Working Paper for the series ‘Governance, State and Development in Sub-Saharan Africa’, Elcano Royal Institute, Madrid.

Introduction: The Political Economy of Taxation and Tax Reform in Developing Countries

The process of tax collection is one of the most powerful lenses in political economy to assess the distribution of power and the legitimacy of the state and of powerful interest groups in civil society. The collection of tax not only requires substantial coercive power, but more importantly requires a state to be legitimate since the vast majority of tax is collected when there is a high level of voluntary compliance (Levi, 1988). Douglass North, for instance, defines the state in terms of taxation powers: ‘… an organization with a comparative advantage in violence, extending over a geographic area whose boundaries are determined by its power to tax constituents’ (North, 1981, p. 21). Long before that, Edmund Burke remarked: ‘Revenue is the chief preoccupation of the state. Nay more it is the state’ (quoted in O’Brien (2001, p. 25).

Taxation is inherently political. In the early 20th century, Joseph Schumpeter once wrote: ‘Taxes not only helped create the state; they helped form it. Schumpeter also famously observed: ‘The fiscal history of a people is above all an essential part of its general history’ (Schumpeter, 1918, p. 1954). Indeed, there is a long history of thinking in political economy and history that links the process of state-building with the capacity of rulers to collect taxation (Tilly, 1990; Brewer, 1990).

Tax collection also reflects basic core capacities of states to collect vast amounts of information which is essential for the formulation of informed policy decisions. The administrative apparatus required to collect and monitor the information required to develop a tax base is one of the most challenging technical and political functions a state can undertake. Thus, taxation has always acted as a key incentive for states to create competent bureaucracies.

In sum, taxation and tax reform are central to state-building for several reasons. First, governments must be able to ensure sustainable funding for social programmes, and for public investments to promote economic growth and development. Because aid generally diminishes over time and is often volatile, domestic resources are necessary to sustain these institutions and programmes. Secondly, taxation is the main nexus that binds state officials with interest groups and citizens. Not only can taxation enhance government accountability, it also provides a focal point around which interest groups (such as producers groups, labour unions and consumer groups) can mobilise to support, resist and even propose tax policies. In other words, taxation is as constitutive of state formation as it is of interest-group formation. Third, taxation, particularly in the form of land and property taxes, customs and border collection can help increase the territorial reach of the state. The diversity of the tax base is a telling indicator of the ability of the state to engage with different sectors and regions and is indicative of the degree to which state authority permeates society. There is a long history of evidence that supports the notion that economic and political development cannot easily happen without a consolidated central state. Fourth, fiscal capacities are needed to build a legitimate state. Democratic elections do not in themselves ensure state legitimacy. Neither do ‘quick-impact projects’ in which aid agencies seek to fill urgent needs. Legitimacy comes in large part from government delivery of services that people want and need. Elections provide an avenue for the citizenry to voice demands; responding to those demands requires capacity to mobilise, allocate and spend public resources effectively.

Much of the work on taxation, and particularly economic and administrative approaches, has been couched in technical, non-political terms. The focus of these approaches has been concerned with how economic structures and levels of development, on the one hand, and administrative capacity, on the other, affect the capacity of states to mobilise tax resources. The technical, non-political approach to taxation is prevalent in IMF and World Bank advice on tax reform. This is part of the larger reform agenda where state capacity-building has been viewed largely as a ‘technical’ exercise in administrative reform (raising wages of civil servants, more training, greater meritocracy).

According to the diagnosis of the capacity approach, ‘poor governance’ is the result of an over-extended state relative to its institutional capacity at a given moment in time (see World Bank 1997, p. 61-75). The analysis of governance crucially assumes that inherited capacity constrains and that this constraint is what should orient the shape of administrative, institutional and policy reform. The policy advice, therefore, for poorly-performing economies generally advocates reducing the state’s role in resource allocation decisions. The main message of the capacity approach is ‘don’t try difficult interventions and reforms at home’. The technical and apolitical nature of the good governance agenda, however, limits an understanding of the political and institutional processes underlying the power and legitimacy a state requires to enforce and change rights and institutions, and to extract and mobilise the resources required to sustain development and growth.

Surprisingly, taxation is not explicitly listed as a separate ‘fundamental’ task of a state (as spelled out in the World Bank Development Report, 1997).[1] This error of omission is indeed remarkable given the centrality of revenue production and resource mobilisation in the historical process of state formation (Schumpeter, 1954 [1918]; Tilly, 1990). The neglect of making tax central to understanding state capacity and governance reflects the decline in the political economy of resource mobilisation as a focal point of development theory and policy.

In the wake of fiscal crises of the state in sub-Saharan Africa and Latin America, designing tax systems that can provide incentives for growth, can meet distributional demands and can increase revenue collection is central to state viability and effectiveness. In post-war economies, reconstruction of the revenue base is essential for the reconstruction of a viable state and sustained peace.

Much of the discourse on governance and state-building has taken place without incorporating analysis as to how states are to finance its even most basic functions. Issues of democratisation and transparency are important, but one has to ask where the domestic resources to finance public goods and services (both crucial for building state legitimacy) can be found, in ways that do not compromise fiscal solvency and economic efficiency.

Finally, taxation is one of the few objective indices we have that measures both the power and legitimacy of the state (in this case, to mobilise resources). Tax data is relatively easy to collect and is generally reliable. Other well known indices of governance such as ‘corruption’ or ‘participation’ are much more indirect and vague as measures and rely in subjective surveys.

The purpose of this paper is to present some key theoretical and policy debates concerning the relationships between taxation, aid, governance and political organisation in the political economy of development in Sub-Saharan Africa. The paper focuses on three main areas: (1) theory and policy debates with respect to taxation in sub-Saharan Africa; (2) the extent to which mineral abundance is a curse or blessing for growth and political stability; and (3) how and why political organisations are central to understanding state resilience in Africa.

Structural Factors Limiting Tax Take

Before discussing the particular challenges of taxation in low-income countries, let us first consider the structural reasons behind the generally lower tax take in such economies. An important component of the applied literature on tax indeed concentrates on why the level and composition of taxes in less developed countries differs from more advanced countries. One important set of factors concerns the economic structure of developing countries. These include:

  1. A large share of (subsistence) agriculture in total output and employment.
  2. Large informal sector and occupations.
  3. Many small establishments.
  4. Small share of wages in total national income.
  5. Small share of total consumer spending made in large, modern establishments.

Combined, these factors mean that the take share as a percentage of GDP tends to be much lower than in countries with greater levels of per capita income.

In comparison with OECD, the most striking differences of LDCs include:

  • Low usage of social security taxes (largest values are in ex-socialist transition economies).
  • High revenues from trade taxes.
  • High levels of non-tax revenue (especially from mineral rents).
  • Higher share of tax revenue from companies rather than individuals (thus much lower personal income tax).
  • Much more narrow base of tax payers (hence the importance of large taxpayers office (LTOs).
  • Higher rates of tax evasion.

The applied theory is generally supported by empirical evidence (see Figure 1).

Figure 1. Tax revenue as a percentage of GDP by GDP/capita category

image001 29 1
Source: Bird & Zolt (2005).

As can be seen, the tax share as a percentage of GDP, on average, increases with increases in GDP. It is, however, important to note that there are substantial variations between countries with similar per capita incomes. Within the OECD, countries dominated by social-democratic parties and labour unions have tax shares of over 45% of GDP (eg, Sweden and the Netherlands), while countries with weaker left-centre parties and labour unions have shares below 40% of GDP (eg, the US and Japan). Within LDCs, there is also substantial variation for both low-income and middle-income countries. South Africa and Brazil collect over 35% of GDP in taxes while Colombia and Mexico collect less than 15% of GDP in taxes. Mineral and fuel abundant LDCs such as the Gulf States, Algeria, Zambia, Chile, Botswana and Malaysia also tend to have higher tax takes than would be predicted by their income per capita levels (although other, such as the Democratic Republic of Congo, have tax ratios below 10% of GDP).

An example of the variation of taxation can be seen within sub-Saharan Africa, as indicated in Table 1.

