Chapter

2. Drivers of Growth in Nonresource-Rich Sub-Saharan African Countries

Author(s):
International Monetary Fund. African Dept.
Published Date:
October 2013
Share
  • ShareShare
Show Summary Details

Africa’s growth takeoff since the mid-1990s is more than a commodity story. Several countries have achieved sustained high growth rates without that growth being driven by the exploitation of natural resource wealth. Reviewing the experiences of the six nonresource-intensive low-income countries (LICs) that registered the highest growth rates over the period 1995–2010 reveals a number of common characteristics that accompanied this growth success. Many of these characteristics have been previously identified as critical in the growth literature and help generate a virtuous circle of growth: improved macroeconomic management, stronger institutions, increased aid, and higher investment in both physical and human capital. Given that these countries started their growth takeoffs with similar or worse initial conditions than the group of low-growth, nonresource-intensive LICs and even some of the countries currently classified as fragile, their success holds important lessons for the rest of the region.

Sub-Saharan Africa has grown strongly since the mid-1990s (Figure 2.1). There is a common perception that this growth has been the result of relatively high commodity prices, particularly for natural resources such as oil and minerals, generating both higher commodity revenues and attracting substantial new investment. Although growth in some countries in the region is heavily dependent on the export of natural resources, many nonresource-intensive LICs have also experienced rapid growth. In fact, 8 of the 12 fastest-growing economies in sub-Saharan Africa since 1995 were LICs considered nonresource-rich during this period, and as a group they have grown slightly faster than the oil exporters (Figure 2.2). More recently, some of these countries have made resource discoveries and will likely become resource producers in the near term.

Figure 2.1.Selected Regions: Real GDP Index

Source: IMF, World Economic Outlook database.

Figure 2.2.Sub-Saharan Africa: Real GDP Index

Source: IMF, World Economic Outlook database.

Note: Low-income countries excludes fragile countries.

This chapter builds on the many studies of growth takeoffs, such as IMF (2008), Commission on Growth and Development (2008), Salinas, Gueye, and Korbut (2011), and Bluedorn and others (2013). The approach taken is to focus on the growth paths followed by the six LICs in sub-Saharan Africa—Burkina Faso, Ethiopia, Mozambique, Rwanda, Tanzania, and Uganda—that experienced the fastest sustained growth during 1995–2010, although they were not natural resource producers during that time.

Although each country’s growth path has been unique, a number of common features emerge from these experiences.1 These common factors are, of course, interrelated—and no attempt is made here to try to establish causality. These factors are also identified in a panel regression analysis on growth in the nonresource-rich LICs (Annex 2.1). The main conclusions follow:

  • All countries managed to create a virtuous circle of sensible and medium-term-oriented policymaking and important structural reforms, which attracted higher aid flows. These flows, coupled with debt relief, created fiscal space enabling rapid expansion in both social spending and capital investment, such as infrastructure, boosting growth. In some cases this occurred following a civil war or a fundamental regime change.

  • Although structural transformation remains limited, agriculture was a contributing force in some countries, and offers much future potential. Services were also important; however, the manufacturing sector remains hampered by small internal markets and limited infrastructure.

  • Ahead, much progress will depend on closing the remaining infrastructure gap and increasing productivity, especially in agriculture, in which much of the population is employed. This would not only help to sustain growth, but also would distribute the benefits of growth more widely across the concerned populations.

The Overall Growth Experience in the Sample Countries

All sample countries—Burkina Faso, Ethiopia, Mozambique, Rwanda, Tanzania, and Uganda—have achieved strong and sustained growth since the mid-1990s, despite not having exploited natural resources on a large scale during this period. The countries were chosen on the basis of having experienced real output growth greater than 5 percent on average during the period 1995–2010, and real per capita GDP growth of more than 3 percent over the same period. Malawi is a borderline case, coming close to meeting these criteria. Burkina Faso and Tanzania have more recently started to produce gold, and Tanzania is now considered a resource-rich country. Mozambique is set to become an important coal, gas, and oil producer. Nevertheless, at the onset of their growth takeoffs in the early to mid-1990s, none of these countries was considered resource rich. However, the sample countries might have benefited indirectly from high commodity prices, as some, such as Mozambique, received large external inflows to finance investment in natural resource production.

At the time of their growth takeoffs, the sample countries had worse initial conditions—in some dimensions—than other similar countries in sub-Saharan Africa that have had weaker growth records during the past 15 years. In the early 1980s, Côte d’Ivoire and Zimbabwe were deemed to have some of the best economic prospects in the sub-Saharan African region, but these forecasts never materialized. Even in the mid-1990s, there were few indications that these six countries would be among the fastest-growing countries in the world a decade and a half later.

The accuracy of economic growth statistics for sub-Saharan African countries has been questioned. Owing to a lack of regular surveys and large informal sectors, national accounts data are sometimes derived from proxies for economic activity and qualitative assessments. As a result, some observers question their meaningfulness and adequacy for statistical analysis and policy conclusions. Alternative indicators, however, appear broadly consistent with high growth in the six sample countries, and indications suggest that the rebasing of GDP estimates currently taking place in many sub-Saharan Africa countries is likely to raise country estimates of the level of output on average (see Box 2.1 and Annex 1.1).

Turning to the salient features of growth in the six sample countries, this analysis finds significant variation in their growth experiences (Figure 2.3, Table 2.1). Each country, though, had a comprehensive policy vision and strategic framework:

Figure 2.3.Sub-Saharan Africa Sample Countries: Real GDP Index

Source: IMF, World Economic Outlook database.

Table 2.1.Sub-Saharan Africa Sample Countries: Real GDP and Real GDP per Capita Growth, Average 1995—2010(Percent)
Real GDP

Growth
Real GDP per Capita

Growth
Burkina Faso6.23.2
Ethiopia7.34.6
Mozambique8.15.9
Rwanda9.65.2
Tanzania5.83.3
Uganda7.44.0
Source: IMF, World Economic Outlook database.
  • Burkina Faso’s growth takeoff was aided by the devaluation of the CFA franc in 1994, together with an early focus on medium-term macroeconomic planning and improved management of the cotton sector.

  • Ethiopia, by far the most populous country in the sample, accelerated growth by actively supporting agriculture and certain export products and services (cut flowers, tourism, and air travel).

  • Mozambique attracted significant foreign direct investment (FDI) and other external capital flows in the late 1990s, which funded the capital-intensive mega-projects to produce and transmit electricity and gas, with the former used to produce aluminum.

  • Rwanda experienced a rebound effect after achieving political stability, underpinned by a national recovery strategy that successfully focused on specific sectors such as tourism and coffee.

  • Tanzania achieved sustained high growth by way of three well-sequenced waves of macroeconomic and structural reforms, which reached across all sectors.

