2. Building Resilience in Fragile States in Sub-Saharan Africa

International Monetary Fund
Published Date:
October 2014
  • ShareShare
Show Summary Details


Fragile states—states in which the government is unable to reliably deliver basic public services to the population—face severe and entrenched obstacles to economic and human development. While definitions of fragility and country circumstances differ, fragile states generally have a combination of weak and noninclusive institutions, poor governance, and constraints in pursuing a common national interest. As a result, these states typically display an elevated risk of both political instability (including civil conflict), and economic instability (through a low level of public service provision, inadequate economic management, and difficulties to absorb or respond to shocks). In addition, crises in such states can have significant adverse spillovers on neighboring countries. At the other end of the spectrum, resilience can be defined as a condition where enough institutional strength, capacity, and social cohesion enable the state to promote security and development and to respond effectively to shocks.

In the early 1990s, much of sub-Saharan Africa—20 out of 44 countries—was regarded as “fragile” (Figure 2.1 and Box 2.1). But the period since then has seen several changes: in some countries societies and leaders have coalesced around a national agenda based on peace and development; the end of the Cold War created a potential for a global peace dividend and an end to surrogate conflicts; the world economy has grown strongly, with emerging markets providing a stimulus to both global growth and global demand for natural resources; the international community wrote off most of the debt of the poorest countries through the HIPC and the MDRI initiatives; and initiatives have sought to reorient aid to be more responsive to recipient country needs and to build needed capacity.

Figure 2.1.Building Resilience in Sub-Saharan Africa

Source: IMF staff calculations.

And progress has been made since the 1990s, including in seven countries—Cameroon, Ethiopia, Mozambique, Niger, Nigeria, Rwanda, and Uganda—that seem to have transitioned out of fragility. These countries, of which two benefited from a natural resource windfall, were able to build more inclusive political arrangements, strengthen institutions, and foster investment.1 They were also able to maintain macroeconomic stability and increase domestic revenues to step up public investment.

Box 2.1.Gauging Fragility in Sub-Saharan Africa

The complex and multidimensional nature of fragility does not lend itself to a simple measure. Even for one of the dimensions, such as “weak institutions,” identifying the most relevant institutions is country-specific, and measuring institutional strength is difficult. A further complication is that most dimensions of fragility (i.e., economic foundations, political instability, capacity constraints) are measured along a continuum—rather than a binary condition—requiring decisions on where to place a threshold to differentiate fragile states from other countries.

Despite these challenges, donor agencies and international financial institutions have worked on operational criteria for measuring and identifying fragility. The World Bank and the African Development Bank regard a state as fragile if it either has an aggregate country policy and institutional assessment (CPIA) rating of 3.2 or less or if it has been hosting a United Nations or regional peace-keeping or peace-building mission in the past three years. The CPIA assesses the quality of a country’s economic and institutional framework and the 3.2 threshold separates the bottom two quintiles of the distribution. Anchoring the assessment on the CPIA score places a greater weight on a country’s economic and institutional framework but does not capture the political or security dimensions of fragility; other indices—such as the OECD-DAC and the Brookings’ Index of State Weakness—place more weight on security and political variables. For instance, the OECD-DAC uses a broader definition of fragility whereby the state is impaired to provide for development and safeguard the security and human rights of its population (OECD, 2013). However, as most indices aim at measuring the degree of state impairment, most countries identified as fragile in one list appear as fragile in other lists (for example, the correlation between the CPIA and the Brookings’ Index of State Weakness is about 0.8).

The analysis in this chapter broadly follows the approach of the World Bank and African Development Bank, with data on CPIA ratings and on conflicts used to identify fragile states in sub-Saharan Africa before 2001 and during 2011–13 (the decade in between is taken as a transition period).

  • Classification of countries before 2001. A country was deemed fragile if its average score on the CPIA ratings during 1991–2000 was 3.2 or less or if it experienced ‘significant conflict,’ the latter defined as either five or more years of lower-level conflict (less than 1,000 deaths per year) or two or more years of severe conflict (more than 1,000 deaths per year). The analysis is based on conflict data compiled by Uppsala University (there are no data on the presence of United Nations forces for that period).
  • Classification of countries in the most recent period. A country is considered fragile if its average score on the CPIA ratings in the period was 3.2 or less or if it had hosted a United Nations/regional peace-keeping or peace-building mission during the three-year period (the results are the same using a five-year average).
  • Countries that were identified as fragile in the 1990s but not in 2011–13 are identified as “became resilient,” and those not identified as fragile in either period as “remained stable.”
Table 2.1.1.Classification of Sub-Saharan African Low-income Countries during 2011–13(Percent)
Remained or became fragileFragile, but progress madeBecame resilientRemained stable
Burundi +Angola + ⊚Cameroon + ⊚Benin
Central African Rep. +Congo, Dem. Rep. of + ⊚Ethiopia +Burkina Faso +
Chad + ⊚Congo, Republic of + ⊚MozambiqueCabo Verde
Comoros +Liberia + ⊚Niger +Gambia, The
Côte d’Ivoire + ⊚Nigeria + ⊚Ghana
Eritrea +RwandaKenya +
Guinea + ⊚Uganda +Lesotho
Guinea-Bissau +Senegal
Madagascar +Tanzania
Malawi +Zambia ⊚
Mali +
São Tomé & Príncipe
Sierra Leone + ⊚
Togo +
Zimbabwe +
Sources: Staff assessment, based on data for the CPIA ratings, the Uppsala conflict database, and information on United Nations/regional peace-keeping or peace-building missions.‘+’ OECD DAC considered these countries to be fragile in 2014.‘⊚’ Resource–rich countries.
Sources: Staff assessment, based on data for the CPIA ratings, the Uppsala conflict database, and information on United Nations/regional peace-keeping or peace-building missions.‘+’ OECD DAC considered these countries to be fragile in 2014.‘⊚’ Resource–rich countries.

The results indicate that 11 countries managed to consistently improve their CPIA rating in the past decade. Of these countries, seven made enough progress to be classified as “resilient” or “stabilized,” while four others, although still displaying features of fragility, also showed significant improvements. Nevertheless, nine countries were not able to make much progress and six actually regressed (Côte d’Ivoire, Eritrea, Madagascar, Malawi, Mali, and Zimbabwe). South Sudan is not included in the analysis as it was not a separate country in the 1990s.