Table 1. Tax collection and composition in selected Sub-Saharan Africanountries

 YearsTax RevenueTrade TaxesGDP/cap
  (as % of GDP)(as % of total taxes)(market prices*)
Lower-tax countries    
Congo (DR)1998-20024.532.0600
Central African Rep.1992-966.139.01,055
Average 9.730.3814
Higher tax countries    
South Africa1998-200225.513.08,764
Ivory Coast1991-9918.040.01,582
Average 21.420.62,420
Average (excl. Botswana, S.Africa)  1,363

(1) At 2000 market prices in US$.
Source: IMF, Government Finance Statistics.

There are several points worth considering with respect to the data in the table. First, as standard theory predicts, low-tax countries tend to have much lower income per capita and tend to be much more reliant on trade taxes, which means that the fiscal consequences of trade liberalisation can be devastating if alternate forms of tax are not quickly increased. However, income per capita is not necessarily associated with higher tax takes. For instance, there are many countries with a lower income per capita than the Central African Republic and Uganda) that collect a much higher share of taxes as a percentage of GDP. Secondly, the level of tax collection does not necessarily indicate that the state has the capacity to promote rapid economic growth. Uganda, Mozambique and Tanzania have been among the fastest-growing African economies in the period 1900-2005 yet have relatively low tax capacity. South Africa and Zimbabwe have higher tax capacity but have not had nearly as impressive growth rates over the same period. Finally, tax levels do not necessarily indicate that a state or government is legitimate. Recent episodes of political violence in Kenya and Zimbabwe, two relatively high-tax states, are examples that show that relatively high tax collection does not preclude violent challenges to state authority. In these two cases, further research is needed to explain if high tax rates were the result of compliance/consent, administrative effectiveness or unsustainable levels of coercion.

The empirical evidence also supports the argument that tax composition changes with increases in per capita income (see Table 2).

Table 2. Tax structure by region, percentage of total tax revenue, 1975-2002

Tax structure by region, percentage of total tax revenue, 1975-2002

Sources: Bird & Zolt (2005).

The most notable challenge for low-income countries, particularly in Africa, is that governments are very dependent on trade taxes. The dependence on trade taxes in low-income/post-war economies presents specific policy challenges. Trade liberalisation in these economies has led to reductions in trade taxes, which are the main source of revenue in weak and low-income states. Moreover, alternative tax revenue (such as from value-added –VAT– and income tax) have risen significantly less than the decline in trade-tax revenue. The overall effect has been a decline in total tax revenues as a percentage of national income in low-income countries. Evidence presented by the IMF (Baunsgaard & Keen, 2005) shows that low-income countries typically recover only 30 cents on each US dollar lost to trade-tax declines.

Taxation and Resource Mobilisation in Broader Perspective

It is important to note that the mainstream economic literature on tax, however, does not consider the wider resource mobilisation question, which was a concern of earlier development economists (eg, Lewis, 1954). As indicated in Table 3, while tax revenues in Sub-Saharan African and Latin American countries from the mid-1980s to 2000 were collected at a similar proportion to GDP as in East Asia, there were dramatic differences in the savings rates between the regions.

Table 3. Resource mobilisation and economic growth in developing countries: regional comparisons

 GDP % growth (1)Tax revenues (% GDP) (2)Gross Savings (%GDP)
Sub-Saharan Africa-0.421.716.313.912.512.7
South Asia3.312.812.213.516.716.8
East Asia & Pacific6.11515.630.831.631.2
Latin America0.815.215.921.718.918.9

Sources: (1) World Bank, World Development Indicators; (2) IMF Government Financial Statistics and author’s calculations.

The East Asian savings-rate averages were more than double as a percentage of GDP compared with South Asia and sub-Saharan Africa and two-thirds higher than in Latin America.

The state’s capacity to mobilise resources beyond taxation is one important feature of developmental success stories that the economic literature on tax misses. In particular, high levels of gross domestic savings have supported robust investment rates. The East Asian economies were in a class of their own in terms of savings rates. This was largely achieved through the coercive power of the state, which was deployed to mobilise resources through various forms of forced savings. Among the coercive elements in East Asian economies were restrictions on consumer credit, financial restraint, mandatory provident pension contributions (used in Singapore and Malaysia) and encouragement of postal savings. Although state actions to increase savings are clear in East Asia, the high and sustained growth rates may have also had an important feedback effect on income growth and therefore on sustaining savings.

This lacuna in the economic approach is important to note because much of the taxation literature assumes that a state’s legitimacy is enhanced when there is a consensus around tax collection. However, economic growth and employment creation are also important sources of legitimacy for a state. Since there is no clear relationship between tax levels and composition and economic growth, it is important to consider the role of taxation in the context of the wider resource-mobilisation challenges of late-developing economies.

Main Challenges in Mobilising Resources in Sub-Saharan Africa

In Sub-Saharan Africa, improving taxation to meet developmental needs is one of the main challenges facing the region (Gupta & Tareq, 2008). The average tax-to-GDP ratio in Sub-Saharan Africa has increased from less than 15% of GDP in 1980 to more than 18% in 2005. But virtually the entire increase in tax revenue in the region came from natural-resource taxes, such as income from production sharing, royalties and corporate income tax on oil and mining companies. Non-resource-related revenue increased by less than 1% of GDP over 25 years. Even in resource-rich countries, non-resource-related revenue has essentially been stagnant (Keen & Mansour, 2008).

Also, in many of Africa’s low-income oil importers, domestic revenue mobilisation has not kept pace with rising public spending. As a result, a growing share of current spending is financed by aid. For example, from 1997-99 to 2004-06, the share of current spending financed by aid (including debt relief) increased fivefold, from 16% to 36% in Ghana, from 22% to 40% in Tanzania, and from 60% to 70% in Uganda (Gupta & Tareq, 2008). Thus, improved taxation is the only route out of aid dependence.

The challenges of tax collection are formidable in low-income and especially in low-income post-war economies. First, the tax base is relatively low, dependent to a large measure on trade taxes and is extremely narrow. Secondly, there is an urgent need to widen the coverage of the tax base in these countries and to examine the political economy of large taxpayer offices in the government. In the post-war economies of the Democratic Republic of Congo, Rwanda and Uganda, for example, the most salient features are that the tax base is relatively low, dependent to a large measure on trade taxes, and is extremely narrow where ‘large’ payers (who are generally in the range of 300-2,000) contribute between 40% and 70% of domestic revenue collection.

Perhaps the greatest challenge facing low-income African economies is how to replace declining trade taxes in the face of economic liberalisation. Trade taxes represent over one-third of all tax revenues in Sub-Saharan African economies. This degree of dependence on trade taxes is substantially higher in Sub-Saharan Africa compared with other regions (Bird & Zolt, 2005). Trade taxes are often the main source of revenue in weak and low-income states. Problems of domestic revenue-raising have been exacerbated by a global shift away from trade taxes as a principal source of revenue. This has been one of the consequences of trade liberalisation policies over the last 20 years. It has posed particular problems for low income countries. IMF research shows that, whereas rich countries have managed to offset the decline with other sources of revenue, notably VAT, the poorest countries have at best replaced about 30% of lost trade taxes (Baunsgaard & Keen, 2005).

Case Study: Uganda
The experience of Uganda provides one exception to the trend of low-income countries experiencing a reduction in tax revenues. Under the Museveni regime, trade liberalisation (that is the decline in import and export tariffs) was imposed gradually over the period 1986-98. Rodrik (2004) classifies Uganda as a case, not of shock therapy liberalisation, but one of moderate and gradual reform. Non-tariff barriers were removed for the first time in 1991, five years after Museveni took power. In 1995 there were still import quotas on beer, beverages and auto parts. In 1999, all non-tariff barriers were eliminated.

It was only in the early 1990s that the structure of trade taxes was switched from export taxation to import taxation, but import tariffs were introduced at a high level. There were few options available for alternative types of taxation, a characteristic of very poor economies with weak fiscal institutions. As a result, import taxes necessarily led fiscal resource mobilisation in the 1990s. In 1996, 10 years after the National Revolutionary Movement (NRM) regime took power, trade taxes still accounted for more than 50% of total tax revenues.

