Journal Issue


International Monetary Fund
Published Date:
August 2012
  • ShareShare
Show Summary Details

I. Inequality and Growth in the Southern African Customs Union Region2

A. Background

1. Botswana has been able to sustain high economic growth for the past five decades and is considered to be one the best performers in sub-Saharan Africa. In effect, real GDP per capita (adjusted using purchasing power parity) grew from around $3,500 in 1980 to close to $12,500 in 2010 (in constant 2005 international dollars), which implies an average annual growth rate of 4.3 percent.

2. However, not all of Botswana’s policies promoted inclusive growth and human development in a broad-based manner. Poverty and inequality remain high, and the shortcomings of labor market policies are evident with an unemployment rate of about 18 percent. Botswana’s income inequality, with a Gini Index in excess of 0.5, is one of the highest in the world especially when compared with other high middle-income countries.

3. Preliminary estimates based on the Botswana Core Welfare Indicators Survey (BCWIS) 2009/10 show a sharp decline in the poverty headcount rate from 30.6 percent in 2002/03 to 20.7 percent in 2009/10. In addition, the absolute number of persons living with income below the poverty line declined from around 500 thousand in 2002/03 to about 373 thousand in 2009/10.

4. In terms of the distribution of poverty between rural and urban areas, the decline in the national poverty rate is fully accounted for by a reduction of the headcount poverty rate in rural areas, where it declined from 44.8 percent in 2002/03 to 25.5 percent in 2009/10. In contrast, the headcount poverty rate increased in urban areas from 10.6 percent in 2002/03 to 14.0 percent in 2009/10.

5. Given that the unemployment rate in Botswana remained nearly constant over the two surveys, at approximately 18 percent of the labor force, the decline in the poverty rate is remarkable. This is because labor income (either in the form of wages or earnings for the self-employed) continues to be the main source of income for most households. Cash earnings (that is, income excluding business profits, unearned cash income, own produce, wages in kind, aid and school meals) account for approximately 73.5 percent of total income, according to the BCWIS 2009/10.

6. This suggests that part of the decrease in the headcount poverty rate could be due to the expansion of social welfare programs, including the feeding programs (the School Feeding Program and the Vulnerable Group Feeding Program), the Old Age Pension program, the Orphan Care program, and the Program for Destitute Persons, among others.

7. Despite a sharp decline in poverty, income inequality remains high in Botswana. According to the two latest Household Income and Expenditure Surveys (HIES), the Gini coefficient was 0.54 in 1985/86 and 0.61 in 1993/94, placing it among the highest levels of inequality in sub-Saharan Africa and in the world, along with a few countries in Latin America.

8. The issue of the inequality-growth relationship is particularly relevant for SACU countries, which face high income inequalities. These inequalities, partly rooted in historical factors, also explain relatively weak growth performances, at least compared to the fast-growing emerging Asian countries. The debate on the relationship between inequalities and growth has led to a vast stream of literature, since the work of Kuznets (1955). Recent studies (Berg and Ostry, 2011, Berg, Ostry and Zettelmeyer, 2012) have renewed the approach, by investigating how income inequalities3 affect not growth itself, but its duration. Their findings underscore how countries with lower income inequalities experience on average longer periods of continuous growth (also called “growth spells”).

9. This chapter applies the work of Berg and Ostry (2011) to the SACU region, to identify how inequalities have played a role on the length of the growth spells in each of these countries, and to elaborate on policy options to reduce inequalities and foster growth. The main findings are as follows:

  • Reducing income inequalities has the potential to lead to significant gains in terms of increasing the duration of growth spells. In particular, SACU countries could almost double the duration of growth periods if they each had the same level of inequality as those selected countries with similar level of development.

  • While reducing inequalities may be desirable, the design of policies to achieve such objective is complex. Policies targeting income inequalities at the sources (for example, through early investments in human capital of the poor) are expected to be most effective to reduce inequalities and promote growth. However, direct redistribution, if carefully crafted, can also be very effective in reducing inequalities while limiting its potentially negative impact on growth.

B. Empirical Analysis

Comparative analysis of inclusive growth for middle-income countries

10. Preliminary estimates based on the Botswana Core Welfare Indicators Survey (BCWIS) 2009/10 show a significant decline in the poverty headcount rate, from 30.6 percent in 2002/03 to 20.7 percent in 2009/10. The expansion of social programs, including the feeding programs (the School Feeding Program and the Vulnerable Group Feeding Program), the Old Age Pension program, the Orphan Care program, and the Program for Destitute Persons, could be responsible for this marked reduction in the poverty rate. Nevertheless, part of the significant decline in the preliminary poverty estimates by Statistics Botswana may be due to the fact that the increase over the period 2002/03-2009/10 in the cost of the consumption basket on which the poverty line data in based (around 43 percent) is smaller than the accumulated consumer price index inflation (around 58 percent).