Why were other countries not able to make similar transitions? While it has long been recognized that the transition from fragility is a complex and long process, could one not have expected more countries to take advantage of favorable external conditions, a decline in the incidence of major conflicts and, in some cases, commodity booms that raised GDP and provided fiscal space even in the absence of effective revenue administrations?

This chapter follows an earlier paper that reviewed the IMF engagement with fragile states and set out changes in policies to better serve these countries (IMF, 2011a). The chapter examines the features that distinguish those countries that became resilient as well as those that could not make progress or regressed by providing an overview of the factors at play, analyzing the performance of countries that began the period as fragile, and reviewing selected case studies. The chapter ends with a few observations on what appear to be the key steps in building resilience.

The Analytics of Fragility

What lessons can be drawn from the literature in terms of effective strategies for exiting fragility? Just as there is no single or common cause of fragility—also in light of the variety of individual country circumstances—there can be no single template for building resilience. However, the analytical work on fragility suggests that solid steps that are part of a long-term vision—with adequate tailoring to the specifics of each situation—are needed to build resilience (this is because of the deep-rooted nature of fragility and the recognition that resilience can take decades to achieve). Such steps should aim at strengthening security, fostering inclusive politics, implementing selected legal, governance and economic reforms, and building capacity.2 In the near term, inclusive politics does not necessarily mean holding free elections but rather implementing a political arrangement that can broadly encompass the interests of society to deter violence while setting the basis for the development of democratic institutions.

Reforms aimed at improving governance and accountability are important, especially in resource-rich countries. Therefore, building resilience involves a mutually reinforcing interaction between state capacity, governance, and growth: income growth (and the structural change that supports it, such as the development of markets) creates the conditions for improving legal and fiscal capacity (including taxation, checks and balances, delivery of services, public investment), which in turn bolsters growth-enhancing structural change. This process needs to be well prioritized and timed on the basis of the assessment of the main factors at play and in line with the level of capacity to implement and absorb reforms.

Recent analytical work has highlighted the role of fiscal institutions in coming out of fragility (Besley and Persson, 2011).3 Public financial management reforms (including revenue management in resource-rich countries) are at center stage as they build the legitimacy of the state by increasing transparency, accountability, and efficiency. On one side, mobilizing revenue in fragile states is essential to support the delivery of public goods and services and foster robust state-society relations. Higher revenue equips the government with the resources it needs for development, but its importance goes beyond that. As noted in OECD (2013), “a transparent and efficient tax system simultaneously bolsters intra-societal relationships and the relationship between citizens and the state,” strengthening governance and the legitimacy of the state. On another side, the main goals of public expenditure are to (i) improve budget execution to establish credibility in the budget and execute development programs; (ii) implement reforms that enhance transparency and accountability, especially through regular publication of fiscal revenues and expenditures; and (iii) work on systems to strengthen financial management in line ministries or subnational governments. In low-capacity environments, careful prioritization and, in some cases, a two-track approach can be considered whereby public services are promoted by the government but initially supervised or delivered by qualified nongovernment entities, while expenditure management and revenue administration reforms are implemented over time alongside improvements in state capacity.

In the medium term, reforms to support the development of the private sector are also critical, particularly those that promote a better enforcement of property rights and facilitate access to credit. In many postconflict cases, targeted policies are needed to promote employment or improve social conditions. At first, embarking on priority reconstruction projects could be useful to create employment and foster economic recovery. Another example of targeted intervention is the assistance provided in many countries to reintegrate demobilized soldiers into peaceful activities after a conflict has ended.4

Although state effectiveness has to be led from within, what is the role for external parties? Studies of the role of aid in fragile states (Chauvet and Collier, 2008; Feeny and McGillivray, 2009) have highlighted that—while growth has been higher than it would have been without aid, building capacity is critical and aid is more effective when it is consistent with national absorptive capacity and is progressively delivered using national systems.5

Factors and Policies Associated with Building Resilience

What factors and policies helped build resilience in fragile states in sub-Saharan Africa? This section looks into this issue by analyzing trends in institutional strength, conflict, macroeconomic growth performance, the role of fiscal institutions and policies, and social outcomes. The analysis focuses on associations and correlations, as the feedback interactions among the different determinants of fragility mar attempts to establish causality.6

These associations suggest that:

  • Fragility is persistent, as the vicious cycle of conflict and political instability and weak growth performance is hard to break. Once sufficient progress is made, however, the achieved resilience is also persistent, supported by a virtuous cycle of stronger institutions, absence of significant conflicts, better economic performance, and improved social indicators. The achieved resilience, however, should not be seen as immutable, considering the possibility that countries can still face renewed political turmoil or severe shocks that can push them back into fragility.
  • Building resilience is associated with economic reforms and sound macroeconomic policies. Countries that built resilience managed to achieve macroeconomic stability and were characterized by better fiscal outcomes and budget institutions. They seem to have been able to mobilize more revenue and make enough room for investment. In addition, support from donors and international financial institutions and an environment conducive to investment and political stability are also associated with resilience. These findings are consistent with the narrative from the case studies in the next section.
  • Resource-rich countries have made significant gains in terms of GDP growth and achieved a measure of economic stability, but some countries have not managed to strengthen institutions and build resilience despite a commodity boom in the past decade.
  • The evidence on social outcomes suggests that building resilience is associated with improvements in social conditions, although there are severe data limitations in this area. Overall, the data show a positive association between building resilience and health and education outcomes.

Assessing progress

How did countries that built resilience perform relative to those that remained fragile? This section uses the classification of countries presented in Box 2.1 based on the country policy and institutional assessment (CPIA) analysis with the exclusion of countries that were labeled as “stable” since the early 1990s.7 Since by construction, the CPIA is correlated with the factors discussed above, the analysis simply aims at taking a closer look at different aspects of the progress made in countries that gained resilience. In the analysis, resource-rich fragile countries are singled out as a distinct group. This distinction is introduced because the commodity boom that many sub-Saharan African countries experienced since 2000 raises the question of whether these countries’ economic fortune has been used to build resilience. Resource-rich countries are defined as those whose primary commodity rents exceed 10 percent of GDP.8

Looking at the CPIA, those countries that became resilient experienced some volatility in the 1990s but started to diverge markedly and consistently from the other groups after 2001. Some fragile resource-rich countries also showed consistent improvement in recent years, while others (including nonresource-rich countries) had a lackluster performance after the mid-1990s.

