A. Introduction

59. There is a growing debate on how best to organize old-age support in developing countries. While research has traditionally focused on contributory pension schemes, recent studies highlight the use of non-contributory or even universal cash transfers for the elderly to combat old-age poverty (Barrientos, 2005; Plamondon et al., 2004).²⁸ Traditionally, it was feared that non-contributory programs would lead to unsustainable fiscal deficits and crowd out inter-generational support. More recently, focus has shifted to the potential advantages of non-contributory pension schemes, which are only found in a handful of developing countries (IDPM, 2003). Namibia features such a pension system, which belongs to the biggest programs of its kind, with a cost of 1¼ percent of GDP.

60. This chapter deals with the sustainability and efficiency of the flat-rate, universal cash transfer scheme available to all Namibians 60 years and older, irrespective of income. The pension scheme is administered through a biometric smart card system at low administrative costs. Due to its ability to reach the poor in rural areas and in the informal economy, the universal pension grant constitutes a very important part of the social safety net in Namibia. It is reported to often provide the only income for a household of three to five people. This chapter analyzes the merits of the universal pension system and assesses its fiscal sustainability. As the high prevalence rate of HIV/AIDS deeply affects Namibian society, this chapter pays close attention to the demographic development and its impact on the sustainability analysis.

61. This chapter concludes that Namibia’s universal pension grant is worthwhile to be maintained as it seems to be sustainable and reasonably efficient. The analysis of Namibia’s population dynamics—not surprisingly—yields the result that HIV/AIDS has a devastating impact on the overall population development by reducing population growth substantially. However, the ratio of old-age persons to working-age persons is projected to change only slightly over the next 20 years. From a demographic point of view, the sustainability of the old-age pension scheme is therefore not endangered. From a fiscal point of view, the same observation holds. The fear traditionally expressed in the literature that non-contributory programs cause unsustainable fiscal pressures does not seem to have merit in the context of the current set-up in Namibia. The financing burden carried by the working-age population, measured as expenditure in terms of GDP, is projected to stay constant or decline. Different assumptions underlying several different scenarios test this observation for robustness. Nevertheless, these conclusions hinge on the assumption that the old-age pension benefit is not made too generous. In addition, while from a conceptual point of view further efficiency gains could be made if means-testing was introduced, this should only be done if the savings from restricting the pension to the needy is not offset by the additional administrative costs necessary to enforce the means-test.

62. This chapter is structured as follows: Section B provides some background information on the institutional status quo. Section C discusses the demographic projections for Namibia available from the U.S. Census Bureau as well as the United Nations and examines the demographic sustainability of the pension system. The population forecasts then feed into the expenditure projections in the following Section D, which are based on different underlying assumptions about the development of the pension grant, the number of pension recipients, and GDP growth. Section E summarizes the results and concludes.

B. Institutional Background

63. Private pension funds and the Government Institutions Pension Fund are important sources of income at retirement age in Namibia (Box III.1). However, these pension schemes only provide benefits to those Namibians who contributed (or to their dependents). The Social Security Act of 1994 envisaged a “National Pension Fund” with Pay-As-You-Go characteristics. However, it has not been established so far, and a detailed discussion of such a scheme is beyond the scope of this chapter.

64. The universal monthly pension grant, available to all Namibians of retirement age, reaches an estimated 75 to 100 percent of the eligible population. While the number of Namibians receiving a social pension is well-known (in November 2005: 119,773), estimates on the number of Namibians 60 years and older deviate. To receive the monthly grant of currently N$300 (about US$45), applications must be submitted, which must include proof of age (birth certificate) and citizenship (identity document, marriage certificate). If applicants do not possess these documents, they can rely on institutions such as churches to provide confirmation of age and citizenship.

65. The universal pension system functions as an important part of the social safety net in Namibia.²⁹ The latest available Household Income and Expenditure Survey for Namibia from 1993/94 indicated that 11 percent of households in Namibia relied on pensions as their main source of income with 35 percent of the population living on an income of less than US$1 a day. The monthly pension grant of N$300 (US$45) provides relief in this respect by contributing US$1.50 to a recipient’s daily income. While the pension transfer by its own does not help to reach the monthly expenditure mean in Namibia (found to be N$462 (US$70) by a more recent survey from 1999), the monthly grant still goes a long way by transferring 65 percent of the expenditure mean.