11. Estimates of the growth incidence curve, based on preliminary data from the 2009/10 BCWIS compared with data from the 2002/03 Household Income and Expenditure Survey (HIES), which depicts the growth rate of real consumption per capita for each percentile of the distribution, suggests that households in the middle of the distribution (between percentiles 15 and 75) experienced more rapid growth than those in the lowest 15 percent or in the highest 25 percent. The results suggest that despite the expansion of the welfare programs in Botswana, they have been relatively less effective in terms of targeting the very poor when compared with other middle-income countries like Brazil, Chile and Indonesia (see Box 1). It is important to underscore that these estimates are provisional, and will be updated once the final official data are made available.

Challenges with Sustained Growth Periods

12. One critical challenge for many emerging or developing economies relates to the capacity to sustain a “growth spell”4over a prolonged period of time (Berg, Ostry, and Zettelmeyer, 2012). Indeed, based on panel data analysis, the authors find that sub-Saharan African countries would typically encounter growth periods that tend to end with prolonged periods of negative growth (about between -3 and -6 percent, on average). It has resulted in overall weaker growth performances. Against this background, they also point out that most countries in that situation have nonetheless demonstrated, episodically, a capacity to generate a growth spell. Thus, the question at stake is to identify what forces would allow some countries to sustain high growth rates over a prolonged period of time.

Box I 1.Lessons for Botswana from Brazil, Chile, and Indonesia for achieving more inclusive growth

While almost no country has been able to sustain high growth without experiencing some increases in inequality at some point during of its development path, Botswana can draw useful lessons from countries that have been able to sustain high growth while maintaining low income inequality or have been able to reduce inequality during the process.

Brazil and Chile have been able to sustain high growth and decrease inequality simultaneously with a significantly lower share of resources allocated to social safety net spending as a percent of GDP than Botswana, as shown in the World Bank’s 2010 Public Expenditure Review for Botswana (see Figure 3, Chapter 5). This suggests that there are potentially large efficiency gains from better targeting of existing welfare programs. According to the World Bank’s 2010 Public Expenditure Review, the poor represented only about one third of the beneficiaries of the various social safety net programs in Botswana, with programs differing significantly in terms of their targeting efficiency. Despite the fact that Botswana spends more on social safety nets as a share of GDP (about 3.2 percent) than many other high-middle income countries, it is covering only about 20 percent of all poor households.

In the case of Brazil, the period of high GDP growth that began in the early 2000s has been accompanied by a remarkable decline in inequality. Brazil’s experience is particularly relevant to Botswana since it started with a similar level of inequality, with the Gini coefficient decreasing from around 0.594 in 2001 to 0.539 in 2009. Part of the decline has been attributed to the expansion of the Bolsa Familia (formerly known as Bolsa Escola) conditional cash transfer program. Begun in the late 1990s, this program targets poor households who have school-age children and provides them with monthly cash transfers conditional on regular school attendance. The distribution of program beneficiaries according to their position in the distribution of per capita income shows that most of the beneficiaries are indeed poor households (see Paes de Barros, Mendonça, and Tsukada, 2012). Bolsa Familia and other government transfers, together with macroeconomic stability and the expansion of access to primary and secondary education, can largely account for the decrease in inequality (see Lopez-Calva and Rocha, 2011).

Chile, a country similar to Botswana since exports and government revenues are highly dependent on a single commodity (Diamonds in the case of Botswana and Copper in the case of Chile), has been able to sustain high growth while decreasing inequality over the 2000s. This has been attributed to a combination of prudent macroeconomic management (through the design and implementation of a fiscal policy rule) and progressive redistribution policies, achieved through a combination of universal and targeted social welfare programs (see Lopez-Calva and Lustig, 2012). While some of the groundwork for the improved targeting of social programs was laid in previous decades with the introduction of the information systems to target poor households (see Castañeda, 1992), new programs targeted to the extreme poor like Chile Solidario have helped to reduce inequality further. Chile has been able to sustain relatively low unemployment partly through developing export-oriented, highly competitive sectors besides mining, including in agriculture (e.g. wineries, fresh fruits, and salmon farming) and forestry. The development of these sectors has been partly supported by a system of selective grants by Innova Chile to promote innovation and credit to small and medium enterprises by the Corporación de Fomento de la Producción, both operating within a highly liberalized and non-discretionary international trade regime.

