Global Integration of Africa versus Regional Integration in Africa

Proposals and schemes for regional integration have abounded in postindependence Afiica, often with disappointing results. Recently there has been a revival of interest in Afiican economic integration. Unlike in the past, regional integration is now motivated as a way to open up Afiican economies. This paper assesses whether regional integration amongst Afiican states is the best way to achieve open economies. The conclusion is that regional integration between Afiican states may be of limited value, apart from possibly contributing to better co-ordination of sound macro-economic policies. Instead, integration between Afiican countries and higher-income regions, such as Euroland, may be a preferred strategy. JELF 15


INTRODUCTION
According to the United Nation's Human Development Index (HDI), 18 of the 20 least developed countries in the world are in Afiica. Afiica's unsatisfactory economic development experience has in recent years come increasingly under scrutiny. The so-called Afiican development "tragedy" is reflected in persistent poverty, low growth and high inequality. During the 1980s, the average growth in GDP per capita in Sub-Saharan Afiica (SSA) was 1.3 per cent. Current estimates put per capita income in Afiica (in PPP terms) at roughly US$ 1045 more than 10 times lower than that in the average OECD country. The result of low growth and low per capita incomes is that more than 40 per cent of the African population may be living below the accepted poverty line ofUS$ 1 per day. Taking into account measurement methods, Afiica also has the most unequal distribution of income in the world -exceeding that of Latin America.
On a political level, calls have increasingly been made for Afiica to reverse this process of economic underdevelopment through "Afiican unity". Calls for and discussions about an "Afiican Century", "The Afiican Renaissance", a "United States of Africa" has all been prominent in recent times. Of course, calls for African unity is nothing new -the Abuja Treaty of 1948 having already called for an African Economic Union by the year 2000. This paper critically assesses, from an economic point of view, the potential of regional economic integration amongst African countries for promoting economic development in Africa. It does so by considering the recent explanations for Africa's poor economic development and by assessing the African experience with regional integration. The insights from these analyses are then combined with recent theoretical advances in understanding the extent to which economic integration schemes leads to divergence or convergence of per capita incomes amongst members states. The paper concludes, with reference to the recent Free Trade Agreement between South Africa and the EU, that integration amongst African states may be sub-optimal to integration between African states and a higher income region (e.g. the EU) and that regional cooperation in terms of economic policy and political stability may be vital in the face of significant neighbourhood effects on economic growth.
Seen in this way, the contribution of this paper is to call on African states to focus on economic integration with countries outside of Africa for expansion of trade and for obtaining technology and investment. In this, the paper sounds a positive note on the advantages of globalisation for African economic development. African countries should furthermore not expect of economic integration agreements amongst themselves to lead necessarily to greater trade. The advantage of regional integration amongst African countries lies rather in its potential to enhance economic policy co-ordination and the adoption of sound macroeconomic policies and governance practices. The economic development case for regional integration between Euroland and Southern Africa can thus be made.
The paper is structured as follows. In section two it is showed that there is a great consensus that a major cause of Africa's disappointing economic performance is that fact that African economies are amongst the most closed economies in the world to international trade. It is argued that regional integration is called for as a way to ensure trade liberalisation but on a limited scale. In section three Africa's experience with regional integration in the past is seen not to have been successful, and that it contributed little to intra-regional trade in Africa. In fact, the argument that regional integration can assist to enlarge the markets for African products is shown to be false, as most African countries have similar revealed comparative advantages and trade in the same basic commodities. In section four, some new contributions to the theory of regional integration are set out. These contributions can be applied to the African situation to argue that regional integration may present Africa with a lose-lose situation and that integra-tion between African economies and a higher-income economy may be preferential. Section five concludes.

