Information about Sub-Saharan Africa África subsahariana

6 Globalization: Implications for Africa

Zubair Iqbal, and Mohsin Khan
Published Date:
December 1998
  • ShareShare
Information about Sub-Saharan Africa África subsahariana
Show Summary Details
Paul Collier

Africa has long been strongly integrated into the global economy. In respect of product markets, the typical African country trades a high share of its GDP, its economy being small and specialized in primary exports. In respect of capital markets, Africa was arguably the first continent to become highly integrated: a higher proportion of African wealth is held internationally than on other continents. Thus, in one sense “globalization” is not new to Africa, but it has meant primary exports and capital flight. However, when economists now talk of globalization they mean something rather different. Specifically, they mean falling trade barriers, integrating financial markets, and transnational corporations. In this chapter, I argue that these changes can potentially integrate Africa into the world economy in a new way, but that whether these opportunities are taken up depends upon African economic policy.

Trade barriers have come down around the world, in particular with the phased elimination of QRs. Although restrictions on textiles and garments will not be fully lifted for a further 7 years, the fact that there is a timetable and a commitment has revolutionary implications for the newly industrializing developing countries. Given Africa’s present pattern of exports, this international trade liberalization is of little consequence since Africa does not face important barriers for its present exports. The reduction in trade barriers is only of significance if Africa changes its comparative advantage. I argue that its present comparative advantage is determined mainly by its policy environment rather than by its factor and natural resource endowments.

Financial capital markets are becoming much more internationally integrated. Financial capital is increasingly flowing to developing countries as pension funds realize the advantages of international diversification. At present, Africa is not attracting significant portfolio capital inflows and so is not benefiting from this globalization of portfolios. I argue that again this reflects its policy environment rather than its endowments.

Further, manufacturing firms are internationalizing their production, shifting investment to lower cost environments. At present, Africa is not attracting footloose manufacturing, except for production for its highly protected domestic markets, and so is not in a position to benefit from this globalization of production. I argue that again this reflects policy rather than endowments.

Thus, on present policies, Africa has little to gain from globalization. At the same time, it has little to lose. This is in sharp contrast to the integrated economies of Europe and East Asia. There, the domestic policy environments are now subject to a new discipline. If governments impose high taxes on manufacturing profits or embark upon reckless fiscal policies, there is a capital outflow. No one is in control of these market forces: even the largest speculators are tiny relative to the size of the market. International capitalism does not mean that “capitalists” have power, but that no one has power. Globalization thus constitutes a much larger loss of power for the governments of Europe and East Asia than for the governments of Africa.

In the next section, I consider the potential of reduced trade barriers for African exports. In particular, I speculate on Africa’s comparative advantage. Two distinguished economists have recently advanced propositions that taken at face value are quite disturbing. Rodrik (1997) argues that Africa already has a level of trade to GDP that is normal given its level of income, and that trade liberalization cannot be an engine for African growth. Wood and Berge (1997) and Wood and Owens (1997) argue that Africa’s natural resource and human capital endowments imply that its comparative advantage is inevitably in unprocessed primary commodities. They argue that Africa will not be able to export manufactures even if it liberalizes its trade policy. An implication that can easily be drawn from these papers is that it is endowments rather than policies that have determined Africa’s present trading pattern, so that there is little that governments can do. I challenge this proposition.

In the third section, I consider the potential of financial market integration. 1 argue that there is scope both for large capital inflows and for risk bearing. In particular, I suggest that Africa is in the anomalous position of facing the highest underlying level of risk in the world, while having the lowest supply of risk-bearing instruments. As a result, it has the highest price of risk.

African Trade in a Globalizing Economy

Since 1970, world trade has risen much more rapidly than world GDP. By contrast, Africa’s GDP (despite growing more slowly than that of other regions) has risen more rapidly than its trade. Further, Africa’s exports have remained overwhelmingly concentrated in the same narrow range of primary commodities; indeed, on some measures, export concentration has actually increased. Even within this narrow range of export products, it has drastically lost market share. Evidently, something has been seriously amiss with Africa’s trade performance.

Rodrik (1997) shows that Africa does not trade less than predicted by its low level of GDP relative to Asia or Latin America: Asia “overtrades,” while Latin America “undertrades.” However, this is quite a misleading statistic. Being on the regression line has no normative significance; it simply shows that on this measure African policies were less anti-trade–oriented than Latin American ones. It does not contradict the proposition that Africa has lost heavily from its antitrade–oriented policies, which have caused its declining share of trade to GDP.

In this section, I investigate how globalization of goods and capital markets will affect Africa’s trade. In particular, I focus on whether Africa has a future as an exporter of manufactures. I use a suite of (informally presented) simple models. Which of these best depicts reality turns out to be critical for analytic prediction of Africa’s response to globalization, both in terms of what will happen to its exports and what should happen to its policies.

Africa’s Comparative Advantage: What Moves Where in a Globalizing Economy?

Reflecting its poverty, the returns to labor are very low in Africa. In a simple two-factor, common-technology world, there are three mechanisms, each of which enables globalization to raise the returns to labor and so reduce African poverty. These three mechanisms are labor emigration, an influx of capital, and the export of labor-intensive goods. If a previously autarkic, poor economy were suddenly opened to the global economy, all three of these mechanisms would operate simultaneously and each would be poverty-reducing. In effect, Africa has three degrees of freedom before globalization becomes ineffective as a poverty-reduction mechanism.

In parts of Africa, the two-factor, common-technology model is an adequate characterization, albeit obviously a radical simplification. An interesting question, then, is which of the three mechanisms is more and which less important. There are some obvious differences within Africa that make each mechanism more or less important for each country. For example, labor migration is probably more sensitive to distance than is trade in products. Hence, Northern Africa is best located to benefit from emigration to Europe. Capital inflows are, as I discuss below, sensitive to perceptions of investor risk. Risk ratings differ greatly among African countries. South Africa and Mauritius, with Investor International ratings in the 40s (on a scale where 0 = maximum risk and 100 = minimum risk), are better placed to attract capital than Nigeria, with a rating below 20.

In the two-factor, common-technology model, the circumstances that give rise to a labor exodus would necessarily give rise to both a capital influx and a comparative advantage in labor-intensive exports. However, once we turn to a more sophisticated model, this need not follow. Two modifications to the model are particularly pertinent for Africa. The first is to introduce additional factor endowments, which unlike labor and capital cannot be traded. Two obvious such endowments are natural resources and public infrastructure. If natural resources are introduced, there is a further modeling choice. A simple approach is to treat natural resources as equally complementary, with both labor and capital in all production activities. A more complex approach is to distinguish between activities that intensively use natural resources and those that do not.

Immobile Natural Resources as a Generalized Productivity Shift

In the first approach, a natural-resource–abundant economy would have higher returns to both labor and capital than one that was resource poor. A resource-poor economy opened from autarky might then experience an exodus of both labor and capital.

In some African countries, the natural resource endowment, broadly interpreted, is indeed so poor that factor exodus will be the main way in which globalization will help to reduce poverty. Since such economies start, in autarky, with abundant labor but little capital, the most important exodus will be their labor. Several African economies are currently at this stage. For example, Eritrea, Burkina Faso, Cape Verde, Malawi, and Lesotho all have huge expatriate communities making remittances. Typically, these economies have a large endowment of labor relative to capital, but also, and more important, have some deep-seated disadvantage (which might be thought of as a natural resource disadvantage) that makes production of an export good very difficult. Eritrea experienced 30 years of war and so has very poor infrastructure, including negative infrastructure in the form of landmines, and is also arid. Burkina Faso is landlocked and arid. Cape Verde is small and arid. Malawi is landlocked, and its international transport lifelines have been partially destroyed by neighboring wars.

In such countries, policy can promote remittances, and there is some evidence that governments could do more. For example, Ethiopia and Egypt have virtually identical populations and similar locations. Earning opportunities within Egypt are better than in Ethiopia, with per capita purchasing power parity income around eight times higher. Yet workers’ remittances are 20 times higher in Egypt than in Ethiopia. One reason for this massive difference is that the earning power of Egyptians abroad is much higher than Ethiopians, reflecting past differences in educational policies. As of 1980, 61 percent of Egyptian males were enrolled in secondary schooling, as opposed to only 11 percent of Ethiopians. A second reason is the exchange rate and financial regimes. Egypt has a convertible currency, and workers abroad can deposit in Egyptian banks at competitive interest rates (Shafik, 1998). By contrast, Ethiopia has an overvalued and inconvertible currency, with a financially repressed banking system. Thus, even when Ethiopian workers have earnings abroad, they have little incentive to repatriate their savings.

