^** Since this article was written in final form, the U.S. Congress has passed (and overridden President Reagan's veto of) a sanctions package that bans all new U.S. investments in, and bank loans to, South Africa. It also bans imports on South African uranium, coal, steel, textiles, and agricultural products, and restricts U.S.-South African airline traffic. Similar but generally weaker sanctions have been adopted by several other states and by the E.E.C. The arguments below suggest that the immediate economic impact of these sanctions is unlikely to be great, although the political significance is considerable.

¹ “President Reagan's Executive Order,” Congressional Digest, October 1985, p. 231.

² “Is the Proposed Anti-Apartheid Act of 1985 Sound Strategy?” Congressional Digest, October 1985, pp. 232–55, at 239. Kemp's discussion is perhaps the clearest analysis by a major political figure of the consequences of sanctions.

³ On the whole, advocates of sanctions appear to be fairly oblivious to the economic consequences of their proposals for nonwhite groups—a point noted by their opponents. Among more thoughtful advocates who recognize that sanctions may well cause severe suffering for nonwhites, Representative Howard Wolpe (D-MI) addressed the issue during congressional debate, albeit somewhat gingerly:

[opponents of sanctions suggest] … that those who are advocating sanctions are really not very sensitive to the prospective suffering that might be created on the part of the black population. Nothing could be further from the truth. The tragic reality is that the policy of constructive engagement… is in fact producing more suffering … (ibid., 250).

⁴ The nonwhite community in South Africa includes “Asians” and “coloureds” as well as “Africans”; for expositional simplicity, I use the words “nonwhite,” “black,” and “African” interchangeably.

⁵ There are, however, several excellent models of the South African economy. In particular, see Lundahl, Mats, “Economic Effects of a Trade and Investment Boycott against South Africa,” Scandinavian Journal of Economics 86 (January 1984), 68–83.CrossRefGoogle Scholar The mathematical model in this paper is closest to Becker, Charles M., “Modeling Southern African Labor Markets: A Multi-Country General Equilibrium Approach,” Modeling and Simulation 16 (1985), 137–43Google Scholar, which, in turn, draws extensively on Lundahl, Mats, “The Rationale of Apartheid,” American Economic Review 72 (December 1982), 1169–79.Google Scholar Also Knight, Jonn B. and McGrath, Michael D., “An Analysis of Racial Wage Discrimination in South Africa,” Oxford Bulletin of Economics and Statistics 39 (November 1977), 245–71.CrossRefGoogle Scholar

⁶ The underlying model assumes a single good produced by skilled labor, unskilled labor, capital, land, and imported intermediates. All factors are gross substitutes; technology exhibits constant returns to scale in this and all other models discussed.

⁷ Lewis, W. Arthur, “Economic Development with Unlimited Supplies of Labour,” Manchester School of Economics and Social Studies 22 (May 1954), 131–91CrossRefGoogle Scholar; Hansen, Bent, “Colonial Economic Development with Unlimited Supply of Land: A Ricardian Case,” Economic Development and Cultural Change 27 (July 1979), 611–27.CrossRefGoogle Scholar In Hansen's two-good, three-factor model, traditional and modern sectors both use unskilled labor, but low-quality land is restricted to traditional farming, and high-quality land is limited to modern agriculture. A further extension incorporates an industrial sector. Most of the long-run results presented here are based on Hansen's comparative statics.

⁸ By 1980, 53% of the African population resided in the homelands, a rise from 37% in 1960. See Lipton, Merle, Capitalism and Apartheid (Totowa, NJ: Rowman & Allenheld, 1985), 380.Google Scholar

⁹ This is particularly true in the mining industry (ibid., 133–34). Also see J. H. Cooper, “South Africa and the Threat of Economic Sanctions,” South African Journal of Economics 52 (No. 4, 1984), 266–81, at 277. For a detailed but slightly dated analysis, see Seidman, Ann and Seidman, Neva, South Africa and U.S. Multinational Corporations (Westport, CT: Lawrence Hill & Co., 1978), 174–92.Google Scholar

¹⁰ Cobbe, James, “The Changing Nature of Dependence: Economic Problems in Lesotho,” Journal of Modern African Studies 21 (No. 2, 1983), 293–310, at 296.CrossRefGoogle Scholar

¹¹ Colclough, Christopher and McCarthy, Stephen, The Political Economy of Botswana (Oxford: Oxford University Press, 1980), 171.Google Scholar Cooper (fn. 9), 276, provides a similar figure for Botswana, and also reports that about one-third of all Swazis in paid employment in 1976 worked in South Africa.