The CPIA rates countries against a set of criteria grouped in four clusters: economic management, structural policies, policies for social inclusion and equity, and public sector management. A decomposition of the CPIA among its four clusters shows interesting insights:

  • The group of countries that became “resilient” in recent years made steady progress across all clusters of performance covered by the CPIA, achieving macroeconomic stability and building institutions (Table 2.1).9 Their CPIA has followed a rising trend and has remained consistently above the 3.2 threshold.
  • Several countries, hampered by inadequate capacity and other constraints, were unable to deliver the required services to their populations and continued in a state of fragility with CPIA scores well below 3.2.
  • Among the group of fragile countries, those rich in natural resources did not fare much better; four of them did make some progress, especially in the area of macroeconomic stability, but further progress is clearly required on institution building.
Table 2.1.Average Change in CPIA Scores by Country Groups
Overall CPIA 1Economic

Management 2

Policies 3
Policies for Social

Inclusion/Equity 4
Public Sector Management

and Institutions 5
Fragile resource-rich0.400.290.170.330.31
Fragile nonresource-rich-0.33-0.20-
Sources: World Bank; and IMF staff calculations.

Changes are measured as the difference between average scores in 2011–13 and 1991–2001.

The economic management cluster includes monetary and exchange rate policy, fiscal policy, and debt policy.

The structural policies cluster includes trade, the financial sector, and the business regulatory environment.

The policies for social inclusion and equity cluster includes gender equality; equity of public resource use; building human resources; social protection and labor; and policies for environment sustainability.

The public sector management and institutions cluster includes property rights and rules-based governance; quality of budgetary and financial management; efficiency of revenue mobilization; quality of public administration; and transparency, accountability, and corruption in the public sector.

Sources: World Bank; and IMF staff calculations.

Changes are measured as the difference between average scores in 2011–13 and 1991–2001.

The economic management cluster includes monetary and exchange rate policy, fiscal policy, and debt policy.

The structural policies cluster includes trade, the financial sector, and the business regulatory environment.

The policies for social inclusion and equity cluster includes gender equality; equity of public resource use; building human resources; social protection and labor; and policies for environment sustainability.

The public sector management and institutions cluster includes property rights and rules-based governance; quality of budgetary and financial management; efficiency of revenue mobilization; quality of public administration; and transparency, accountability, and corruption in the public sector.

Conflict and political instability

The incidence and severity of conflicts in sub-Saharan Africa have declined gradually since the early 1990s. While about nine countries experienced conflict in any given year in the 1990s, only five experienced conflict at any time between 2011 and 2013. The incidence of severe conflicts also fell, from an average of about three countries affected by major conflicts (more than 1,000 deaths per year) each year in the 1990s to almost no country in the recent period. Furthermore, there was a substantial decrease in the incidence of conflict among fragile resource-rich countries and those that became resilient. Notwithstanding the trend toward improved security in the region, threats have emerged in recent years, especially in Central African Republic, Mali, Nigeria, South Sudan, and, on a much more limited scale, Kenya and Mozambique.

In parallel with the gradual improvement in security conditions, political stability has also improved, especially in countries that became resilient and in fragile resource-rich countries. For example, between 1996 and 2012, the World Bank index of political stability shows improvements of 14 percent and 42 percent, respectively, in the groups of resilient and fragile resource-rich countries, and a 47 percent deterioration in the index for fragile nonresource-rich countries.

Macroeconomic performance

Since the early 2000s, different country groups have exhibited a markedly different growth performance. Countries that have become “resilient” and resource-rich fragile countries have displayed stronger real GDP growth compared with non-resource-rich countries that have remained fragile or regressed (Figure 2.2).10

Figure 2.2.Macroeconomic Indicators

Sources: World Economic Outlook database; and World Penn Tables.

The group of resilient countries, which was less dependent on commodity exports, managed to implement good economic policies and reform programs supported by a favorable regulatory and institutional environment, which in turn contributed to higher investment, including through better access to credit. The resilient group also experienced a marked decline in inflation, which fell from more than 20 percent per year in the early 1990s to single digits in recent years. These countries strengthened the capacity of their central banks, which enabled them to maintain a predictable foreign exchange regime and to develop successful monetary and exchange rate policy frameworks to anchor inflation. Furthermore, they also managed to strengthen and develop their financial markets (IMF, 2014a).

Improvements in external conditions helped but did not determine the extent of progress (or lack thereof) in resource-rich countries, while a deterioration of such conditions posed a drag for other countries. On the one hand, resource-rich countries benefited from persistent improvements in their terms of trade (which rose at an annual rate of 4 percent since the early 2000s) that resulted in a steady export boom (on average, their export receipts increased from 30 percent to 45 percent of GDP). This commodity boom enabled them to achieve better growth and lower inflation. As indicated above, however, four of these resource-rich countries have improved their fiscal institutions (as reflected in the CPIA clusters), while other resource-rich countries do not show such progress. Moreover, private investment has not yet picked up in resource-rich countries. On the other hand, we observe that nonresource-rich countries (many of these oil importers) that remained fragile or regressed experienced an average annual decline of 2 percent in their terms of trade, which constrained their export earnings and may have contributed to their fragility.

Looking at growth volatility, fragile resource-rich countries exhibited the highest volatility compared with the other country groups in both the 1990s and 2000s (Figure 2.3). At the same time, resilient countries experienced a significant decline in volatility in the past decade, while fragile nonresource-rich countries showed little volatility over both subperiods.

Figure 2.3.Average Growth Volatility in Real GDP per Capita: 1990s versus 2000s


Sources: World Economic Outlook database; and IMF staff estimates.

A deeper look at the growth performance reveals that resilient countries have achieved periods of sustained GDP growth and successfully avoided growth breakdowns (formally called accelerations and decelerations).11 The group of countries that remained fragile or regressed has experienced, on average, fewer years of growth accelerations than other country groups during 1990–2011 (Figure 2.4). And while fragile countries have often experienced growth downturns, the data point to the virtual absence of growth decelerations among resilient countries.12 In addition, fragile countries, particularly resource-rich ones, have experienced on average larger contractions per episode. In other words, countries that managed to build resilience through better institutions and policies not only experienced more and longer growth accelerations but they also managed to avoid sharp and sustained periods of weak or negative growth.