This gradualism of trade liberalisation proved crucial to maintaining fiscal revenues until the political and administrative problems of introducing VAT could be overcome. The tax revenues in Uganda increased from 7% of GDP in 1986 to nearly 11% by the mid-1990s. While this is still below the Sub-Saharan African average, the fiscal consequences of more rapid trade liberalisation could have been devastating. The case against rapid tariff reduction as a means for maintaining and increasing fiscal resources, a key element in state consolidation and state-building, is one of the main lessons in the political economy of the Ugandan post-war reconstruction.

It is important to consider however that trade taxes can create disincentives for production and distortions in the economy and thus the impact of trade taxes on economic performance need to be carefully monitored. Collier & Reinikka (2001), for instance, argue that the substitution of export with import taxes created greater inefficiencies in Uganda because import taxes were subject to greater dispersion of tax rates since the latter were subject to more tax rates than the former.

In theory, this could have proved to be a problem, but there were several factors that allowed the Ugandan economy to overcome this. First, the replacement of export taxes was important in improving incentives for exports. Secondly, the substitution of export taxes with import taxes (however much dispersion) was essential for maintaining resource mobilisation, which was central to state-building. Third, a dispersion of import taxes allows the state to provide selective rents (and therefore incentives) for the development of particular sectors. A uniform import rate provides much less scope for industrial and agricultural strategies. Fourth, tariffs provide a fiscally more sustainable mechanism to promote domestic industry in low-income countries. While export subsidies may be less distorting than tariffs, fiscal constraints in low-income countries prevent the extensive use of subsidies as a tool of industrial policy. Finally, the argument that trade policy created static inefficiencies does not explain why Uganda achieved one of the fastest growth rates in the developing world over the period 1986-99. Tariffs on commodity exports, for example, while potentially providing some disincentives to production, were the only mechanism to tax the incomes of wealthy farmers.

Export tariffs can thus provide a functional substitute to weak income-tax capacity in low-income/post-war economies. In the Ugandan case, such tariffs did not coincide with a decline in export growth, but rather were compatible with relatively rapid export and production growth in commodities (Di John & Putzel, 2005). The Ugandan strategy ultimately favoured a greater reliance in import tariffs rather than on high export tariffs although this emerged as a result of trial-and-error. To understand the political economy dynamics of this, it is important to consider the initial conditions of the economy in 1986.

Cross Country Evidence in Africa and Other Low-income Countries
IMF (2005) examines the experience of a sample of eight low-income countries. They have in common a decline in the collected tariff rates over the past 20 years, but differ in the extent of revenue recovery.

In Kenya, Sri Lanka, Egypt and Cote d’Ivoire lost trade-tax revenues were not replaced. In Malawi, Uganda, Senegal and Jordan, they were. The conclusions of this study were as follows:

  1. Those countries which did recover total tax revenue also increased domestic consumption tax revenue, often by an amount broadly corresponding to the loss of trade tax revenue.
  2. The presence of a VAT does not in itself appear to enhance the ability to recover revenue, a result similar to the econometric evidence provided by Baunsgaard & Keen (2005).
  3. In those countries with high recovery, there has also been a strengthening of income-tax revenues, suggesting that the burden of adjustment has not been borne solely by shifting to taxes on consumption. This result is important since it contradicts the conventional wisdom that consumption taxes are the main source offsetting trade-tax revenue.
  4. Reductions in tax/GDP ratios in low- and middle-income countries are not confined to those undertaking trade reform. Of the 14 low-income countries in which collected tariff rates did not decline over the past two decades, nine experienced a decline in the tax ratio. This suggests that while trade liberalisation poses particular challenges to maintaining revenue collection, there are other political economy factors that need to be researched.

In sum, trade liberalisation needs to be purposively sequenced with domestic tax reform and donors need to focus on this issue. This is especially the case since high levels of informality in post-conflict economies may make the collection of taxes from value-added taxes particularly difficult in the short run (Emran & Stiglitz, 2005; though see Keen, 2008, for an opposing view). While tariff protection may not necessarily create much productive capacity given the weak state of domestic business capacity, the role of moderate tariffs in preventing a collapse in fiscal revenues may be a reasonable ‘second best’ solution to the problem of tax collection in post-war/low-income contexts, at least in the short- to medium-run. As discussed above, further research is needed to explain why low-income countries find it difficult to replace lost trade taxes with domestic revenues, and the condition under which VAT is potentially more conducive to tax capacity-building in LDCs.

Taxation and Commodity Booms: Missed Opportunities?
Notwithstanding the potential danger of an oil boom for growth and governance, recent commodity booms do offer an important opportunity for mineral-abundant countries to generate significant tax revenues and increase their policy space. The potential revenue capture from such booms far outweighs aid flows. However, recent experience suggests that, in Sub-Saharan Africa at least, this potential is not being realised. Two recent examples that illustrate the challenges of mineral-based development are Zambia and Mozambique.

Case Study: Zambia

Zambia is one of the poorest countries in Sub-Saharan Africa. It is a land-abundant but sparsely-populated country of 11 million inhabitants. Copper is the dominant export industry and the development of export diversification has been further hampered by the fact that the country is landlocked and is surrounded by five countries which have experienced civil wars and political disorder. By any conceivable measure, the growth performance of Zambia has been dismal, a chronicle of decades of relentless economic decline as indicated in Table 4.

Table 4. Zambia’s per capita growth rates in comparative perspective, 1961-90

Sub-Sahara Africanaverage2.
Zambia ‘s rank16/2620/3122/3230/3229/3626/4032/4134/41

Source: World Bank, World Development Indicators.

The reasons for the decline in Zambia’s economic performance are complex, but include a combination of the disruption of regional trading routes, the nationalisation of the copper industry before the development of skilled workers and managers emerged on the domestic scene, and mismanagement of the state-owned copper industry (see Weeks et al., 2004). Copper production declined from 600,000 tons in the 1960s to just over 300,000 tons by the end of the 1990s.

The government’s response in the late 1990s was to privatise the copper industry and lower mineral royalties in order to attract foreign investment. This was undertaken in the context of desperation, namely historically low world copper prices, declining copper production and an unsustainable debt burden. Its privatisation strategy for copper included the a reduction in the corporate tax rate from 35% to 25%, exemption from customs duty on inputs up to US$ 15 million, reduction of the mineral royalty from 2.0% to 0.6%, exoneration from excise duty on electricity, an increase in the period for which losses could be carried from 10 to 20 years and exemption from the withholding tax on interest, dividends, royalties and management fees (Fraser & Lungu, 2007).

Indeed, the mining sector contributes less to government revenues than either the finance or telecoms sectors. In sum, the mining companies effectively paid almost no income taxes in the period 2000-06. The effect of these so-called incentives was that it would be decades before the government received substantial revenue from the new mining companies.

While the government in 2008 has considered raising the royalty rate to 2.5% with the support of the IMF, this is still low by the standards of Zambia’s neighbours –an IMF survey of tax and royalty rates in developing countries found no other African country charging royalties below 2% and some with royalties as high as 20% (Baunsgaard, 2001). As a result, taxes as a percentage of GDP declined from 18.4% in 1996 to 17.0% in 2005. In 2006, the government received just US$25 million in copper royalties out of a US$2 billion turnover in copper sales. This substantially hampers the extent to which the government can finance improvements in physical infrastructure which are essential for reviving productive capacity and growth in non-copper sectors in agriculture and light manufacturing.

Case Study: Mozambique

Mozambique is considered one of the success stories of post-war reconstruction. A turbulent post-independence period and long civil war coincided with declines in economic activity. In the period 1974-86, real GDP per capita declined by one-third. Economic reforms, begun in 1987, and the end of the civil war in 1992, helped revive the economy. In the decade from 1987, annual growth averaged 5.3%, and accelerated further to over 8% per year in the period 1996-2006. Growth has been fuelled by substantial levels of foreign aid, which has financed approximately one-half of government expenditures over the period 1985-2005 (Virtanen & Ehrenpreis, 2007, p. 17), which has coincided with an increase in the tax take, which has risen from 11.7% of GDP in 1995 to 14.6% of GDP in 2004 (USAID 2004, Table I-1, p. 1-13).

