Macroeconomic Impact of Capital Flows in Sub-Saharan African Countries, 1980-2008

by John Weeks

Read the PDF of this article here: pdf

The financial crisis of the late 2000s generated the deepest global recession since the end of the Second World War. The recession resulted in declines in international private capital flows, including those from developed to underdeveloped countries. The fiscal stress in developed countries provoked by the recession gave rise to speculation that the fall in private flows might be followed by reductions in official development assistance. The recession-generated changes in the tempo and pattern of capital flows have major implications for the national economies in the sub-Saharan African region. Prominent among these is the need for capital account management, that is, “capital controls”.

In mainstream (“neoclassical”) economic analysis, an international regime of unregulated currency movements facilitates capital inflow, which can contribute to funding investment and faster growth. Even if one accepts this analytical generalization, however, one should carefully distinguish between stable and unstable components of flows, in specific between direct foreign investment and “portfolio” capital. In practice the absence of regulation allows for outflows of foreign exchange which is not matched by inflows. In some countries “portfolio” flows are no more than an euphemism for capital flight. The focus of this article is the probable impact of the most important components of capital flows, including capital flight, on the macroeconomic performance of sub-Saharan African countries over several decades leading up to the global financial crisis.

The empirical evidence used to make my case is for 31 countries included in the Ndikumana and Boyce (2011) database. Inspection of the statistics strongly suggests that in the absence of effective regulation of the external account, as is the case in most of the sub-Saharan African countries, capital flight is quite substantial, both absolutely and compared to other types of resource flows. In many countries, capital flight plus official debt service outweigh positive flows from development assistance and direct foreign investment. The comparison of inflows and outflows leads to the conclusion that the impact of capital inflows on sub-Saharan African countries has been exaggerated.

The evidence suggests that loss of foreign exchange through debt service and capital flight may in part explain the relatively weak growth of the countries of sub-Saharan Africa. In the 2000s, outflows from debt service declined for most countries as a consequence of long-delayed debt relief, which contributed to improved growth performances in the years before the global financial crisis. National measures to limit capital flight would result in further improvement in economic performance.

Capital Flow, Growth and Instability

Categorizing Countries
To empirically assess the likely impact of capital flows I divide the 31 countries into three groups, exporters of petroleum, conflict-affected, and “others.” The categories are exclusive. A consensus exists that conflict and an export sector dominated by petroleum affect macroeconomic performance. Difficulty arises in assigning countries to these categories in an exclusive and non-arbitrary manner. For the petroleum-exporting group arbitrariness arises if a country qualifies for only a part of the time period, either because of recent discovery or substantial decline in production. But this is not a major problem for the 31 countries in the data set. Another objection to the classification might be that some mineral producers, such as Botswana, have macroeconomic characteristics similar to petroleum exporters. I take a “minimalist” approach, adhering strictly to the petroleum category and including only those whose exports have been oil-dominated for over a decade: Angola, Cameroon, Chad, Republic of Congo, Gabon, Nigeria, and Sudan.

The “conflict-affected” category presents analytical and practical difficulties. Because few countries of the world are completely free from conflict, distinctions must be made on the basis of degree. For the 31 countries over the years 1980-2008, few would object to the inclusion of Burundi, Democratic Republic of Congo, Mozambique, Rwanda, and Sierra Leone. I also include Ethiopia, whose internal conflict lasted throughout the 1980s, formally ending with Eritrean independence and a new government in Addis Ababa in 1991. In subsequent years the ebb and flow of tensions between the two countries resulted in armed hostilities during 1998-2000. In addition I include Zimbabwe, due to the severity of its internal strife, despite the absence of some formal characteristics of civil war.

A more serious categorical objection is that at least two of the petroleum-exporting countries also clearly qualify as conflict-affected, Angola and Sudan. For this analysis of capital flows, petroleum-exporting status here takes priority over conflict status. Arguments could be made to include at least three other countries: Côte d’Ivoire (civil war during 2002-2007, rekindled in 2011), South Africa (anti-apartheid conflict until the early 1990s), and Uganda (civil war until about 1985, conflict in the north since the late 1980s). I omit Côte d’Ivoire because its conflict affects less than a third of the years covered by the statistics. The economic effects of the conflict in South Africa were not sufficiently substantial for inclusion, though the human cost of apartheid was enormous beyond measurement. Uganda is omitted for the same reason as Côte d’Ivoire. The substantial economic impact of its conflict lasted less than a third of the time period under review.