While the case of Indonesia is somewhat different from Brazil and Chile in the sense that its income inequality is relatively low and it also has a significantly lower level of GDP per capita than Botswana and a population more than 100 times larger, it is still relevant since its growth has been partly based on the extraction of non-renewable resources (oil in particular). Indonesia has been able to avoid the resource curse, successfully diversify its economy and develop a competitive manufacturing sector. More importantly, it has been able to grow while reducing poverty and maintaining a relatively low level of inequality. Indonesia’s ability to diversify its economy and avoid the natural resource curse has been attributed to macroeconomic stability, fiscal discipline and trade liberalization. Part of Indonesia’s success in facilitating structural transformation from agriculture into industry has been a series of predictable agricultural policies, including moderate input subsidies (e.g. fertilizers) and stable farm-gate price for export commodities and rice prices for consumers. In addition, investments in labor-intensive infrastructure and rural development helped boost labor demand in rural areas (see Maehle 2012).

13. In order to determine the impact of inequality on growth, Berg et al (2012) employ a proportional hazard model specification in which the dependent variable is the duration of growth spells, and estimate the impact of several economic and political variables on the probability that a growth spell will end. Higher income inequalities are not only associated with shorter growth durations but also appear to be a major contributing factor (Figure I.1). Improvements in income distribution, namely a reduction in the Gini coefficient from the 50th to the 60th percentile, would typically be associated with 50 percent longer growth period. In other words, controlling for other factors such as terms of trade, FDI received, price competitiveness, it appears that income inequalities, as measures by the Gini coefficient, do play a very significant impact on growth spells duration. Other factors are also found statistically significant in explaining the duration of growth spells, but to a lower degree, such as (i) investment in infrastructure, (ii) external shocks (for instance, changes in terms of trade or nominal US interest rate), (iii) quality of public institutions, notably as measures by the autocratic degree of political regimes, and (iv) financial sector development. In contrast ethnic, linguistic and religious heterogeneity do not seem do have significant association with length of growth spells. Similarly, human capital measures (education,5 health6) are associated with improved predicted duration of growth spells.

Figure I.1:Effect of Increase of Different Factors on Growth Spell Duration

Source: IMF Staff estiamtes and computations.

For each variable, the height of the figure shows the percentage increase in spell duration resulting from an increase in that variable from the 50th to the 60th percentile, with other variables at the 50th percentile, except autocracy, which is not a continuous variable. For autocracy, the figure shows the effects of a move from a rating of 1 (the 50th percentile) to 0 (the 73th percentile.)

Analyzing Growth and Inequalities within SACU

14. The overall growth performances in the SACU region over the past decade have been mixed, when comparing each country to an average of countries with similar level of income per capita (Figure I.2). With the exception of Lesotho, the poorest country in the region, all exhibit weaker growth performance, in terms of GDP per capita growth. Despite the proximity to a large emerging economy (South Africa), as well as significant natural resources (Botswana, Namibia, South Africa), translating growth potential into improved GDP per capita growth remains difficult. The finding highlights the need for additional investigation of growth sources, with a specific focus on the interaction between growth and inequalities.

Figure I.2.Average GDP per Capita Growth Rate, 2000–10

(percent, data based on GDP per capita in US$)1

Citation: 2012, 235; 10.5089/9781475504897.002.A001

Source: WEO; and IMF Staff computations.

1 Comparisons are based on GDP per capita, in constant US$, averaged over the period 2008–10. Each of the SACU country is then compared to the countries having the next three higher income, and the next three lower income.

15. Cross-country comparisons underscore growth spells vulnerabilities for most SACU countries, except Botswana, and to some extent, South Africa (see Appendix I and Table 1). In the case of Lesotho and Swaziland, the finding is particularly acute, as the past few years also indicate a low growth, particularly compared to the other countries. South Africa’s growth path is contrasted, with a continued growth spell that started in 1998, i.e. after the democratic reforms of 1994 (Table 1). Similarly, Namibia, since its independence in 1990, has also shown a capacity to generate high growth rates of income per capita, with a growth spell that started in 1995. However, even if both countries have experienced a growth spell, performances during that spell have been relatively weak, with the average growth of GDP per capita of 2¼ percent (Namibia) and 3 percent (South Africa). 7 Thus, the hazard ratios reported are specific to each growth spell, and are computed at the end of each available spell. A ratio above 1 indicates a higher, and country-specific, risk of growth spells ending, compared to the sample average. All countries have a hazard ratio significantly above 1, except Botswana, with a hazard ratio of 0.7. It indicates Botswana’s stronger resilience during the spell ending in 1988, compared to the average growth spell in the sample.

Table 1.Contributing Factors to the End of Growth Spells in the SACU Region1
Spell DatesGDP per capita growthHazard RatioContributing factors

(Percent of total hazard)
startEndDuring SpellNext decadeInequalityLow FDI InflowIncreased External DebtMore AutocracyOver-valuationTrad openess
Lesotho 2197219784.92.67.548.
Namibia 319952.22.672.
South Africa 319983.02.673.
Swaziland 2197119797.2-0.27.445.
Source: IMF Staff estimates and computions.