The main pole of growth and exports has been generated through foreign-owned mega-projects in mining and natural-resource-based industrialisation. The leading project in this is Mozal, a large aluminium smelter (completed in 2000) on the outskirts of the capital city, Maputo. Mozal cost US$2.4 billion to build and produces 512,000 tons of aluminium ingots. South African mining interests control two-thirds of the project, as is the case in most mega-projects in Mozambique. As of 2004, Mozal contributes 75% of Mozambique’s manufacturing exports and 42% of its total export revenues (Castel-Branco, 2004). Aluminium represents nearly half the total manufacturing output.

Tax policy has been central in attracting foreign investment in mega-projects. Mozal was given Free Industrial Zone (FIZ) status. This means that it is exempted from paying duties on imports of material inputs and equipment. It is also exempted from valued-added taxes and corporate income taxes are limited to 1% of sales! The failure of the government to develop a more revenue-enhancing tax package was the result of it not seriously considering the offers of rival aluminium producers (Kaiser, a US multinational, made initial offers in the late 1990s but was rejected by the Mozambican government on the grounds that it did not have enough influence on world markets to succeed). Irrespective of the reasons for rejecting the Kaiser bid, an important policy lesson is that governments can use competition among multinationals to produce more lucrative tax packages out of mineral-based investments. The increased interest of Chinese corporations in mineral development in Africa may provide an opportunity for governments to reap the fiscal rewards of competitive bidding among multinationals.

While Mozal has undoubtedly contributed to the export and production capacity of the Mozambican economy, there are several issues that are of concern for the prospects of economic development in the long run and productive capacity-building. First, the negligible tax payments Mozal makes to the government limits the fiscal linkage such projects can generate (Castel-Branco, 2004). This limits the extent the government can invest in developing productive capabilities and infrastructure elsewhere in the economy. Secondly, the mega-projects have focused FDI and manufacturing production around the capital city, inducing a substantial regional concentration in manufacturing production (in 2003, 81% of industrial activity was generated in Maputo Province [USAID 2004, Table 12-3, p. 3]). Manufacturing production outside the capital is negligible. Third, most of Mozal’s economic links are with firms in South Africa, not in Mozambique. This is mainly because Mozambican firms do not have the technical capacity to provide inputs that Mozal needs, but also because there is not a wider industrial strategy to provide either carrots or sticks for Mozal to develop important supplier contracts with Mozambican firms.

There are several policy implications that it is possible to draw from the Zambian and Mozambican cases. First, there is an urgent need for mineral-abundant states to enter into a renegotiation of mining contracts when they are unfavourable. Secondly, there is a need for governments to develop productive strategies that exchange mineral rights for local content conditions, whereby foreign investors are obliged to use domestic suppliers on an increasingly greater scale. Local content management has been one of the main ways in which FDI can be used for the benefit of national productive capacity. Finally, capacity-building in the geological survey capacity in Sub-Saharan Africa needs to be developed in order to improve the bargaining power of states vis-à-vis multinationals. This is an area where the international financial institutions can play a leading role.

Export Taxes, State Territorial Reach and Production Strategies

Export taxes on agriculture are generally inadvisable for developing countries because of the well-known disincentives they provide for producers. However, there are some examples of the developmental role these taxes can play when they are explicitly part of a production strategy to improve agricultural productivity. For such taxes to work, they need to the earmarked directly to finance infrastructure investment in agriculture. Apart from this, such taxes have played an important role in expanding the territorial reach of the state and the territorial dimension of state-society relations. Let us examine some country examples.

In the case of Mauritius, export taxes on sugar, the main export commodity in the 19th and most of the 20th century, had several positive effects on state-society relations and in increasing the productive capacity of the sugar sector (Bräutigam, 2008). First, the tax was an effective substitute for income taxes, and was generally progressive as it shifted the burden of taxation and redistributive spending on the wealthy and middle classes. This contributed to the public sense of fairness and solidarity and thus enhanced state legitimacy. Secondly, the tax was used by the state to finance research and development, infrastructure and marketing which enhanced production and productivity growth in the sugar sector. An often neglected aspect in tax analysis is to explain how tax reform can be linked to productive strategies (which Grabowski [2008], for instance, argues was central to successful agricultural development in Japan, South Korea, and Taiwan). Third, the export tax helped the private sector organise, and it built their capacity to interact with the government over time. Fourth, it helped both the state and society to solve collective action problems they faced in building skills and in supporting research on sugar. Finally, the export tax helped develop the territorial reach of the state since the tax affected the main employer in the countryside and promoted mutually beneficial rights and obligations between the state and farmers, both large and small.

A second important example concerns the role agricultural marketing boards have played in some countries in expanding the territorial reach of the state and in linking rural interest groups to the state. Marketing boards were also an important source of state resource mobilisation through the mechanism of monopolising the purchase of cash crops at below world market prices and selling such crops abroad at world market prices. The surplus generated was often of similar magnitudes to formal total tax collection levels, particularly in Sub-Saharan African economies in the 1960s and 1970s. Marketing boards were effective in some countries such as in Taiwan, South Korea, Indonesia and India because the state gave something in return to producer groups such as services, infrastructure, research and price stability.

By the 1980s, however, an extensive critique of marketing boards developed in the wake of worsening agricultural performances, particularly in Sub-Saharan Africa (Bates, 1981). It was generally viewed that the system worsened the terms of trade by paying farmers less than what the state received for the products at the world market. This often created disincentives for farmers to produce and/or led to smuggling –both of which reduced the resource mobilisation capacity of African states–. Economic liberalisation of agriculture was promoted as the cure for the growth-retarding effects marketing boards had in many contexts.

Despite these concerns, there are other important factors to consider in terms of the role marketing boards played in state-building. A principal task of policy-makers is to understand why some marketing boards performed better than others. The historical evidence suggests that the political power of the state and the nature of the political coalitions underpinning the central state are significant factors determining the effectiveness of marketing boards. For instance, in Taiwan during the 1960s the ability of the state to undertake land reform removed the power of large landowners who historically resisted state penetration of the countryside (Amsden, 1985). This state penetration allowed the state to tax rice farming in return for financing inputs that improved the productivity of rice production.

To take a Sub-Saharan African comparison, Bates (1995) argues that the Kenyan coffee board was, in the 1970s and 1980s, more effective than the Tanzanian coffee board because the nature of the political coalition in power differed in the two countries. In Kenya, large and medium-sized coffee farmers were a powerful interest group, whereas in Tanzania coffee farmers were not a powerful group in the national government’s support base. As a result, policies in Kenya were developed in ways that extracted much fewer net resources from coffee producers than in Tanzania.

Even where marketing board policies were relatively ineffective, such as in Tanzania and Zambia, they have played an important role in increasing the territorial reach of the state, developing state-rural interest group links, and in providing social infrastructure and services. In these two countries the reach of the state was a by-product of the development of nationally-based political parties which developed an inclusive system of patronage across all agricultural regions (see Hesselbein, Golooba-Mutebi & Putzel, 2006, on Tanzania, and Di John, forthcoming on Zambia). There is also evidence that the inclusive reach of marketing boards contributed to the maintenance of political stability and nation-building in both these cases. Further comparative historical work is required to assess the differential impacts marketing boards have had in state-building and in enhancing the territorial reach and legitimacy of the state.

International Obstacles to Tax Collection: The Problem of Capital Flight and Off-Shore Financial Centres

Another important concern for many countries, developed and less developed, is the extent to which international financial liberalisation has facilitated capital flight to onshore and offshore financial centres. The Tax Justice Network has estimated that capital flight from all countries, including funds undeclared in the country of residence, is approximately US$11.5 trillion (Spencer, 2006; Christensen 2009). Annual global income from such sources is conservatively estimated at US$860 billion, and the annual world-wide tax revenue lost is approximately US$255 billion, which equals the funds estimated to meet the UN Millennium Development Goals (ibid.).