Growth and Investment

As one would expect, the group of conflict-affected countries had the lowest rate of economic growth in every decade since 1980, with twenty years of declining per capita income overall followed by weak recovery slightly above population increase. The growth rates for the petroleum exporters and the group of other countries were very close on average for the three decades as a whole, but growth among the petroleum-exporting countries was considerably more variable over time. Much of the greater variation resulted from the swings in world petroleum prices.

Also as should be expected, the share of gross investment in GDP is greatest for the petroleum-exporting countries and least for the conflict-affected group. The most important characteristic of the investment statistics for all three groups is how low they are. If the typical aggregate net capital-output ratio for sub-Saharan African countries is about four, the gross would be close to five, implying a quite modest potential growth rate of slightly above four percent for petroleum exporters and slightly below that for the group of other countries. These rates, close to the actual ones for the three decades, mean extremely modest per capita growth, requiring roughly fifty years for income per person to double.

The three-year moving average growth rates for the three groups are shown in Figure 1. This highlights the greater variability among the petroleum-exporting countries. The chart also shows especially poor growth performance in the first half of the 1990s. My working hypothesis, demonstrated elsewhere for these countries (Weeks 2012), is that the share of investment is the major driver of the potential rate of growth in the region. This hypothesis becomes relevant below, when I link the investment rate to capital flows.

Figure 1. GDP Growth Rates, 31 sub-Saharan countries, 1980-2008 (3 year moving average)

Overview of Capital Flows

Sub-Saharan African countries receive substantially less private capital inflow than other developing regions. This is the case for portfolio flows because no country except South Africa has a developed capital market. South Africa also has the only substantial domestic market, implying limited opportunities in the other countries for direct investment except for resource extraction (which is also important in South Africa).

Figure 2 provides IMF statistics on private capital flows to all countries of the region, divided among direct investment, portfolio, and other flows. In contrast to direct investment, portfolio and other flows show extreme volatility. In addition, these two types of flows were severely affected by the global recession. During the 1990s, the aggregate non-FDI inflow to the region was US$19 billion at prices of 2008,1 which fell to a negative US$49 billion for 2000-2007. The negative flow almost doubled to US$94 billion for the next four years, 2008-2011. Direct investment was a positive US$118 billion during 2000-2007, and hardly changed over the next four years (US$116 billion). The volatility of non-FDI flows compared to FDI is not difficult to explain. The former is itself not linked to any productive activity. If governments allow it, non-FDI flows generate severe balance of payments instability. The dramatic difference in behavior between FDI and non-FDI flows makes a strong circumstantial argument for regulation of short-term capital flows.

Figure 2: Net private capital flows to all sub-Saharan countries, 1980-2011 (billions of constant US dollars of 2008)

Notes: FDI is direct foreign investment, PrPortf is private portfolio flows, and PrOther is other private flows. The deflator is the US GDP price index.
Source: IMF, World Economic Outlook 2012, data tables.

The pattern is quite different for official development assistance and net official flows (Figure 3). The former, taken from the World Bank (World Development Indicators) and using the OECD definition of official assistance, increased continuously in constant prices from its low point in 2000 (US$ 15 billion in prices of 2008, lowest total since 1984). The movement of net official flows, which includes debt servicing, was a negative US$ 42 billion over the eight years 2000-2007. The cancelling of most sub-Saharan African official debt, especially that to the IMF and World Bank, led to a dramatic shift to a positive net flow of US$ 80 billion during 2008-2011.

Figure 3. Official flows to all sub-Saharan countries, 1980-2011 (billions of constant US dollars of 2008)

Notes: Nt Off(IMF) is net official flows, and ODA(WB) is official development assistance.
Sources: Net official flows is from IMF, World Economic Outlook 2012, data tables; and ODA is from World Bank, World Development Indicators.

The rest of this paper seeks to disaggregate the private flows at the national level for the 31 sub-Saharan African countries. As noted already, the financial structure in most countries of the region is underdeveloped and in some cases non-existent. As a result, non-FDI flows fall overwhelmingly into the category of capital flight, in the sense that they are unrelated to any domestic financial asset at any stage of their life cycle.

The Ndikumana and Boyce data allow a closer look at capital flight for these countries. Figure 4 presents those statistics, again applying a three-year moving average. Measured capital flight is negative (indicating net outflows) for each group and every year except one (the “other” country category for 2001 is marginally positive at 0.1). From the end of the 1980s to 2000-2001, capital flight shows a diminishing trend for the seventeen “other” countries, and also for the petroleum exporters, though for the latter group the tendency is weaker. This trend was sharply reversed after the early 2000s for each group, a reversal that pre-dated the Global Financial Crisis by at least five years.