16. While inequalities clearly play a crucial role, other factors are also at play, and may have strong impact on policy options. As shown in Table 1, deteriorating trade openness and increasing degree of autocracy have also played a significant role in some countries. As such, improving growth spell duration should not be viewed as a simple question of the growth/inequalities nexus, but a broader challenge of institutional set-ups.

  • For smaller SACU members, insufficient trade liberalization contributed from 14 percent up to 28 percent to the end of growth spells. A relatively low degree of trade liberalization in smaller SACU members is also shown as contributing to the hazard ratio of growth spells. However, the indicator8 developed by Wacziarg and Welch (2008) considers these countries relatively closed, until the revision of the SACU agreement of 2002, largely because external tariffs, common to each member, were set by South Africa (WTO, 2003). SACU countries are currently highly integrated with the world economy, with various trade agreements, including non-reciprocal preferential treatment with the EU and the US that benefit smaller SACU members.

  • Similarly, the status of political institutions matters greatly.9 According to the measure of the degree of autocracy of political regimes given by Polity IV, two countries stand out: Lesotho during the pre-democracy period (which started in the 1990s), and Swaziland which is an absolute monarchy. During Lesotho’s first growth spell period, autocracy was responsible for 24 percent of the end of the growth period, whereas in Swaziland autocracy contributed 26 percent and 36 percent to the end of the two spells, respectively. These findings can also be understood in the context of the general literature on inequalities, growth and political systems. A key risk for growth and its sustainability is the concentration of political powers, which cannot only increase inequalities, but also increase the risks of political crises (Saint Paul et Verdier, 1996, Engerman and Sokoloff, 2000, Bénabou, 2002). Even if an actual political crisis does not occur, the risk of social unrest can be sufficient to lower investment and growth. However, it should also be stressed that democracies are not exempt from risks, where, for example, pressures for redistribution can also lead to higher taxation and lower growth (Alesina and Rodrik, 1994, Person and Tabellini, 1994). More generally crises can be generated when the poorer segments of the population compensate income disparities with access to financing (Kumhof and Rancière, 2010). An overabundance of such financing can precipitate financial and/or political crises.

17. While high income inequalities are associated with the growth spell vulnerabilities, the causal relation remains complex. For almost all countries, income inequalities have remained high throughout the sample.10 Thus, changing the level of income inequalities cannot systematically be associated with a prolonged growth spell. Crafting an active policy to jointly reduce inequalities and foster growth is further discussed in the next section. Additionally, other factors not included in the model, and specific to SACU countries, can affect both income inequalities and growth performances. For example, Botswana’s economy is heavily dependent on mineral extraction (mostly diamonds). Mining activities are not labor intensive, and are thus less prone to reduce income inequalities.

Reducing Inequalities to Strengthen Growth Performances

18. Reducing income inequality could potentially lead to significantly improved growth performances (Table 1). SACU countries exhibit higher income inequalities than countries with similar level of GDP per capita (Figure I.3). In particular, Botswana, Lesotho, Namibia, and Swaziland (BLNS) all exhibit much higher Gini coefficients than their respective peers.11 In contrast, South Africa has a Gini coefficient only slightly higher than the average country at the similar income level. However, even in the post-apartheid period (i.e. since 1994), South Africa Gini has shown signs of slight worsening (+1 percent). While redistribution among ethnic groups took place, the overall impact on income inequalities has remained, at best, marginal.12 Applying the methodology of Berg and Ostry (2011)13 to SACU countries help quantify the potential gains of lower income inequalities. Two experiments are then considered: one where, for each SACU country, income inequalities are set to their lowest historical level, and another, where inequalities are set to the average level encountered in countries of similar level of developments. Table 2 summarizes the results: for all countries, the gains from this hypothetical improvement in inequality could be quite significant. For most spells, the average duration could have been increased from about 5–8 years, up to 15 years and above (Namibia, Botswana). The result for Namibia is largely driven by the high degree of income inequalities encountered during its growth spell. In Botswana, the unusual potential gain is largely driven by the length of the estimated growth spell. Thus, even marginal improvements in income distribution could result in a much prolonged growth spell.

Table 2.Comparative Impact of Income Inequalities on Growth Spell Duration
Gini coefficentsAverage


Increased duration resulting from a lower Gini
Average 1Historical 2Cross country 3Historical 4Cross country 5
Botswana52.248.542.723+5 years+32 years
South Africa52.741.743.28+6 years+7 years
Namibia73.942.143.35+17 years+3 years
Swaziland61.743.647.04+6 years+3 years
Lesotho61.050.341.25+4 years+10 years
Source: IMF Staff estimates and computations.

