made more secure and also made it clearer to whites that payments to the homelands were transfers of funds to which blacks had legitimate claims, not burdensome handouts. Vorster was careful to point out that although the formula did not directly affect the amount of money available to homelands, it would place "homeland governments in a much better position to project revenues ahead and plan their expenditure programmes." The change involves a transfer of some indirect taxes, such as customs, excise, and sales taxes on goods consumed by blacks. Further grants would be made from taxes paid by companies or branches of companies in the homelands (and possibly in the border industrial areas), whether controlled by blacks or not.^[29]

The Bantu Laws Amendment Act of 1975 embodied these modifications. The statutory grant was changed to a 31 March 1974 base for all the homelands, in order to reflect changes in the costs of functions that had been handed over earlier and to permit all of the homelands, some of which had acquired functions earlier on a lower cost basis, to be treated equally. This amount will continue to be supplemented by the costs of any new programs subsequently transferred to the homeland governments. Except for African-controlled companies, the revenues of which are already subject to homeland taxation, homelands will also be credited with taxes paid by companies operating in their areas. In addition, allowance will be made for customs, excise, and sales duties derived from the area of the homeland. Payments to white

[29] Vorster to Federated Chamber of Industries, 2 Oct. 1974, Star Weekly, 19 Oct. 1974; Comment and Opinion (Pretoria), 14 Oct. 1974.

personnel will no longer be separated from the statutory grant. The effect of these changes will be to raise the regular or normal amounts paid to each homeland while reducing the discretionary additional grant. In fiscal 1975/76, the annual grant to the homelands was R163.6 million, and the supplementary grant totalled R86.5 million, together equalling R250 million. In addition, the revenue vote for Bantu administration provided R49 million for health services and R14 million for the salaries of seconded whites. Significantly, from the Loan Account the Bantu Trust received R116.4 million, up from R90.4 million the previous year, for developmental projects. Much of this money will be channelled through the investment and mining corporations.^[30]

The new allocation may help to reduce the dependency of the homelands. One approximate global estimate asserts that the homelands' share of customs and excise taxes comes to about R100 million annually, and that mining taxes and royalties will bring a further R100 million, the two sums together being roughly equal to the R199 million voted for all the homelands in fiscal 1973/74. However, the new law does not empower a homeland government to tax South African companies operating within its borders; South Africa will continue to tax such companies and remit revenue to the homelands. Mineral royalties will presumably still be paid over by companies to the tribal and regional authorities that have received them in the past and to the Bantu Trust. The new law may conceivably benefit Bophuthatswana if mining profits are treated as industrial earnings. There is considerable mining there, and potential for more.^[31] Mangope now can at least employ a line of argument not previously available to him.

The sources of revenue for Bophuthatswana and KwaZulu (for the years for which they are available) are shown in table 6.4. In both instances there are clear changes in the relative importance of the different categories of funds. From fiscal 1972/73 to fiscal 1974/75 Bophuthatswana's statutory grant remained more or less constant, then in fiscal 1975/76, under the older accounting practices, the transfer of a new Department of Health brought with it added revenue equivalent to that which the Republic had been spending at the time of the transfer. Similarly, for KwaZulu, the statutory grant remained around R20 million from fiscal 1972/73 to fiscal 1975/76. Although the statutory grants remained nearly constant, the level of grants from Parliament and revenues from internal sources rose. The additional grant to Bophuthatswana increased six-fold and that to KwaZulu twenty-fold. Internal revenues have almost doubled in Bophuthatswana since fiscal 1972/73 and have climbed by nearly 70 percent in KwaZulu during the same period. The proportionate importance of these homeland receipts and taxes

[30] Survey of Race Relations, 1975, 121–125.

[31] Star Weekly, 19 Oct. 1974.

has not risen however, because of the more rapid growth of the supplementary appropriation. The statutory grants declined relatively as well. The effects of the fiscal reforms can be seen in the separate calculations for fiscal 1975/76. It is probable, however, that the statutory amounts will once again diminish in weight—although not so precipitously as before—because the new excise and profit contributions will not be elastic enough over time to provide sufficient new revenues to permit homeland budgets to grow as rapidly as they have been growing, and as they will need to grow in the future.

The homelands will continue to receive much of their revenue from the Republic for at least the next two decades. Only a major shift in policy that would allow the homeland governments the fiscal autonomy to tax white businesses and mines, and would also allow them to collect income taxes from their absentee populations, could create a situation in which fiscal independence would prove feasible. The discretionary amount remains important even after the recent reforms, and there is no guarantee that an independently-minded homeland would not be subject to the sanction of budget cutting, including part or all of the statutory amount.