¹² Cooper (fn. 9), 270–71.

¹³ In 1983, about 24% of South Africa's manufactured exports (beyond SACU) were bought by other African nations (calculated from Cooper, fn. 9,268); in manufactures (excluding textiles and pulp, paper and paperboard), the African share amounts to 37%.

¹⁴ Lipton (fn. 8), 303.

¹⁵ The model that serves a basis for this analysis is the multisector computable general equilibrium simulation model of the Indian economy. See Becker, Charles M.Mills, Edwin S., and Williamson, Jeffrey G., “Modeling Indian Migration and City Growth, 1960–2000,” Economic Development and Cultural Change 34 (October 1986), 1–33.CrossRefGoogle Scholar The Indian economy, as portrayed there, has strong parallels to the South African one: it has a modern urban services and manufacturing sector, accompanied by a large urban informal services sector; there is also a large rural sector—overwhelmingly peasant-populated—that corresponds to the South African homelands. The main adaptation that a long-run model of South Africa needs is to incorporate Hansen's assumption of “infinite marginal land” availability that leads to constant real unskilled wages, plus an assumption of restricted labor mobility (both skilled and unskilled) across some sectors (fn. 7). A more detailed discussion of a closely related model is provided in Kelley, Allen C. and Williamson, Jeffrey G., What Drives Third World City Growth? A Dynamic General Equilibrium Approach (Princeton: Princeton University Press, 1984).Google Scholar Short- and long-run consequences of shocks similar to those discussed here are presented in detail in Kelley-Williamson.

¹⁶ The statements regarding unemployment and bankruptcies assume that a Keynesian open-economy model with flexible exchange rates operates in the very short run; the equilibrium models of Hansen (fn. 7) and Becker, Mills, and Williamson (fn. 15) are more appropriate in the medium and long term.

¹⁷ This scenario does not hold in the unlikely event that the contraction causes an even greater decline in import demand than the initial fall in exports.

¹⁸ The ability of South Africa's manufacturing sector to expand exports in the face of devaluation is unclear; certainly its short-run response to the September 1983 exchange decontrol was not promising. See “Adjustment and Recovery: 1981–1987,” Barclay's Business Briefs (Johannesburg), January 1986, pp. 1–8.Google Scholar

¹⁹ In a simple two-good, two-factor model (the Heckscher-Ohlin world), a rise in the price of the labor-intensive good will cause the wage rate to rise relative to both commodity prices (Stolper-Samuelson theorem). For a readily accessible text, see Ethier, Wilfred J., Modern International Economics (New York: W. W. Norton, 1983)Google Scholar, chap. 3. For a more rigorous discussion, see Chacholiades, Miltiades, International Theory and Policy (New York: McGraw-Hill, 1978)Google Scholar, chap. 10.

Ambiguity ensues in this case because the prices of key intermediate imports have risen in both sectors as well, causing the demand for labor to decline. As demand for South African exports is imperfectly elastic, and as net effects on the demand for labor depend on the price rise of the labor-intensive good relative to that of input costs, the O.E.C.D.'s attitude toward South Africa's labor-intensive exports could well determine whether real wages rise or fall. In the very long run, the wages of unskilled labor will be roughly constant, as they are determined by migrants' opportunity costs, which in turn depend on the (roughly constant) average product of peasant labor in the surplus land states surrounding South Africa (see Hansen, fn. 7).

²⁰ All of the models upon which the analysis has thus far been based assume fixed labor supply. Extension of the analysis to consider variable labor supply is in Barnum, Howard N. and Squire, Lyn, “Predicting Agricultural Output Response,” Oxford Economic Papers 32 (November 1980), 284–94.CrossRefGoogle Scholar

²¹ This has been established so frequently as to be a commonplace. The most conclusive evidence, for a large number of countries, is given by Lluch, C.Powell, A. A. and Williams, R. A., Patterns in Household Demand and Savings (New York: Oxford University Press, 1977).Google Scholar

²² The relative import intensity of high-income groups is due to their high expenditure shares of durable and luxury goods and low proportion of food shares. See ibid.