EXPLAINING AFRICA'S ECONOMIC DEVELOPMENT CRISIS
The economic crisis in Africa has led scientists and policy makers to use a number of approaches to gather evidence as to the causes of the crisis. These approaches included use of cross-country growth regressions, case studies as well as micro-economic survey evidence on households, markets, firms, labourers and investors.
Micro-economic longitudinal surveys have been conducted with increasing regularity in African countries since the early 1990s. These include household surveys and so-called Living Standard Measurement Surveys that have indicated that macro-economic reforms and greater openness of African economies can lead to improvements in living standards (see e.g. Demery). They also include the World Bank's Regional Programme on Enterprise Development (RPED) that surveyed firms in eight African countries. These surveys supported recent policy implications of the endogenous growth literature 2 but also supported the macro-economic evidence of the cross-country regressions as far as suggesting the importance of lowering the riskiness of the African environment and of opening up African economies to foreign trade.
Cross-country growth regressions have been popular in economics since the revival of growth theory and the rise of the so-called endogenous growth theory, see for instance Barro (1991) and Sala-I-Martin (1997). In these regressions, which are typically run using some adaptation of the Solow-model, the dummy variable for "Africa" as a region was initially consistently significant. A number of authors (e.g. Easterly & Levine, 1997) expanded the set of variables in order to account for this "African dummy". The found that six broad sets of variables were significant in explaining the African dummy, namely The lack of social capital in African communities coupled with high levels of ethnic diversity; The lack of openness to trade; Deficient public services; Africa's geography and high riskiness of investment in Africa; The lack of financial depth of African economies; and Africa's high aid dependence.
From the perspective of regional integration, many calls for regional integration amongst African countries in recent years have been motivated by the finding that African economies need to be more open to international trade. Thus, regional integration is seen as a reciprocal way of lowering trade barriers, but without subjecting previously sheltered industries to the full force of international competition.
Although it will be shown that this policy prescription is wrong, the diagnosis is correct and there is now a significant consensus that one of the most serious causes of Africa's slow growth has been its lack of openness to trade. As noted by Jenkins and Thomas (1999: 4) with reference to the cross-country regressions "Almost all of these concerned exclusively with Africa find that impediments to trade have been detrimental to growth perfonnance, reducing the annual growth rate by between 0.4 and 1.2 percentage points".
In the regressions of Sachs and Warner (1995) openness to international trade is based on a composite indicator consisting of the size of tariffs, the extent of nontariff barriers, the difference between the black market and official exchange rates, the type of economic system, and the extent of state control over major exports. Sachs and Warner (1995) fmd that most of the countries in Africa were effectively closed to international trade for the 25 years from 1965 to 1990. Only Botswana and Mauritius were "open" for any length of time.
The lack of openness was especially negative as far as it contributed to the failure of manufacturing in Africa. This is firstly because closed economies meant that the input prices of capital and intennediate inputs were above world market prices. Secondly, the closed domestic markets were small, preventing scope for either economies of scale or domestic competition. As a result, African manufacturing !inns remained inefficient, which in turn caused reduced investment and skills fonnation (Jenkins & Thomas, 1999: 9). The overall impact of this vicious cycle is the low growth rates in Africa and the inability of African countries' per capita incomes to convergence to levels of other countries. Ng and Yeats (1999) found that in Africa, countries that adopted less restrictive governance and trade policies were in fact able to achieve significantly higher levels of per capita GDP as well as higher growth rates for exports.
Many countries have embarked on regional integration as a "reciprocal" manner to liberalise trade as opposed to unilateral trade liberalisation. In this light, it can be understood that from an economic point of view, calls for African regional integration in 2000 are different from the calls that were made for African regional integration in the years immediately after independence. Then, the calls were made based on arguments for larger markets, as well as political arguments. Today the calls are made to ensure the opening up of African economies (Collier & Gunning, 1995: 387).