Figure 2.4.Sub-Saharan Africa: Real GDP per Capita Growth Acceleration and Deceleration, 1990–2011

Sources: World Penn Tables; and IMF staff estimates.

Budget institutions and fiscal space

Fiscal policy plays a critical role in delivering public goods and services and catalyzing private investment. As such, an important angle in transitioning out of fragility is the ability of countries to build stronger fiscal institutions, mobilize domestic revenues, and carry out growth-enhancing fiscal spending. In fact, those countries that became resilient do display relatively stronger fiscal institutions as measured by indices of the quality of budget institutions (Gollwitzer, 2011; Dabla-Norris and others, 2010). In contrast, although some fragile resource-rich countries have gained fiscal space in recent years, the quality of their budget institutions remains relatively low (Figure 2.5).

Figure 2.5.Sub-Saharan Africa: Fiscal Institutions and Policy Space

Disaggregated fiscal space indicators suggest substantial progress in both resilient and resource-rich fragile countries (Figure 2.6). Both groups of countries succeeded to better control their fiscal deficits compared with other fragile countries, even though countries in all three groups benefited substantially from debt relief during this period. Furthermore, resilient countries have also generated more fiscal policy space—in terms of lower public debt, stronger government financial positions, higher revenue-raising capacity, and expenditure flexibility—compared with other groups of countries. During and shortly after periods of conflict or acute instability, most of these countries increased current spending to fund oversized armies, yet, they managed to control current spending over time by embarking on civil service reforms or demobilization programs while channeling additional resources to fund public investment. However, nonresource-rich fragile countries experienced debt relief somewhat later than other countries, partly because of a slow pace in reaching the HIPC completion point. In addition, fragile countries were relatively less successful than other countries in raising tax revenue and in containing current spending, and hence had less fiscal space to implement investment programs.

Figure 2.6.Sub-Saharan African Fragile States: Fiscal Space Indicators

Source: IMF World Economic Outlook database.

Social outcomes

Despite the paucity of data on social indicators, there is evidence that most countries have made progress toward achieving the Millennium Development Goals (MDGs), although progress in fragile states has been more modest. Progress in under-five infant mortality rates and primary enrollment rates (probably the best measured MDGs besides per capita GDP), has been more subdued in the fragile group of countries compared with the resilient group (Figure 2.7). Countries that have become resilient, which had the highest infant mortality rates in the early 1990s, managed to reduce infant mortality drastically in the late 2000s.13 Other countries have also made progress, but at a slower pace. Similarly, countries that have become resilient have made faster progress in raising primary school enrollment compared with countries that have remained fragile or regressed. Progress in expanding access to improved water was similar across all country groups.

Figure 2.7.Sub-Saharan African Fragile States: Social Indicators

Source: World Bank, World Development Indicators.

The evidence on poverty reduction is somewhat mixed, possibly because of limited data and measurement problems, although as noted earlier there has been considerable progress in raising GDP per capita in resilient and resource-rich countries. Although poverty rates are consistently higher in the group of fragile countries compared with countries that have become resilient, they have remained relatively high in all country groups since the 1990s. Resilient countries and some of the resource-rich fragile countries, however, show improvements in the social inclusion/equity cluster of the CPIA, although there has not yet been a corresponding decline in poverty rates in some of these countries.14 In any case, it is difficult to draw conclusions because of scant data and measurement errors.

Econometric take on factors linked to resilience

The analytical work on the factors behind fragility has highlighted the following elements: scant constraints on executive power (David and others, 2011; Collier and Hoeffler, 2004), poor economic and social indicators such as low economic growth, high inflation, and high infant mortality (Fearon and Laitin, 2003; Jakobsen and others, 2013), a history of conflict (Collier and Hoeffler, 2004), and weak governance and institutions (David and others, 2011).

Looking at the data, the previous analysis indicated that, among sub-Saharan African countries that were deemed fragile in the 1990s, those that became resilient improved pretty much across all dimensions, be it macroeconomic and growth outcomes, political stability and conflict, or institutions and social outcomes. The analysis also highlighted the important role of fiscal institutions and fiscal space.

The relative strength of those conclusions can also be assessed with a simple econometric model. The model does not aim at revealing causality given that the factors involved are closely intertwined and interact with each other. Instead, a probabilistic regression framework is used to identify factors that are significantly associated with the odds for any of the countries to be deemed resilient in any period (Box 2.2).

The results are consistent with the earlier analysis. Comparing the contribution of each factor in the three country groups, the following implications can be drawn. First, although fragility is highly persistent, resilience appears persistent as well: if a country was resilient at a point in time, it would most likely remain resilient in a subsequent period; all else equal. Second, macroeconomic indicators, namely private investment and terms of trade, also contribute to the odds of becoming resilient. Third, fiscal policy space, particularly measured as the ability to raise public investment, is associated with a higher probability of becoming resilient. Fourth, international support is associated with a better chance of being resilient. Note that the probabilistic model also captures the “curse of natural resources,” where the median resource-rich fragile country is less likely—all else equal—to become resilient than the median resource-poor fragile country.

Case Studies

This section reviews the experience of four countries (Rwanda, Mozambique, Democratic Republic of the Congo, and Central African Republic) that were or still are deemed fragile. Rwanda (a resource-poor landlocked country) and Mozambique (a coastal country that was resource-poor during the period under review) emerged from conflict in the early to mid-1990s, rebuilt capacity and institutions in the following decade, and managed to build resilience as evidenced by CPIA scores consistently above 3.2 since the mid-2000s (Figure 2.8).

Figure 2.8.Overall CPIA Scores

Source: World Bank, Country Policy and Institutional Assessment.

Democratic Republic of the Congo and Central African Republic have had far more difficulties in building resilience. In Democratic Republic of the Congo, a resource-rich coastal country, conflict ended with a peace accord in 2001 and general elections in 2003. While the improvements in its CPIA score in the early period were encouraging, the country has not yet been able to break through to a zone of nonfragility. Lastly, Central African Republic, a resource-rich landlocked country, has been mired in repeated spells of political and civil conflicts since independence in 1960, with a long string of political instability, coups, and civil conflicts. The 2007 peace agreement started the latest period of stabilization but the country fell back into conflict in 2012, erasing much of the progress made in previous years.

All four country cases have some similarities in policies and priorities, but several elements set them apart as shown in Figure 2.9 (the classification in the figure is subjective, based on the assessments of different factors below).