19. Despite the potential positive impact on growth from reducing inequalities, redistributive policies need to be carefully crafted in order to avoid a negative impact on work and investment incentives. There are two main considerations to bear in mind when implementing redistributive policies.

  • Reducing inequalities in human capital should be at the core of policy intervention aimed at reducing future income inequalities and promoting growth.

  • In parallel to promoting human capital investment, policies could also help private sector development, so that eventually new skills available are matched with corresponding vacancies. Otherwise the economy could well be trapped in structural imbalances between labor supply and demand.

  • Carefully crafted direct redistribution of income is also desirable, especially to alleviate extreme poverty.

Figure I.3:Comparing Gini Coefficients Between SACU countries

(percent; higher Gini = higher income inequalities)1

Citation: 2012, 235; 10.5089/9781475504897.002.A001

Source: UN-WIDER (; UTIP (; and IMF Staff computations.

1 Comparisons are based on GDP per capita, in constant US$, averaged over the period 2008–10. Each of the SACU country is then compared to the countries having the next three higher income, and the next three lower income.

Promoting Human Capital Investment for All

20. Income inequalities are primarily related to disparities in human capital (health, education). On the health side, the HIV/AIDS epidemic is a major contributor, as SACU countries share the highest prevalence rates in the world (Table 3). Groups relatively more affected by HIV/AIDS complete fewer years of education (Fortson, 2011). Eventually it contributes to increasing inequalities, as the less educated and therefore less-paid individuals are also more vulnerable to the disease (Chicoine, 2012). This inequality is also passed-on to future generations, as affected people are more likely to leave behind orphans, who, in turn, are less likely to attend school (Case and Ardington, 2006, Evans and Miguel 2007). On the education side, the individual return of education is negatively affected by health issues (HIV/AIDS).

Table 3.Highest HIV/AIDS Prevalence Rates1 in the World, 2009
RankRatePop. Growth 2
South Africa417.80.8
World average1.9
World median0.4
World st. dev.4.4
Source: UNAIDS; IMF staff computations. Data available on the internet:

21. Policies geared towards the provision of health and education services can be effective to reduce inequalities (Bourguignon, Ferreira, and Menéndez, 2007) and increase growth. Market failures can explain why a sub-optimal investment in human capital can occur and be detrimental to both inequalities and growth. The poor face typically two constraints for their personal investment: their resources are so limited that they simply can’t invest (Galor and Zeira, 1993, Piketty, 1997), or the return of human capital investment is too low, largely because of externalities14 (e.g., Glomm and Ravikumar, 1992, Bénabou, 1993, 1996, 2002). Facilitating the investment in human capital by the poor, would typically involved the public provision of education and health services, which in turn would imply taxing relatively richer individuals to finance it. Thus, an arbitrage takes place between disincentives created by higher taxation, versus higher income generated by the provision of those services. However, by fostering human capital investment for all, the economy can generate more benefits for all, primarily because it can correct market failures and externalities, thus achieving – at the level of the economy as a whole – a higher level of investment, and higher growth, despite the negative impact of the added fiscal pressure.

22. The cost of public intervention would nonetheless need to be carefully assessed. Public systems can be very costly (e.g., programs of education for all), in terms of buildings, teachers trainings, and overall cost for the budget. Donors assistance, including financial, could be desirable to mitigate fiscal risks and facilitate quality goals. Additionally, the quality of the education system is not a simple function of funding allocated to it. For instance, the pupils-to-teachers ratio plays an important factor, and so do the skills of teachers. It could thus be very well the case where growth and inequality reduction gains would be maximized with targeted improvements in the quality of the education system.

23. SACU countries have made significant efforts to improve the access to education. For example, Lloyd and Hewett (2009) show, using data from Demographic and Health Surveys (DHS),15 as well as UNICEF’s Multiple Indicator Cluster Surveys (MICS)16 how SACU countries as a whole have among the best scores in primary school completion rates in sub-Saharan Africa. Although active policies have helped improve the access to education and therefore reduced inequalities in terms of “quantitative parameters” (years of education), qualitative differences in education contribute to explain income disparities (van der Berg, 2009, Keswel, 2009). These qualitative differentials also contribute to explain why overall income inequalities have not been reduced yet in South Africa.

Increasing Employment Opportunities

24. The very large unemployment rate in SACU, ranging from about 20 to 50 percent (Leigh and Flores, 2011), underscores the need to increase employment opportunities in parallel to human capital development. Overall, there is a need for broad development policies where human capital investment would be coupled with strategies to develop the private sector.