Capital flight incurs many economic, political and social costs. Particularly when capital is scarce, capital flight results in a loss of resources to finance investments in infrastructure and social spending. Capital flight also lessens the resources available for investment more generally. This contributes to declines in growth rates, which results in growing unemployment, informalisation of economic activity and poverty. Declining investment also harms the technological upgrading required to keep exports competitive. In many countries, particularly in Sub-Saharan Africa and Latin America, capital flight has been accompanied by increases in foreign borrowing –that is, increased indebtedness has been used not to finance investment or even consumption, but to finance capital flight itself (Boyce & Ndikumana, 2005)–. The resulting debt burdens are likely to most hurt the poor, as social spending and infrastructural spending needs to be cut in the face of debt repayments.

Despite the global nature of the capital flight problem, there are important regional differences between developing regions. Consider Table 5 below.

Table 5. Capital flight as a share of private wealth in Latin America and East Asia (%)

 1980-89 (a)1990-98 (a)1980-89 (b)1990-98 (b)
Sub-Saharan Africa27.630.127.430.3
Latin America &Caribbean8.
East Asia and Pacific4.

Note: (a) all observations; (b) full data points only.
Source: Collier et al. (2004, Table 1A, p.22).

Capital flight as a share of private wealth has been estimated to be between two to three times higher in Latin America compared to East Asia in the 1980s and 1990s. For Sub-Saharan Africa the situation is even worse. In the region where capital is most scarce, capital flight as a percentage of private wealth was, on average, six times higher than in East Asia in the 1980s and over 10 times higher than East Asia in the 1990s.[2] It is likely that capital flight both caused and was caused by lower growth, macroeconomic instability and political instability in Latin America and Sub-Saharan Africa. Whatever the mechanisms, capital flight in both regions has severely lowered the tax base and with it, the domestic resources available to finance public investment in infrastructure and social services. Capital flight may also weaken political elite interest in local economic growth and development, creating a vicious cycle.

Policy proposals to address the tax revenues lost due to capital flight include selective use of capital controls, overriding bank secrecy in onshore and offshore financial centres, improvements in tax administration in less developed countries and further implementation of tax information exchanges between countries. Exchange of information on capital flight between governments was advocated by John Maynard Keynes and Harry Dexter White, the principal architects of the Bretton Woods institutions in 1944. This proposal was opposed by the US financial community, which had benefited from capital flight. Another more radical solution would involve selectively repudiating past loans, invoking the doctrine of ‘odious debt’ in international law as well as historical precedent (Boyce & Ndikumana, 2005).[3] The idea here is to repatriate funds that were illegally transferred out of the country by state leaders. Some analysts have also suggested that the IMF, World Bank and OECD should take the lead in implementing an international financial architecture to reduce the incentives and means for engaging in capital flight (Spencer, 2006). As in the past, the financial community in the advanced industrial countries, as well as wealthier individuals and corporations in the poorer countries, would likely oppose such policy proposals. Nevertheless, it would be possible to make the case that taxpayers in the OECD countries would shoulder less of the burden of financing international aid if tax revenues lost from capital flight (both from developed and less developed country residents) were re-captured and/or prevented.

The Non-monolithic and Historically Specific Nature of State Capacity and the Prospects for Growth

While taxation is a useful objective indicator of state capacity and legitimacy, it would be a mistake to assume that effective tax capacity translates into similarly effective capacities to intervene in other spheres.

There are numerous examples of this. South African tax collection capacity is much greater than its ability to undertake industrial policy or tackle HIV/AIDS. Botswana’s democratic institutions are among the most robust in the developing world yet it has also been very poor at controlling HIV/AIDS. Brazil has among the highest levels of tax take but is not (politically) capable of collecting personal income and property tax. Brazil’s industrial policy is also very uneven: success stories in autos and aerospace stand out, while many other sectors have been less successful. The Colombian state is known for being among the best in macroeconomic management but has one of the lowest take takes in Latin America and is unable to contain decades of guerrilla and paramilitary political violence. Venezuela has long maintained a stable democratic system but has been unable to promote export diversification. Tanzania and Zambia have had relatively poor records on economic performance but have been able to prevent large-scale political violence, unlike most of their neighbours. This variation in capacity is not picked up by aggregate measures and our understanding of why capacity varies so much within polities is limited. Detailed historical analyses of the political coalitions and settlements underpinning specific state capacities are essential to increase our understanding of variable state capacity within a polity.

Taxation, Governance and Growth

Inherent in much of the recent work on taxation is that a broader-based taxation system will consolidate state-interest group bargaining which will, in turn, generate a greater degree of legitimacy which supposedly will generate more effective governance. Good governance, in turn, is seen as central for sustained rapid economic growth (World Bank, 1997, 2002).

The problem with much of the new literature, however, is that it identifies the tax nexus as the main source of a state’s legitimacy. This is problematic in the context of economically underdeveloped countries. Because tax rates and composition are not systematically correlated with economic growth, it is not helpful to focus on taxation in isolation of other factors that affect capital accumulation, the efficiency of investment and economic growth. For instance, national savings and particularly public savings (which in part come from the efficient operation of state-owned enterprises) may be as –if not more– important to the growth prospects of an economy.

Also, it is important to keep in mind that the governance structures of low-income countries differ from rich ones because of underdevelopment, which implies a limited fiscal base of the state. The implication is that while improving tax capacity in less-developed countries is important, one should not expect state-society relations to resemble OECD countries. The evidence also suggests that the wider resource mobilisation capacity of states is as important –if not more– to the expansion of the tax nexus. Donor and government policy needs to link tax reform with productive strategies and aid policy needs to focus much more on facilitating capital accumulation. One suggestion is that aid needs to shift back towards the financing of directly productive economic activities and particularly physical infrastructure. The advocacy of improving tax capacity as a way to help construct states and state legitimacy is well placed, but needs to be done in a way that is more informed by history and political economy dynamics that accompany low levels of per capita income. It is important not only to bring politics back into issues of taxation and governance but also the realities of the stage of economic development.

Resource-Curse Argument

One influential strand of the political economy literature is the rentier state model, or resource-curse argument. The main premise of the rentier state model of governance is that when states gain a large proportion of their revenues from external sources, such as mineral-resource rents or aid, the reduced necessity of state decision-makers to levy domestic taxes causes leaders to be less accountable to individuals and groups within civil society, more prone to engage in and accommodate rent-seeking and corruption and less able to formulate growth-enhancing policies. In addition, it is posited that the greater abundance of unearned income makes such economies more prone to violent political conflict, including civil war. There are two variant of the resource curse argument:

  1. ‘Honey pot’, or rent-seeking argument, which suggests that oil-abundant less-developed countries generate valuable rents and that the existence of these rents tends to generate violent forms of rent-seeking that take the form of ‘greed-based’ insurgencies.
  2. Rentier state argument: oil states are more likely to have weak state structures because they have less need to create strong bureaucracies to raise revenue. Weak state structures, in turn, can make the state more vulnerable to insurgency.

The core argument of the second variant, which has been called ‘political Dutch Disease’, is that rentier state leaders, by relying on ‘unearned’ income (in the form of mineral rents and/or aid), do not develop a set of reciprocal obligations with citizens via the nexus of domestic taxation. As a result, mineral rents (particularly oil and gas) can, in lower-income countries, coincide with weak or illegitimate state institutions and thereby trigger conflict.

According to the rentier state model, mineral abundance generates: (a) low levels of government legitimacy; (b) slow economic growth; and (c) higher levels of political violence. This is because of three factors: (1) a growing independence of states from citizens due to high levels of unearned income (from mineral rents) and low levels of domestic taxation; (2) a potentially retarding effect on state capacity of unearned income is the decline in bureaucratic capacity; and (3) mismanagement of resource wealth can create grievances that, when combined with a history of ethnically-based secessionist tendencies, can increase the likelihood of organised armed rebellion.

The logic as to why oil economies are subject to greater political violence is worthy of particular attention since the onset of such violence is the greatest expression of an illegitimate government. In the rentier state model, the reliance on unearned income can have several negative effects on a regime’s legitimacy and capacity to combat or prevent rebellion. These include:

  1. Increased autonomy of states from citizens can increase the ability of state leaders to act in predatory ways, or at the very least reduces the need for state leaders to develop long-run political bargains with interest groups. This, in turn, makes taxation and revenues more unpredictable, which may increase arbitrary confiscation when volatile mineral rents suddenly collapse.
  2. With little bureaucratic presence in tax collection and limited information about what goes on at the grassroots level, states may be vulnerable to organised predators including guerrillas and private armies.