Figure 2.9.Factors in Building Resilience

Source: IMF staff estimates.

The results suggest that political inclusion that leads to peace and avoids major political turmoil is a precondition for building resilience, while fiscal policy space is important for the government to deliver visible results to the population. In particular, mobilizing domestic revenue goes beyond the fiscal aspect as it creates an implicit contract between the citizens and the government. Donor support appears beneficial if provided in sufficient volume and for a long enough period to sustain the buildup of resilience. In addition, debt relief was critical for debt sustainability and fiscal space, but the key is how well the freed-up resources were used. The successful cases consistently expanded their priority spending and investment, while support from the international community, including the IMF, played an important role.


Political inclusion and checks on power. An inclusive political settlement is an essential foundation for peace and building resilience. In this context, ‘inclusive’ denotes primarily the degree to which previously unrepresented or competing groups have been included; it does not necessarily mean that the system is inclusive in the sense of a well-functioning mature democracy. In Mozambique and Rwanda, broad-based governments defined early on their political, economic, and social objectives and established sufficient institutional provisions to be held accountable for them (the General Peace Agreement for Mozambique in 1992; and, in Rwanda, the formation of a government of national unity in July 1994 comprising five political parties and incorporating the principal provisions of the 1993 Arusha Accord). Judging by the political stability observed in both countries since the 1990s, these efforts have so far been successful (Figure 2.10), though both countries have yet to experience a political transition. In Democratic Republic of the Congo, the political settlement has been holding so far, but still needs to stand the test of time (Inter-Congolese National Agreement in 2003). In Central African Republic, the return to conflict in 2012–13 highlights the feeble implementation of the power-sharing agreements reached in 2007–08 and in 2012.

Figure 2.10.Political Stability, 1996–2011

Sources: World Bank, Country Policy and Institutional Assessment; and IMF staff estimates.

Capacity and institutions. Albeit with different results across countries, efforts at rebuilding economic capacity and institutions focused on three areas: public financial management (PFM), in particular the budget process; mobilizing revenue; and strengthening the central bank and the banking sector. Rebuilding PFM systems was important not only for transparency, accountability, and inclusiveness, but also for the gradual routing of donor support through national budgets. Along with other international financial institutions and bilateral donors, the IMF supported these efforts through technical assistance and training in its core areas of expertise. Most successful among the four countries were Rwanda and Mozambique (Figure 2.11) although they still have some way to go to fully implement their reform agenda. Rwanda reinstituted the budget process with parliament adopting annual budget laws since 1998, and had broadly rebuilt its PFM system by the mid-2000s. Tax administration was strengthened and has remained a priority for the authorities. The central bank’s effectiveness to run monetary policy was improved quickly, but reforming the banking sector proved difficult and took longer than anticipated. In Mozambique, revenue administration reforms were instrumental in achieving a steady increase in government revenue since 1999, and the 2002 PFM law paved the way for increased transparency in budget execution. Central bank functions were streamlined in the early-2000s although central bank independence and restructuring of the banking sector took more time to materialize. Democratic Republic of the Congo made progress in the first two years after the peace accord, but has regressed since then. The initially good economic performance proved difficult to sustain owing to political instability, recurrent conflicts, and lack of reforms, including support for ex-combatants. Fiscal space limitations and revenue shocks resulted in low pro-poor spending and investment. In 2005, a new election cycle and fiscal loosening led to high inflation and a loss in foreign reserves as well as delays in the implementation of reforms, with pervasive poverty and other vulnerabilities, leaving the country exposed to crises and reversals. Central African Republic also improved somewhat in these areas at first, but fell back again with the onset of renewed conflict.

Figure 2.11.CPIA Public Sector Management and Institutions Cluster, 2005–13

(Three-year moving average)

Source: World Bank, Country Policy and Institutional Assessment.

Macroeconomic stability. Macroeconomic stability was lost in periods of conflict in all countries and in most cases restored within two to four years after the conflict. Mozambique, Rwanda, and Democratic Republic of the Congo all moved quickly to liberalize prices, control monetary growth, and remove other state controls on the economy and the financial sector, facilitating a swift consolidation and the regrouping of their economies. In parallel, economic policymaking and capacity were gradually strengthened. Mozambique and Rwanda then set off on a strong postconflict rebound with prolonged high growth before stabilizing in later years. In contrast, the long-term erosion of the economy and the state in Democratic Republic of the Congo impeded its ability to catch up quickly. Although macroeconomic stability was restored and growth resumed at about 5–6 percent, there was no postconflict rebound comparable with Mozambique and Rwanda, and inflation remained high for several years. Similarly, in Central African Republic, limited progress was made toward macroeconomic stability and growth remained weak before the country was caught in conflict again.

Delivering to the population

Policy space. The return to liberal market systems in Democratic Republic of the Congo, Mozambique, and Rwanda not only helped to regain macroeconomic stability and growth, but also to create policy space. The liberalization of prices drove up inflation temporarily, but as market incentives and stabilization policies started to work, inflation abated and real incomes increased. Moreover, the liberalization of the trade regime helped bring in much-needed goods at lower prices. Liberalizing the foreign exchange system also increased policy space and helped bolster foreign exchange reserves. In contrast, while the exchange rate peg provided a much-needed anchor, Central African Republic had difficulties building sufficient policy space as fiscal policy could not be adjusted enough, notably through revenue mobilization and reforms, with the recurrence of conflicts making progress even more difficult.

Fiscal space

Mobilizing revenue. All four countries placed emphasis on mobilizing domestic revenue, but the results varied. Mozambique, Rwanda, and Democratic Republic of the Congo made impressive progress (albeit Democratic Republic of the Congo from a very low base and as a result of hydrocarbon revenues). In contrast, Central African Republic made little or no progress (Figure 2.12).

Figure 2.12.Government Revenue, 1990–2013

Source: IMF, World Economic Outlook database.

Donor support. Aid levels to the four countries were significant, especially following conflict. Aid flows to Democratic Republic of the Congo, Mozambique, and Rwanda averaged about 50 percent of GDP in the immediate years after conflict, leveling off to about 20 percent of GDP annually since then. At about 10 percent of GDP, aid levels to Central African Republic were much smaller, with fluctuations reflecting recurrent instability and conflict. Although there is concern whether countries can fully absorb drastic surges in aid flows, high levels of aid seem to be needed until the country has managed to build some resilience.

