1 A typical example: The three goals of Cote d’Ivoire’s Five Year Plan in 1971 were: “i. Strong economic growth; ii. Greater participation by Ivory Coast citizens in the economy; iii. An improvement in the standard of living of the rural population” (EIU 1971). Note that Cote d’Ivoire was among the most market-friendly regimes in Africa at that time.
2 Africa is full of examples of state enterprises that were unproductive until they were bought by foreign private capital, such as the Bonagui soft drinks factory in Guinea, which was installed under Sekou Toure with an annual capacity of 2 million cartons of soft drinks but never entered production until it was bought by Coca Cola Corporation of the USA and Stella Artois of Belgium in 1986 (EIU 1987).
3 Guinea had 181 public enterprises in 1980 (Nellis 1994).
4 It was eventually nationalized in 1973 by Colonel Acheampong.
5 For example, in 1974-75, the government of Benin nationalized petrol stations, insurance firms, banks, telecommunications firms, transport companies, the brewery, and one hotel; for all this, compensation was a mere $8m (Allen 1989, p. 80).
6The short-lived democratic government in Nigeria from 1979-1983 quickly became a “colossus of corruption” (Othman 1984) whose overthrow by the military was welcome. Corruption is a common excuse for militaries to overthrow democratic regimes in Africa (Decalo 1990). Politicians in Nigeria’s current democracy are also succumbing to the temptation to buy votes to win elections (Collier 2009, p. 39).
7 “Through patronage, popular economic policies, and astute cabinet appointments, attention to the demands of all segments of the population, … Eyadema adroitly nibbled at Ewe opposition to his rule…” (Decalo 1990, p. 223).
8 An alternative would be to measure capital inflows as a percentage of GDP. Since government policies affect GDP, I chose to use a denominator (population) that depends less on government choices.
9 The nine countries that are thereby excluded are Cape Verde, the Comoros, Djibouti, Equatorial Guinea, Gabon, the Gambia, Guinea Bissau, Sao Tome and Principe, and Swaziland. Most of these countries (all but Gabon, the Gambia, and Swaziland) are missing many data for many years. They also tend to have extreme outliers for many of the data. For example, while the average FDI inflows per capita for sub-Saharan African states was $11 for countries with average populations over one million, it was $105 per year for smaller countries. Equatorial Guinea had FDI inflows per capita as high as $2,897 per year, while Gabon had inflows as high as $502 per year and as low as -$411 per year. Similarly, the average level of public investment for countries with average populations below one million was 15.3% of GDP, almost double the average level of public investment for countries with larger populations, 8.6% of GDP. See the descriptive statistics in Table 1 for more evidence of outliers.
10 One rejected candidate, for example, is Kornai’s (1992) list of socialist countries. Most of Kornai’s socialist countries are in Asia and Eastern Europe. Those in Africa (as of 1987) are limited to Congo-Brazzaville, Somalia, Benin, Ethiopia, Angola, Mozambique, and Zimbabwe (Kornai 1992, pp. 6-7). The reason why Kornai’s list is so restricted is because of his operationalization: “The undivided power of the Communist party is the sole criterion for inclusion in this table. From now on those included will be referred to in this book as socialist countries” (Kornai 1992, p. 5). The problematic nature, for my purposes, of this operationalization is highlighted by the fact that China, the developing world’s largest recipient of FDI, continues to be coded as socialist.
11 I haven’t finished coding all country-years for the SEI measure. I’ve done about 800 and have about 400 left to do.
12 Other relatively robust variables are tariff rates and wage rates are not included because of lack of data. The most complete wage data set (Freeman and Oostendorp 2001) has 95% of the country-years missing data, and 75% continue to be missing data after using linear interpolation to impute for missing values. Inclusion of the data including imputations indicates that wage levels have an insignificant (negative) effect, which may reflect the fact that Africa does not tend to be a destination for efficiency-seeking FDI. Inclusion of wage data eliminates some of the findings, but it is unclear whether this is due to 75% of the cases dropping out
13 However, this economic prediction rarely plays out in practice. According to Lucas (1990), this can be explained with levels of productivity via human capital, but Markusen (1995) argues that this is because most FDI is market-seeking rather than efficiency seeking.
14 The FDI stock data from UNCTAD begins in 1980, while the flow data begins in 1970. I used the net inflow data to calculate stock data, and replaced negative values with zeroes.
15 Until recently, twelve African countries had stock markets: Botswana (opened 1989), Cote d’Ivoire (1976), Ghana (1989), Kenya (1954), Malawi (1996), Mauritius (1989), Namibia (1992), Nigeria (1961), South Africa (1887), Swaziland (1990), Zambia (1994), Zimbabwe (1946) (Kenny and Moss 1998).
16 The mean portfolio equity inflows per capita for all SSA countries is $1 per year; for South Africa it is $28. Most of this took place in 1997-2000.
17 Running the regression using a GLS model with random effects returns the same coefficients but with much smaller standard errors. For example, the interaction terms using the variables of interest that are significant at the 90% level using OLS with PCSE are significant at the 99% level using the random effects model.
18 There is a 76% correlation between total portfolio flows into Africa and its lag, and so I will use the lag to generate predictions for chapter four.
19 This is an adaptation of the rule used by Fearon and Laitin (2003). I use the value of fuel and mineral and ore production as a percentage of gross national income (GNI) rather that fuel primary commodity exports as a percentage of GDP because of superior data availability. Five oil-producing countries in Africa are heavily endowed by this criterion, two with small populations (Gabon and Equatorial Guinea) and three with populations above one million (Angola, Nigeria, and the Republic of Congo). Two countries met this criterion with mining activity from 1970-1976: Liberia (primarily iron) and Zambia (primarily copper). Namibia also met the criterion from 1984-1989 but is missing data for all four dependent variables for all years. Chad meets the criterion after 2003.
20 The Gwartney measure is calculated using private investment (public investment divided by pulblic + private investment).
21 I use the Economist Intelligence Units quarterly reports and the World Bank’s Privatization Database to count sector investment and divestment activities.