66. The existence of a universal cash transfer to the elderly does not seem to lead to a loss of inter-generational support in Namibia. On the contrary, the pension income seems to be shared within the household, which typically consists of multiple generations. This intra-household, inter-generational support is especially important in the context of the HIV/AIDS epidemic, which causes many among the working-age generation to die or to become too sick to work.

67. The universal pension income provides the resources to support children and keep them in school. The household survey indicated that 79 percent of households with a member aged 60 or more had also school-age children. IDPM (2003) found evidence for South Africa’s non-contributory pension program that a common motivation for sharing the pension income with relatives was to help with education costs.³⁰ Also for South Africa, Duflo (2000) found a statistically significant link between the pension transfer to the grandmother and nutrition levels of her granddaughter.

68. The authorities’ estimates suggest that introducing a means-test would reduce the number of recipients by 30 percent. A distinguishing feature of Namibia’s old-age grant system is its universality. Individuals can apply and receive the transfer irrespective of their income. While, from a conceptual point of view, this approach seems to be suboptimal, the authorities motivate the universality with the lack of bureaucratic infrastructure to enforce a means-test. It is unclear how administrative costs would need to be adjusted if such a procedure was successfully introduced.

69. Monthly payments can be received through a bank account, or can be picked up in cash at a post office or a mobile unit. In November 2005, 11 percent of pensioners chose to receive their monthly grant electronically through a bank account, 18 percent picked up their transfer at a post office, while the majority of 71 percent opted to receive their cash grant through a mobile unit. The cash disbursements by mobile units are offered through a tendered private contractor at pay-points countrywide, especially in the rural areas. Grant recipients use a smart card, which includes a picture and an ID number, to identify themselves. In addition, a fingerprint is used instead of a PIN. If the database, which is updated monthly and carried by the mobile unit, matches name and fingerprint, cash is disbursed. The contractor is paid by transaction depending on the total number of transactions. The more transactions, the lower the fee he can charge. At the current number of transactions, the contractor charges the government N$9.75 for one disbursement of N$300 at a mobile unit.

70. The administrative costs constitute only around 4 percent of total costs for the pension scheme and are relatively small compared to the costs of other countries’ welfare programs. Administrative costs actually decreased in 2005, as the transaction fee charged by the private contractor for the cash disbursement was lowered. As a comparison, welfare programs can have administrative costs of 30 percent—as Mexico’s TORTIVALES program (Grosh, 1994). Coady (2003) cites Mexico’s PROGRESA program with administrative costs constituting less than 10 percent of total program costs as a benchmark. The proportionally low administrative costs would speak in favor of using the smart card system when introducing a more targeted cash transfer scheme as discussed in Chapter II.

71. Since independence, the costs of the pension scheme have declined marginally, from 1½ percent of GDP in 1990/91 to 1¼ percent of GDP in 2004/05. Relative to total government spending, pension costs fell from 5 percent to 3¼ percent during that period. This moderation mainly reflects prudent increases of the pension level (see Section D) and changes to the eligibility criteria. At independence, the value of the pension depended on a person’s race and ranged from N$382 to N$55. This procedure was only reversed in 1994 when all Namibians of retirement age became eligible to receive a monthly transfer of N$135. This amount has steadily increased up to the current amount of N$300. Future expenditure depends on the level of the pension transfer and on the number of future grant recipients. This figure again depends on the population dynamics, which will be discussed in the following section.

C. Population Dynamics

72. The high HIV/AIDS prevalence rate of about 20 percent substantially influences population growth in Namibia. This chapter uses population projections from the U.S. Census Bureau (USCB, 2004) which has compiled population forecasts for the focus countries of the President’s Emergency Plan for AIDS Relief (PEPFAR), among them Namibia.³¹ The assumptions underlying the USCB model do not explicitly incorporate the authorities’ efforts to combat HIV/AIDS as summarized in the “Third Medium-Term Plan (MTP III)”. The USCB projections can be interpreted as a scenario in which the implementation of MTP III largely fails. As a contrast, this chapter also uses demographic projections from the United Nations Populations Division (UN, 2005). The UN model explicitly takes into account that the authorities are expanding treatment of AIDS patients with antiretroviral drugs.³² Therefore, the UN projections can be used as a proxy for the scenario in which the authorities successfully implement MTP III. Up-to-date detailed population projections from the Namibian authorities were not available.