Immobile Natural Resources as an Alteration in Comparative Advantage

A more complicated model with natural resource endowments distinguishes between products that are differentially intensive in them. Thus, primary products are more intensive in natural resources than are manufactures. Adrian Wood has used variants of such a model to predict the composition of Africa’s exports. In a series of papers, Wood argues that Africa cannot industrialize because of its factor endowments: large amounts of natural resources and little human capital. Here I focus on the natural resource argument, since the human capital argument can be subsumed into it.

Wood establishes two important facts. First, he shows econometrically on a global sample that countries that are well endowed with natural resources tend not to export manufactures. He interprets this as showing that countries that are abundant in natural resources do not have a comparative advantage in manufactures, so the capital needed to produce them goes elsewhere. Second, he shows that Africa has an atypically large endowment of natural resources and does not export manufactures. His explanation for the lack of African manufactured exports follows: Africa is simply an instance of the general phenomenon.

The Wood thesis is theoretically entirely correct within its own terms. In the simple common-technology, mobile-capital, immobilenatural–resources model, countries well endowed with natural resources will not export manufactures. However, it is important to understand the mechanism whereby, in the natural-resource–augmented model, abundant natural resources would indeed crowd out manufactured exports.

Were the only difference between Africa and Asia in the endowment of natural resources, then Africa would be uncompetitive in manufactures because its labor would be more expensive than that of Asia. In effect, Africa would experience Dutch disease. The high natural resource endowment would raise the productivity of African labor and thereby crowd out manufactures. However, this would not be a problem because, after all, Africa would have higher wage levels than Asia, so that its labor force would earn more. Below, I will argue that Wood’s apparently easy passage from a statement about factor endowments to the inference that Africa cannot export manufactures via an inference from its comparative advantage is too hasty. However, I first continue my taxonomic review of the globalization models. One of these becomes my alternative explanation for Wood’s results.

Immobile Public Infrastructure

I now introduce the other type of immobile factor, namely public infrastructure. Findlay (1996) sets out a simple political economy model of the effects of globalization in a model with infrastructure. Private capital is complementary with publicly provided infrastructure. In this sense, infrastructure is like natural resources in the first of the two models discussed above: it makes the whole environment better for other factors. However, the interest of the Findlay model is that, whereas natural resources are exogenous, the amount of infrastructure is a choice variable of the government. In the model, the government has a choice between using its revenue to provide infrastructure and using it to redistribute income to its supporters.

Clearly, in an autarkic world the government can choose to spend its revenue on redistribution without dire consequences for the private capital stock. However, in a globalized world in which capital is mobile, if it persists with this choice, then private capital will relocate to countries in which the government makes the opposite choice. If the redistributive expenditures favor labor, then the exodus is confined to capital: unlike the previous model, the lack of the immobile factor does not produce a generalized factor exodus (which in poor countries would be predominantly labor).

If, as a result of globalization, the government changes its policy, then globalization hurts the social group that was previously benefiting from redistributive expenditures. More generally, the model would predict that as a result of globalization governments around the world would switch their expenditures toward infrastructure and away from redistribution.

Governments that wished to maintain redistributive expenditures would attempt to prevent their economies from globalizing by retaining restrictions on capital outflows. An example of this might be Zimbabwe.

The Findlay model may have some applicability to Africa. In Africa, the composition of public expenditure is, on one measure, more skewed toward redistributive expenditures and away from infrastructure expenditures than in Asia (Table 1).

Table 1.Infrastructure versus Redistributive Government Spending(Percentage points of GDP)
CategorySub-Saharan AfricaSouth AsiaEast Asia
Economic services5.77.26.1
General public services6.96.26.3
Source: Collier and Gunning (1997), Table 11.Note: the proxy for infrastructure is “economic services” and for redistributive expenditure “general public services” (taking the main redistribution mechanism as being the public sector payroll).
Source: Collier and Gunning (1997), Table 11.Note: the proxy for infrastructure is “economic services” and for redistributive expenditure “general public services” (taking the main redistribution mechanism as being the public sector payroll).

However, the differences shown in Table 1 are hardly dramatic. Further, there are grounds for doubting whether infrastructure expenditures are more productive than redistributive expenditures. Devarajan and others (1996) found that in developing countries as a whole government recurrent expenditure was conducive to growth whereas government capital expenditures actually reduced growth (an interpretation being that capital expenditures were more associated with rent seeking).

Policy Technology as a Generalized Productivity Shift

I now introduce the second modification to the simple-mobile–factor, common-technology model, relaxing the assumption of common technology. Technology here needs to be understood in its proper technical sense as defined by the model rather than by reference to popular imagery about machinery. Of course, it is true that producers in Africa often tend to use different machinery from that used in industrial countries. This is, however, endogenous; that is, it is the result of other differences rather than itself being the problem. For Africa to have a technological disadvantage would mean that the same bundles of inputs, machines, and labor, when used in Africa, produce less than when used elsewhere. I have already considered differences in the endowments of immobile factors as a possible explanation for such differences in the productivity of mobile factors. Here I want to abstract from this explanation and focus instead on the policy environment. As with natural resources, there is a modeling choice. Differences in policy technology can have a generalized productivity effect across the economy, or they can affect activities differentially. I first consider the former.

Assume that countries differ in their policy environments as well as in their labor forces and private capital stocks. Initially, assume that both of these are immobile. Africa has a worse policy environment than the rest of the world, and as a result the returns on both labor and private capital are reduced. Imagine, first, that goods cannot be traded internationally. Real wages will differ between countries, reflecting differences in policy and in capital: Africa will have low wages both because it has little capital and because it has poor policies. Now open economies up to trade in goods. Trade will raise average incomes but redistribute from the scarce to the abundant factor. In Africa, labor will gain and capital will lose, with labor gaining more than capital loses.

Now open up private capital to international mobility. Suppose that the African policy environment is so much worse than those of other regions that the return on capital is lower than elsewhere despite the low ratio of capital to labor. Capital will leave those countries in which it has a low return and move to where it has a high return. This will be globally Pareto-efficient: the gainers could compensate losers. The gainers are owners of private capital in initially low-return-on-capital countries, and wage earners in initially high-return-on-capital countries. The losers are owners of capital in initially high-return-on-capital countries. When U.S. financial capital shifted to the Republic of Korea because returns were higher, the beneficiaries were U.S. pensioners and Korean workers, and the losers were Korean owners of capital and U.S. workers. When Nigerian financial capital moved to the United States, the gainers were Nigerian owners of capital and U.S. workers and the losers were Nigerian owners of labor.

Even in a poor policy environment, globalization of the capital market is Pareto-efficient: the gainers (owners of African capital) could compensate the losers (owners of African labor) and still be better off. However, the mechanisms for such compensation are obviously liable to be weak. Hence, there may appear to be a presumption that unless African governments improve their policy environments, the new globalization of world capital markets will make most Africans worse off. This is wrong for a simple reason: African-owned wealth globalized many years ago. I return to this below.

Policy Technology as an Alteration in Comparative Advantage

If factors are differentially mobile, then a common reduction in factor productivity across both factors and all sectors will still change the composition of activity. If, for example, capital is mobile but labor is immobile, then capital will leave poor policy environments while labor is locked in.1 In turn, the alteration in the factor endowment will change comparative advantage. If Africa loses capital (and human capital), then it becomes more heavily concentrated in labor-intensive activities.

However, policy differences are not neutral across activities. The core point of this chapter is to argue that Africa has turned its comparative advantage away from manufacturing because its poor policy environment has been particularly hostile to manufacturing activity. The normal focus of discussion on the African policy environment vis-à-vis manufacturing is on trade policy. This is not my focus. Rather, I will discuss a constellation of policies, all of which have the affect of raising the cost of transactions.

Manufacturing is a transactions-intensive activity, much more so than natural resources and agriculture. Manufacturing involves the purchase of a wide variety of inputs from multiple sources, their storage, and the storage and sale of the output to a variety of customers in multiple destinations. While the image of manufacturing is to classify it as a “productive” activity like agriculture, it is in fact much more akin to shopkeeping. As in shopkeeping, there is a high ratio of purchased nonfactor inputs to value added. By contrast, natural resource extraction, like agriculture, has a much lower ratio of inputs to value added, and a much narrower range of suppliers and customers. Storage of inputs and outputs is much less central to the operation.