  • Uganda started to carry out significant macroeconomic and structural reforms just before 1990, stimulated private investment, and launched a policy to diversify its export base to include nontraditional products.

Turning to the relative contributions of individual sectors to growth, the evidence from the sample countries suggests that services and, in some cases, agriculture were important driving forces (Figure 2.4).

Figure 2.4.Sub-Saharan Africa Sample Countries: Contributions of Sectors to Real GDP Growth, Average 1995–2010

Source: IMF, World Economic Outlook database.

Agricultural growth has provided a strong impulse to aggregate growth in Ethiopia and Rwanda, yet it has provided much less of an impulse in the other countries, especially Uganda. Services have grown rapidly among the sample countries, averaging between 8 percent and 10 percent per year in the fastest-growing countries of the sample. In Uganda, construction and services have contributed more than 60 percent to the growth rate.

Although manufacturing output grew in all countries, it started off from such a low base that its contribution to overall growth remained limited during the sample period (Table 2.2). Ethiopia had the lowest share of manufacturing in total GDP—at 5 percent—whereas the shares for Rwanda and Uganda at 8 percent and 7 percent, respectively, were comparable with the median for LICs in sub-Saharan Africa. The development of the manufacturing sector remains hampered by structural impediments such as high production input costs and limited infrastructure. In particular, limited and expensive electricity supply remains a major obstacle.

Table 2.2.Sub-Saharan Africa Sample Countries: Sectoral Growth in Real GDP(Percent)
AgricultureManufacturingConstruction
Share of total GDP 1995Growth 1995–2010Contribution to total growth 1995–2010Share of total GDP 1995Growth 1995–2010Contribution to total growth 1995–2010Share of total GDP 1995Growth 1995–2010Contribution to total growth 1995–2010
Burkina Faso34.65.301.6013.94.330.514.77.700.36
Ethiopia73.56.144.265.411.510.323.217.000.58
Mozambique40.27.022.326.512.261.192.819.890.69
Rwanda43.88.373.407.86.620.464.921.190.89
Tanzania133.04.321.269.07.860.735.79.420.59
Uganda47.21.180.227.19.080.707.116.611.62
Industry Excluding Manufacturing and ConstructionServices
Share of total

GDP1995
Growth

1995–2010
Contribution to

total growth

1995–2010
Share of

total GDP

1995
Growth

1995–2010
Contribution to

total growth

1995–2010
Burkina Faso1.414.640.3345.46.933.21
Ethiopia3.323.960.5514.59.101.59
Mozambique1.830.960.5548.76.763.00
Rwanda1.211.860.1442.310.124.22
Taizania14.48.310.4147.97.303.59
Uganda1.219.940.4237.211.825.35
Source: IMF, African Department database.

Despite sustained growth, structural transformation remains limited in the sample countries. As argued in Chapter 3 of the October 2012 Regional Economic Outlook: Sub-Saharan Africa, a basic definition of structural transformation is the movement of labor from sectors with low levels of labor productivity (generally agriculture) to sectors with above-average levels of productivity (industry and a subset of services). A new employment database, based on household surveys for about 30 sub-Saharan African economies, helps to analyze the degree of structural transformation in the sample of countries and compare them with other sub-Saharan African countries (Fox and others, 2013). This analysis suggests that the most significant labor reallocations have occurred in Rwanda and Tanzania, with sizable declines in the agricultural employment shares, combined with large increases in services employment (Table 2.3). In general, workers in the sample countries have moved out of agriculture mostly into the service sector, and this profile is similar to that of Nigeria, the region’s most populated country and its largest oil producer, as well as to other, more advanced economies on the continent.

Table 2.3.Sub-Saharan Africa Sample Countries: Annual Change in Relative Employment Shares through 2010(Percent)
AgricultureIndustryServices
Burkina Faso−0.20.00.2
Ethiopia−0.70.20.5
Mozambique0.40.1−0.5
Rwanda−1.30.31.1
Tanzania−1.50.41.2
Uganda−0.30.10.2
Nigeria−4.10.13.9
Sources: IMF, African Department database; and IMF staff calculations.Note: Initial period for Burkina Faso and Mozambique is 2003; Rwanda, 1995;Tanzania and Uganda, 2001; and Ethiopia and Nigeria, 2005.

Box 2.1.Is It Real? Alternative Benchmarks Evaluating GDP Estimates

In the majority of sub-Saharan African countries, economic growth statistics are characterized by considerable weaknesses. Recent work (Jerven, 2013; Young, 2012) has focused on the issue, and suggests that GDP estimates in many African countries lack accuracy and reliability. Limited capacity and financial means hamper the quality and coverage of national accounts data because surveys are not conducted in a regular and reliable fashion (Chapter 1, Annex 1.1). Large informal sectors, including the continued prevalence of subsistence agriculture, also pose a distinct obstacle for data accuracy. As a result, some observers question its meaningfulness for statistical analysis and policy conclusions.

A broader perspective on the growth takeoff in the sample countries can be obtained by comparing the GDP data with alternative quantity indicators from different sources collected in the sample countries (Figure 2.1.1). The six sectors chosen for the comparison include agriculture, construction, utilities, hotels and restaurants, transportation and communications, and financial services. These sectors constitute the lion’s share of output in the countries examined (Burkina Faso, Ethiopia, Mozambique, and Rwanda). The red bars in each of the sectors correspond to the growth rate from national accounts on a value-added basis, and the green bars correspond to physical output of some of the components of each sector. The components are not expected to match the sectoral aggregates because of the different basis—production versus value added—but we would expect to see a pattern of the components exceeding the national accounts data.

Figure 2.1.1.Sub-Saharan Africa: Comparison between Sectoral Growth Rates and Underlying Indicators, 2000–111

Source: IMF staff calculations based on data from the country authorities.

1 Data for Rwanda are 2000–12.

Comparing agricultural production data for major crops with national accounts estimates for the agricultural sector suggests that staple crops have grown much faster than the aggregate agricultural growth rate in Ethiopia and Rwanda, whereas maize has been very weak in Mozambique, and millet has been weak in Burkina Faso. For Burkina Faso, though, the overall sectoral growth rate does not seem out of line. The official agricultural growth rate is difficult to reconcile in Mozambique, because even cash crops have not grown very fast. For construction, the growth rate in cement consumption is similar to the construction growth rate in Mozambique and Rwanda, but considerably higher in Ethiopia. The estimates for utilities do not seem to be too high for any country. Ethiopia’s electricity production has been very strong in recent years, and it has also been fairly robust in Mozambique and Rwanda. In Rwanda, the strong growth rate in electricity production is not reflected in the utilities growth rate, because a large amount of imports have been used in the generation process, so the value-added component is low.