²² Once again, the Stolper-Samuelson theorem applies.

²⁴ The simulated effects of a comparable shock (the 1973 OPEC oil price increases) on a “representative developing country” are described in detail in Kelley and Williamson (fn. 15), 128. Although the price shock does cause an increase in the growth rate of agricultural output along with a pronounced decline in manufacturing, the impact of the oil cost increase appears to swamp the Stolper-Samuelson effect of a relative price increase of the most labor-intensive good.

²⁵ I am grateful to William Kaempfer and Anton Lowenberg for this point.

²⁶ In addition to raising the cost of intermediates to many sectors, the embargo effectively raises the relative prices of capital-intensive goods; Stolper-Samuelson effects now work to raise skilled wages and capital rents, and to reduce unskilled wages.

²⁷ South Africa's foreign debt rose from $18.0 billion in mid-1981 to $23.7 billion in late 1985—a period of stagnant GNP. See “Adjustment and Recovery: 1981–1987,” Barclays Business Brief, January 1986, p. 6.

²⁸ This complementarity result is the key finding from the seminal cross-country multi-skill production functions estimated by Fallon, P. R. and Layard, P.R.G., “Capital-Skill Complementarity, Income Distribution and Output Accounting,” Journal of Political Economy 83 (April 1975), 279–301.CrossRefGoogle Scholar

²⁹ In fact, even if a particular firm doubted that the sanctions campaign will directly effect investments, should it nonetheless conclude that other rand holders believe that it will, it would have an incentive to reduce rand holdings.

³⁰ The assumptions stated here prevent unemployment of the Harris-Todaro type, in which migration, and hence unemployment, responds to a rise in the average modern-sector wage rate. See Blomqvist, Ake, “Urban Job Creation and Unemployment in LDCs: Todaro vs. Harris and Todaro,” Journal of Development Economics 5 (March 1978), 3–18.CrossRefGoogle Scholar

³¹ Indeed, if skilled labor is a substitute for unskilled labor, as Fallon and Layard (fn. 28) find, then efforts to conserve on unskilled labor will cause demand for skilled labor to increase further.

³² Lipton (fn. 8), 383–84.

³³ Ibid., 386.

³⁴ South Africa supplies all of the BLS states' and much of Zimbabwe's and Zambia's petroleum products. In the long run, only Lesotho is completely dependent on South Africa for its petroleum, but disruptive military operations by South Africa may keep this long run perpetually on the horizon for the other peripheral states as well. See Cooper (fn. 9), 279.

³⁵ South Africa already has been attempting to reduce the foreign component of its miners. This share peaked at 79% in 1973 (297,000 of 379,000), but fell to 42% by 1982 (190,000 of 448,000). See Lipton (fn. 8), 385. Ironically, further reductions may make the periphery less susceptible to being held hostage by South Africa in case of sanctions.

³⁶ This rise is reported in internal World Bank documents. Average real wages in the formal sector declined in the 1970s, but this appears to reflect massive entry at the low end of the job scale, and hence a diminution in average skill level. See Christiansen, Robert E. and Kydd, Jonathan G., “The Return of Malawian Labour from South Africa and Zimbabwe,” Journal of Modern African Studies 21 (No. 2, 1983), 311–26CrossRefGoogle Scholar, for a detailed analysis of Malawi's absorption of returned migrants. Many, but not all, of the returnees came from South Africa. Malawian employment in the South African mines peaked in 1972 at 120,000; this number fell to zero by 1976 in response to the Malawian government's prohibition of contract employment in South Africa after a plane crash in 1974 that killed 74 returning miners. In all, Christiansen and Kydd (p. 321) estimate that net Malawian male in-migration between 1966 and 1977 was 175,000, over 18% of the total male labor force in 1966.

³⁷ Similar employment expansion in agriculture may accompany a boycott of South African gold and diamonds. However, since mining wages are nearly twice as high as agricultural ones, remittances are likely to fall in this case.

³⁸ Lewis (fn. 7).

³⁹ Unpublished Speech at the South African Benefit of the American Committee on Africa Protest Rally (New York, December 10, 1965), cited in Baldwin, Lewis V., “The Vision of Martin Luther King, Jr. and the Apartheid System of South Africa,” mimeo (Nashville: Vanderbilt University Department of Religious Studies, 1986).Google Scholar