REGIONAL INTEGRATION IN AFRICA
Most of the current regional integration groupings in Africa have been in existence for over 30 years. The major regional economic integration agreements in Africa is summarised in Table l.  Table I shows that there are currently 9 major regional integration schemes in SSA. In order to assess the feasibility of a 'United States of Africa", one must judge the success of these schemes in contributing to the economic development of their member states. A large literature exists on regional integration in Africa (see e.g. Yeats, 1998). From this literature, it seems fair to draw out the following "stylised facts" of African regional integration.
First, despite decades of regional integration agreements, intra-regional trade still comprises only 12 per cent of total African exports.
Second, in the trade patterns amongst African countries, trade between only 5 countries completely dominates intra-African trade (excluding South Africa). These are Cote d'Ivoire, Nigeria, Kenya, Zimbabwe and Ghana. Together the trade between these 5 countries account for over 75 per cent of intra-regional trade.
Third, African intra-regional trade is dominated by four products. None are manufacturing goods; all are commodities namely petroleum, cotton, maize and cocoa. These commodities are responsible for 50 per cent of intra-African trade.
Fourth, despite decades of regional integration, the percentage of intra-regional trade in Africa has not increased significantly. In other words, there is little evidence that regional integration in Africa has lead to trade creation. Table 2 summarises the percentage share of intra-regional trade in the various regional integration groupings in 1970, 1985 and 1993.  Table 2 indicates that in over 30 years, the movement towards regional integration has done little to contribute to trade amongst African countries. In major groupings such as the PTA and SADC, trade has actually declined in percentage SAJEMS NS Vol 4 N03 tenns. This suggests that African countries' trade with the rest of the world has been increasing must faster that intra-regional trade. Other evidence seems to suggest that this integration of Africa with the rest of the world (outside Africa) is not only a pattern in trade in goods, but also in tenns of capital markets -for instance through capital flight from the continent. Collier for instance, argues that up to 30 per cent of the wealth of Africa could currently be held outside the continent.
Fifth, if one considers the imports of SSA countries, almost 75 per cent of this consists of capital goods such as machinery and transport equipment. The current economic structure of SSA countries is such that they have very little capacity to meet these needs and many African countries show a Revealed Comparative Advantage (RCA) in the same products. In other words, -there seem to be very little motivation for African countries to trade with one another given the lack of complementarity in trade patterns.
The above overview of regional integration has shown that regional integration in Africa is unlikely to lead to substantial trade between African countries. As supported by Yeats (1998: 2) this makes less compelling arguments that regional trade can help overcome problems of small domestic markets in African countries. Moreover, if the contributions from the new theory of regional integration are correct, then indeed regional integration may present Africa with a lose-lose situation. These will be outlined in the next section.

AFRICA AND THE THEORY OF REGIONAL INTEGRATION
The classic contribution towards understanding the economics of regional integration had been made by Viner (1950). Viner used the terms trade creation and trade diversion to detennine in which instances it would be beneficial for countries to engage in regional integration. Baldwin and Venables (1995) contain a good overview of the subsequent literature on understanding trade creation and trade diversion.
The practical experience of various regional integration agreements has supported the implication in Viner's and other's work that the effect of regional integration on the welfare of a participating country may be ambiguous. Venables (1999) point to the experience of the EU where regional integration has led to convergence of per capita incomes amongst member states. However, in ECOW AS and other agreements, regional integration has led to divergence in per capita incomes.
To predict who gains and who loses in a regional integration agreement, Venables (1999) has recently integrated two strands of research, namely research on the comparative advantage of countries and research on the importance of agglomeration forces. With reference to the former strand of research, Venables (1999: ii) concludes: Typically, the country in the free trade agreement (FTA) that has comparative advantage most different from the world average is most at risk from trade diversion. Thus, if a group of low income countries form an FT A, there will be a tendency for the lowest income members to suffer real income loss due to trade diversion. In contrast, if an FT A contains a high income country (relative to other members and to the world average) then lower income members are likely to converge with the high income partner.
Venables' contribution to incorporate agglomeration forces into the understanding of regional integration is a useful contribution that contains vital insights from the so-called new economic geography. Given the great concern in Africa about the continent's lack of manufacturing development, it is necessary to provide a more detailed explanation.
Regional integration may lead to welfare-inducing changes in the agglomeration of economic activities in member countries. This is because regional integration (through for example an FT A) makes it easier to supply consumers from a few locations. This would lead to the relocation of industries in either one of two ways, each with different implications for convergence of divergence of per capita incomes between countries. The first is when the regional integration causes particular sectors to become more spatially concentrated -e.g. a certain country experiences an agglomeration on financial services, another of manufacturing enterprises, another of commercial agriculture, etc. In such a case, despite adjustment costs, the net effect of regional integration may not necessarily be greater inequality between countries. Ifhowever, regional integration leads to manufacturing as a whole to cluster in a few locations, it may lead to the de-industrialisation of less-favoured regions. This will lead to divergence in per capita incomes amongst member states. As shown by Venables (1999: 17) this is more likely to occur if manufacturing as a whole is a small share of the economy -as in most African economies.
The agglomeration forces descnbed above may interact with competitive advantage forces. Specifically, agglomeration may accentuate the competitive advantage force in the direction of divergence in the case of developing countries (Venables, 1999: 18).
The conclusion from the above analysis, when combined with the stylised facts on African regional integration is that regional integration schemes amongst African countries could lead to divergence in incomes due to agglomeration forces, as well as the comparative advantages that will be enjoyed be only a few countries. Schiff (1999) extends the theory of regional integration by asking not whether regional integration should be pursued or not, but that if it is pursued, what criteria should be used by countries to decide on a partner country or block. This question is relevant in the present context, since if it is the case that regional integration amongst African countries could lead to divergence in per capita income, and regional integration with higher income countries is preferred, a decision should be made about which country to integrate with. Schiff (1999) shows that pre-FTA (Free Trade Area) volume of trade is not a useful criterion for selecting a partner bloc. Instead, he concludes that: A home country will be better off with a large partner country (or bloc). This is because a large partner country is more likely to satisfy the home country's import demand at world prices. Thus, for SADe countries, it will be more beneficial to integrate with Euroland than with South Africa. The FT A as a whole is likely to be better off if each country imports what the other exports (rather than each country importing what the other imports, as was shown in section 3 to be the case in Africa).
The recommendation from the above is that African countries should pursue unilateral trade liberalisaton on a most favoured nation (MFN) basis as a preferred strategy to regional integration, with a FT A with a higher income region the second option. However, it must be added here that this does not mean that regional integration groupings such as those listed in section 3 should necessarily be abandoned. There is a strong case to be made for the maintenance and expansion of these. This is because there is empirical evidence that significant "neighbourhood" effects exist in Africa. Easterly and Levine (1995) found that there are significant spillovers of growth performance between neighbouring countries in Africa. They conclude that if neighbouring countries in Africa act together to improve economic policies and political stability, the effects on the growth rates of all are much greater. Specifically, the results suggest that the effect of neighbours adopting a policy change (e.g. opening up trade) is 2.2 times greater than if a single country acted alone.