Growth in services as reported in national accounts data has also been in line with alternative indicators. Hotels and restaurants’ growth has generally followed the growth in tourist arrivals, but more use should be made of length of stay to benchmark the sectoral growth rate, because length of stay can vary significantly from arrivals. Currently, only Mozambique uses information on length of stay to benchmark this growth rate. In transport and communications, mobile telecom penetration has been very strong in all four countries, which helps explain why the national accounts growth rates for this sector are higher than the modest growth rates for road and rail transportation and cargo. Except for Ethiopia, real growth in domestic credit has been comparable to the estimate for growth in financial services, with real deposit growth more variable (in Burkina Faso and Rwanda, in particular).

Thus, overall, the underlying indicators appear broadly consistent with the high aggregate growth estimates for these countries.

This box was prepared by Alun Thomas.

Evidence From Country Cases

This section synthesizes the findings of the detailed background case studies from the six sample countries. A common theme emerging from this analysis is that macroeconomic stabilization and decisive structural reforms were crucial in laying the foundations for sustained growth. Several of the countries emerged from armed conflict or had been characterized earlier by African socialism and state-led development strategies stemming from post-independence periods. Therefore, countries had to address major macroeconomic imbalances, often with the help of IMF-supported programs.

Moreover, the six countries all saw higher aid flows, higher FDI, significant debt relief, and higher public capital expenditure than other nonresource-intensive LICs and fragile states (Figure 2.5).2 Although the nexus between investment and economic growth is not fully understood, this sustained high investment is likely to have supported high growth in the sample countries.3 This link is also confirmed by the results of the growth decomposition discussed in Box 2.2, and the econometric evidence presented in Annex 2.1.

Figure 2.5.Sub-Saharan Africa Sample Countries: The Virtuous Circle, Average 1995–2010

Sources: IMF, African Department database; and IMF, World Economic Outlook database.

Note: LIC = low-income country.

Against this background, the following sections detail a range of dimensions of the growth experience in the sample countries, highlighting the roles of the overall macroeconomic environment and structural reforms, governance and institutions, human capital formation, and the financial sector; and the roles of aid, investment, and driving forces at the sectoral level.

Box 2.2.The Role of Total Factor Productivity in the Growth Acceleration of Nonresource-Rich Low-Income Countries

Growth can come from two sources: using more factors of production or inputs (labor and capital) to increase the number of goods and services that an economy is able to produce, or combining inputs more efficiently to produce more output for a given amount of inputs. Decomposing growth into these two sources yields insights into the proximate causes of growth.

This box presents the results of a human-capital augmented growth accounting exercise for the six nonresource-intensive low-income countries. It is based on the standard decomposition proposed by Solow (1957), which decomposes the growth rates of real GDP into a fraction attributed to the growth rates of factor inputs and a residual, commonly labeled total factor productivity (TFP).1 TFP should be broadly understood to encompass the use of new, more efficient technologies, as well as any improvements in the way inputs are combined, which can include improvements in governance, regulatory quality, business climate, and so on.

Given that educational attainment in the sub-Saharan Africa region has increased significantly during the past two decades as a result of the reallocation of resources toward public education, it is important to take into account the improvement in the skills of the labor force. Otherwise, improvements in the quality of the labor force would be incorrectly attributed to increases in TFP. To do so, the accounting exercise adjusts the quality of the labor input using data on educational attainment from the Barro and Lee (2012) data set.2

Table 2.2.1 shows the results of the decomposition exercise. The first column shows the growth rate of real GDP; the next three columns show the contributions of capital, labor, and education, respectively, to growth, and the last column shows the growth rate of TFP. As can be seen, the growth rate of real GDP accelerated in the 1990s relative to the 1980s (except in Tanzania and Rwanda, where it remained nearly constant), and in all cases except for Uganda it also accelerated in the 2000s relative to the 1990s.

Table 2.2.1.Sub-Saharan AfricaSampleCountries: Output Growth Decomposition(Contribution to annual growth rates, percent)
CountryPeriodReal GDPCapital StockAdjusted LaborEducationTotal Factor Productivity
Burkina Faso1980–902.32.61.00.6−1.8
1990–20004.61.71.30.61.0
2000–105.51.92.80.40.4
Ethiopia1980–902.21.70.80.7−1.1
1990–20002.90.71.00.40.9
2000–108.12.42.10.33.4
Mozambique1980–90−0.10.70.6−0.1−1.4
1990–20005.42.33.90.0−0.8
2000–107.92.41.90.43.2
Rwanda1980–901.82.80.90.2−2.1
1990–20001.70.5−2.20.33.1
2000–107.72.42.30.22.8
Tanzania1980–903.5−0.21.80.11.6
1990–20003.31.21.40.10.6
2000–106.82.51.60.22.6
Uganda1980–903.10.81.50.30.5
1990–20007.22.11.20.23.7
2000–106.43.11.70.21.4
Sources: IMF staff calculations based on Penn World Table, Version 8; the Barro and Lee (2012) education attainment data set; and country household surveys.

Moreover, in almost all countries, TFP accelerated between the 1980s and the 1990s, suggesting a rebound growth effect. This acceleration is also likely to reflect innovation and structural changes, which supported higher real GDP growth for Ethiopia, Mozambique, and Tanzania through the 2000s.

The exercise also shows a higher contribution of capital accumulation in the 2000s in all countries, which is consistent with the fact that investment rates in the sample countries increased significantly, even when compared with the group of nonresource-rich low-income and fragile countries. The contributions of capital accumulation are not uniform across countries: physical capital contributed to the acceleration of growth of real GDP during the 1990s in all cases, but only strongly in Burkina Faso, Mozambique, and Uganda.

In sum, the growth acceleration among nonresource-intensive low-income countries has been driven by higher TFP growth and faster factor accumulation. To the extent that the labor force is expected to continue growing in most countries in the region, and that relative to other fast-growing countries the investment rate is still low in many countries (Ethiopia, Mozambique, Rwanda, Uganda), productive inputs could continue to be sources of growth in the future.

This box was prepared by Rodrigo Garcia-Verdu and Alun Thomas.1 To construct the time series for capital, this exercise uses data on investment rate from the Penn World Table (PWT), Version 8, and applies the perpetual inventory method, according to which capital is built by assuming a value for the initial stock of capital and adding up investment and subtracting a constant depreciation in subsequent years. For labor, the exercise uses the recent estimates by the IMF and the World Bank of the number of workers constructed from household surveys from the last two decades for each of the six countries. Given that almost no household survey data are available for the 1980s, these are taken from the implicit PWT series.2 In particular, the exercise uses the production function proposed by Bosworth and Collins (2002), in which labor (L), measured by number of workers, is multiplied by a measure of human capital (H), which in turn is obtained by combining an estimate of the rate of return to schooling (assumed to be 14 percent per year) with the average years of schooling for the population ages 25 and older from the Barro and Lee (2012) data set.