Supporters of the rentier-state model suggest that reducing a state’s ‘unearned income’ from mineral rents will enhance the prospects of peace. Policy recommendations include advocating greater transparency in the payments multinationals in extractive industries make to host governments in poor countries, or avoiding extractive industries altogether and concentrating efforts in order to diversify mineral-dominant economies towards agriculture and manufacturing. The plausibility of these arguments depends on the extent to which oil wealth necessarily generates the aforementioned problems.

Rentier-State Model: Methodological Problems
First, leaders are implicitly assumed to ‘own’ the natural resources –they are assigned the ‘property rights’ over resources–. How rulers appropriate and maintain power is not adequately analysed. Secondly, leaders are assumed to have predatory as opposed to developmental aims. The neglect of the political processes through which a leader appropriates power limits our understanding of the motivations of state leaders. In sum, the state is not a thing, such as ‘a predator’, but a set of social relations. The existence of oil abundance does not preclude the possibility that state leaders share income from resource rents with groups that comprise their political support base.

Oil Abundance and Political Violence: Is there a link?
Comparative work on oil states (Smith, 2004) has found that, in the period 1974-99, oil wealth is robustly associated with increased regime stability, even when controlling for repression, and with a lower likelihood of civil war. As Smith notes:

‘Durable regimes in oil rich states are not the outliers that both rentier state and resource curse theorists have assumed them to be. Regimes such as Suharto’s in Indonesia, which lasted 32 years, Saddam Hussein’s Ba’athist regime in Iraq, which lasted 35 years (and was only ousted by a full-scale U.S.-British invasion in March, 2003), and long-lived monarchs of the Persian Gulf appear to be more representative of regime durability than do the favourite cases of Iran, Nigeria, Algeria, and Venezuela –the ‘big four’. Moreover, the durability effect has been independent of the consistent access to rents with which regimes can buy legitimacy, since the busts created no trend toward regime crisis or instability in exporting states’. (p. 242).

Is there Convincing Evidence in Support of the Rentier-State Model?
Ross (2004) suggests that there are at least three mechanisms through which oil dependence might generate legitimacy problems for governments: (1) high corruption levels; (2) resulting from the first, slow economic growth; (3) the trend towards authoritarianism.

First, let us consider the extent to which mineral abundance is likely to generate higher levels of corruption. The main assumption of rentier-state theorists is that oil abundance tends to centralise production in the state and thus increase state control over the economy. For mainstream rent proponents, the root of the negative impact of corruption is the discretionary centralised authority that accompanies state intervention. Discretionary centralised authority is characterised by the power vested in top decision-makers to intervene in activities and welfare of subordinates with impunity. The very existence of such authority and intervention make possible its inappropriate use. In mainstream models, the costs of corruption include the waste of resources in attempts to influence public authorities and the reduction in the security of property rights, since corrupt transactions need to be kept secret.

At first glance, a comparison of all the developing countries for which there is corruption data also seems to indicate an indeterminate relationship between levels of mineral abundance, and corruption and growth. Consider Table 6:

Table 6. Growth and corruption in mineral-abundant and non-mineral-abundant developing countries, 1965-2000

1965-19901. Mineral-Abundant2. Non-Mineral-Abundant
 Developing Countries (2)Developing Countries (2)
 (13 observations)(19 observations)
Median GDP Growth4.35.6
Rate 1965-90(2.5 – 12.4)(1.5 – 9.5)
Median Corruption3.93.6
Index 1980-85 (1)(0.2 – 6.5)(0.7 – 8.8)
1990-20001. Mineral-Abundant2. Non-Mineral-Abundant
 Developing CountriesDeveloping Countries
 (13 observations)(19 observations)
Median GDP Growth4.03.7
Rate 1990-2000(1.6 – 7.0)(-0.6 – 10.3)
Median Corruption3.33.2
Index 1996(0.7 – 6.8)(1.0 – 5.0)

Note: (1) a corruption index of 10 indicates minimum corruption, an index of 0 indicates maximum corruption; (2) mineral-abundant is defined as those economies where mineral/fuel exports in total exports in 1980 is equal or greater to 35%; non-mineral abundant is defined as those economies where mineral/fuel exports in total exports is less than 35% in 1980.
Sources: World Bank, World Development Indicators; Subjective Corruption indices from Transparency International.

In sum, two important mechanisms where the resource curse hypothesis may explain increased conflict, namely low growth and high corruption, are not supported by the evidence (see also Di John, 2007). The fact that subjective corruption rates are similar in mineral-resource abundant and mineral-scarce economies suggests that the existence of mineral rents does not necessarily generate higher forms of illegal rent-seeking in the former type of economy. This has two implications. First, mineral-resource states do not generate greater legitimacy problems because of corruption. Secondly, corruption, a generally non-violent form of influencing, is not systematically less prevalent in mineral-resource-abundant economies. If this were the case, it could be argued that the absence of non-violent, illegal forms of rent-seeking were suppressed, then violent forms of rent-seeking such as rebellion were more likely. This is not the case.

Due to limited data on corruption rates across all oil exporters, the above analysis compared mineral-abundant economies to non-mineral-abundant economies to demonstrate the indeterminate relationship between mineral abundance, corruption and growth. However, even if we assume that corruption rates are higher on average in oil economies, is it the case that their growth rates are systematically worse than non-oil economies?

Table 7 compares the growth rates of oil-export economies in the period 1965-98 to the growth rates of four developing-country regions (Latin America and the Caribbean, East Asia and the Pacific, Sub-Saharan Africa and South Asia), all developing countries and the world economy.

Table 7. Economic growth of selected oil-export developing countries in comparative and historical perspective, 1965-98

 Fuel Exports, 1980GDP Growth
Oil-Export Economies(% total exports)(average annual change, %)
Latin America     
Trinidad & Tobago93 5.0-0.82.4
South-East Asia     
Middle East     
Kuwait100 1.61.3n.a
Saudi Arabia99
Iraq>40 7.3-8.8n.a
Angola78 n.a3.7-0.4
Transition Economies     
Kyrgyz Republic93*
Average (excluding transition economies)80.2
Overall Average80.2
   GDP Growth
Regions  1965-801980-901990-98
East Asia and Pacific
Latin America
Sub-Saharan Africa
South Asia
Lower and middle-income countries

Note: * in 1990: ** in 1996.
Source: World Bank, World Development Indicators.

Indeed, there is little evidence for the rentier-state hypothesis. In the period 1965-80, the average annual growth in GDP of oil-export economies was 6.7%, which was faster than the average annual growth of all developing regions except East Asia and the Pacific, faster than the growth of all lower- and middle-income economies, and considerably faster than growth in the world economy. The period 1980-90 saw a partial reversal of this trend. In this period, the average annual growth in GDP of oil-export economies slowed down considerably to 1.8%, which was considerably lower than the average annual growth of East Asia and the Pacific (8.0%) and South Asia (5.7%), and only one-half the growth rate of all lower- and middle-income economies (3.5%), and considerably lower than growth in the world economy (3.2%). However, the growth rate of oil-exporters slows to the same rate as in all of Latin America and the Caribbean (1.6%) and Sub-Saharan Africa (1.8%).

Oil exporters are not particularly poor performers: they have plenty of company. Finally, in the period 1990-98, the average annual growth of oil exporters (now including the transition oil economies) imploded to 0.6%, considerably below the growth rates in all developing regions, the lower- and middle-income economies and the world economy. However, this result turns on the particularly poor growth performance of the four transition oil economies. In this period, the average annual growth rate of the four transition oil economies was minus 7.3%. The average annual growth rate of oil exporters (excluding the transition economies), in the period 1990-98, increases to 2.9%, which while considerably slower than growth in East Asia and the Pacific (8.1%) and South Asia (5.7%), is similar to the growth of all lower- and middle income countries (3.3%), Latin America and the Caribbean (3.7%), and indeed faster than growth in the world economy (2.4%). In sum, there is little evidence that oil abundance is necessarily a ‘curse’. Moreover, the downturns in economic growth in oil exporters in the period 1980-98 follows closely the growth slowdowns in both Latin America and Sub-Saharan Africa.