Figure 4: Capital Flight as share of GDP, 31 sub-Saharan African Countries, 1980-2008 (3 year moving average)

Notes: See annex for country groups. Numbers in legend are the averages for the entire period.
Source: Ndikumana and Boyce data base on capital flight.

The reversal is all the more striking when one inspects the data for the conflict-affected group. The conflicts in five of the seven countries in the group had ended or became less intense as the 2000s proceeded (Burundi, Ethiopia, Mozambique, Rwanda, and Sierra Leone) and in a fifth was no worse (Democratic Republic of Congo). Despite this, capital flight in the mid-to-late 2000s, averaged across the seven countries, was the greatest for the three decades both absolutely and as a portion of GDP (this was also the case for the petroleum exporters).

The other major non-trade element of external outflow in sub-Saharan African has been debt service. The absolute and relative burden of debt service increased into the early 1990s, then began to ease. When the two major outflows, capital flight and debt service, are combined (Figure 5), there appears to be an improvement for the 2000s in total outflow for the 17 “other” countries. No improvement occurs for the conflict-affected countries compared to the level of the early 1980s, though an upward (less negative) trend can be seen since the mid-2000s. The overall impression is of a similarity of the patterns for the three country groups. This similarity may reflect underlying forces common to the three groups, with the intra-group characteristics determining levels rather than patterns of movement.

Figure 5: Capital Flight plus Debt Service as share of GDP, 31 sub-Saharan African Countries, 1980-2008 (3 year moving average)

Notes: Numbers in legend are the averages for the entire period. Source: Ndikumana and Boyce data base on capital flight.

The statistics on direct investment indicate a positive trend since the 1980s in all three groups of countries (see Figure 2), although the petroleum exporters, recipients of the largest amounts, suffered a sharp decline in the second half of the 2000s. Nevertheless, direct investment is far less than capital flight for all three groups. This is especially the case for petroleum exporters, which received the largest inflows and the largest outflows (plus 3.4 percent of GDP for FDI and minus 6.0 percent for capital flight). It is not possible to assess what part of direct investment was accounted for by mergers and acquisitions as opposed to new investment.

The upward trend in direct investment contrasts with the results for official development assistance (Figure 6). For the average share in GDP across all 31 countries for three decades, there is no significant trend. If one arbitrarily splits the time series at the mid-point (1995), for the group of 17 other countries the mean is slightly lower in the second period, and slightly higher for the petroleum exporters, but neither difference is statistically significant. The pattern for the conflict-affected countries is one of extreme fluctuations, reflecting the large inflows that come with the end of hostilities.

Figure 6: Official Development Assistance as share of GDP, 31 sub-Saharan African Countries, 1980-2008 (3 year moving average)

Notes: See Figure 3.
Numbers in legend are the averages for the entire period.

The last time series chart, Figure 7, combines inflow and outflow to obtain net flow. The sum of direct investment, development assistance, debt service, and capital flight produces positive values for all years but one for the group of seventeen “other” countries. From that negative value for 1987, one finds an almost continuously upward movement. However, the statistically significant positive trend occurs only during 1987-1995. From 1995 onward the pattern appears cyclical. The pattern for conflict-affected countries also appears cyclical. The petroleum exporters show a relatively long upward tendency for 1982-2004, followed by a sharp decline.

Figure 7: Net Capital Flow as share of GDP, 31 sub-Saharan African Countries, 1980-2008 (3 year moving average)

Notes: See Figure 3.
Numbers in legend are the averages for the entire period.

In assessing the net capital flow statistics, it must be stressed that only part of development assistance enters the recipient country. Leakage occurs not only due to the recycling of ODA into debt service (here netted out), but also due to payments directly to donor country suppliers that involve no material inputs, such as consultancy fees. If over the three decades in these 31 countries leakages averaged a modest twenty-five percent of gross aid, the net capital flow as portion of GDP for the 31 countries in Figure 7 would be zero. While it is not justified to draw conclusions from information we do not have, it is probable that net flows were considerably less than estimated in this study.

This review of the major flows of funds produces the following conclusions for the 29 years 1980-2008:

1. Capital flight appears strongly cyclical for all three country groups, with a striking reduction in the early 2000s that was reversed during the following years.

2. Debt service declined continuously from the mid-1990s, although prior to debt relief in the 2000s the measured decline may overstate the actual because of recycling of development assistance.

3. Direct foreign investment shows an upward trend for all three country groups, though for all it is substantially less than capital flight.