Macroeconomic policymaking and structural reforms

All sample countries were able to carry out comprehensive macroeconomic reforms and conduct policies geared toward stability throughout the 1990s. In this fashion they often turned around an economy that had been characterized by stop-and-go cycles. A prime example is Tanzania, which went through three phases of comprehensive liberalization reforms (Nord and others, 2009; and Robinson, Gaertner, and Papageorgiou, 2011). After dismantling state control of the economy and taking the first steps toward a market economy, the country liberalized its exchange rate and its trade and agricultural marketing regimes, and also conducted significant reforms of the domestic financial system through the mid-1990s. These reforms began to bear fruit, though only after 1996, when the country also privatized some of the remaining parastatals and pursued fiscal consolidation and a more independent monetary policy. Growth accelerated markedly from the mid-1990s onward, and remained sustained despite external shocks in the form of fluctuating international commodity prices and plummeting global demand, as well as domestic shocks such as droughts. Burkina Faso maintained macroeconomic stability through a period of external and domestic shocks, such as international food and energy price shocks, climatic shocks, and political unrest in neighboring Côte d’Ivoire, reflecting responsible policymaking and political stability.

Most of the sample countries rebuilt their nations after ending violent conflicts. These countries had experienced armed conflicts and unrest in the 1980s or early 1990s, and thus transformed their economies starting from a low base with damaged infrastructure into growing and well-managed countries ahead of their peers. In the early 1990s, Mozambique launched a first wave of structural reforms, and started attracting significant donor flows and FDI that were channeled into infrastructure investment and mega-projects.

In Uganda, well-sequenced macroeconomic reforms in a postconflict environment focused on loosening price controls, liberalizing and eventually floating the exchange rate, reforming the civil service, and privatizing parastatals.

Because the unstable years before the turnaround in the mid-1990s were characterized by the destruction of physical capital, low or no growth, or declines in output per capita, some of the growth takeoff can be attributed to a rebound effect (Box 2.2). However, the sample countries have experienced sustained and accelerating growth for more than 15 years, suggesting structural shifts in their economies that transcend a rebound effect. Furthermore, other nonresource-rich countries that also experienced civil war have not yet managed to attain high sustained growth (for example, Burundi and Eritrea).

Another important factor in promoting macroeconomic stability was debt relief. Most sample countries qualified for debt relief under the Heavily Indebted Poor Countries (HIPC) Initiative and the Multilateral Debt Relief Initiative (MDRI) in the early 2000s. As a result, their debt burdens were reduced significantly, freeing up considerable fiscal space that countries were encouraged to use for poverty-reducing expenditure, bringing about a shift in the composition of expenditure toward the education and health sectors. The extra fiscal space also enabled the building of fiscal buffers, which were critical in enabling countries to pursue countercyclical fiscal policies during the global financial crisis, and to adopt a more medium-term strategic view of expenditure policies.

Better macroeconomic management has been accompanied by accelerated growth, lower inflation, narrower fiscal deficits, and higher international reserves (Table 2.4). By contrast, current account deficits widened in some countries, reflecting the high level of investment they carried out during the sample period, often financed by both aid and FDI.

Table 2.4.Sub-Saharan Africa Sample Countries: Economic Indicators
1980–941995–2010
(Percent change)
Real GDP Growth
Burkina Faso3.36.2
Ethiopia1.77.3
Mozambique1.78.1
Rwanda−2.09.6
Tanzania2.85.8
Uganda3.57.4
Group average1.87.4
Group weighted average2.47.0
Inflation, period average
Burkina Faso5.03.1
Ethiopia7.48.8
Mozambique43.014.2
Rwanda9.510.1
Tanzania29.79.6
Uganda92.16.3
Group average31.18.7
Group weighted average27.88.5
(Percent of GDP)
Fiscal Balance
Burkina Faso−2.2−2.7
Ethiopia−4.4−4.0
Mozambique−6.7−3.3
Rwanda−5.8−1.4
Tanzania−4.6−2.4
Uganda−1.9
Group average−4.7−2.6
Group weighted average−4.4−2.9
Current Account Balance
Burkina Faso−3.0−8.9
Ethiopia−2.0−3.5
Mozambique−13.3−13.3
Rwanda−7.2−3.9
Tanzania−5.1−6.7
Uganda−2.9−6.0
Group average−5.6−7.1
Group weighted average−4.3−6.5
(Months of imports)
Reserves Coverage
Burkina Faso4.35.2
Ethiopia1.93.2
Mozambique2.24.4
Rwanda3.75.0
Tanzania0.74.7
Uganda1.36.1
Group average2.44.8
Group weighted average1.84.6
Source: IMF, World Economic Outlook database.Note: … = not available.

Governance and institutions

Governance and institutional quality can be measured along numerous dimensions. The sample countries registered improvements in governance relative to the comparator group in four out of the five dimensions considered in the World Bank’s Worldwide Governance Indicators: control of corruption, government effectiveness, political stability and absence of violence, and regulatory quality (Figure 2.6). This analysis suggests that improvements registered in these dimensions of governance and more effective institutions relative to the comparator group could be among the driving forces of the higher growth rates registered among the case study countries. Indeed, the panel regression (see Annex 2.1), as well as earlier empirical work (Salinas, Gueye, and Korbut, 2011), confirms the significance of such institutional factors.4 All sample countries also benefited from extensive technical assistance directed at improving the quality of their institutions.

Figure 2.6.Sub-Saharan Africa: Worldwide Governance Indicators

(Higher values correspond to better governance)

Sources: IMF staff estimates based on World Bank, Worldwide Governance Indicators.

Note: LIC signifies low-income country. The composite measures of governance of the Worldwide Governance Indicators are expressed in units of a standard normal distribution, with mean zero, standard deviation of one, and running from approximately −2.5 to 2.5, with higher values corresponding to better governance.

Decisive structural reforms in the sample countries and efforts to upgrade institutional capacity were accompanied by a better business environment (Table 2.5). Various measures of business regulations from the World Bank’s Doing Business database show that the group of high-growth, nonresource-abundant LICs does indeed have better regulations than the average for the comparator group, based on most of the indicators. For reference, the averages for all of sub-Saharan Africa and middle-income countries are also presented in Table 2.5.