Rather surprisingly, the rapid growth of spending on the homelands—matched only by increases in defense spending—has not been challenged by white public opinion or politicians. The government has been successful in defending the amounts as necessary for internal security and for the alleviation of the pressures of black migration toward white cities. The spending levels can also be justified as ways of meeting foreign criticism of racial segregation and as humanitarian necessities. Although the homelands will have to rely upon funds that are subject to decisions beyond their control, it is unlikely that such support will be withdrawn capriciously. Nonetheless, dependence on the Republic for over 80 percent of revenues is a major factor constraining the political and economic independence of the homelands.

Since the autonomy of the homelands will depend vitally upon the capacity of their governments to finance themselves from internal sources, their present tax bases merit examination. Table 6.5 shows the sources of revenue under the control of Bophuthatswana and KwaZulu in two recent years. In both homelands the general personal tax is the preponderant source of income. The respective homeland assemblies have imposed new per capita taxes, R2.50 in the case of the Tswana and R3.00 for the Zulu; the assemblies will depend upon the Republic to collect these amounts from permanently absent persons. The increases appear to have added several million rand to the homeland revenue accounts, and the share of taxes in total revenue has been rising. The other important revenue item is township rental income. The remaining income consists of licensing fees and income from small charges for public services rendered by government

departments. In KwaZulu, forest revenues and income from sugar cane and fibre estates have amounted to over R1 million in recent years (included in "Other"). Since in both homelands over four-fifths of revenue comes from a fixed personal tax and from housing and hostel rentals, it is clear that raising revenue from these sources will be painful and politically unpopular. Given the already low incomes of many homeland citizens, tax or rental increases will have deleterious effects on welfare. Absentee workers and those in the homelands who do not feel any particular allegiance to the new political entities are not likely to greet new taxes—particularly when governmental services are so woefully inadequate in rural locations and in the burgeoning homeland cities—with any degree of enthusiasm. Probably the most that can be expected is that homeland internal revenues will rise rapidly enough to avoid even greater proportionate reliance upon the Republic's subventions.

Both homelands face practically insurmountable obstacles in increasing their autonomy by bringing a greater proportion of revenue under their

control. Even if every man, woman, and child in Bophuthatswana were to pay the per capita tax, the yield would be only in the order of R4.25 million per year — only 9.4 percent of total revenue for fiscal 1975/76. Similarly, even if all the de jure population of KwaZulu were to pay the tax, it would raise something in the order of R12 million — 13.9 percent of total revenue in fiscal 1975/76. In fact, the new tax in KwaZulu will raise about R1 million in fiscal 1975/76, about 33 percent of the net gain in Part I revenues. Similarly, Bophuthatswana's new tax will raise R1.2 million in fiscal 1975/76, 73 percent of the net gain coming in Part I revenue. Although these proportionate increases are impressive, revenue realized directly by the homelands fell in the case of Bophuthatswana from 25 percent in fiscal 1972/73 to 17 percent in fiscal 1975/76, and in the case of KwaZulu from 31 percent to 19 percent during the same period. Every delegation of function to the homelands and every increase in cost due to inflation only makes more difficult the gaining of greater control over revenue, and, therefore, greater freedom to initiate policies that would provide increased services, not merely maintain existing ones.

The fiscal 1974/75 budget totals and estimated populations show that Bophuthatswana spent R28 per resident and R16 per member of the de jure population; KwaZulu spent R27 per resident and R14 for each de jure citizen. The amounts are up by about four-fifths in nominal terms since 1970, perhaps by half allowing for inflation. In contrast, the Republic was spending R296 per white resident in 1970, or R53 for each person in the Republic. By 1975 these averages had risen considerably, but these lower figures suffice to indicate how far the homelands currently lag behind South Africa in public spending.

The funds that the homelands have at their disposal are disbursed by their several departments: seven in Bophuthatswana, six in the case of KwaZulu. Table 6.6 shows the allocation of funds among these departments. The addition of a Department of Health in Bophuthatswana in fiscal 1975/76 sharply reduced the share of the Department of the Interior, which had previously contained some health functions. With the bias in the pattern of spending toward the Departments of Works and Education, and the absence of a defense account, these budgets resemble those of states (provinces) or cities more closely than those of national units. As such, they delineate the regional character of the homeland governments.

