South African Divestment: The Investment Issues

Literature › Works

Wayne H. Wagner; Allen Emkin; Richard L. Dixon (1984). South African Divestment: The Investment Issues. Financial Analysts Journal, Vol. 40, No. 6. CFA Institute.

Type	journal article
Scope	Wider apartheid literature
Method	descriptive quantitative
Period studied	1979–1984
Regions	United States, South Africa (as the object of divestment)
Themes	sanctions divestment
Citation key	wagner1984

Argument

The authors take no moral or political position on divestment. They ask what it would mean in practice for US public pension funds to be barred from investing in companies doing business in South Africa. They build an 'Alternative Universe' by replacing each of the 152 S&P 500 companies with South African links with the largest unrestricted company in the same industry. They compare the two universes on capitalization, beta, diversification, Value Line quality ratings, trading costs, analyst coverage, manager style, corporate debt and commercial paper. They conclude that divestment would raise risk, shrink opportunities and increase costs, and that the burden grows with fund size. This challenges earlier studies that found little effect.

Evidence and method

S&P 500 and Wilshire 5000 capitalization data (end 1983 and Q1 1984), Wilshire return estimates, Value Line Timeliness and Safety Ratings, Thomas Loeb's dealer-quote trading cost table, I/B/E/S analyst estimate counts, Moody's bond and commercial paper ratings, and a Wilshire manager-style comparison. Opinions of investment managers gathered for the Washington, D.C. Retirement Board are also cited.

Place in the debates

• Is divestment from South Africa costless or low-cost for pension portfolios? (see Sanctions, disinvestment and the anti-apartheid movement). Argues against Rudd, Calderwood and Schwartz. It says restrictions have little effect only on funds of about $50 million or less, and that large funds face higher risk and costs and may need to restructure their portfolios.
• Does divestment hurt or improve investment returns? (see Sanctions, disinvestment and the anti-apartheid movement). The 152 South African-linked stocks returned less than their replacements over five years. The authors attribute the gap to the small-firm effect, not to a benefit of divesting, and say higher returns carry higher risk.

Findings

• The scale of exposure: 229 US companies with South African ties made up a large share of US equity capitalization. (p. 3)

  includes 229 companies with a market capitalization on December 31, 1983 of over $600 billion.

• The Alternative Universe is much smaller than the S&P 500 in capitalization terms. (p. 4)

  The Alternative Universe thus came to $728 billion-less than 62 per cent of the value of the S&P 500.

• Divestment is agreed to be harmless only for small funds. (p. 3)

  Investment officers, advocates of divestment and previous studies all agree that divestment restrictions have minimal practical effects on portfolios of $50 million or less.

• Replacement stocks earned higher past returns, but this reflects the small-firm effect and not any benefit of divesting. (p. 4)

  A dollar invested in the 152 replacement stocks would have grown to $2.60-a difference in annual rate of return of over 7 per cent.

• Small-cap outperformance in the period is documented using Wilshire estimates. (p. 4)

  Wilshire Associates estimates that the 500 largest U.S. companies returned 9.7 per cent per year over the last 10 years, whereas the second largest 500 returned 17.9 per cent annually.

• The Alternative Universe is well diversified but carries higher market risk. (p. 4)

  Although well diversified, it is riskier than the market, having a beta of 1.08.

• Divestment sharply reduces the supply of high-safety stocks available to institutions. (p. 5)

  Companies rated highly for safety represent 59 per cent of the market capitalization of the S&P 500 but only 43 per cent of the Alternative Universe.

• Replacement stocks cost more to trade, and the cost rises with fund size. (p. 6)

  The replacement companies are significantly more expensive to trade than the original companies because of company size.

• Replacement companies get less analyst coverage, implying higher research costs. (p. 6)

  but only 10.9 estimates on each replacement company.

• The investment-grade corporate bond universe shrinks substantially. (p. 7)

  This represents a 35 per cent reduction in the amount of corporate debt securities.

• Commercial paper quality is lower among the replacement companies. (p. 8)

  Of the 116 eliminated companies, 78 per cent are rated Prime-1, whereas only 67 per cent of the replacement companies achieve that rating.