Transactions costs are high in Africa for several distinct reasons. First, transport costs are high (Amjadi and Yeats, 1995). In some countries, this is because of poor location: a higher proportion of the African population lives in countries that are landlocked than do the Asian and Latin American populations. Landlocked countries and those with poor port facilities tend to grow more slowly (Sachs and Warner, 1997). However, perhaps a more important reason is that the transport sector is insufficiently competitive. This is most obvious for air transport. The privileges given to national airlines have raised prices and reduced reliability. For example, the landlocked francophone economies have handicapped themselves by maintaining Air Afrique as a high-cost operator. In Zimbabwe, the diversion of planes of the national airline to presidential use delayed the delivery of roses in Europe for Christmas Eve, when the price was high, to the dead period between Christmas and New Year’s, when they were much less valuable. Private airlines take advantage of this situation to overprice their African routes (Okeahalam, 1996). In sea transport, preferences for nationally registered shipping have reinforced shipping cartels, raising freight rates (Amjadi and Yeats, 1995). Internally, railways are sometimes given legal monopolies over certain categories of traffic. For example, in Uganda until the early 1990s the main export, coffee, could only legally be transported by rail. The small scale of road transport also lends itself to cartels. For example, once Uganda legalized the transport of coffee by road an unexpected benefit was that the expansion of road transport broke the cartel so that road freight changes approximately halved.

Transport is not only expensive; it is unreliable. As a result, firms must keep higher levels of inventories. Efficient East Asian producers using just-in-time production methods are able to reduce inventories of inputs to only 20 minutes of production. By contrast, African firms are commonly holding 2 or 3 months worth of inputs. This is all the more striking because it is much more costly to hold inputs in Africa than in Asia, because interest rates are usually much higher, reflecting Africa’s cartelized banking system and lack of financial integration into the global economy. The unreliability of supply “ripples through the system” (Fafchamps, 1996). Let down by one supplier, a firm in turn lets down its own customers. The level of inventories necessary fully to protect a firm from unreliable supplies would be prohibitively expensive.

A second reason why transactions costs are high is that contract enforcement is difficult. Paradoxically, this is more of a problem in the modern sector than in the traditional economy, which has evolved institutions such as the kin group. Modern sector firms are usually not able to use kin groups because too few members of the group are modern sector entrepreneurs. The exception is the ethnic minority communities whose kin groups are heavily specialized. For example, in Kenya African-owned and Asian-owned firms use different means of assessing whether a potential new business customer is creditworthy (Biggs, Raturi, and Srivastava, 1996). The Asian firms largely rely upon information from their kin-based social network. The African firms, lacking such a business-oriented network, rely predominantly upon the visual inspection of the premises of the potential customer. Lacking informal enforcement mechanisms, firms need reliable courts, yet the courts function badly (Widner, 1997). In much of Africa, firms are very reluctant to resolve disputes through the courts. Where the courts are perceived as more reliable, as in Zimbabwe, around three times the proportion of disputes are settled in them as in the rest of Africa.

A third reason why transactions costs are high is because of the high cost of information. Partly, this is because of the small scale of the African business community. For example, in most countries the market for a financial press is too small to sustain the fixed costs of information gathering. A further reason is the high cost, unreliability, and low density of telecommunications. The high cost and unreliability reflect the restriction of telephone services to public monopolies.2 Telephone charges, especially for international calls, are the highest in the world, and Africa has three times the rate of faults per line as in other developing regions. Being on the telephone conveys externalities; for example, if nobody else is on the telephone, there is no point in being on oneself. Hence, the high cost and unreliability, which directly discourage use, have an additional indirect discouraging effect as a result of the reduction in telephone density. Africa has the lowest telephone density in the world. The poor phone system is an obvious direct impediment to manufacturing exports. However, it also feeds back onto the formation of social capital. Social capital is formed by the interaction of business managers. A bad phone network raises the cost of such interaction and so reduces it. Social networks are important mechanisms for the gathering of information about manufacturing techniques as well as reducing problems of contract enforcement. For example, Barr (1996) in a study of Ghanaian manufacturing firms finds that the social network of the firm substantially and significantly affects its productivity: the larger the network, the more productive the workforce. She further shows that foreign-owned firms in Ghana are significantly more productive than Ghanaian-owned firms only because they have larger networks of foreign contacts. The years of antitrade-biased policies in Ghana reduced the opportunities for businesses to make foreign contacts, so that foreign ownership provided an important social advantage. In a more outward-oriented policy environment, firms would have been able to make foreign contacts through trade, so ownership would probably have been less important.

A final reason why transactions costs are high is because of the poor quality of ancillary public services. For example, one of Africa’s exports with considerable scope for growth is fish. However, fish exports require health inspection and certification; otherwise, they are not allowed into the importing country. Fish inspectors are public officials. In Uganda, fish exporters complain that they may have to wait several days before an inspector visits, the resultant delay making the fish unexportable. Since modern retailing in European markets itself depends upon the complete reliability of supply, such interruptions may make the product completely unexportable. A second example is the working of certification of truck transportation. While load certification can have some value, its actual operation is a source of competitive corruption (Schleifer and Vishny, 1993). If many enforcement agencies and officers each take the opportunity to exact a rent, the total level of charges can be transaction-prohibiting. In the Cote d’Ivoire in 1997, the syndicate of transport operators went on strike against the predation of the police force. In response, the government confined the police to barracks; it is reported that transport costs fell. In this instance, a public service that should be reducing transactions costs was in fact significantly increasing them. A final example is the failure of duty-drawback schemes to work properly. This is important because tariff levels on inputs for potential manufactured exports are often high. Dutydrawback schemes in Africa, unlike in parts of Asia, do not make use of shortcut procedures such as standardized input coefficients. Instead, every particular drawback has to be established by documentation. This is a recipe for delay, and in the context of high inflation, delay substantially reduces the value of any eventual duty refund.

With high transactions costs, a transactions-intensive activity such as manufacturing is disadvantaged both absolutely and relative to the transactions-extensive exports of agriculture and natural resources. Recall that Wood and Berge (1997) find a negative econometric relationship between the natural resource endowment and manufactured exports. However, this may be due to an entirely different mechanism from that which he assumes. It is now established that there is a relationship from the natural resource endowment to the policy environment: abundant natural resources appear to induce a poor policy environment (Sachs and Warner, 1995). Hence, the empirical result may have arisen not primarily because of the Dutch disease mechanism, but because of the greater sensitivity of manufactures to the poor policy environment that natural resources induce. Thus, Africa has had both poor policies and large natural resources. Natural-resource exports can survive poor policies better than manufactures. For example, once tree crops are planted, it is costly to uproot them, and they stay productive for around 50 years; harvesting, and hence exports, will persist even during severe deteriorations in the policy environment. The economics of mineral extraction is analogous. Africa has thus experienced increased concentration in natural resource exports, and in some instances absolute contraction in manufactured exports. However, this outcome reflects the policy environment rather than the resource endowment. Although Wood controls for trade policy, this is only one aspect of the policy environment affecting the productivity of manufacturing. To summarize, on this thesis, Africa’s problem is not that its endowments make it intrinsically unsuited to manufacturing, but that its current policy environment makes it unsuited by raising transactions costs.

Distinguishing Between Endowment and Transactions-Cost Theses

Both Wood’s natural-resource-endowment thesis and the above transactions-cost thesis predict that Africa would have low levels of manufactured exports relative to natural-resource exports. However, they are by no means identical in their implications or predictions. In their implications, they are clearly radically different. On Wood’s thesis, Africa can forget about manufactures; there is nothing to be done. On the transactions-cost thesis, there is everything to be done. Policymakers can, by reducing transactions costs to world levels, make Africa into the most competitive region in the world for labor-intensive manufactures because of Africa’s low and relatively declining real incomes.

Since the implications are so drastically different, it is important to see how the theses can be distinguished. Wood’s thesis and the transactions-cost thesis have two testably different predictions. The first testable corollary of the Wood explanation for Africa’s failure to export manufactures is that the returns to labor would be higher in natural-resource-abundant environments, such as Africa, than in Asia. Clearly, there are instances in which the Dutch disease mechanism is sufficiently powerful to yield this result. Thus, Saudi Arabia does not export manufactures because its wage levels are too high for it to be competitive, and in Africa the same is likely to be the case for Botswana. Globally, the Dutch disease effects of natural-resource abundance will partially account for the econometric relationship that Wood establishes. However, in Africa it is evidently not the predominant explanation for the failure to export manufactures: income levels are now lower than in Asia, and the gap is widening rapidly. Africa’s greater quantity of natural resources evidently has not mapped into a higher price of labor.3 China already has a purchasing power parity per capita income higher than all sub-Saharan African countries except Botswana, Namibia, and South Africa. During the period 1990–94, it was growing 12 percent annually more rapidly than the average for sub-Saharan Africa. At such growth rates, every 6 years per capita income in China will double relative to that in Africa. By around 2006, even the higher-income African countries, such as Cote d’Ivoire, Kenya, and Zimbabwe, would have per capita incomes less than one-fifth those of China. This evidence must be qualified because there is no necessary mapping from differences in real incomes in the economy as a whole to differences in real product wages in manufacturing. First, the real exchange rate might be higher in Africa than in Asia, so that real incomes (which reflect mainly the price of nontradables) could be higher in Asia, while real product wages could be higher in Africa. However, this itself may be an outcome of policies that keep the exchange rate appreciated. Second, the wage in African manufacturing may be higher relative to the returns to labor in the economy as a whole than in Asia. This might reflect labor market policies, such as minimum wages, or the greater power of African labor to extract rent-sharing wage levels. Again, such wage premiums might be amenable to policy.