Debt relief. Debt relief under the enhanced HIPC and MDRI initiatives was successful in restoring debt sustainability in all four countries, but in terms of supporting the buildup of resilience, the decisive factor was how the additional fiscal space was ultimately used. Debt service reductions freed up resources in the order of about 1.5–2 percent of GDP per year intended to help increase social and priority spending. However, as noted below, there are differences in the degree to which this was reflected in actual budgets.

Priority spending. All four countries developed Poverty Reduction Strategy Papers (PRSP) through a participatory consultative process in which they laid out their developmental priorities. The amount spent on these economic, institutional, and human development priorities are a good measure of the government’s commitment to them and, more broadly, to building a more inclusive society. Mozambique saw the largest and most sustained increase in priority spending, rising from 6 percent of GDP in 1999 to about 15 percent of GDP in the early 2000s and to about 20 percent more recently. Similarly, Rwanda expanded its priority spending from 4 percent of GDP in 1999 gradually but steadily to about 12–14 percent of GDP in 2008–12. In contrast, priority spending in Democratic Republic of the Congo has hovered around 6 percent of GDP, and in Central African Republic, it has remained about 2–3 percent.

Public investment. Public investment plays an important role in rebuilding infrastructure, attracting private investment, and boosting growth. During the period under review, Mozambique outperformed the other countries while Rwanda has been catching up since the early 2000s, with both countries’ investment ratios now in the range of 12–15 percent of GDP. Democratic Republic of the Congo and Central African Republic have remained well below these levels. Democratic Republic of the Congo has recently climbed up from near zero to more than 5 percent of GDP, whereas Central African Republic has been in a gradual long-term trend decline, reaching about 3 percent of GDP recently. In addition to the volume, the quality of public investment is also important, both in terms of project selection and implementation and in terms of the quality of the outcome. Judging by the latter, Mozambique and, in particular Rwanda displays a higher quality of infrastructure compared with other sub-Saharan African countries (Chapter 3, Figure 3.4).

International support

Donor coordination. As countries progressed from emergency aid to development aid, donor coordination became stronger in all cases except in Central African Republic. In Mozambique, close donor coordination began in the mid-1990s and was formalized in 2000, coordinating support in several areas (that is, tax reform, financial sector, trade, poverty reduction, private sector development, and health and education). In Rwanda, donor coordination began in 1998 and was formalized in 2003, with donors funding an aid coordination unit in the finance ministry and coordinating public expenditure reviews, macroeconomic reviews, and poverty reduction monitoring. In Democratic Republic of the Congo, donor coordination was strengthened since 2005 and a country assistance framework was established in 2008, covering 95 percent of all external assistance. In contrast, donor coordination in Central African Republic has remained informal despite several attempts to form a consultative group.

IMF-supported programs. The IMF has been closely engaged with Mozambique and Rwanda, supporting the authorities’ economic strategies through early and continued programs to the present day. In addition to providing direct financial and technical support for countries’ strategies, IMF programs play a catalytic role in unlocking support from other donors. The IMF supported Mozambique from before the end of its conflict—the country successfully implemented five medium-term programs between 1987 and mid-2007 before moving to a policy support instrument (PSI). In the aftermath of the genocide in 1994, Rwanda was supported through emergency facilities (1995, 1997), while capacity was being rebuilt to implement an upper-credit tranche program. Since 1998, the IMF supported a series of medium-term economic programs with structural adjustment facilities and, more recently, with a PSI.15 In contrast, Democratic Republic of the Congo did not have an IMF-supported program until 2002 and was in arrears to the IMF. After arrears were cleared, Democratic Republic of the Congo had two structural adjustment programs during 2002–06 and 2009–12, although performance under these programs was uneven due to political uncertainty and social tensions coupled with low levels of priority spending. Finally, Central African Republic’s involvement with the IMF was characterized by large gaps within and between programs, reflecting recurring crises.

Private sector

The private sector does not appear to have played a significant role in the early stages of recovery, except for foreign direct investment in the resource-rich countries and in Mozambique (triggered by the onset of peace and stability). During the early stages of recovery, private domestic investment may have been affected by lingering uncertainties.

Foreign direct investment (FDI). FDI played a significant role in the resource-rich countries, particularly in Democratic Republic of the Congo and Mozambique. In Mozambique, FDI into aluminum production rose after the peace accord and, after a slump in the early to mid-2000s, picked up rapidly to above 10 percent of GDP. In Democratic Republic of the Congo, FDI had been hovering since the 1970s at low levels, but took off in 2002–03 reflecting the end of the civil war and ensuing political stabilization. FDI in Central African Republic displayed a similar pattern, though the increase after the onset of peace was much smaller, peaking at 6 percent in 2009. In contrast, Rwanda has not attracted significant amounts of FDI.

Private domestic investment. In Mozambique, private domestic investment was relatively low at an average of about 7 percent of GDP in the 1990s and 2000s. Following the discovery of large gas and coal deposits, it surged to more than 30 percent since 2009 (part of the increase is related to the influx in FDI). In Rwanda, private investment recovered gradually and has recently been at about 12 percent of GDP. In Democratic Republic of the Congo, it fluctuated strongly, reflecting the volatility of the political and security situation. Finally, private investment in Central African Republic also suffered from political instability; and even though it rose following the 2007 peace accord, it has not surpassed 8 percent of GDP during the past two decades.


Security, political stability, and governance. Both Mozambique and Rwanda followed an approach that led to political stability and avoided conflict. Both countries also strove to improve their governance systems, as evidenced by their significantly higher ratings on indicators of governance effectiveness, regulatory quality, control of corruption (Figure 2.13), and rule of law (Figure 2.14). In contrast, both Central African Republic and Democratic Republic of the Congo have struggled to make progress in these areas.

Figure 2.13.Control of Corruption, 1998–2012

Source: World Bank, Governance Indicators.

Figure 2.14.Rule of Law, 1998–2012

Source: World Bank, Governance Indicators.