25. Promoting employment opportunities in rural areas would be essential to provide long-term answers to inequalities and poverty. A vast fraction of the populations (about ¾ of the population in Namibia and Swaziland, ½ in Botswana) depend on subsistence agriculture. Their productivity is low and low investment in turns limit the individual profitability of human capital investment (World Bank, 2006). The poor in rural areas are also typically affected by two self-reinforcing factors: lack of access to financial services, and land ownership.17 The persistence of inequalities underscores the need for continued improvements in access to land asset inequalities inherited from the apartheid area. Successful land reforms would notably provide fair, transparent, and long-lasting rights, so as to foster the use of land for collateral (World Bank, 2006).

26. Private sector development would be essential to strengthen existing – or develop new – comparative advantages. Not surprisingly, human capital alone is rarely found as having a significant impact on growth (Mankiw, Romer, Weil, 1992, Pritchett, 2001, 2006). It’s positive impact on growth comes when it complements technological efforts, notably by favoring technological transfers to emerging and developing economies. From this perspective human capital investment complement private sector development to generate, jointly, a positive impact on growth (Benhabib and Spiegel, 1994, del Barrio, Lopez, and Serrano, 2002, Engelbrecht, 2002, Frantzen, 2002).18 Similarly, without demand for educated labor, the marginal return on education would also decrease rapidly (Pritchett, 2001).

Fighting Poverty

27. Redistributive policies (cash and in-kind transfers, progressive taxation) have potential benefits in terms of addressing poverty – and thus inequalities. Cash and in-kind transfers have been increasingly used by SACU countries, as by other sub-Saharan African countries (Garcia and Moore, 2012). They are also a part of social protection for vulnerable groups, including elderly people and orphans who are trapped in poverty. They can also mitigate market failures, by bringing resources to the poor to undertake investments, which otherwise could not be financed. Cash transfers can also be conditioned to reduce child labor and promote schooling (e.g. Bolsa Família program in Brazil,19 see also World Bank, 2006). It is also noteworthy that the South African Child Support Grant, introduced in 1998 as an unconditional cash transfer to eligible caregivers of children has had some positive impact on overall health and education attainment by children (Coetzee 2011).

28. Thus, measures could also be considered such as improving labor market flexibility and developing in-work tax credits, so as to preserve work incentive and an effective labor matching process. Such measures would typically imply a trade-off between reducing inequalities and increasing employment.

C. Conclusions

29. Estimates based on preliminary data from the 2009/10 BCWIS, compared with data from the 2002/03 Household Income and Expenditure Survey (HIES), suggest that there was a decrease in inequality in the intervening period, as measured by a decline in the Gini coefficient. This said, staff’s analysis suggest that despite the expansion of the welfare programs in Botswana, they have been relatively less effective in terms of targeting the very poor when compared with other middle-income countries like Chile, Brazil,, and Indonesia.

30. Botswana and the SACU region as a whole exhibit a high degree of income and non-income inequality. Reducing these inequalities is a major challenge, but evidence suggests that lowering income inequality has the potential to extend the length of growth spells and offer durable solutions to poverty and long run growth. As shown in this chapter, the design of policies to achieve these goals is complex. Policies have to carefully balance the effects of added fiscal pressures on businesses and individuals, with benefits stemming from greater investment in human capital and reduced poverty. While there is a clear role for public intervention, the overall fiscal impact would have to preserve the soundness of public finances. Timing would therefore be of essence: gains expected from step up investment in human capital would typically imply trade-off between short-term fiscal costs, with more long-term benefits. While desirable, these benefits would need to be carefully assessed and complemented by policies geared towards private sector development, so that fiscal sustainability remains preserved.

Appendix I.1

The following charts summarize the estimated growth spells in the SACU Region during the period.

Figure 1.Growth Spells in the SACU Region, 1950–20101

Source: Penn World Tables 6.0; IMF Staff estimates.

1 Minimum spell period set at 5 years, with a p-value of 25 percent.

Appendix I.2

The duration of a growth spell is estimated with a duration model, where the time since growth accelerated, t, has a probability Γt of seeing the end of the growth period at the next period. Formally, Γt depends on a set of variables, some time-dependent, Xt, and some time-independent, z, according to the following functional form: Γt = λteβ[Xt, z], where β is a vector of coefficients to be estimated. Γt also depends on an autonomous factor, λt = ptp-1, which quantifies a “natural” duration of the growth spell, i.e. the extent to which a growth spell is bound to end regardless of explanatory variables. The time-dependence of the growth spell duration is positive when p > 1, in which case spells would become more sustainable over time; or negative when p < 1, in which case spells would become less sustainable. The parameters p is also estimated.