While both real exchange rate differences and labor market differences probably considerably reduce the differential in real manufacturing product wages relative to that in incomes, as the income differential widens and as policies change, Africa should become competitive. Africa’s favorable endowment of natural resources alone does not make it intrinsically uncompetitive relative to Asia’s. On the contrary, Africa will look increasingly competitive, as long as it can offer an environment in which factors can be as productive as in China.

The second testable corollary of Wood’s thesis concerns the flow of capital. Natural-resource–based activities may not need as much capital as manufacturing (although some of them clearly need more), so that Africa would not attract a large capital inflow. However, it seems hard to believe that it would, on account of having natural resource abundance, experience a capital outflow. Yet this is what has happened. Table 1 shows that by 1990 Africa had had far more capital flight than any other region relative to its total wealth, despite having a far lower stock of capital per worker.

Table 2.Capital Flight and Factor Proportions
RegionRatio of Capital Flight

to Private Wealth
Private Capital per

Worker (in U.S. dollars)
Sub-Saharan Africa0.391,069
Middle East0.393,678
Latin America0.1017,424
South Asia0.032,425
East Asia0.069,711

The remarkable exodus of capital demonstrates two things. First, it shows that African governments have never really had captive capital. During the 1970s and 1980s, when they behaved as though they had such power, they faced a massive capital exodus. Between 1970 and 1990, 39 percent of African private wealth came to be held outside the continent. Africa was thus the first continent to experience the globalization of its capital market. It was better integrated into world capital markets than other continents because its economies were small and heavily internationalized as a result of the investment pattern under colonialism. A corollary of this is that Africa has little further capital to lose from globalization. Rather, it has enormous potential benefits should globalization develop into a two-way instead of a one-way street.

The second thing that the exodus of capital demonstrates is that the African policy environment has been much worse than elsewhere. Such an exodus cannot be explained in terms of an abundance of natural resources.

Where does this leave Wood’s thesis? Evidently, in its initial form it is not sufficient to explain observed phenomena; however, it can be rehabilitated by a modification.4 The modification combines the assumptions of the Wood model with those of the generalized policy technology model: across all sectors uniformly, Africa is less productive than Asia. This produces a capital exodus and lowers African incomes, but it does not, by assumption, differentially disadvantage manufacturing. Manufacturing then gets hit again by Dutch disease from natural resources, and it is this that determines Africa’s comparative advantage.

Wood “Mark II” is consistent with the observed stylized facts of trade, wages, and capital flows, and it carries the same policy punchline as Mark I. Although in Mark II African governments can do a lot to improve the policy environment, this will not bring about manufactured exports; rather, it will lead to more productive natural-resourcebased exports. However, Wood Mark II still faces problems with the detailed facts of the African policy environment. As discussed above, this environment is peculiarly hostile to transactions. This is why, in the more extreme cases of poor policy, such as Uganda during 1971–86, the structure of the economy shifted heavily out of transactions-intensive activities and back into subsistence agriculture (Collier, 1997). Manufacturing is transactions-intensive. If this is accepted, then Wood Mark II simply falls to Occam’s razor: it is unnecessary as an explanation because the phenomenon has already been explained.

What Will Move Where in Africa?

I now return to the notion that, in a globalizing economy, Africa potentially has three degrees of freedom: labor emigration, capital mobility (in each direction), and trade. Different African countries are best suited to different ways of integrating into the global economy.

I have already suggested that one group of economies, which might be thought of as the Burkina Faso-type economies, are most likely to integrate through labor exodus. The policy issue is then to ensure that, instead of whole families leaving the country permanently, workers leave it temporarily and then repatriate their savings.

A second group of economies might be termed the Wood economies. These are the economies in which natural resource endowments are so large that Dutch disease will preclude manufacturing. I think that this applies to Angola, Botswana, and Namibia.

A third group of countries are the Findlay economies. The composition of public expenditure is hostile to capital, so there will be a capital exodus. Zimbabwe and South Africa are the potential candidates for this category. The recent decision of President Mugabe to make large unbudgeted public expenditures on war veterans is a classic instance of such behavior and of the genuine dilemmas that give rise to it.

A fourth group of countries are the high-transactions-cost economies. I would categorize this as applying to most African economies. For these economies, whether globalization achieves anything will depend upon the radical reduction in transactions costs. In the countries that succeed in doing this, there will be a capital inflow, manufacturing exports, rapid growth, and falling poverty.

Trade Policy

So far I have said nothing about trade policy and globalization, either Africa’s trade policy or that of the rest of the world. Implicit in what I have said is that trade policy is no longer the central policy aspect of Africa’s trade performance, and in this I agree with Rodrik (1997). However, it is evidently of some importance.

Much the most important aspect of trade policy is that African governments should continue to reduce trade barriers. Despite Rodrik’s evidence to the contrary,5 the balance of evidence remains heavily that trade restrictions reduce growth and that they have done so more catastrophically in Africa than any other region (Sachs and Warner, 1997; Collier and Gunning, 1997). Africa continues to have higher trade barriers than other regions (Dollar, 1992), and it participated least in the negotiated reductions in barriers during the Uruguay Round. Fortunately, as Wang and Winters (1997) show, Africa does not face very significant trade barriers in its major markets, so in one sense the lack of participation in the WTO has had little cost. However, highly informed observers such as Richard Blackhurst (the former chief economist of the WTO), think that Africa would have been able to achieve reductions in the barriers that it faces were it to have offered something during the round. Since the reductions in African trade barriers would directly benefit Africa, any further benefit in terms of reduced barriers facing Africa would be an additional bonus rather than the rationale for trade liberalization. I argue below that participation in the WTO is potentially important for African countries, but not primarily because of their opportunities for improved market access.

Probably the most important Uruguay Round change that will affect Africa is the phasing out of QRs, and in particular the demise of the MFA during the next 8 years. This is both a threat and an opportunity. It is a threat because it removes the value of Africa’s current MFA quotas. Hence, if by 2005 Africa still needs protection against Asia to be able to sell its garments and textiles, then it will lose its markets. This is very likely in parts of Africa. Conversely, in those parts of Africa that become internationally competitive by 2005, the end of the MFA is a huge opportunity. The world market in garments and textiles is so large relative to Africa’s current level of manufacture that being competitive would permit explosive growth even faster than that seen during the 1990s in Bangladesh.

What, if any, role has trade policy had in making African manufacturing more competitive? I have argued above that the core of this process concerns public policy toward reducing transactions costs. I now briefly consider whether there is a case for “industrial policy.” There is, in fact, an important role for policy toward industry. That is to confront firms with as much competition as possible, by reducing trade protection, by prosecuting cartels and removing backdoor protection through government procurement privileges. There is some evidence that trade liberalization can raise productivity both by a learning-through-exporting effect, and through increased competition. Current research by Bernard Gauthier on a panel sample of manufacturing firms in five African countries establishes that those firms that were exporting in the first round of the survey experienced significantly faster growth in productivity in the subsequent two years. This is consistent with a learning process from exporting analogous to Barr’s result on the efficacy of foreign networks discussed above.6Hay (1997) shows that the sharp increase in competition resulting from trade liberalization in Brazil raised total factor productivity by 40 percent. In Africa, the most protected industrial sector was that of Zimbabwe, which had enjoyed blanket QRs from 1965–91. A survey of 1992 found that only 7 percent of manufacturing firms regarded competition as a significant problem. Two years later, after liberalization had started to have an effect, the proportion had risen to 37 percent (Ncube and others, 1997). This strategy of raising productivity through competition is the opposite of “selective protection” to build “capabilities,” which has been advocated by some policy commentators.

African Integration into World Financial Markets

The feature of globalization on which I now focus is the increasing integration of the world’s financial markets, both for capital and for risk. I consider them in turn.