The budgets of both homelands show how rapidly their functions and spending authority have expanded. Between fiscal 1972/73 and fiscal 1975/76 expenditures have almost exactly tripled — from R16 million to R50 million in Bophuthatswana and from R31 million to R93 million in KwaZulu. Bophuthatswana's rate of increase accelerated from 22 to 60 percent per year,

while KwaZulu's expenditure went up by about 40 percent per year. The proportion of the budget spent by each department shows differences from one homeland to another, but there are broad similarities. Authority Affairs and Finance, Community Affairs, Interior and Justice are essentially administrative departments and have maintained steady proportions of the budgets. Surprisingly, Agriculture also shows a constant rather small share (about 9 percent in each case) or expanding budgets. The rise in expenditures on new agricultural activities and continued spending on physical planning have not given Agriculture an increased share. The two largest spending shares are those of Works and Education, together taking 61 percent of the budget in Bophuthatswana in fiscal 1975/76 and 71 percent of the budget in KwaZulu that year. In Bophuthatswana, Works maintained a constant high share and Education fell a little. In KwaZulu the share of Works rose by 9 percent between fiscal 1972/73 and fiscal 1975/76, while that of Education fell by 5 percent.

The heavy increase in expenditures of the Departments of Works was made up in part by the cost of the "establishment of townships," to quote the budget item. In Bophuthatswana the figures also show increases in expenditures on townships, but less of a preoccupation with them than in KwaZulu. From fiscal 1972/73 to fiscal 1973/74 expenditures on townships actually fell by 8 percent but rose by 29 percent from fiscal 1973/74 to fiscal 1974/75 and by a further 16 percent in fiscal 1975/76. Increases in expenditures on housing took a much smaller proportion of the total increase in the Works vote, only 22 percent in fiscal 1974/75 and 17 percent in fiscal 1975/76. In KwaZulu 47 percent of the increase in the Works vote between fiscal 1972/73 and fiscal 1974/75 went to townships; between fiscal 1973/74 and fiscal 1974/75, 69 percent; and between fiscal 1974/75 and fiscal 1975/76, 57 percent

It is possible that these expenditures raise an ideological issue. The establishment of townships in the homelands is a highly controversial issue, being related to the resettlement policy that is aimed at reducing the numbers of Africans in white farming areas and the cities, especially those not employed there. It has been policy, at least until very recently, to compel as large a proportion as possible of the African labor force to migrate and to provide housing and facilities for dependents and the unemployed primarily in the homelands. Also, "black spot" removals have returned Africans to Bophuthatswana and KwaZulu. It is not possible to establish which expenditures are based on these policies. Nor is it clear why expenditures on new townships are relatively so much more important in KwaZulu than in Bophuthatswana. What is patent is that a great deal of the expanded expenditures on Works is going into social capital in the form of townships with electricity, water, and sanitation, and into the expansion of hospitals and schools. Whether or not this is the most efficient way to provide housing and urban services for

African populations, a major portion of homeland budgets will continue to be committed in this fashion.

The needs of the homeland populations for education, housing, and other urban services remain great. There are grave deficiencies to be made up even if the homelands want merely to ensure that every resident obtains a rudimentary education and has an adequate place to live. Highways, streets, power, water, and schools need to be provided in most areas. The homelands are also responsible in part for agricultural and commercial development. They must contend with inflation and with the expansion of their populations due to natural growth and the influx of persons from white farms and "black spots." Under these pressures, the homeland governments have sought increased funding from Parliament and have also searched for additional internal sources of funds.

Other Expenditures on the Homelands

Homeland governments are not the only agencies involved in heavy expenditures affecting the lives of homeland citizens, nor are they in a position to control many aspects of developmental policy. The Bantu Trust, set up in 1936, is generally responsible for land acquisition and consolidation for a homeland as well as for a number of social services. The Bantu Investment Corporation was established in 1959 by the Bantu Trust to promote investment, especially by Africans, in the homelands. The Xhosa Development Corporation performs the same functions in the Ciskei and the Transkei. A separate development corporation for the Tswana is now being created. The Zulu and other groups are expected to obtain their own investment institutions in the near future. The Bantu Mining Corporation is responsible for overseeing the activities of white mining firms in the homelands and for developing African mining operations. Table 6.7 sets out estimates of expenditures for the homeland governments, the Bantu Trust, and the development corporations. These core expenditures amounted to over R221 million in fiscal 1972/73 and rose to R418 million in fiscal 1974/75. The homeland governments have accounted for about 60 percent of current expenditures, but the Bantu Trust has spent large sums on land planning, housing, and infrastructure. Prior to fiscal 1975/76, substantial Bantu Trust funds were also spent on social services, including payments to the Department of Health for clinical and medical services. The decline in the trust's spending in fiscal 1975/76 apparently reflects, at least in part, the transfer of these activities to the homelands.