Summary of Regressions Coefficients
Inequality (GINI)-0.05
Income per capita at the begining of the growth spell-0.11
Debt liabilities0.00
FDI liabilities0.02
Change of inflation within spell-0.01
First lag of US interest rate change-0.24
Overvaluation of exchange rate0.00
Polity IV autocracy measure-0.13
Trade liberalization0.66
Terms of trade growth0.01

    Alesina, A., and A., Rodrik, 1994, “Distributive Politics and Economic Growth,” The Quarterly Journal of Economics, 109(2), pp. 465490.

    Bénabou, R., 1993, “Working of a City: Location, Education, and Production,” The Quarterly Journal of Economics, 108, pp. 619652.

    Bénabou, R., 2002, “Tax and Education Policy in a Heterogeneous-Agent Economy: What Levels of Redistribution Maximize Growth and Efficiency?Econometrica, 70(2), pp. 481517.

    Berg, A., and J., Ostry, 2011, “Inequality and Unsustainable Growth: Two Sides of the Same Coin?”, IMF Staff Discussion Note SDN/11/08. Available via the internet:

    Berg, A., Ostry, J., and J.Zettelmeyer, 2012, “What makes growth sustained?Journal of Development Economics,. Available via the internet:

    Bourguignon, F., Ferreira, F., and M., Menéndez, 2007, “Inequality Of Opportunity In Brazil,” Review of Income and Wealth, International Association for Research in Income and Wealth, 53(4), pp. 585618.

    Coetzee, M., 2011, “Finding the Benefits: Estimating the impact of South African Child Support GrantStellenbosch University, Department of Economics, Working Paper # 16/2011.

    Cowell, F., 2000, “Measurement of Inequality” in Atkinson, A., and F., Bourguignon (eds.), Handbook of Income Distribution, (Amsterdam: North Holland).

    Del Barrio Castro, T., Lopez Bazo, E., and G., Serrano Domingo, 2002, “New evidence on international R&D spillovers, human capital and productivity in the OECD,” Economics Letters, 77: pp. 4145.

    De la Fuente, A., and R., Doménech, 2006, “Human capital in growth regressions: How much difference does data quality make?Journal of European Economic Association, 4, pp. 136. Available via the internet:

    Engerman, S., and K., Sokoloff, 2000, “History Lessons: Institutions, Factors Endowments, and Paths for Development in the New World,” Journal of Economic Perspectives, 14(3), pp. 217232.

    Engelbrecht, H., 2002, “Human capital and international knowledge spillovers in TFP growth of a sample of developing countries: an exploration of alternative approaches,” Applied Economics, 34, pp. 831841. DOI: 10.1080/00036840110061947

    Frantzen, D., 2002, “Intersectoral and international R&D knowledge spillovers and total factor productivity,” Scottish Journal of Political Economy, 49, pp.280303. DOI: 10.1111/1467-9485.00232

    Garcia, M., and C., Moore, 2012, The Cash Dividend, The World Bank, Washington DC. Available via the internet: 0020Box367844B09953137.pdf?sequence=1

    Galor, O., and J., Zeira, 1993, “Income Distribution and Macroeconomics,” Review of Economic Studies, 60, pp. 3552.

    Galor, O., Moav, O., and D., Vollrath, 2009, “Inequalities in Land Ownership, the Emergence of Human-Capital Promoting Institutions, and the Great Divergence,” Review of Economic Studies, 76(1), pp. 143179.

    Glomm, G., and B., Ravikumar, 1992, “Public versus Private Investment in Human Capital: Endogenous Growth and Income InequalityJournal of Political Economy, 4, 818834.

    Kumhof, M., and R., Rancière, 2010, “Inequalities , Levelrage and Crises”, IMF Working Paper 10/268, (Washington: International Monetary Fund). Available via the Internet:

    Kuznets, S., 1955, “Economic Growth and Income InequalityThe American Economic Review, 45(1), pp. 128.

    Leigh, L., and I., Flores, 2011, “Closing the Jobs Gap in the Southern Africa Customs Union (SACU)”, in Mongardini, J., (ed.) Building a Common Future in Southern Africa: Challenges and Opportunities, forthcoming, IMF, Washington DC.

    Lloyd, C., and P., Hewett, 2009, “Educational Inequalities in the Midst of Persistent Poverty: Diversity across Africa in Educational Outcomes,” Journal of International Development, 21(8), pp. 11371151.

    Lopez-Calva, Luis Felipe and SoniaRocha (2012), “Declining Income Inequality in Brazil: Learning and Looking Forward,” Policy Note, Poverty Reduction and Economic Management Unit, Latin America and the Caribbean region, Washington DC: The World Bank.

    Maehle, Nils (2012), “Boosting Employment: Lessons from Natural Resource Abundant Countriespresentation delivered at the IMF conference on Natural Resource Management in Lusaka, Zambia, May2012.