Table 2.5.Sub-Saharan Africa: Cost of Doing Business Indicators
Ease of Doing Business IndexCost to Build a WarehouseCost to Enforce a ContractCost to Get ElectricityCost to Register PropertyCost to Start a BusinessStrength of Investor Protection Index
(1 = easiest; 185 = most difficult)(Percent of income per capita)(Percent of claim)(Percent of income per capita)(Percent of property value)(Percent of income per capita)(0 = no protection; 10 = most protection)
2011–122009–122009–122009–122009–122009–122009–12
Sample countries120.4500.162.95,5495.765.54.9
Other nonresource-intensive LICs and fragile countries158.7894.149.98,0249.6105.44.1
Middle-Income countries89.8499.229.51,5498.220.45.3
Sub-Saharan Africa139.0998.051.35,7699.993.54.5
Credit: Strength of Legal Rights IndexDepth of Credit Information IndexTotal Tax Rate1Labor Tax and Contributions2Minimum Paid-in Capital Required to Start a BusinessTrade: Cost to ExportTrade: Cost to Import
(0 = weak; 10 = strong)(0 = low; 6 = high)(Percent of profit)(Percent)(Percent of income per capita)(U.S. dolars per container)(U.S. dollars per container)
2009–122009–122009–1220122009–122009–122009–12
Sample countries5.42.537.110.7125.62,1822,961
Other nonresource-intensive LICs and fragile countries4.50.886.413.0169.81,4771,846
Middle-income countries6.53.229.98.023.41,5521,810
Sub-Saharan Africa5.31.864.413.1130.61,9642,495

Human capital formation

Some of the sample countries have taken advantage of their expanding fiscal space to invest in education, as reflected in both the amount of resources allocated to education (Figure 2.7) and the increase in the average number of years of schooling per individual. The East African countries stand out as having the most education, with the average years of schooling at between five and six for Rwanda, Tanzania, and Uganda, compared with slightly more than one year of schooling for Burkina Faso.

Figure 2.7.Sub-Saharan Africa Sample Countries: Expenditure on Education

Sources: IMF, African Department database; and IMF, Fiscal Affairs Department, Public Spending on Health and Education database.

Note: LIC = low-income country.

Financial sector development

The financial sector experience in the six sample countries has been mixed. All countries—although to different degrees—have made progress in deepening their financial sectors in the past decade, including modernizing regulatory and supervisory frameworks, restructuring banking systems, and developing microfinance institutions. Banking systems have steadily expanded and remained well capitalized in general, accompanied by significant increases in credit to the private sector as a percentage of GDP, except in Ethiopia (Figure 2.8).5 However, except for Mozambique, the levels of intermediation by commercial banks have remained well below the levels of countries with similar economic structures, and access to financial services has been relatively limited (Figure 2.9).

Figure 2.8.Sub-Saharan Africa Sample Countries: Credit to the Private Sector

Sources: IMF, African Department database; and IMF, Fiscal Affairs Department, Public Spending on Health and Education database.

Note: LIC = low-income country.

Figure 2.9.Sub-Saharan Africa Sample Countries: Adults with an Account at a Formal Financial Institution, 2011

Sources: IMF, World Economic Outlook database; and World Bank, Global Financial Development database.

Note: LIC = low-income country.

Aid

The improved macroeconomic management attracted significant donor financing to the sample countries. Grants through the central government were equivalent, on average, to 6.4 percent of GDP over 1995–2010 for the group, compared with a sub-Saharan Africa nonresource-intensive low-income and fragile country average of 2.1 percent of GDP. Program and project loans added a further 3.5 percent of GDP of donor financing to the fast growers, compared with 1.5 percent of GDP for the comparator group. This aid financed a relatively large share of expenditure in the sample countries, compared with other nonresource-intensive LICs (Figure 2.10).

Figure 2.10.Sub-Saharan Africa Sample Countries: Donor Financing, Average 1995–2010

IMF, African Department database; and IMF, World Economic Outlook database.

Note: LIC = low-income country.

Ethiopia was the lowest recipient of donor flows at 7 percent of GDP, still more than double the level of the comparator groups. Mozambique was the highest at 14.5 percent of GDP, whereas Burkina Faso and Rwanda ranged between 11.5 and 12.5 percent. As the regression analysis presented in Annex 2.1 shows, the aid-to-GDP ratio was significantly associated with growth in the sample countries.

Higher levels of aid and changes in the method of aid delivery have placed an increased focus on achieving better donor coordination to enhance the predictability of aid flows both within a budget cycle and in a multiyear planning environment. Many countries have introduced a continued and structured dialogue between donors and governments designed to facilitate the connection of aid delivery to national poverty-reduction strategies, but this has also come with considerable transaction costs in some cases for both donors and governments.

Increased levels of aid may pose challenges to macroeconomic management stemming from volatility of inflows and the ability of economies to absorb large volumes of aid. This can lead to overheating, exchange rate appreciation, and loss of competitiveness (see IMF, 2013b). Although the sample countries may appear vulnerable to these adverse (or Dutch disease) effects from aid—rigidities in the supply of nontradables owing to scarcity of skilled workers, limited access to ports, and infrastructure bottlenecks (four out of the six are landlocked)—there is little evidence of major misalignments of the real exchange rate in the sample countries. In some cases, such as Tanzania, the closing of macroeconomic imbalances and the adoption of market-oriented reforms resulted in the erasure of large exchange rate misalignments (Figure 2.11). The ability of the sample countries to manage the increase in aid may reflect improved use of donor financing and the fact that most countries have opened up their construction sectors to foreign companies, particularly, but not just, from China, thereby reducing a potentially binding capacity constraint. However, for Uganda, Atingi-Ego (2006) finds some real exchange rate appreciation pressures in 2003–04, reflecting donor-funded increases in government spending.

Figure 2.11.Sub-Saharan Africa: Misalignment with Equilibrium Exchange Rate

Source: Sub-Saharan Africa: Regional Economic Outlook, April 2011.

Investment

Investment rates outpaced the region’s upward trend. Excluding Burkina Faso, where investment lagged relative to the rest of the group, the sample group’s investment rate averaged about 22 percent of GDP in 2000–10, compared with an 18 percent average for low-income and fragile countries in sub-Saharan Africa (Figure 2.12). This rate increased by 4 percent of GDP between 2000 and 2010, reaching 25 percent. In the sample countries, financial intermediation and the level of domestic savings remain relatively low, even though the mobilization of domestic resources increased in the sample period.

Figure 2.12.Sub-Saharan Africa Sample Countries: Saving and Investment, 1995–2010

Sources: IMF, African Department database.

Note: LIC = low-income country.

Early and determined efforts by the sample countries to strengthen public financial management resulted in improvements in capital expenditure planning and execution that helped sustain public investment and channel aid flows toward infrastructure (Dabla-Norris and others, 2011).6 Burkina Faso was one of the first LICs in sub-Saharan Africa to introduce medium-term planning tools such as rolling medium-term budgetary and expenditure frameworks. Ethiopia and Rwanda also conducted important public financial management reforms, placing fiscal policy into a medium-term budgetary framework. These measures helped to link budgetary allocations with policy priorities as identified in countries’ poverty-reduction strategies. Finally, all countries are characterized by advanced public investment management in project selection and implementation.

Moreover, enhanced donor coordination protected sample countries from aid shortfalls within a given budget cycle, thus avoiding cutbacks in public domestic investment, which is often one of the few discretionary expenditure items (Figure 2.13). Celasun and Walliser (2006) find that shortfalls in budget aid typically lead to cutbacks in domestically financed investment spending. They contrast the experience of Burkina Faso, where enhanced donor coordination led to more predictable aid flows and a smoother budget execution, with that of Mali, where donors were less coordinated and aid disbursements more erratic, hampering budget execution (although the situation has now changed see Annex 1.1).