The expenditures of the three development corporations have grown proportionately more rapidly than those of the homelands and of the trust, their share of the total having risen from 8 to 18 percent since fiscal 1972/73.

Most of the corporations' activities are aimed at creating income and jobs, and their rising share of the totals shows a shift in emphasis in the bureaucracy from routine governmental activity toward growth-inducing activities. But these organizations are not black-controlled and development decision-making has not on balance gravitated toward the homelands to the same degree as has the performance of routine administrative tasks. The Board of Directors of the new Tswana Development Corporation is, however, equally divided between homeland-appointed black and Republic-named white members, including officials of the Bantu Investment Corporation. This change resolves a controversy over the management of development banks and may bring together homeland governments, private economic interests, and the Investment Corporation.

To these expenditures in the homelands must be added part of the budget of the Department of Bantu Education. In fiscal 1972/73, R28 million was spent on African education in the Republic and in fiscal 1973/74, over R35 million, an increase of one-quarter.^[32] Most funds were spent upon urban and rural black schools and their supporting systems, but some financed the homeland universities, and part covered examination, inspection, and advisory services in the homelands. The department spent in the neighborhood of R60 million in fiscal 1974/75, and, if one-fifth of this total can be assigned to the homelands, then about R10 million could be added to the core spending of the agencies listed in table 6.7. This would raise spending in fiscal 1974/75 to R430 million. In fiscal 1975/76, the Department of Bantu Education spent R10.4 million in the homelands, but this figure may not include general overhead expenses. Health absorbed R50.6 million, explicitly apart from the Bantu Trust accounts.^[33] Added to the core disbursements, these amounts yield a comprehensive sum of R591 million.

Adding the relevant expenditures directly affecting the homelands made by other departments of the Republic, the fiscal 1974/75 total could be as high as R450 to 460 million, and that of fiscal 1975/76 could be in the range of R610 to 620 million.^[34] If allowance is made, however, for expenditures connected with policies such as "black spot" removals that are of no clear benefit to Africans, the figure would be smaller, although how much smaller depends upon judgments on the merits of aspects of the homeland development program. Whether one uses the core spending figures, the overall totals, or some corrected interpretation of both, it is plain that spending in and on the homelands has climbed steeply since the late 1960s. The data are too fragmentary and not comparable enough, and the time span is too short,

[32] Survey of Race Relations, 1974, 342.

[33] BENBO, personal communication (xeroxed document), 19 March 1976.

[34] An incomplete estimate of R13 million for South African departmental spending in the homelands is given in Survey of Race Relations, 1975, 125.

to allow computation of meaningful growth rates or to permit projections. Homeland government spending rose by over 168 percent from fiscal 1972/73 to fiscal 1975/76; the corporations' disbursements rose by 403 percent. Total core spending rose by 27 percent, 49 percent, and 27 percent in the three year-to-year intervals between fiscal 1972/73 and fiscal 1975/76, implying an average compound growth rate of about one-third. At the same time, South Africa's cost of living rose by about 10 to 12 percent per year. Using this figure as a rough guide, the real value of these funds grew by only about 20 to 22 percent per year. African population growth of 3.5 percent would reduce this growth to about 18 percent on a per capita basis. Real increases in teachers' salaries or the pay of homeland civil servants, or the transfer of new functions to the homelands, would further shrink the available incremental funds. Nonetheless, even with these allowances, it is obvious that the trend of spending on the homelands is strongly upwards.

The homeland governments have not issued debt instruments or received foreign loans. Under the Second Bantu Laws Amendment Act of 1974, the minister of Bantu administration and development is empowered to guarantee the interest and principal of homeland debts. The question of foreign aid has been explicitly raised by Chiefs Mangope and Buthelezi following overseas trips to the United States and Europe. The South African government has approved the idea of external aid in principle, but still insists that it must enter the homelands through the Republic's channels. It is difficult to imagine exactly how external assistance would fit into existing financial practices and the developmental projects proposed by the white agencies. Aid grants from outside could lead to a reduction in the effort of the Republic by an equivalent amount. If foreign sources are able to transfer funds to the homelands, the preferred form would be grants, since the simple and inelastic tax systems of the homelands are unlikely to provide a basis for the interest payments and principal amortization required by loans. Mangope and Buthelezi may fare better by cultivating their own tax resources and affecting the Republic's allocative process rather than by pursuing the chimera of foreign grants. By campaigning for external funds, however, black leaders may induce the Republic to raise its own levels of expenditures for the homelands.