The Attraction of Capital to Africa

Before 1914, the capital surplus countries, notably the United Kingdom, were running capital outflows of around 9 percent of GDP, and the economies with new investment opportunities and low domestic savings capacity, such as Canada, were receiving capital on a similar scale. The various catastrophes of the twentieth century interrupted this process. However, the world financial system now has far better information than it did before 1914, and proportionately much higher levels of securitized assets under professional management. It also faces much lower risks of major warfare, because the precarious balance of power guarantee of peace that prevailed before 1914 has given way to the pax Americana. Hence, it should be possible to reach levels of capital flows to capital-scarce economies that proportionately exceed these earlier levels. Capital flows to developing countries are still proportionately much lower than pre–1914, but they are growing rapidly.

However, flows to Africa other than South Africa are negligible (Bhattacharya, Montiel, and Sharma, 1997). It might appear that Africa has yet to integrate into the world capital market. More properly, Africa has become detached from private capital inflows: 20 years ago, it accounted for around 9 percent of the total private capital flows to developing countries, whereas by the first half of the 1990s its market share had fallen to only 1.6 percent. Even this was an improvement upon the late 1980s; during the period 1989–94, net investment from the OECD economies was $8 billion, a fourfold increase over the period 1983–88.

The investment of the 1990s has been inward looking and confined to a small section of the world’s companies that happen to have an informational advantage about Africa. Evidence on the inward-looking nature of foreign investment to Africa comes from a survey conducted of foreign investors to five East African countries conducted in 1994. It found that virtually none had considered a location outside East Africa for their project. Presented the other way round, this shows that East Africa is not on the shortlist for internationally footloose investment projects; it was considered only for those projects that could not be located elsewhere, presumably either because of natural resources or because the investment was intended to serve the local market. While this inward orientation continues, foreign investment will be growthdependent, rather than growth-generating. Evidence on the informational segmentation of corporate investment in Africa comes from its geographic composition. More than 60 percent of OECD investment in Africa during 1989–94 came from just two medium-sized OECD economies, France and the United Kingdom. That investment to Africa was so skewed toward the two former main colonial powers suggests that other potential investors were deterred by a lack of knowledge. Investment in Africa appears to have been only for the cogniscenti: those firms that had long-established links with Africa, were overwhelmingly British or French, and were in a position to detect the improved investment climate and respond to it. Other foreign firms did not invest because of a lack of knowledge. In these two senses, Africa has yet to globalize. It is not attracting investment from a global pool of investors, and it is not attracting investment that is globally, rather than domestically, oriented.

Until recently, the explanation for this was that, in terms of the rate of return on capital, Africa was not capital-scarce: the poor policy environment had lowered returns below world levels. Collier and Gunning (1997) estimate that during the period 1960–90 the return on capital in Africa has been around one-third less than the world average. However, with the policy reforms of the 1990s this is changing: during 1990–94 the rate of return on FDI was around 60 percent higher in Africa than in other developing regions (Bhattacharya, Montiel, and Sharma, 1997). The phenomenon to be explained is that Africa is not attracting capital despite these reforms.

Africa has failed to attract private capital during the 1990s because it is perceived as a high-risk environment. One measure of this is the Institutional Investor risk ratings, which estimate country risk on a scale from 0 (high risk) to 100 (safe). On this scale, Africa deteriorated on average from 30 in 1980 to 21 in 1995, by when it was the most risky region in the world. Jasperson and others (1998) establish that there is a significant econometric relationship globally between the risk ratings and investment. Presumably, the risk ratings are proxying concerns that genuinely deter investment. We can learn more about these concerns from investor surveys. Surveys of investor intentions find that the single main deterrent to investment in Africa is the perceived high risk (Blakey, 1992; World Bank, 1994). Among risks, the most important is the fear of policy reversal: African policy environments are seen as unpredictable. Hence, the task of integrating into world capital markets, on the receiving end of capital instead of in the past as suppliers of capital, is primarily that of African governments building reputations for policy reliability.

Clearly, the most important means by which a government can build reputation is through its performance. However, elsewhere, governments have used commitment arrangements to accelerate the building of reputation beyond that which would occur purely through good performance. In the limiting case, a government that is resolved to maintain good performance simply uses commitment mechanisms as a signaling device. That is, the commitment mechanism merely signals to private agents that the government has resolve, rather than in any way changing government behavior. If the signal is effective, the commitment mechanism may be useful in accelerating the building of reputation and hence in raising investment. However, a government might wish to use a commitment mechanism not only as a signal but also as a discipline that changes its own behavior. For example, a government might resolve to avoid large budget deficits, but nevertheless find that the promulgation of a cash budget procedure provides a commitment mechanism without which it is unable to implement its resolve. In general, commitment mechanisms have both these features: they are signals to private agents about government intentions, and they are also specific procedures for converting good intentions into realized policies. In turn, a mechanism that converts intentions into realized policies will enhance credibility. The cash budget not only signals to private agents that the government is resolved to avoid large fiscal deficits; it also reassures them that a viable means has been put in place for achieving this resolve.

I refer to commitment mechanisms as “agencies of restraint.” I classify such agencies according to their domicile and the source of their power. Their domicile can be purely domestic, purely external, or a hybrid. For example, a national central bank is a purely domestic agency, donor conditionality is purely external, and regional agreements are partly domestic and partly external. Agencies of restraint can derive their power through being based on penalties, through the devolution of authority, or through some combination. For example, by establishing an independent central bank, a government is able to restrain itself from interfering with the setting of interest rates. It achieves this through shedding authority. However, an independent central bank thereby also acquires a role as a restraint over the fiscal policy of the government. If the government chooses to run a large fiscal deficit, the central bank can raise interest rates. In turn, this inflicts political costs on the government since electors dislike the increase in interest rates. Hence, an independent central bank is both a restraint upon monetary policy, based upon shedding authority, and a restraint upon fiscal policy, based upon penalties. I now consider various agencies of restraint that African governments can use to reassure investors. I first consider the use of commitment mechanisms to allay investor fears of macroeconomic instability. I start with two purely domestic agencies of restraint, an independent central bank and a cash budget, then introduce two external agencies, capital account convertibility and donor conditionality, and finally the hybrid of regional convergence criteria.

Independent Central Bank

Granting independence to the central bank is the classic way in which the government can restrain itself. The perceived effectiveness of the Bundesbank, the Federal Reserve, and the Reserve Bank of New Zealand has induced a widespread move toward the institution. In 1997, the British government gave the Bank of England independence over monetary policy, and the planned European Central Bank will extend independence to most of Europe. Many African governments have adopted legislation that devolves power onto central banks and protects the governor from dismissal. However, the evidence on the effectiveness of central bank independence in developing countries is not encouraging. Cukierman, Webb, and Negapti (1992) found no relationship between their degree of legal independence and the rate of inflation. Rather, it was the tenure of the governor that was important, suggesting that power relationships remained personalized rather than being embedded in the legal framework. The recent shift to democracy in Africa is likely to change this, making central bank independence a reality rather than a legal fiction. However, it will take time for legally independent central banks to establish their own credibility with private agents. In the short term, central banks face the same credibility problem as the government and so are not an important means for credibility enhancement.

Cash Budget

To date, the most effective domestic agency of restraint upon governments has been the cash budget. The power of this rule derives from its simplicity and its conversion of what is intrinsically a continuum into a specific quantitative target, namely a zero deficit. The principle that spending should not exceed income is readily defensible in cabinet, even though as an economic proposition it is not in fact correct. The economically correct proposition, that a deficit should not over the course of a business cycle exceed the level consistent with sustainable debt accumulation, does not produce a precise target. As a result, it does not provide a defense in cabinet against incremental spending requests.

While cash budgets have proved effective in the short term, given their lack of economic rationale it seems unlikely that they can persist as long-term restraints. They therefore gradually need to be replaced by more sustainable restraints, conceivably by independent central banks.

Capital Account Convertibility

Capital account convertibility provides a restraint upon the government because unsustainable policies are punished by capital flight. The experiences of the United Kingdom in 1992, of Mexico in 1994, and of East Asia in 1997 demonstrate the power of this mechanism to inflict punishment. However, the very fact of these currency crises also demonstrates that governments have sometimes failed to take the threat sufficiently seriously. A few African governments have recently adopted full convertibility. Since capital controls have been ineffective, the adoption of convertibility is usually unlikely to produce transitional capital flight, although this occurred in Zambia, where the adoption of convertibility preceded fiscal stabilization. Rather, the expectation would be that convertibility would lead to an initial increased inflow of capital. The potential withdrawal of this capital would then constitute the penalty-based restraint mechanism. In several economies, private capital inflows have become bunched into euphoric surges that have appreciated the exchange rate and financed consumption booms. There is clearly a need for macroeconomic management to prevent such unsustainable behavior. One possibility is for the taxation of short-term capital flows. However, this carries a high cost. It signals to potential investors that the government reserves the right to tax them, and reduces the liquidity of their investment, thereby confirming investors in their fears about African governments. Further, it is likely to be highly ineffective in practice because investors can find many illegal routes for removing capital.