Does Scaled-up Aid Substitute for Domestic Tax Mobilisation?

The rentier-state model has influenced ideas on the relationship between aid and taxes. In particular, this line of thinking has generated concern that the scaling-up of aid will crowd out domestic tax efforts and thus generate patterns similar to what supposedly occurs in ‘petro-states’ (Gupta et al., 2003). Another claim is that increases in central government transfers will crowd-out tax mobilisation at the state or municipal level. This is an empirical question and there is no robust evidence to support the claim that aid or central government transfers crowd-out domestic tax effort.

A recent econometric study, however, has arrived at the opposite conclusion. Gambaro, Meyer-Spasche & Ashikur (2007) find evidence that there is a positive association between aid inflows and tax revenue, which is primarily driven by the positive relationship between grants and tax revenue over the period 1990-2000.

Gambaro et al. (2007) emphasise that their conclusions only hold for the period 1990-2000, which is both a more recent and shorter time frame than the Gupta et al. (2003) study. One possible reason for these results may be that the role of development policy post-1990 has had a stronger focus on institutions. The positive correlation between grants and tax revenue lends some support to the interpretation that development aid since the 1990s, through its stronger focus on institutions may have led to an improvement in the tax administration and revenue collection in recipient countries. An important conclusion Gambaro et al. highlight is that ‘both donors and recipient countries should try to identify the pivotal set of policies that influenced the response of tax revenue to the inflow of aid after 1990’.

Matching Aid Funds and Domestic Tax Effort
Despite the lack of conclusive evidence that increased external funds reduce domestic tax efforts, international donors can nevertheless improve the incentives of government leaders to increase domestic tax efforts. It makes sense for donors to enter into a multi-year compact with post-conflict host governments to provide matching funds for direct-budget support purposes. The current arrangement in many post-conflict countries is that donors provide budget support when the government specifies its expenditure needs and calculates what its financing gap is.

This system can create several problems. First, the incentive for the government to raise revenue may be diminished. Secondly, the capacity of the government to identify and assess macro-level expenditure revenue trade-offs are reduced as ministers are not forced to prioritise spending based on what revenues they can collect; instead, they simply present a wish list. Third, there is considerable uncertainty and volatility in the actual aid flows that are dispersed, creating problems for macroeconomic management and planning.

The matching-funds approach can address these concerns if donors could agree to match a percentage of the funds collected by the government up to a fixed limit. The matching percentage could be reduced over time, reflecting the increased capacity of the government to raise revenue. The main advantage of this approach is that it increases the incentives for revenue collection since state officials will know that raising extra revenue will result in additional inflows of donor resources.

For this system to work, it is necessary that donors commit for the medium- and long-term through the development of trust funds. The major challenge involves the willingness of donors to make their aid flows predictable and reliable, putting into practice agreements on good donor practice made in such statements as the 2005 Paris Declaration. The matching-funds approach will mean that donors need to find a higher level of coordination, planning and discipline than has been demonstrated in post-conflict situations in the past.

Political Organisation and State Resilience

In the political science literature there is a tendency to model ‘African polities’ monolithically as dysfunctional states where corruption, clientelism and patrimonial rule predominate. Examples would include single characterisations of African politics as ‘personal rule’ (Sandbrook 1985), as the ‘politics of the belly’ (Bayart 1989), as the ‘politics of chaos’ (Kaplan 1994) or as ‘disorder as political instrument’ (Chabal & Daloz 1999). As Allen (1995) points out, however, a more careful reading of African political history reveals a much greater variation and change in the nature of African polities.

There are two main problems with this approach. First, since clientelism and corruption are prevalent in nearly all Sub-Saharan African states, theories focusing on patrimonialism cannot explain why economic growth rates vary across (clientelist) Sub-Saharan African polities, or why many countries in Sub-Saharan Africa achieved rates of growth close to East Asia and Latin America in the period 1960-80. Secondly, such theories cannot explain why some states like Tanzania, Zambia, Botswana and Ghana have avoided degrees of political instability, state collapse and political violence that have occurred in countries such as the Democratic Republic of the Congo, Sierra Leone or Uganda in the 1970s and 1980s. A key analytical challenge, therefore, is to explain why some countries are able to create more developmental outcomes in the context of clientelism and corruption and why other states do not.

Nature of Elite Bargains Matter
A brief examination of what may account for greater political stability in some patrimonial states points to the importance of elite bargains. The principal solution through history to the classic Hobbesian problem of endemic violence is the creation of what North et al. (2007) call limited-access orders (as opposed to the much rarer open-access orders, which characterise advanced market economies).The limited-access order creates limits on the access to valuable political and economic functions as a way to generate rents. When powerful individuals and groups become privileged insiders and thus possess rents relative to those individuals and groups excluded (and since violence threatens or reduces those rents), the existence of rents makes it in the interest of the ‘privileged insiders’ to cooperate with the coalition in power rather than to fight.

One particular factor that facilitates elite bargains is political organisations and the structure of patronage. Recent research by the Crisis States Research Centre ( suggests that the degree of centralised rule and patronage matters for political stability. A cursory examination of relatively peaceful polities (Tanzania, Zambia) and those where the state survived even during civil war (Mozambique, Colombia) suggests that the construction of political organisations, particularly political parties, has been central to providing the institutional mechanisms of distributing patronage to regional elites and to important political constituencies in ways that either prevent challenges to authority and/or maintain cohesion of the ruling coalition.

Further evidence of the importance of political party organisation and centralised patronage in the maintaining state resiliency can be seen in the cases of South Africa, Botswana and Mauritius. These countries all have strong centralised national parties. Uganda under Museveni would be another example of the construction of centralised patronage backed by a strong political organisation. Beyond the African context, there is a substantial literature on the role that political party pacts have played in maintaining peaceful transitions to democracy in less developed countries. Moreover, this line of inquiry will help establish why ‘horizontal inequalities’ become more politically salient in some contexts as opposed to others.

This line of research also suggests that countries with more fragmented state structures and ‘spoils systems’ (Zaire/Democratic Republic of the Congo, Nigeria) involves a more personalist and narrow presidentialist rule without extensive political party support and the implementation of ‘divide and conquer’ strategies to more selectively accommodate ethnic and regional interests. The latter strategies tend to result in a more ‘winner takes all’ (or indivisible) distribution of resources which increases the amount of grievances in society, often resulting in armed rebellions.

In conclusion, national political organisations and the centralisation of patronage can achieve political stability through several means. First, itenables the executive to have an encompassing interest in the maintenance of political stability. When national political parties dominate, ethnic and regional elite interests are likely to be accommodated in the distribution of patronage. It is also more likely that the development of cross-ethnic coalitions will prevent the emergence of horizontal inequalities which can contribute to political violence. Secondly, it limits the extent to which the executive engages in predatory behaviour because political parties generally provide endogenous enforcement by coalition members/party cadres. Third, it enables the creation of a loyal and unified military. If powerful elites receive a fair share of rents, they are less likely to mobilise and promise future rewards to factions of the military. This is more likely in ‘divide and conquer’ strategies where ethnic and regional boundaries are more salient. Fourth, it makes the cost of elite exit and rebellion higher, since centralised political patronage controls the distribution of most of the valuable resources in the economy. Finally, because leaders of national political parties have an encompassing interest in the polity, it allows the management of adverse economic shocks and crises in ways that do not generate state breakdown.

Jonathan Di John
School of Oriental and African Studies (SOAS), University ofondon


Allen, C. (1995), ‘Understanding African Politics’, Review of African Political Economy, nr 65.

Amsden, A. (1985), ‘The State and Taiwan’s Economic Development’, in P. Evans, D. Rueschemeyer & T. Skocpol (Eds.), Bringing the State Back In, Cambridge University Press, New York.

Bates, R. (1995), ‘Social Dilemmas and Rational Individuals: An Assessment of the New Institutionalism’, in J. Harriss, J. Hunter & C. Lewis (Eds.), New Institutional Economics and Third World Development, Routledge, London.

Bates, R. (1981), Markets and States in Tropical Africa, University of California Press, Berkeley and Los Angeles.