Economic growth. Mozambique and Rwanda enjoyed sustained increases in real per-capita income since the mid-1990s, which accelerated in the last decade (Figure 2.15). Since the early 2000s, both countries embarked on second-generation economic reforms that helped sustain growth beyond the postconflict rebound. In Mozambique, reforms to strengthen revenue mobilization and PFM continued, as did efforts to strengthen governance and the anticorruption framework. In addition, the country embarked on reforms to strengthen the monetary and financial sectors, the framework for managing natural resources, and the business and investment climate. In Rwanda, reforms also focused on the latter, including financial sector and legal reforms, boosting trade and diversification, and raising agricultural productivity. Unfortunately, both Central African Republic and Democratic Republic of the Congo were not able to accomplish much in these areas.

Figure 2.15.Real GDP per Capita

Source: World Penn Tables.

Social progress. Political and macroeconomic stability, growth, higher social spending, and investment have led to significant improvements in social indicators in both Mozambique and Rwanda (Table 2.2). In both countries, poverty rates were reduced substantially (though they are still high), enrollment rates increased, and the under-five mortality rate declined. Central African Republic appears to have made progress in poverty reduction and net enrollment rates, although the under-five mortality has not declined much. In contrast, poverty and under-five mortality rates remained at high levels in Democratic Republic of the Congo.

Table 2.2.Sub-Saharan Africa: Social Indicators
EarliestLatestData Vintage
A. Poverty headcount ratio at $1.25 a day (PPP, percent of population)
Central African Republic83631992, 2008
Congo, Democratic Republic of then.a.882006
Mozambique81601996, 2008
Rwanda75632000, 2011
B. Net enrollment rate, primary and secondary education
Central African Republic59691990, 2011
Congo, Democratic Republic of the70791992, 2011
Mozambique73911990, 2012
Rwanda951031990, 2011
C. Mortality rate under five years (Per 1,000 live births)
Central African Republic1681641990, 2012
Congo, Democratic Republic of the1811681990, 2012
Mozambique2281031990, 2012
Rwanda180901990, 2012
Source: World Bank, World Development Indicators.
Source: World Bank, World Development Indicators.

The Way Forward

The analysis in this chapter highlights that, while many fragile states in sub-Saharan Africa have made progress since the 1990s, there are still too many countries that have not been able to break out of fragility despite a supportive external environment. At the same time, there are cases of countries that have regressed, highlighting the need to ensure that efforts to build resilience and related reforms are sustainable. Moreover, new challenges have surfaced; for example, the recent emergence of violent groups operating across borders creates new threats to the cohesiveness of states and a need for these states to work cooperatively.

On the one hand, the analysis shows that a number of countries have managed to build resilience by setting up more inclusive political arrangements, strengthening the quality of their economic policies and key economic institutions (especially through better fiscal policy and by building budgetary capacity), and improving the environment for investment. International financial and technical support, including from the IMF, have also played a supporting role. Debt relief helped restore debt sustainability and, together with increased aid and domestic efforts, helped expand priority spending and investment on development. The evidence in terms of social outcomes is not entirely conclusive, but some gains are evident, especially in health and education.

On the other hand, fragility has been persistent in several countries where the factors of state weakness were all at play, namely hesitant leadership and lack of political cohesion; weak capacity and poor commitment to build economic institutions and to implement pro-growth policies and reforms; and inability to generate and appropriately use policy space, all leading to recurrent crises and/or conflict. The countries caught in this trap include several countries that are rich in natural resources and that, while experiencing windfall export gains in recent years, still need to translate those gains into concrete development outcomes.

The results indicate that a combination of reinforcing factors is most likely needed for countries to overcome fragility. A first condition for building resilience is political inclusion that helps sustain peace and prevents major political turmoil. A second necessary requirement appears to be an effective leadership capable of driving the adoption of policies and reforms that promote good governance, transparency, and accountability. These policies and reforms would foster economic stability, strengthen fiscal institutions, and generate fiscal policy space to deliver improvements in living standards. In this regard, domestic revenue mobilization can play a double role in enhancing the government’s financial strength and in establishing an implicit contract between the citizens and their government that promotes good governance and accountability. In resource-rich countries, it is also important to establish effective frameworks for the transparent management of natural resource wealth. In the medium term, fostering an environment that promotes the expansion of the private sector also seems necessary to achieve sustained growth. Third, international stakeholders should be prepared to engage with fragile countries on a long-term basis, providing financial assistance in ways that can improve the effectiveness of the state, coordinating their efforts closely, and focusing capacity development efforts on economic institutions, especially fiscal ones.

Exiting fragility remains a difficult challenge for several countries, and additional work is needed to better understand the processes of state building and capacity building. In recent years, the international community has sought to respond to these challenges by refining its modalities of engagement with fragile states, including through an increased focus on capacity building. The analysis is moving ahead in various areas, including the role of domestic natural resources and revenue mobilization (OECD, 2013) and the specific challenges related to harnessing natural resource wealth (AfDB, 2014; Africa Progress Panel, 2013; and Collier, 2012). In addition, a new international dialogue has been established in which development partners, multilateral agencies, and the G7 group of 18 fragile and conflict-affected states cooperate to promote ownership and use best practices under the “New Deal”.16 It is recognized that the pursuit of resilience involves a transition through a number of phases, ranging from complete state failure and conflict to less extreme symptoms of weak governance and institutions, with an evolving set of challenges as countries move along this spectrum.17 The World Bank and the African Development Bank (AfDB) have joined in this dialogue contributing to the analytical work, notably with the World Bank’s 2011 World Development Report on overcoming fragility. At the same time, the IMF has adopted a revised framework for engagement with fragile states, focused on greater flexibility to support economic programs, due regard for political economy issues, and stronger capacity-building efforts to strengthen economic institutions and policymaking (IMF, 2011a).18

Box 2.2.Logit Framework

The probabilistic regression framework assesses whether and by how much a set of explanatory variables affects the odds for a country to become “resilient,” with resiliency approximated by a time-varying indicator variable that takes a value of 1 when the CPIA score is above 3.2 and there are no significant conflicts, and zero otherwise. It should be noted that, by construction, the logit model does not imply causality and that most independent variables are not exogenous. Nevertheless, the results can help identify factors that are significantly associated with resiliency.

The explanatory variables include the first-order lag for the resiliency indicator (due to high persistence),1 macroeconomic indicators (growth rate of GDP per capita, double-digit inflation, terms of trade, and private investment), political stability (presence of severe conflicts and constraints on executive power), development indicators (infant mortality), fiscal factors (tax revenue and public investment), international support (development aid and the presence of a medium-term program supported by the IMF as a proxy for engagement with international financial institutions), and interactions with a dummy variable for resource-rich countries.