Donor Conditionality

Donor conditionality has been used extensively as a macroeconomic policy restraint mechanism. Governments reach agreement with the IMF on monetary targets, which are then monitored. Breaches of the targets in principle attract aid penalties both directly through IMF ESAF financing and indirectly through cross-conditionality between IMF programs and the aid programs of other donors. In practice, donor conditionality has been quite ineffective as a means of enhancing credibility. One measure of this is that, despite being the major recipient of such conditionality, Africa is rated as the riskiest continent for investment. One reason for this is that fully three-quarters of IMF ESAF programs have missed their targets sufficiently badly for the program to be canceled or interrupted. As a result of this history, donor conditionality now faces its own credibility problem as a restraint.

Regional Stability Pacts

A hybrid agency of restraint that has yet to be used in Africa is regional agreements on a “stability pact.” Such criteria, encapsulated in the “convergence criteria,” have recently proved highly effective in Europe. European governments agreed on targets for fiscal deficits (3 percent of GDP) and for government debt (60 percent of GDP). As with cash budgets, there is no particular economic rationale for these precise numbers. However, their promulgation has a powerful political effect for two reasons. First, the quantitative target converts something that is intrinsically a continuum into something precise. This changes the politics of incremental public expenditure. While ever the cost of deficits is viewed as a continuum (which is correct in terms of economics), no single proposal for additional spending can be vetoed on the grounds of the clear and unacceptable cost of an increased deficit. By contrast, once there is a quantitative target, the political costs of incremental expenditure cease to rise as a continuum but become discrete: any proposal that would push the deficit over the target incurs high political costs and so can be rejected. Second, the power of the convergence criteria rests on the fact that governments that are accepted as peers have reached agreement on them and that many of these governments will meet them. Failure to achieve the target thus signals to the domestic electorate that its government has failed on its own terms. The European agreement on convergence criteria will eventually include fiscal penalties. However, these are probably incidental to effectiveness. To date there have been no fiscal penalties, yet the criteria have already radically changed fiscal behavior as governments make painful fiscal decisions to achieve the targets. This demonstrates that it is the existing political penalties rather than the prospective fiscal penalties that give the institution its power.

There is scope for African governments to use their regional groupings, and indeed their continental groupings (ECA, ADB, OAU) to forge stability pacts. The fiscal and debt targets need not be the same as the European criteria, nor indeed need the coverage of an agreement include or be confined to fiscal deficits and debt. However, the principle should be that African governments would themselves agree on what constituted unacceptably bad economic policies. By delineating such policies and agreeing to avoid them, governments would build political penalties against their adoption that would reduce the perceived risk facing investors. This process has dramatically reduced interest rates across Europe in the past year, and it can have a similar effect on investor confidence in Africa.

I now move from general fears about the macroeconomy to specific fears about investor rights. I consider three agencies of restraint: public insurance arrangements, credit syndication, and investment charters.

Public Insurance Agencies

The most direct way to cope with risks is to insure against them. However, the political and policy risks that concern investors in Africa are only to a very limited extent covered by commercial insurers. Instead, several OECD governments have created public insurance schemes for investment in and exports from their countries, notably the Overseas Private Investment Corporation (OPIC) in the United States and the Export Credit Guarantee Department in the United Kingdom. More recently, the World Bank has created its own insurance arm, the Multilateral Investment Guarantee Agency (MIGA). Although ostensibly insurance arrangements, these agencies are in fact better thought of as commitment mechanisms that can be used by the host government. That is, they are agencies of restraint.

The public insurance schemes act as agencies of restraint because the government of the country receiving the investment is a party to a legal agreement that commits it to compensate the insurance company for any payouts. For example, OPIC recovers 80 percent of the money it pays out to firms on claims from host governments. OPIC thus acts as an intermediary more than an insurer. Its function as an intermediary is nevertheless valuable. Firms gain comfort from the fact that their claim will be assessed impartially and payment will not be delayed. Thus, however suspect is the legal system in the host country, or however impecunious the government, if the firm has a legitimate case it will receive swift compensation. The shifting of the task of recovering compensation from the firm to the insurance agency also increases the chances that the government will indeed pay the compensation. This works in two ways: investor coordination and cross-conditionalities.

Whereas a single firm is likely only to be in dispute once with a host government that (say) arbitrarily confiscates its assets, the insurer and the government are in a long-term relationship. If the government fails to adhere to its agreement with the insurer, the insurer can refuse further business in the country. This acts as a substitute for direct coordination among future private investors. Clearly, such coordination would be impossible because most firms that will become future investors do not know that they will do so. Private firms cannot, therefore, make credible threats of investment strikes in response to government behavior. The threat of withdrawal of insurance for future investors thus acts as an indirect coordination mechanism. Were the risks instead covered by a private competitive insurance market, the insurance companies would not be able to achieve the same degree of investor coordination unless they themselves acted collectively. However, collective action among competitors encounters problems of free riding. With political risk insurance effectively supplied only by MIGA and a few national agencies such as OPIC and ECGD, the scope for collusion is much greater.

Although this coordination of investors weakens the power of the host government, the government can find it valuable. The government faces a problem brought about by the lack of credibility of its own assurances to individual investors, since ex post of investor commitment, the government is in such a strong position. The main consequence of this is not that the government is able to benefit from exploiting naive foreign investors, but that it forfeits the potential benefits of investment as streetwise investors go elsewhere. Investor coordination can thus be in the interests of the host government. An analogy is how the French government of the eighteenth century rebuilt its reputation with potential creditors after a history of bond defaults. The government itself initiated bondholder coordination by creating an officially recognized association with which it dealt. With bondholder collective action facilitated (and thereby bondholder power increased vis-à-vis the government), French government bonds became less risky, and so the government was able to raise more finance.

While the public insurance agencies work in part by indirect investor coordination, their main power probably comes from crossconditionality. Behind OPIC stands the U.S. government with its multiple powers of enforcement, including both the financial flows from the U.S. Agency for International Development and diplomatic channels. Behind MIGA stands the World Bank. Cross-conditionality between insurance claims and these other relationships does not have to be explicit for it to be effective.

The combined effect of investor coordination and cross-conditionality is that the public insurance agencies function as effective agencies of restraint. Because they recover around 80 percent of claims, their premiums are correspondingly much lower than the underlying nature of the risks involved and so much lower than would be charged by a competitive private insurance market.

However, the public insurance agencies have three disadvantages from an African perspective. First, they do not provide cover for domestic firms. Second, they are highly selective in the business they accept. Third, partly because of highly conservative capital requirements for its guarantees, MIGA is reaching the limits of its capital.

The restriction of coverage to foreign firms clearly places domestic firms at a disadvantage. In the limit, this would give rise to a situation in which African wealth was held in the OECD economies, while the African capital stock was owned by OECD-based firms. Nevertheless, African firms are able to benefit indirectly from the public insurance agencies in two ways. First, they commonly seek joint enterprises with foreign firms. Association with a foreign firm increases the security of the domestic firm vis-à-vis its own government. Second, the presence of foreign firms that have protection from government abuse of power provides some defense for domestic firms. The standards that the government must adopt toward foreign firms may spill over toward its treatment of domestic firms. More interestingly, foreign firms may regard the way in which the government treats its domestic firms as a signal of its true intentions toward the business sector as a whole, and thus as a guide to their own long-term security. As a result, the government may feel constrained in its behavior toward its captive domestic firms by the response of noncaptive foreign investment. Despite these indirect benefits, if public insurance were the main agency of restraint in Africa, it would clearly disadvantage domestic firms. Although this pro-foreign bias is not Africa-specific, since risk is much more important for investors in Africa than in other continents, its effect in Africa will be disproportionately great.

The public insurance agencies have been accused of cherry picking, in effect restricting their cover to the lowest risk propositions. One reason for this may be that their capital base does not permit them to expand their business very rapidly. The present practice is for an investment guarantee to require capital backing to the full extent of the guarantee. This highly conservative practice has led to MIGA rapidly reaching the limit of its capital base.

The combination of restriction to foreign companies and selectivity among them leaves much potential African investment uncovered by political risk insurance. Hence, it remains important to reduce the risks facing investors through the use of other agencies of restraint.