4. Official development assistance shows much the same pattern for the petroleum exporters and the group of 17 countries: upward movement until the mid-1990s, then stagnation or decline. For the conflict-affected group, development assistance has extreme peaks and troughs.

In summary, on average across all 31 countries, official development assistance was the largest of the four flows, but it was barely large enough to equal the sum of debt service and capital outflow. Direct investment was relatively small, contributing only a bit over two percent of GDP as a positive element in external flows.

Policy Measures

Reduction of capital flight is essential to increase the resources available in sub-Saharan Africa for both consumption and for public and private investment. With development assistance likely to decline as a proportion of recipient national income, stemming capital flight may be the most important growth-generating policy available to governments of sub-Saharan African countries.

Part of the responsibility for reducing capital flight lies with the governments of developed countries (Ndikumana and Boyce 2011, Chapter 5). Governments in sub-Saharan Africa can hope for but not rely on the implementation of effective measures by developed country governments. Even if several major countries introduced effective measures, there would for the foreseeable future remain other “offshore” money centers beyond the reach of regulation.

Therefore, while pressing for effective action by external actors, sub-Saharan African governments should individually and collectively pursue their own solutions. The minimum first step would be for governments to require all foreign exchange transactions to be registered with the central bank in order to be legal. A second and broader step would be to require all businesses, donors, and non-governmental organizations to channel their foreign exchange through the central bank. While this measure need not in itself involve currency controls, it lays the basis for such controls if necessary. Thirdly, governments could legally require their citizens to provide details of foreign bank accounts they hold.

Finally, governments in sub-Saharan Africa could consider a measure implemented in Argentina in recent years with great effectiveness. In Argentina, any person, company or institution wishing to send abroad funds exceeding a specified minimum must provide the central bank with proof that the appropriate taxes have been paid on the income generating the funds to be remitted. In addition to its other virtues, this policy conforms to the global fight against the laundering of drug money.

Capital flight is a blight that has seriously undermined growth and development in sub-Saharan Africa. Although it might not be possible to eliminate it, it should be possible to substantially reduce it. While African governments should encourage the uncertain process of reform in the developed world, there are effective measures they can take themselves to reassert control over their external capital flows.

John Weeks is professor emeritus of economics of the University of London, School of Oriental and African Studies. He is author of many scholarly articles and books on sub-Saharan Africa. His recent policy work includes advising the government of Sierra Leone on macroeconomic policy and the Central Bank of Kenya on its short-term forecasting model.


Actionaid. 2011. Real Aid: Ending Aid Dependency. London: Actionaid.
Boyce, James K, and Léonce Ndikumana. 2001. “Is Africa a net creditor? New estimates of capital flight from severely indebted sub-Saharan African countries, 1970-1996.” Journal of Development Studies 38, 2, 27-56.
_____. 2005. “Africa’s Debt: who owes whom?” In Capital Flight and Capital Controls in Developing Countries, edited by G A Epstein, 224-40. Northampton and Aldershot: Edward Elgar.
Ndikumana, Léonce, and James K Boyce. 1998. “Congo’s odious debt: external borrowing and capital flight in Zaire.” Development and Change 29, 2, 195-217.
_____. 2003. “Public Debts and private assets: explaining capital flight from sub-Saharan African countries.” World Development 31, 1, 107-30.
_____. 2010. “Measurement of capital flight: methodology and results for sub-Saharan African countries.” African Development Review 22, 4, 471-81.
_____. 2011. Africa’s Odious Debts: How Foreign Loans and Capital Flight Bled a Continent. London: Zed Books.
United Nations Conference on Trade and Development (UNCTAD). 2000. The Least Developed Countries Report 2000: Aid, Private Capital flows and external debt, The Challenge of Financing Development in the LDCs. Geneva: UNCTAD.
_____. 2012. “Capital Flows in sub-Saharan Countries with special reference to Mozambique, 1980-2008,” paper presented to the IESE Conference, Mozambique: Accumulation and Transformation in the Context of International Crisis, September 2012, Maputo.
Weeks, John. 2012. The Irreconcilable Inconsistencies of Neoclassical Macroeconomics: A False Paradigm. Oxford: Routledge.
Weeks, John, Victoria Chisala, Alemayehu Geda, Hulya Dagdeviren, Alfredo Saad Filho, and Carlos Oya. 2006. Economic Policies for Growth, Employment and Poverty Reduction: Case Study of Zambia. Lusaka: United Nations Development Programme. Zambia Country Study.pdf.


  1. This base year is chosen because it corresponds to that used by Ndikumana and Boyce. [Back to Text ↩]