Figure 2.13.Sub-Saharan Africa Sample Countries: Public Capital Expenditure, 1995–2010

Source: IMF, African Department database.

Note: LIC = low-income country.

Uganda relied more on the private sector to boost its investment rate (Figure 2.14). Uganda’s private gross capital formation is estimated to be the strongest in the sample for 2000–10, reaching 18 percent of GDP, more than 4 percent of GDP above Mozambique, which had the second highest private investment rate by 2010. Although all sample countries managed to establish business-friendly environments, Uganda’s business environment compares favorably with the rest of the group. The country also undertook a significant privatization program of public enterprises in the 1990s, and liberalized markets and lending rates.

Figure 2.14.Uganda: Private Investment, 1996–2010

Source: IMF, African Department database.

Note: LIC = low-income country.

Overall, however, infrastructure investment and rehabilitation needs remain massive in all sample countries. Poorly targeted energy subsidies often keep energy prices artificially low, hampering investment in the sector. At the same time, limited energy supply and frequent blackouts severely impede the development of a manufacturing sector, and create significant extra costs for existing businesses (Figure 2.15)(World Bank, 2011). Poor port governance and difficult trade logistics are also important obstacles. Even in Mozambique, which focused on priority districts and corridors (Box 2.3), the connectivity between urban and economic clusters remains poor—corridors link these clusters to ports but not to one another. In Tanzania, coverage of electricity and road networks is also low, even by regional standards (Ter-Minassian, Hughes, and Hajdenberg, 2008). In Rwanda, transport costs remain high, hampering even sectors that have recently been transformed and contribute to poverty reduction, such as the coffee sector (World Bank, 2011). Similarly, some sample countries, in particular Burkina Faso, have some of the most underdeveloped power systems in sub-Saharan Africa, with insufficient electricity production to meet increasing demand. However, this situation is projected to change in Ethiopia with the carrying out of massive new investment to take advantage of the low generation costs of hydropower.

Figure 2.15.Sub-Saharan Africa Sample Countries: Infrastructure Indicators

Sources: World Bank, World Development Indicators; and World Economic Forum.

Driving forces of growth at the sectoral level

Agriculture remains an important sector in the sample countries, for its contribution to output and the large share of the labor force that is concentrated in this sector. As a result, even modest growth rates may have a significant impact on overall growth, because the sector accounts for a high share of GDP (Table 2.2). Ethiopia, and Rwanda saw productivity and output gains in the sector above the average for sub-Saharan Africa in 1995–2010, whereas in Uganda, agriculture growth was low in relative terms (Figure 2.16).

Figure 2.16.Sub-Saharan Africa Sample Countries: Share of Agriculture in GDP

Source: IMF, African Department database.

In Uganda, this is mostly attributed to slow productivity growth in the sector, owing to already elevated cereal yields by regional standards at the beginning of the period. In other countries, in particular Burkina Faso, agricultural production is hampered by the increasing frequency of weather shocks, land degradation, and desertification. Overall, agricultural growth was more volatile than growth in other sectors, reflecting the continued dependence on weather conditions, and a lack of irrigation systems and advanced farming practices.

Sources of agricultural growth varied across the group and over time. Burkina Faso relied mainly on increases in factors of production, particularly the farming area for cotton, with limited gains in productivity. However, it was also the first country to experiment successfully with genetically modified cotton seeds to increase cotton yields dramatically, albeit toward the end of the sample period. Although Rwanda initially also relied on factor growth, it later achieved significant improvements in yields and productivity (Box 2.4).

Agriculture in Ethiopia and Mozambique, however, relied equally on improved yields and increased farming area. Uganda had the least increase in arable land within the group, which together with near constant yields and a declining share of labor in agriculture contributed to the relatively weak performance in the sector.

Where they occurred, increases in agricultural productivity in the sample countries were supported by public policy, either through structural reforms or more direct interventions. In Burkina Faso, decisive and well-sequenced market reforms helped expand the cotton sector, which was the main driver of growth in agriculture. The introduction of a stable producer price throughout the planting season, backed by the cotton parastatal, created the right incentives for farmers to cultivate cotton. Other important steps were enhancing the role of producer organizations, transferring some economic functions to these groups, and allowing the entry of private operators into the sector (Table 2.6). These reforms led to a significant increase in cotton production in the mid-2000s.7 In Ethiopia and Rwanda, national programs to improve or subsidize access to fertilizer and offer extension services to poor farmers helped boost yields (Box 2.4).

Table 2.6.Burkina Faso: Cotton Reforms, 1992–2008
YearDevelopment
1992Formal commitment made by Societé Burkinabé des Fibres et des Textiles (SOFITEX), the national cotton parastatal company, to let producers’ representatives participate in reform debate.
1999State partially withdraws from the sector through partial privatization of SOFITEX.
2000–06Economic activities—including provision of cereal input credit; management assistance of cotton groups; and participation in quality grading, financial management, and price bargaining—progressively delegated from SOFITEX and the government to the cotton union.
2002–06New players—including private input providers, new regional private cotton monopsonies (SOCOMA, FASOCOTON), and private transport companies—begin operating in the sector.
2006–08Price-setting mechanism changed to better reflect world price levels; new price-smoothing fund managed by an independent organization becomes operational in 2008.

Box 2.3.Investment in Mozambique

Mozambique managed to attract the highest foreign direct investment (FDI) flows among the sample countries, reaching 14 percent of GDP in 2010. After stabilizing the macroeconomy and launching the first wave of structural reforms in the early 1990s, Mozambique started attracting FDI, which was channeled into mega-projects and infrastructure development.1 The World Bank (2011) estimates that during the late 2000s, Mozambique spent, on average, 10 percent of GDP a year on infrastructure. A prime example is the Mozal aluminum smelter, which is the key anchor project of the Maputo Development Corridor, linking the port of Maputo with major South African cities. This project is considered a success and is estimated to have supplied about 7 percent of total global aluminum production in 2005 (Economic Commission for Africa, 2004). To attract FDI, the authorities launched a wide range of tax incentives for foreign investors, and also established industrial zones. They also established several specific investment promotion institutions, and issued investment guarantees for foreign investors.

Mozambique was also one of the first low-income countries in sub-Saharan Africa to make use of public-private partnerships (PPPs) to develop its infrastructure. These arrangements offer opportunities to share the costs and risks associated with big infrastructure projects between the public and the private sectors, provided that they are properly structured and carried out to increase efficiency. Moreover, the private operator can contribute significant technical know-how and better execution skills than governments typically have. After some initial difficulties with individual PPPs, related to lower-than-projected user volumes, the authorities used PPPs successfully to operate major railway lines and to improve the performance of ports. Six of Mozambique’s seven seaports are operating with the involvement of the private sector. The country also developed its road network with the help of concessions, but the quality of the road system remains poor, except for the toll road between Maputo and South Africa.