Credit Syndication

Credit syndication provides a further opportunity for a restraint mechanism. It relies for its effectiveness upon the same two processes used by the public insurance agencies: investor coordination and cross-conditionality. For example, when large mining houses make investments in high-risk locations, they typically build syndicates of as many as 60 banks from different countries. The rationale for this is not that the mining companies need the finance, since they are often among the largest financial entities in the world. Rather, by involving the banks of many OECD countries in their investment, they seek to maximize the international damage that would be done to the host government in the event of default. First, the large network of banks provides a degree of indirect coordination among the unknown population of prospective investors. Second, there is an element of crossconditionality, both because the banks that are party to the default are less likely to lend to the government, and because they are liable to lobby their own host governments for redress. Syndication functions to orchestrate the response to a default, thereby increasing the penalties and so reducing the underlying risk. Thus, while ostensibly a risk-sharing mechanism, it is better regarded as a risk-reducing mechanism.

Investment Charters

Recently, African governments have started to promulgate investment codes. Warner (1998) proposes the essential elements that any such charter should contain: basic obligations, specific obligations, and a dispute resolution mechanism. The basic obligations are of transparency, the right of establishment, and the right to equal treatment. The specific obligations are of the adoption of international standards with respect to compensation for expropriation; the avoidance of double taxation; prohibitions on corrupt practices; and a commitment to principles of competition, including the prohibition of cartels. There should be a binding settlement mechanism for disputes between investors and states, a model being that set out in the NAFTA treaty. At present, African investment charters do not usually meet these requirements. Several national charters have recently been issued by investment authorities. However, they tend not to have independent dispute resolution mechanisms, leaving the power of interpretation with either the government itself or with national courts. There have been some investment charters at the regional level. For example, both ECOWAS and the PTA include investment codes. However, these are very general, and the basis for their enforcement is unclear. An alternative advocated by Warner is for African governments to sign up to multilateral investment charters. There have been schemes for an Asia-Pacific Investment Code and an OECD-sponsored Multilateral Agreement on Investment (MAI) that might be lodged with the WTO. Warner suggests that African governments would gain more credibility by signing one of these charters than from producing one that was Africa-specific.

Currently, the proposed MAI is making significant progress and could well become adopted as a part of the architecture of global economic relations. Properly designed, this would most benefit those countries with the most severe credibility problems. Ironically, at present the group of low-income developing countries is seeking to oppose the MAI on the grounds that they do not want to be locked in to good treatment of foreign investors. This is a failure to recognize the time inconsistency problem. A refusal to build lock-in mechanisms does not lead to successful national predation of foreign investors but rather to a low level of foreign investment. Finally, I consider two measures for locking in to liberalized trade policy.

North-South Reciprocity

The model for North-South reciprocity as a lock-in mechanism for trade policy is NAFTA. In principle, the penalty of loss of access to export markets that enforces adherence could be achieved within an African regional grouping. However, the potential intra-African trade (other than that involving South Africa) is still quite limited, so such regional groups as yet cannot impose sufficiently severe penalties. Although there is currently a proposal within the United States for a NAFTA equivalent with Africa, since United States-Africa trade is quite small, the obvious equivalent for Africa is not an arrangement with the United States but with Europe. The revision of Lome would provide a negotiating context for such reciprocity.

World Trade Organization

The WTO is an external, participatory agency of restraint; a member country binds itself by accepting the GATT-WTO provisions. The most useful provision is the ability to bind tariff rates. Bound tariffs are more credible and so reduce investor risk. To date, African governments have made little use of this opportunity; few tariffs are bound, and even where they are bound they are at levels so far above actual (applied) tariffs that the binding is more likely to signal an intention subsequently to raise tariffs than to maintain a liberalized regime. For example, the government of Zimbabwe undertook to bind its tariffs at the levels in force on the day it signed the Uruguay Round. On that day, it momentarily raised all its tariffs to 100 percent. Not only are African governments failing to make use of the credibility-enhancing opportunities of the WTO, but in an important respect the WTO rules are too weak for Africa’s needs. WTO rules do not restrain trade policy in the area that in the past African trade restrictions has been most prominent: foreign exchange rationing.


Africa currently faces a credibility problem, which is deterring a capital inflow. This is particularly severe with respect to macroeconomic policy, investor rights, and trade policies. If an African government maintains good policies in these three areas, it will eventually overcome the credibility problem. I have reviewed a variety of agencies of restraint that may enable Africa to accelerate this process of reputation-building. There is particular scope for hybrid agencies that involve cooperation between African governments, sometimes in conjunction with OECD governments. There is also scope for the gradual evolution of purely domestic agencies from the present heavy reliance upon cash budgets to reliance upon independent central banks. There is also scope for greater use of instant restraints, notably capital account convertibility.

Integration into Global Risk-Bearing Markets

I now turn to potentially the most important aspect of globalization for Africa: the market in risk. It is important for Africa in two senses. First, Africa is more marginalized in the market for risk than in the markets for products and capital. Second, in many dimensions it is the riskiest continent, and yet it probably has the least (and arguably diminishing) domestic risk-bearing institutions and organizations.

I have noted above that, on the international risk ratings, Africa is the riskiest continent. However, this reflects only a small part of why economic agents in Africa face atypically high risks. In agriculture, Africa is unusually risky, both because of its semi-arid climate and lack of irrigation and because of the high prevalence of pests and disease. In business, risks are high because of the problems associated with the high transactions costs discussed above.

If information is sufficiently abundant and transactions costs are sufficiently low, then risks can be almost entirely pooled and diversified. In OECD economies, the modern corporation need not display risk averse behavior: it can hedge many of its risks at virtually no cost, and residual risks can be borne because its equity is held in portfolios that are highly diversified. As portfolios globalize, the scope for the pooling of incompletely correlated risks is increasingly being utilized.

Since traditional African economies were characterized by high risks with few liquid assets, there was a very high premium upon developing insurance mechanisms. Hence, institutions such as the extended family developed to provide a low-cost solution to the standard moral hazard and adverse selection problems of insurance. In some respects, these traditional institutions are now being undermined by the modern economy; it may now be advantageous for better-off households to opt out of their traditional obligations despite the penalities of exit. That is, there is an increasing problem of adverse selection. Perhaps more important, the modern sector itself, while facing much higher risks than its counterparts on other continents, has been unable to develop either informal insurance arrangements akin to those in the traditional economy or well-functioning formal insurance. The limited nature of formal insurance reflects the high transactions costs that have been a theme of the chapter; insurance is one of the most transactions-intensive activities. As a result, the typical household in Africa, and more especially the typical firm, must behave in a riskaverse fashion in its production and investment decisions, even if these are very costly in terms of loss of income. Thus, in Africa there is much more risk than elsewhere, and it is more expensively priced.

Yet the risks that Africa faces would be almost entirely poolable or diversifiable if the transactions costs were lower. For example, African stock markets are almost completely uncorrelated with each other, let alone with OECD stock markets. OECD pension funds could add Africa to their portfolios without any addition in overall risk despite the high riskiness of individual African stocks and markets. At present, few African stock markets even have unit trust funds that permit probably the single largest reduction in the risk of equity holdings. This makes participation in the local market prohibitively risky for the African small investor. Such investors are far safer purchasing unit trusts of OECD stocks. Again, the main reason for this is that unit trusts become inefficient as transactions costs increase.

A second set of risks that is capable of being borne by modern derivatives markets is that of commodity price fluctuations. A third is that of exchange rate fluctuations. Africa is atypically dependent upon commodities, and it has more currencies relative to GDP than any other region. Yet very little use is made of derivative markets in commodities, and no African country has a fully operating forward exchange market with associated derivatives. Virtually all African trade is at spot prices. This dependence upon spot prices is made more problematic by the atypically poor transport connections facing African businesses. There are inevitably long delays between the sale of an output by one firm and its receipt by another firm. Normally, in OECD economies, these delays would be far shorter, yet they would be covered by a range of derivatives markets; neither firm would itself take the risk that market prices would change during the delivery interval. (Risk reduction gives a further reason for the just-in-time production system.) In Africa, where producing firms are financially much weaker, they are faced with either bearing these risks or off-loading them onto traders. In the event, manufacturing firms in Africa operate to a remarkable extent upon cash-spot transactions. In effect, they buy and sell to traders rather than to each other. The trading sector thus performs an insurance function in addition to its more normal functions of transportation and information. Yet the trading sector is itself ill-equipped to perform this role, and it may help to account for the widespread perception in Africa that traders operate on wide margins.

Globalization of risk-bearing financial markets offers Africa the gains of risk-pooling. In principle, nearly all these gains would accrue to Africa, because the price of risk in Africa would fall to the world price of risk rather than the two prices meeting midway.