Baunsgaard, T. (2001), ‘A Primer on Mineral Taxation’, IMF Working Paper WP/01/139.

Baunsgaard, T., & M. Keen (2005), ‘Tax Revenue and (or?) Trade Liberalization’, IMF Working Paper 05/112.

Bayart, J.F. (1989), The State in Africa: the Politics of the Belly, Fayard, Paris.

Bird, R.M., & E.M. Zolt (2005), ‘Redistribution via Taxation: The Limited Role of the Personal Income Tax in Developing Countries’, UCLA Law Review, nr 52 (6), p. 1627-95.

Boyce, J., &L. Ndikumana (2005), ‘Africa’s Debt: Who Owes Whom’, in G.A. Epstein (Ed.), Capital Flight and Capital Controls in Developing Countries, Edward Elgar, Cheltenham.

Bräutigam, D. (2008), ‘Contingent Capacity: Export Taxation and State-building in Mauritius’, in Deborah Bräutigam, Odd-Helge Fjeldstad & Mick Moore (Eds.), Taxation and State-Building in Developing Countries: Capacity and Consent, Cambridge University Press, Cambridge.

Brewer, J. (1990), The Sinews of Power: War, Money, and the English State, 1688-1783, Harvard University Press, Cambridge, MA.

Castel-Branco, C.N. (2004), ‘What is the Experience and Impact of South African Trade and Investment on the Growth and Development of Host Economies? A View from Mozambique’, paper presented at the Conference on Stability, Poverty Reduction and South African Trade and Investment in Southern Africa, sponsored by the Southern African Regional Poverty Network and the EU, Pretoria, South Africa, 29-30 March.

Chabal, P., & J.P. Daloz (1999), Africa Works: Disorder as Political Instrument, James Currey, Oxford.

Christensen, J. (2009), ‘Africa’s Bane: Tax Havens, Capital Flight and the Corruption Interface’, Working Paper nr 1/2009, Elcano Royal Institute, Madrid.

Collier, P., A. Hoeffler & C. Pattillo (2004), ‘Africa’s Exodus: Capital Flight and Brain Drain as Portfolio Decisions’, Journal of African Economies, vol. 13, supplement 2.

Collier, P., & R. Reinikka (2001), Reconstruction and Liberalization: An Overview’, in R. Reinikka & P. Collier (Eds.), Uganda’s Recovery: The Role of Firms, Markets and Government, The World Bank, Washington DC.

Di John, J. (forthcoming), ‘Zambia: State Resilience against the Odds: An Analytical Narrative on the Construction and Maintenance of Political Order’, Crisis States Research Centre Working Paper, London School of Economics, London.

Di John, J. (2007), ‘Oil Abundance and Violent Political Conflict: A Critical Assessment’, Journal of Development Studies, vol. 43, nr 6, August.

Di John, J., & J. Putzel (2005), ‘Institutional Change for Growth and Poverty Reduction in Low Income Countries: The Case of Uganda’, paper presented at the Conference When Institutions are Weak: Strategies for Change, International Monetary Fund, Washington DC, 6-7 July.

Emran, M., & J. Stiglitz (2005), ‘On the Selective Indirect Tax Reform in Developing Countries’, Journal of Public Economics, nr 89.

Fraser, A., & J. Lungu (2007), ‘For Whom the Windfalls? Winners & Losers in the Privatisation of Zambia’s Copper Mines’, Mine Watch Zambia, manuscript funded by Christian Aid, UK,

Gambaro, L., J. Meyer-Spasche & R. Ashikur (2007), ‘Does Aid Decrease Tax Revenue in Developing Countries?’, London School of Economics, processed.

Grabowski, R. 2008. Modes of Long-run development: Latin America and East Asia. Journal of Institutional Economics, 4 (1).

Gupta, S., B. Clemens, A. Pivovarsky & E. Tiongson (2003), ‘Foreign Aid and Revenue Response: Does the Composition of Aid Matter’, in S. Gupta, G. Inchauste & B. Clemens (Eds.), Helping Countries Develop: The Role of Fiscal Policy, IMF, Washington DC.

Gupta, S., & S. Tareq (2008), ‘Mobilizing Revenue’, Finance and Development, vol. 45, nr 3, pp. 44-47.

Hesselbein, G., F. Golooba-Mutebi and J. Putzel (2006), ‘Economic and Political Foundations of State-making in Africa: Understanding State Reconstruction’, Crisis States Research Centre Working Paper 3 (series 2), London School of Economics, London.

IMF (2005), ‘Dealing with the Revenue Consequences of Trade Reform, Background Paper for Review of Fund Work on Trade’, IMF.

Kaplan, R. (1994), ‘The Coming Anarchy’, Atlantic Monthly, February

Keen. M. (2008), ‘VAT, Tariffs, and Withholding: Border Taxes and Informality in Developing Countries’, International Monetary Fund, mimeo.

Keen, M., & M. Mansour (2008), ‘Revenue Mobilization in Sub-Saharan Africa: Key Challenges from Globalization’, paper presented at the Conference Globalization and Revenue Mobilization, Abuja, Nigeria, February.

Lewis, A. (1954), ‘Economic Development with Unlimited Supplies of Labour’, The Manchester School, vol. 22, nr 2.

North, D. (1981), Structure and Change in Economic History, Norton, New York.

North, D., J. Wallis, S. Webb & B. Weingast (2007), ‘Limited Access Orders in the Developing World: A New Approach to the Problems of Development’, World Bank Policy Research Working Paper nr 4359, The World Bank, Washington DC.

O’Brien, P. (2001), ‘Fiscal Exceptionalism: Great Britain and Its European Rivals’, Working Paper nr 65/01, Department of Economic History, London School of Economics.

Rodrik, D. (2004), ‘Rethinking Growth Strategies’, WIDER Annual Lecture 8, WIDER, Helsinki.

Ross, M. (2004), ‘What Do We Know about Natural Resources and Civil War?’, Journal of Peace Research, nr 41 (3).

Sandbrook, R. (1985), The Politics of Africa’s Economic Stagnation, Cambridge University Press, Cambridge.

Schumpeter, J. (1954 [1918]), ‘The Crisis of the Tax State’, International Economic Papers, nr 4.

Smith, B. (2004), ‘Oil Wealth and Regime survival in the Developing World, 1960-1999’, American Journal of Political Science, nr 48 (2).

Spencer, D. (2006), ‘The IMF and Capital Flight: Redesigning the International financial architecture’, The Tax Justice Network.

Tilly, C. (1990), Coercion, Capital and European States: AD 990-1992, Blackwell, Oxford.

USAID (2004), Removing Obstacles to Economic Growth in Mozambique, US AID, vol. 2, Washington DC.

Virtanen, P., & D. Ehrenpreis (2007), ‘Growth, Poverty, and Inequality in Mozambique’, International Poverty Centre Country Study, nr 10, UNDP, Brasilia.

Weeks, J., V. Chisala, A. Geda, H. Dagdeviren, T. McKinley, C. Oya & A. Saad-Filho (2004), Economic Policies for Growth, Employment and Poverty Reduction: Case Study of Zambia, Centre for Development Policy and Research, School of Oriental and African Studies (SOAS), University of London, and United Nations Development Programme Bureau for Development Policy, London and New York.

World Bank (1997, 2002), World Development Report, Oxford University Press, Oxford.

[1] According to the World Bank (1997, p. 41-60), the five ‘fundamentals’ that lie at the core of good governance for a state are: (1) establishing a foundation of law; (2) maintaining a non-distorting policy environment, including macroeconomic stability; (3) investing in basic social services and infrastructure; (4) protecting the vulnerable; and (5) protecting the environment. While tax is not explicitly mentioned as a core function of governance, tax capacity is implicitly behind items [3] and [4].

[2] These estimates, of course, do not indicate variations within regions.

[3] Boyce & Ndikumana (2005) note: ‘At the end of the 19th century, the US government repudiated the external debt owed by Cuba after seizing the island in the Spanish-American War. The US authorities did this on the grounds that Cuba’s debt had not been incurred for the benefit of the Cuban people, that it had been contracted without their consent and that the loans helped to finance their repression by the Spanish colonial government’ (p. 338).