This box was prepared by Isabell Adenauer.1 FDI in Mozambique increased from an average of 1.5 percent of GDP in 1993–98 to 5.4 percent of GDP in 1995–2010 (World Bank, 2011; and IMF staff estimates).

Box 2.4.Examples of National Programs Aimed at Raising Yields

In Ethiopia and Rwanda, the government played an active role in raising agricultural yields, including by providing extension services to small farmers and intervention in the fertilizer market (Figure 2.4.1).

Figure 2.4.1.Sub-Saharan Africa: Cereal Yields, 1995–2010

Source: World Bank, World Development Indicators.

Note: LIC = low-income country.

In 1994, Ethiopia undertook a national program to promote the use of improved seeds and fertilizer. The National Agricultural Extension Intervention Program (NAEIP) provided integrated access to seeds, fertilizer, and credit to nearly 40 percent of farming households over a 10-year period. As a result, the use of improved seeds and fertilizer increased by nearly 50 percent and 30 percent, respectively, between 1998 and 2008 (Speilman, Kelemwork, and Alemu, 2011). However, while NAEIP was initially successful, gains in yield have been volatile, and the scale up fell short of expectations for increasing yields (the target was a doubling of yields). This was caused by inefficiencies stemming from government crowding out of the private sector via the extension services. The challenge ahead is to provide agricultural inputs in a more targeted manner, while motivating the private sector to take over the role of intermediary, to improve efficiency and limit costs for the government.

Rwanda has enjoyed significant growth in the agricultural sector since 2006, driven by higher crop yields and increased cultivation of arable lands. This was partly due to a comprehensive reform package, the Crop Intensification Program (CIP), focusing on crop regionalization (matching eligible crops for subsidy to agro-ecological conditions across regions), increasing use of fertilizer, land consolidation and protecting arable lands from erosion, as well as providing more productive seed varieties and extension services. As a result, yields have increased significantly, from being among the lowest to among the highest in sub-Saharan Africa. Unlike other fertilizer programs, the CIP had an exit strategy regarding fertilizer subsidies, and the subsidy was gradually phased out. Over the course of the program, fertilizer use increased 300 percent (Druilhe and Barreiro-Hurlé, 2012).

This box was prepared by Kareem Ismail.

Most of the labor force in the sample countries remains concentrated in the agricultural sector, and more significant productivity increases will be crucial to promoting inclusive growth and lifting more people out of poverty. It is estimated that the bulk of the sample countries’ active labor force is engaged in the agricultural sector—about 80–81 percent in Mozambique and Burkina Faso, 71 percent in Uganda, and 65 percent in Tanzania (2010 estimates) (Fox and others, 2013). Moreover, most of the extremely poor rely on subsistence farming for their livelihoods. The limited increases in labor productivity and in total factor productivity of the agricultural sector in the sample countries discussed above therefore hampered more inclusive growth and faster reduction of poverty. Important structural obstacles include limited access to markets because of the lack of roads; limited crop research; lack of access to quality inputs such as seeds, fertilizer, and finance; and lack of property rights, which discourages farmers from investing in and planting on significant parts of available land.

However, all countries have significant potential to increase their agricultural production in the future. The two coastal sample countries—Mozambique and Tanzania—have the greatest potential, because they have large areas of uncultivated land, coupled with relatively low cereal yields.

Where agricultural growth was driven by performance of cereal crops, growth proved more inclusive, and rural poverty declined more significantly. Growth in agriculture for Burkina Faso and Mozambique was primarily driven by cash crops such as cotton, sugar, and tobacco, whereas growth in Ethiopia and Rwanda was driven by increased production of staples. Growth in cash crops brought dividends from exports, whereas growth in staples helped support domestic food supply and lower poverty. Despite similar average growth rates in the agricultural sectors, rural poverty in Burkina Faso and Mozambique remained high, at about 50 percent and 70 percent, respectively, whereas it declined in Ethiopia by 15 percent during 1995–2010 to 30 percent; and in Rwanda, by 15 percent during 2006–11 to 49 percent.8

The service sector has been the strongest component of growth in the sample countries except in Ethiopia. A further decomposition shows that this contribution has been fairly equally shared between the various subcategories of services (Figure 2.17). Public administration and trade have made the largest contributions to the overall services sector, which might reflect higher government wages due to hiring in health and education, and estimates for the large informal sectors, in which trade plays a role. Growth in finance and insurance contributed the least, except in Mozambique, which may be related to the very large FDI flows that the country has experienced. Transportation and telecom has also made a large contribution—particularly the mobile phone market, which has grown exponentially in all countries and supported overall growth (Annex 2.1).

Figure 2.17.Sub-Saharan Africa Sample Countries: Contributions of Sectors to Services Growth, Average 2000–11

Source: IMF, African Department database.

Concluding Remarks and Outlook

The country cases presented in this chapter highlight both the commonalities and the differences in the paths taken by the six fastest-growing nonresource-rich sub-Saharan African economies. The experience demonstrates that a shift in macroeconomic policymaking, combined with comprehensive structural reforms and sustained external financing, can create the fiscal space to finance productive investment and generate a growth dividend. These components are clearly interconnected and have been present in all these successes, but each of these economies reached this virtuous circle in somewhat different fashions.

However, some general observations can be drawn from the analysis of the six country cases:

Macroeconomic stability. All countries in the sample show that improved macroeconomic policies, structural reforms, and a greater degree of overall stability are crucial for economic growth. Greater fiscal space—reflecting expenditure prioritization and a combination of debt relief, aid, and greater domestic resource mobilization—can translate into higher levels of social and investment spending, supporting growth in both the near and medium terms. Lower inflation and a more predictable economic environment reduce risks and transaction costs, encouraging private sector activity. Improved macroeconomic policies also help to attract foreign financing, in the form of both foreign aid and private resources, such as FDI.

Effective use of foreign aid. The increased level and predictability of foreign aid enabled a greater focus on medium-term planning and greater opportunities for alignment with national poverty-reduction strategies.

Strong national policymaking institutions. Identifying and translating increased fiscal space into sustainable growth requires effective macro policymaking and strong institutions, in particular regarding public financial management capacity.

High investment levels. Sustained high investment levels are crucial to increasing the capital stock and boosting overall productivity levels. Investment in infrastructure is particularly important because it promotes private sector development by lowering overall business costs. The lack of an adequate transport infrastructure can bar large segments of the population from access to markets, particularly in rural areas, where poverty is concentrated.

Deeper financial markets. Deeper financial markets support growth by enhancing domestic savings to finance investment. Financial inclusion has tended to lag, and could—including through use of new technologies—be prioritized to make the growth process more inclusive.