The regression results2 (Figure 2.2.1) point to the difficulties that fragile countries face in becoming resilient and the fact that, once such state is achieved, it has persistence. At the same time, better terms of trade, higher private investment, and public investment are associated with resiliency. The presence of a medium-term program supported by the IMF is also associated with resiliency. Comparing the total probability of being resilient and each significant component for country groups, it is clear that for the median resilient country, higher investment, more favorable terms of trade, and more engagement with international institutions lead to a higher predicted probability of becoming resilient. Moreover, the estimation also illustrates the “curse of natural resources,” with resource-rich countries displaying—other things equal—a lower probability of becoming resilient than resource-poor countries.

Figure 2.2.1.Key Factors Associated with Building Resilience: Simple Logit Model

Sources: IMF, World Economic Outlook database; and IMF staff estimates.

Note: The model is estimated by a population-average panel logit regression using the panel of 26 fragile sub-Saharan African countries from 1992–2013. Contribution of each factor is calculated with the coefficient estimates at the median value of each factor in three different country groups.

1 The model is estimated by a population-average logit estimator, which specifies the marginal distribution of the population rather than the full distribution.2 The columns in the figure represent odds ratios from the logit regression. The interpretation of the odds ratio is standard. For instance, if a variable has an odds ratio of 2, it means that increasing the variable by 1 unit will double the odds of a country becoming resilient.

In the remainder of the chapter, the concept of institutions is applied in the narrow sense that denotes structures and rules governing specific areas of public intervention, such as fiscal operations.


The need for such an approach is rooted in the notion that fragility is a pervasive condition whereby political instability and violence, weak enforcement of contracts, bad governance (corruption), and capacity constraints are mutually reinforcing factors that can keep countries trapped in a low-investment, slow-growth equilibrium with high risk of recurrent crises or violence (Andriamihaja and others, 2011).


According to this analysis, basic determinants such as common interests in society or cohesive political institutions help drive investment in state building and improvements in fiscal management and legal capacity. Conversely, in the absence of common interests and/or cohesive political institutions, government revenues from natural resources or flows of aid can trigger political instability and violence. States with elite control and poor governance tend to be predatory, with poor incentives to invest in state capacity, creating conditions for important needs of citizens to remain unmet. Thus the same conditions that lead to low investment in state capacity also lie at the root of political instability—repression of the opposition by incumbents and, at the limit, civil conflict.


This task is particularly important given the risk that demobilized soldiers lacking a rewarding activity could reengage in counterproductive activities, including conflict.


A careful prioritization of policies and reforms is critical, as failure to deliver results can also compromise the momentum for reforms. As noted in Pritchett, Woolcock, and Andrews (2013), trying to force the pace of institutional development in these conditions can encourage the phenomena of isomorphic mimicry—laws and organizational structures in appearance resemble those of developed states but in practice fail to perform the functions that they are supposed to fulfill.


Poor data quality and availability, which are likely to be nonrandomly associated with countries’ institutional capacity, is another challenge.


Papers that use the CPIA score to define fragility (the opposite of resiliency) include Bertocchi and Guerzoni (2010) and Chauvet and Collier (2008). As noted in Box 2.1, using other classifications would not lead to meaningful changes in the country groups.


The CPIA methodology has changed over time. After 1997, coverage was expanded to include governance and social policies, and the ratings scale was changed from a 5-point to a 6-point scale. In 2004, a second revision streamlined the evaluation criteria. For the analysis in this chapter and to allow for comparisons over time, the CPIA scores were rebased to a 6-point scale for the whole period under consideration.


Per capita incomes have also increased substantially in countries that became resilient and resource-rich countries. For the resilient and resource-rich groups, real GDP per capita grew from less than 1 percent per year in the 1990s to 4 percent and 3.5 percent per year, respectively, during the last decade. However, the group of nonresource-rich fragile countries barely grew during the past two decades.


We use an approach similar to that in Arbache and Page (2007) to define growth accelerations and decelerations. An acceleration (deceleration) occurs in a year when: (a) forward-looking three-year average per capita GDP growth rate is above (below) the backward-looking three-year average growth rate; (b) forward-looking three-year average per capita GDP growth rate is above (below) the country’s overall average growth rate; and (c) the forward-looking three-year average per capita GDP level is above (below) the backward-looking three-year average GDP per capita level. Only when this acceleration (deceleration) occurs at least for three consecutive years, it becomes a growth acceleration (deceleration) episode. The real GDP per capita is measured in PPP constant U.S. dollars and available until 2011 (PWT version 8). Other studies that analyzed countries’ ability to sustain growth accelerations and managed shorter recessions include Abiad and others, 2012; Pattillo and others, 2005; and Berg and others, 2008.


An earlier report (IMF, 2011a) reached similar conclusions for a larger set of fragile and nonfragile countries. In addition, a forthcoming report (IMF, 2014b) also concludes that African countries that had negative or no growth since 2000 are all fragile.


This is consistent with the case studies in the next section, which find that resilient countries such as Mozambique and Rwanda were able to significantly increase poverty-reducing expenditures.


Since the early 2000s, social safety nets have been developed in a number of countries (most notably in Cameroon, Mozambique, and Rwanda) with support from donors. Although the scale of these programs is not large, they constitute a promising tool for reducing poverty.


The PSI offers advice and supports policies in low-income countries that do not borrow from the IMF.


The New Deal for Engagement in Fragile States was adopted in 2011 and has been endorsed by 45 countries and organizations (AfDB, ADB, EU, OECD, United Nations, and the World Bank). It seeks to promote five peace-building and state-building goals, namely legitimate politics, security, justice, economic foundations (generating employment and improving livelihoods), and revenues and services (managing revenue and building systems for accountable and fair service delivery). The New Deal entails conducting country-specific fragility assessments and development compacts with dedicated donors, as well as an inclusive policy dialogue. The IMF’s work focuses on economic foundations and revenues and services, and it has been working closely with the New Deal parties to coordinate support.


The spectrum could be divided into five stages: crisis, rebuilding, transition, transformation, and resilience. See The Fragility Spectrum (2013).


Regarding the latter, the IMF has recently established topical or country-specific medium-term programs of capacity building and expanded its activities through regional technical assistance and training centers, especially in sub-Saharan Africa.

    Other Resources Citing This Publication