    Mankiw, N., Romer, D., and D., Weil, 1992, “A Contribution to the Empirics of Economic Growth”, The Quarterly Journal of Economics, 107(2), pp. 407437.

    Paes de Barros, Ricardo, Rosane Mendonça, Raquel Tsukada (2012), “The determinants of the sharp improvement in income distribution in Brazil,” Presentation, Secretaria de Assuntos Estratégicos, Brasília, 27March2012.

    Person, T., and G., Tabellini, 1994, “Is Inequality Harmful for Growth?,” The American Economic Review, 84(3), pp. 600621.

    Piketty, T., 1997, “The Dynamics of the Wealth Distribution and the Interest Rate with Credit Rationing,” The Review of Economics Studies, 64, 173189.

    Pritchett, L., 2001, “Where Has All the Education Gone?The World Bank Economic Review, 15(3), pp. 367391. Available via the internet:

    Pritchett, L., 2006, “Does Learning to Add Up Add Up? The Return to Schooling in Aggregate Data”, in Hanushek, E., and F., Welch, (eds.), Handbook of the Economics of Education, (Amsterdam: Elsevier), 1(06), pp. 635695. Available via the internet:

    Rajan, R., 2009, “Rent Preservation and the Persistence of Underdevelopment,” American Economic Journal: Macroeconomics, 1(1), pp. 178218.

    Saint-Paul, G., and T., Verdier, 1996, “Inequalities, Redistribution and Growth: A Challenge to the Conventional Political Economy Approach,” European Economic Review, 40(3–5), pp. 719728.

    Van Der Berg, S., 2009, “The Persistence of Inequalities in Education”, in South African Economic Policy under Democracy, Aron, J., Kahn, B., and G., Kingdon, (eds.) Oxford University Press, New York.

    Wacziarg, R., and K., Welch, 2008Trade Liberalization and Growth: New Evidence.World Bank Economic Review, 22(2), pp. 187231.

    World Bank, 2006, World Development Report, The World Bank and Oxford University Press, Washington DC.

    WTO, 2003, “Trade Policy Review – Southern African Customs Union”, available on the internet:

Initial results for the two chapters were presented to the Botswana authorities in Gaborone in January and May 2012.

This paper is largely based on the forthcoming IMF Working Paper “Inequalities and Growth in the Southern African Customs Union (SACU) Region” by Olivier Basdevant, Dalmacio Benicio, and Yorbol Yakhshilikov.

Measured by the Gini coefficient, which can vary from 0 (perfect equality: all agents have the same income) to 100 (the extreme case of perfect inequality, where only one agent receives all the income generated by the nation, and all the others receive nothing).

Berg, Ostry and Zettelmeyer (2012) define growth spells as periods of real GDP per capita growth of at least 5 years, identified as beginning with an upbreak of per capita growth in excess of a minimum of 2 percent and ending wither with a downbreak followed by a period of an average growth of less than 2 percent, or simply the end of the sample.

Improvement in primary education enrolment rate.

Child mortality rate level and change.

Additionally, all countries, with the exception of South Africa, exhibit higher growth spell vulnerability compared to the sample average. This relative vulnerability is measures by the hazard ratio, reported in (Table 1). The hazard ratio predicts probability that a spell would end during the five years prior to its actual end, as a ratio to the predicted probability of a spell ending for the average observation in the entire sample.

This indicator equals 1 if current year is greater than the year of trade liberalization and no reversal of the trade policy reforms have occurred and 0 otherwise.

In the estimated model political institutions were measured by Polity IV variable on autocracy (, scaled from 0 to 10 (most autocratic society).

Income inequality, measured by Gini coefficient here, has low variation across time.

The result for Namibia has to be taken with cautious, as there are only two data points for its Gini coefficient.

The Gini coefficient, as most measures of inequalities, complies with the “anonymity axiom” (see Cowell, 2000) that is, the measure does not depend on who has which income level.

See Appendix for a short description.

Externalities are generated because the individual profitability of human investment is actually dependent on decisions made by other individuals. It would typically result in richer people congregating to generate–and benefit from–higher level of externalities for themselves, but leaving lower opportunities for the vast majority of the population.

Unequal access to land ownership can lead to lower education investment, as a result of pressures for preserving a low-skilled labor force in the agricultural sector. On the contrary, economies with a more equal access to land ownership have led to more incentives for higher investment in education, eventually leading to the emergence of skilled-labor intensive industrial activities (Galor and others, 2009, Rajan, 2009).

The findings should nonetheless be taken with caution, as De la Fuente, and Domenech, (2006) show that the results are sensitive to the quality of measurement of human capital.

Details can be found on the World Bank website,

Other Resources Citing This Publication