Evidence ledger

Sources used

Source	Kind	Period	How used
Wilshire Associates list of 229 companies with employees in South Africa or business relations with the South African government (market capitalization at Dec 31, 1983)	company records or reports	December 31, 1983	Defines the set of 'prohibited' companies whose weight in the S&P 500 and Wilshire 5000 is measured and whose 152 S&P members are replaced.
Standard & Poor's 500 index constituents, capitalization and industry weights (Q1 1984)	stock market financial data	End of first quarter 1984	Serves as the benchmark universe against which the Alternative Universe is compared for capitalization, beta, R-squared and industry weights.
Wilshire 5000 index	stock market financial data	1983-1984	Used to show the South Africa-linked companies make up almost 35 per cent of all common stock weight.
Wilshire Associates return computations for the 152-stock South Africa-linked portfolio and the 152 replacement stocks (five-year)	stock market financial data	Approximately 1979-1984	Shows the replacement portfolio's higher five-year return, attributed to a small-firm effect.
Wilshire Associates estimates of returns on largest 500 vs second largest 500 U.S. companies (cited via Nesbitt, Pension and Investment Age, 5 Sept 1983) and 1983-84 market decline	stock market financial data	Ten years to 1983; July 1983-June 1984	Supports explanation that the higher return of the replacement portfolio reflects the small firm effect and that smaller firms carry more downside risk.
Beta and R-squared calculations for the Alternative Universe and manager-style portfolios (Tables II, VI)	stock market financial data	Around 1984	Quantifies the higher market risk (beta 1.08) and diversification (R-squared 0.968) of the restricted universe and its effects on manager styles.
Value Line Investment Survey, Timeliness and Safety Ratings	newspapers periodicals	c. 1984	Compares S&P 500 and Alternative Universe by distribution of timeliness and safety ranks and institutional holdings (Tables III, IV).
Institutional holdings of S&P 500 stocks	stock market financial data	c. 1984	Quantifies how much institutional money in top-rated stocks would be displaced by divestment.
Thomas F. Loeb, 'Trading Cost: The Critical Link Between Investment Information and Results', Financial Analysts Journal, May/June 1983 (table of actual dealer quotes)	secondary literature compilation	c. 1983	Loeb's cost-by-company-size and trade-size figures are applied to the 152 original and 152 replacement firms to estimate transaction costs (Table V).
I/B/E/S (Institutional Brokers Estimate System), Lynch, Jones & Ryan	stock market financial data	c. 1984	Average number of analyst estimates per company shows replacement firms get less research coverage.
Moody's corporate debt ratings and outstanding debt amounts (end 1983)	stock market financial data	End of 1983	Shows divestment shrinks the investment-grade corporate bond universe and lowers its quality (Table VII).
Moody's commercial paper ratings and bank CD/lending exposure to South Africa	stock market financial data	c. 1984	Shows reduced quantity and quality of commercial paper and bank CDs available to cash managers (Table VIII).
Opinions gathered for the Washington, D.C. Retirement Board from professional investment managers	interviews oral history	c. 1983	Shows professional managers think divestment would reduce performance, hamper diversification and impair quality.
Prior divestment studies: Rudd (1979), Calderwood (1983), Schwartz (1983 testimony)	secondary literature compilation	1979-1983	Provides the claims of minimal impact that the authors contest.

New data

An 'Alternative Investment Universe' constructed by replacing each of the 152 South Africa-linked S&P 500 companies with the largest unrestricted U.S. company in the same industry, with comparative statistics on capitalization, beta, R-squared, ratings, trading costs and debt for the two sets.

Inference

Quantification: Descriptive tables.

Comparative simulation: a constructed counterfactual universe is compared with the S&P 500 on capitalization, beta, R-squared, industry weights, Value Line ranks, transaction costs (borrowed Loeb schedule), analyst coverage and debt/commercial paper ratings. Descriptive tables and arithmetic, with no statistical testing.

Key claims

Claim	Basis	Strength
Replacing the 152 South Africa-linked S&P 500 firms with the largest unrestricted industry peers shrinks the investable universe to under 62 per cent of the S&P 500's capitalization. (p. 4)	Authors' constructed Alternative Universe from the S&P 500 capitalization data ($554bn removed, $107bn added).	Systematic data
The Alternative Universe is riskier (beta 1.08) and slightly less diversified than the S&P 500, and shifts weights toward domestic industries. (p. 4)	Beta and R-squared calculations and the industry weight comparison in Table II.	Systematic data
The replacement portfolio's higher five-year return reflects the small-firm effect rather than a benefit of divestment, and carries higher downside risk. (p. 4)	Authors' return computations plus Wilshire large-firm vs second-500 return estimates and cited small-firm literature.	Partial data
Divestment would reduce the supply of high-quality (top Value Line timeliness and safety) stocks available to institutions. (p. 5)	Value Line ranks applied to both universes with institutional holdings (Tables III and IV).	Systematic data
Trading, research and administrative costs would be higher for restricted portfolios, with the burden rising with fund size. (p. 6)	Loeb's dealer-quote cost schedule applied to the firm size distribution; I/B/E/S analyst estimate counts.	Partial data
Divestment would shrink and lower the quality of fixed income and cash-equivalent universes (corporate debt, commercial paper, bank CDs). (p. 7)	Moody's ratings and outstanding debt amounts for original and replacement companies (Tables VII, VIII).	Systematic data
Divestment restrictions have a substantial impact on large funds but minimal effect on funds of $50 million or less. (p. 8)	Synthesis of the above comparisons; the claim about small funds rests on the agreement of officers and previous studies, with no separate test.	Partial data

The author's caveats

The authors note that the transaction cost estimates may understate the problem because insider holdings reduce the float of small replacement firms, and that strategic risks of divestment are harder to quantify. They also say they do not address moral, political or ethical issues and cannot speculate on how any particular manager would adjust.

Gaps in the evidence

The article does not document how the list of 229 South Africa-linked companies was compiled or its dates, and it offers no out-of-sample or time-series test of the restricted portfolios' risk and return, relying on a single five-year window dominated by the small-firm effect. Transaction cost estimates borrow Loeb's schedule rather than observed fund execution data. Actual pension fund trading records, tracking-error simulations for specific funds and the later experience of divested funds would test the argument better; the latter exist now but not in 1984.

Replicability: Data poorly documented.