However, the obstacles to these gains from globalization are far greater than those to reaping the further gains from trade or the gains from capital market integration. Consider, for example, the obstacles to the creation of derivatives in African currencies. First, there are very few professional players in African currency markets. The scale of the market is too small to justify the high fixed costs of information needed to support professional speculation. Second, there is a scale threshold for the enforcement of proper conduct on the part of market participants. For example, banks require a hierarchy of supervision to prevent rogue trading. Third, derivatives contracts depend upon enforceability, which as discussed above is problematic in Africa. It is well understood that transactions in the formal credit market in Africa are limited by the low creditworthiness of potential borrowers; firms are small, they have little collateral, and there is little reliable information about their performance. Like credit, derivatives involve intertemporal contracts, in which one party will have a strong incentive to default. In effect, the constraints upon participation in the global market for risk are similar to those in the global market for manufactures. While transactions costs are high, transactions-intensive activities, whether manufacturing or financial services, will be curtailed.


I have argued that Africa has much to gain from globalization. Potentially, Africa can expand its exports into manufacturing, attract capital inflows that increase its capital stock, and participate in risk-bearing mechanisms that reduce costly risk averse behavior. However, none of these gains will accrue on present African policies. While trade and exchange rate policies have to date been at the center of African policy debates and might appear to have become even more central as a result of globalization, they are often no longer the policies most in need of reform. I have suggested two types of policies that are now important.

First, African firms operate in an environment in which transaction costs are much higher than elsewhere, directly or indirectly because of government policies. These costs differentially handicap those activities that are transactions-intensive. Both manufacturing and financial services are considerably more transactions-intensive than agriculture or natural resource extraction. These have consequences, both for the comparative advantage of Africa in product markets and for its participation in the markets at risk. In product markets, Africa has a comparative advantage in agriculture and resource extraction, and this is becoming more pronounced as transactions costs continue to fall elsewhere in the world. In the risk market, the price of risk is much higher in Africa than elsewhere: African firms face more risks and yet make the least use of risk-management financial instruments.

Second, foreign capital is deterred from Africa by perceived political risks. There are a variety of agencies of restraint that can be used to reduce these risks. Usually, they involve governments choosing to lock themselves in to policy reforms, either by building penalties against themselves or by shedding authority. To date, African governments have tended to rely upon donor conditionality as their restraint. The poor risk ratings of most African governments indicate that they need to build agencies that have greater credibility.


    AmjadiA. and A.J.Yeats1995“Have Transport Costs Contributed to the Relative Decline of Sub-Saharan African Exports?”Policy Research Working Paper 1559 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    BarrA.1996“Entrepreneurial Networks and Economic Growth” (Ph.D. thesis, Oxford UniversityOxford).

    BhattacharyaA.PJ.Montiel and S.Sharma1997“Private Capital Flows to Sub-Saharan Africa: An Overview of Trends and Determinants” (unpublished; Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    BiggsT.M.Raturi and P.Srivastava1996“Enforcement of Contracts in an African Credit Market,”RPED Discussion Paper (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    BlakeyG.G.1992“Economic Integration in Sub-Saharan Africa: Implications for Direct Foreign Investment” (Brussels: European Commission).

    • Search Google Scholar
    • Export Citation

    CollierP.1997“The Economic Consequences of Civil War,”Discussion Paper 97–18 (Oxford: Centre for the Study of African Economies).

    • Search Google Scholar
    • Export Citation

    CollierP. and J.W.Gunning1997“Explaining African Economic Performance,” Discussion Paper (Oxford: Centre for the Study of African Economies).

    • Search Google Scholar
    • Export Citation

    CollierP.J.W.GunningA.Hoeffler and C.Pattillo1998“Capital Flight as a Portfolio Choice” (unpublished; Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    CukiermanA.S.B.Webb and B.Negapti1992“Measuring the Independence of Central Banks and Its Effects on Policy Outcomes,”World Bank Economic ReviewVol. 6 pp. 35398.

    • Search Google Scholar
    • Export Citation

    DevarajanS.V.Swaroop and H.Zon1996“The Composition of Public Expenditure and Economic Growth,”Journal of Monetary EconomicsVol. 37 pp. 31344.

    • Search Google Scholar
    • Export Citation

    DollarD.1992“Outward-Oriented Developing Economies Really Do Grow More Rapidly; Evidence from 95 LDCs 1976–86,”Economic Development and Cultural ChangeVol. 40 pp. 52344.

    • Search Google Scholar
    • Export Citation

    FafchampsM.1996“The Enforcement of Commercial Contracts in Ghana,”World DevelopmentVol. 24 pp. 42748.

    FindlayR.1996“Modelling Global Interdependence: Centers, Peripheries and Frontiers,”American Economic Review Papers and ProceedingsVol. 86 pp. 4751.

    • Search Google Scholar
    • Export Citation

    HayD.A.1997“The Post–1990 Brazilian Trade Liberalisation and the Performance of Large Manufacturing Firms: Productivity, Market Share and Profits” (unpublished; Oxford: Institute of Economics and Statistics).

    • Search Google Scholar
    • Export Citation

    JaspersonF.Z.A.H.Aylward and A.D.Knox1998“The Effects of Risk on Private Investment: Africa Compared with Other Developing Regions,” in Investment and Risk in Africaed. byP.Collier and C.Pattillo (London: Macmillan).

    • Search Google Scholar
    • Export Citation

    NcubeM.P.CollierJ.W.Gunning and K.Mlambo1997“Trade Liberalisation and Regional Integration in Zimbabwe,” in Regional Integration and Trade Liberalization in Sub-Saharan Africa Volume 2: Country Case Studiesed. byA.OyejideB.Ndulu and J.W.Gunning (London: Macmillan).

    • Search Google Scholar
    • Export Citation

    OkeahalamC.C.1996“An Analysis of Costs and Pricing of Airline Services to Africa” (unpublished; Gaborone: University of Botswana).

    • Search Google Scholar
    • Export Citation

    RodrikD.1997“Trade Policy and Economic Performance in Sub-Saharan Africa” (unpublished; Cambridge, Massachusetts: John F. Kennedy School of Government, Harvard University).

    • Search Google Scholar
    • Export Citation

    SachsJ.D. and A.M.Warner1995“Natural Resources and Economic Growth,”Development Discussion Paper No. 517a (Cambridge, Massachusetts: Harvard Institute for International Development).

    • Search Google Scholar
    • Export Citation

    SachsJ.D. and A.M.Warner1997“Sources of Slow Growth in African Economies,”Journal of African EconomiesVol. 6 pp. 33576.

    SchleiferA. and R.W.Vishny1993“Corruption,”Quarterly Journal of EconomicsVol. 108 pp. 599617.

    ShafikN.1998“La Dolce Vita and the Open Door,”Trade Shocks in Developing Countriesed. byP.Collier and J.W.Gunning (Oxford: Clarendon Press).

    • Search Google Scholar
    • Export Citation

    WangZ.K. and L.A.Winters1997“Africa’s Role in Multilateral Trade Negotiations,”Policy Research Working Paper No. 1846 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    WarnerM.A.A.1998“A Charter for FDI?” in Investment and Risk in Africaed. byP.Collier and C.Pattillo (London: Macmillan).

    WidnerJ.1997“The Courts as Restraints,”Risk and Agencies of Restraint: Reducing the Risks of African Investmented. byP.Collier and C.Pattillo (London: Macmillan).

    • Search Google Scholar
    • Export Citation

    WoodA. and K.Berge1997“Exporting Manufactures: Human Resources, Natural Resources and Trade Policy,”Journal of Development StudiesVol. 34 pp. 3559.

    • Search Google Scholar
    • Export Citation

    WoodA.K.Berge and T.Owens1997“Export-Oriented Industrialisation through Primary Processing?”World DevelopmentVol. 25 pp. 145370.

    • Search Google Scholar
    • Export Citation

    World Bank1994“East Africa Survey of Foreign Investors” (unpublished; Washington: World Bank).

1Were the analysis to be extended to human capital, it might be regarded as being of intermediate mobility, less mobile than capital but more mobile than unskilled labor. Hence, Africa’s shortage of human capital might be seen as partly endogenous to its poor policy environment. There are, for example, reputedly more Malawian doctors in Manchester than in Malawi.
2Even when these are privatized, the monopoly is sometimes retained, as recently in Côte d’Ivoire.
3Similarly, even though the quantity of human capital in Africa is less than in Asia, it does not necessarily follow that its price is higher than in Asia, because the demand for human capital might be lower than in Asia if transactions-intensive activities are also intensive in human capital.
4Wood has proposed this modification to me in discussion, though I do not wish to imply by this that he necessarily accepts that such a modification is necessary.
5This relies only upon an African sample, and since intra-African variation in trade policy is much more limited than global variation, it is to be expected that trade policy would appear to be less significant as a determinant of growth.
6It is, however, not conclusive. Firms with better managements could both be better at exporting and have more rapid efficiency growth.

    Other Resources Citing This Publication