Initial conditions are not so important. The majority of the countries in the sample are landlocked, and many were just emerging from conflict situations in the early 1990s. Postconflict and other fragile states can move onto rapid growth paths once a clear vision is formulated and consistently implemented.

Looking ahead, the question is whether the robust growth achieved in these six countries is sustainable. Even after a period of high growth, the six sample countries are still far from attaining their full potential—productivity and capital stocks remain relatively low; significant infrastructure gaps remain; exports are concentrated in a small number of products; cross-border trade within sub-Saharan Africa remains low; and there has been limited progress on the structural transformation of the economies. Thus, a large agenda remains, providing significant new growth opportunities.

Annex 2.1. Quantitative Assessment of Factors Influencing the Growth Surge

To assess the importance of the factors identified in the case studies in a quantifiable manner, this annex discusses a basic econometric specification of the determinants of the growth of real GDP per capita for low-income countries that are identified as nonresource-intensive (Table 2.7). The sample comprises countries in the low- and lower-middle-income country group as defined by the level of income per capita in 2010 (World Bank definition), for which natural resource exports account for less than 20 percent of total exports. Small states are excluded. This covers about 40 countries, 20 of which are in sub-Saharan Africa, and includes the six countries that are the basis of this chapter. The sample period is 1990–2011.

Table 2.7.Determinants of Real per Capita Output Growth
Dependent variable: Per capita real output growth
[1][2][3][4][5][6][7]
Initial output per capita−0.0030.001−0.003−0.006*−0.004−0.006*−0.008**
Initial agriculture output per capita−0.045***−0.043***−0.038***−0.039***−0.043***−0.039***−0.042***
Real exchange rate (log)−0.014*−0.011−0.009−0.015*−0.013*−0.014*−0.006
Schooling (log)0.0050.0030.007*0.008*0.0040.008*0.006
Public investment rate0.046
Investment rate0.049***−0.010.0230.0230.042**0.024−0.02
Aid ratio0.0310.082***0.066**0.061*0.0350.060*0.102***
FDI ratio−0.035−0.136−0.152**
External debt ratio−0.005***−0.01*−0.012**−0.014***−0.005***−0.014**−0.013**
Real manufacturing output growth (Mozambique)0.25***0.279***0.275***0.192***
Cell phone subscriptions0.0020.002
Government effectiveness0.013**
Number of observations708331564591737536398
R-squared0.220.240.270.50.460.340.27
Source: IMF staff calculations.Note: *, **, and *** indicate significanceat the 10 percent, 5 percent, and 1 percent levels, respectively.FDI = foreign direct investment.

The case studies identified a few characteristics common to each of the six countries that stimulated their recent growth rates. These include initial conditions, as measured by GDP per capita or agricultural output per capita, the level of schooling, the investment rate, foreign aid, and mega-projects (in particular for Mozambique).

The results, based on ordinary least squares regression, show an inverse relationship between the growth of real GDP per capita and the initial levels of agricultural productivity so that countries that began the period with low levels of agricultural productivity had faster growth rates (similar results hold when the random effects model is used). This effect comes from both strong agricultural performance during the two decades (convergence in the agricultural sector is faster than for the whole economy), and from spillovers between the agricultural sector and other sectors. This finding may appear surprising, given the small sample of countries with similar levels of GDP per capita.

However, in 1990, a few countries had relatively high levels of agricultural output per capita, but their growth prospects have faltered since (Burundi, Haiti, Madagascar, and Moldova). Moreover, a number of countries that have grown rapidly subsequently had relatively low initial levels of agricultural productivity (Cambodia, Egypt, Ethiopia, Ghana, and Malawi).

The investment rate, especially public investment, appears to be a driving force that propels output in most low-income countries, although aid also plays a determining role. Significant other determinants of per capita growth include the real exchange rate and the external debt ratio. Perhaps surprisingly, FDI does not appear to have played a determining role in the growth surge.

The manufacturing sector is an important determinant of output growth in Mozambique, dominated by mega-projects, such as the Mozal aluminum plant, and is consistent with the view that the manufacturing sector plays an important role in the growth performance of some countries (Figure 2.18, which shows regression specification 6). The very large effect for Mozambique from manufacturing is partly a catch-all effect because the other variables cannot fully explain its growth surge. Depreciated real exchange rate levels have also benefited output growth, as have reductions in external debt.

Figure 2.18.Selected Countries: Contributions to Real GDP per Capita Growth, 2000–11

Source: IMF staff calculations.

Telecommunications, measured by cell phone usage, also has an effect on growth. This variable could also be capturing more generally the impact of technological change on output growth, a fact that would be consistent with the role of higher TFP growth in accounting for GDP growth shown in Box 2.2. However, the residual is fairly large for most countries, showing that not all country-specific factors are being picked up by the model specification. Finally, an indicator of government effectiveness based on surveys of perception has significant explanatory power, although its inclusion results in a considerable loss of observations. This outcome shows that the quality of public institutions can make a significant difference to the growth profile of individual countries.

Previous studies of the growth experience during this period for the individual countries include: IMF, Regional Economic Outlook: Sub-Saharan Africa, April 2011, for Rwanda, Tanzania, and Uganda; further work on Tanzania by Nord and others (2009) and Robinson, Gaertner, and Papageorgiou (2011); work on manufacturing in Ethiopia by Dinh and others (2012); and on Mozambique by Jones and Tarp (2013).

The comparator group is the average of sub-Saharan African countries in the low-income group (including fragile countries) with natural resource exports less than 25 percent of total exports, weighted by purchasing power parity.

Bond, Leblebicioglu, and Schiantarelli (2010) find empirical evidence of a positive relationship between investment and the long-term growth rate.

The fact that the voice and accountability dimension is not among the potential causal factors should not be interpreted as implying that voice and accountability are not important determinants of growth, as high-growth LICs could have registered even higher growth had they improved in this dimension. The results simply suggest that they cannot account for the higher growth over the period, given that the gap with the comparator group remained relatively constant.

In Ethiopia, lending to public enterprises by the dominant state-owned bank and the requirement for private banks to purchase treasury bills have crowded out credit to the private sector.

All sample countries score highly on the Index of Public Investment Efficiency (Dabla-Norris and others, 2011), with Rwanda scoring highest for sub-Saharan Africa LICs, closely followed by Burkina Faso.

Kaminski, Heady, and Bernard (2011) estimate that two-thirds of the threefold increase in production from 1996 to 2006 is explained by these reforms, and underline that the Burkinabe reform model is unique in addressing government failures and local realities within the current institutional framework.

Burkina Faso, Poverty Reduction Strategy Papers (2009); Ethiopia and Rwanda National Household Surveys; and World Bank analysis on Mozambique.

    Other Resources Citing This Publication