SUMMARY AND CONCLUSIONS
This paper critically assessed, from an economic point of view, the potential of regional economic integration amongst African countries for promoting economic development in Africa. It did so by considering the recent explanations for Africa's poor economic development and by assessing the African experience with regional integration.
The insights from this analysis were combined with recent theoretical advances in understanding the extent to which economic integration schemes leads to divergence or convergence of per capita incomes amongst members states.
It may be concluded, with reference to the recent Free Trade Agreement between South Africa and the EU, that integration amongst African states may be sub-optimal to integration between African states and a higher income region (e.g. the EU). Specifically, regional trade agreements between SSA countries could be of limited value and could even lead to trade diversion and a divergence of per capita incomes amongst member states.
Older arguments for African regional trade integration agreements centred round the lack of market size in individual African countries. However, given the WTO's GATT, the Lome Agreement with the EU and the Generalised System of Preferences (GSP), few external obstacles to fmding markets for African countries' products still exist.
The recommendations following from this paper are therefore that instead of first and foremost promoting regional integration and/or a United States of Africa, that African policy-makers focus on the following: The liberalisation of import barriers on a most favoured nation (MFN) basis to increase the openness of their economies; The concluding of a Free Trade Agreement (FT A) with high-income country (-ies) in the form of North-South regional integration. The South African FTA with the EU and possibility of extending this agreement to SADC offers a starting point. The use of regional integration agreements not with the primaI)' expectation of leading to increased trade or regional convergence, but to achieve mechanisms of co-operation between African countries in terms of macroeconomic policy making and for providing a politically and economically stable environment to facilitate the repatriating of the huge amounts of African wealth currently held outside of the continent.

SAJEMS NS Vol4 (2001) No 3
The advantages of the above strategy are that trade creation occurs, which could facilitate convergence between African per capita incomes and that of the rest of the world, that Foreign Direct Investment and technology transfers are raised, that cheaper inputs needed can be sourced and that a lock-in (commitment) mechanism is created for reducing the risk of African countries reneging on trade liberalisation. It also alleviates the problems of negative neighbourhood effects in Africa.