73. The difference between the actual demographic situation and a counterfactual No-AIDS scenario demonstrates the devastating effect HIV/AIDS has on population dynamics. The factual AIDS situation is marked by a dynamic increase in the crude death rate (number of deaths per 1,000 persons) and a sharp drop in life expectancy in the middle of the 1990s (Figures III.1 and III.2). The resulting population pyramids differ vastly from those produced by hypothetically assuming that AIDS did not exist (Figure III.3).

View Full Size

Namibia: Crude Death Rate (number of deaths per 1,000 persons), 1980-2025

Citation: IMF Staff Country Reports 2006, 153; 10.5089/9781451828429.002.A003

Source: USCB (2004), UN (2005).

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Namibia: Life Expectancy at Birth, 1980-2025

Citation: IMF Staff Country Reports 2006, 153; 10.5089/9781451828429.002.A003

Source: USCB (2004), UN (2005).

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Namibia: Population Pyramids, 2005, 2015, and 2025

Citation: IMF Staff Country Reports 2006, 153; 10.5089/9781451828429.002.A003

Source: USCB (2004).

• Download Figure
• Download figure as PowerPoint slide

74. HIV/AIDS cuts life expectancy by a third. As a result of HIV/AIDS mortality, Namibia has at present lost about 25 years of life expectancy at birth. Instead of expecting to live until 70—if AIDS hypothetically did not exist—life expectancy is estimated to be currently between 44 and 46 years, according to USCB and UN estimates, respectively.³³ The loss in life expectancy due to AIDS is much smaller in Nigeria (a country with a much lower HIV/AIDS prevalence rate of 5 percent) and comparable in South Africa (with a similar prevalence rate).

75. The dependency ratio—the number of children and people of retirement age divided by the adult population available to support them—is projected to decline (Table III.1). While the mortality among adults increases with HIV/AIDS, which alone would lead to an increase in the dependency ratio, the mortality among infants and children also increases due to the transmission from mother to child, which produces a counter effect on the ratio. In addition, fewer people reach retirement age due to the epidemic. These two effects explain why, for Namibia, both the UN and the USCB model project the dependency ratio to decline in the future. Economically, this trend means that the same number of working people will have to finance grant schemes for fewer dependent people.

View Full Size

• Download Figure
• Download figure as PowerPoint slide

76. From a demographic point of view, the old-age pension scheme seems to be sustainable. Narrowing the focus to the number of people at retirement age relative to the working-age population (thereby excluding children from the dependency ratio), both the UN and the USCB model yield the following analysis. In 2005, 10 old-age persons faced 100 working-age persons. This ratio is projected to marginally change to 11 old-age persons per 100 working-age persons in 2015 and to 12 in 2025. Therefore, both demographic models predict that the ratio of working-age persons financing old-age pension recipients is expected to change only slightly.

D. Fiscal Sustainability of the Universal Pension Grant

77. In general, HIV/AIDS exerts an upward pressure on the fiscal deficit. The high prevalence of HIV/AIDS reduces population growth and destroys human capital. In the hypothetical No-AIDS scenario, the labor force would be significantly larger and more productive (see Box III.2). As a consequence, AIDS negatively affects the tax base and government revenue. The government is also confronted with the need for higher spending on social services to take care of orphans and treat the sick. Overall, HIV/AIDS exerts an upward pressure on the fiscal deficit. However, the analysis in this section will show that the old-age pension system does not contribute to increasing deficits over time. Using the demographic projections discussed above in the fiscal sustainability analysis, different simulations indicate a constant or declining cost in terms of GDP over the next 20 years, provided that the benefit is not made too generous.

78. Five expenditure scenarios are computed, each based one a differing combination of assumptions (Figure III.4) for (i) the number of persons receiving the old-age pension, (ii) the development of the level of the old-age pension, and (iii) the growth of real GDP. To keep the analysis concise, the following combinations of assumptions are used to simulate five expenditure scenarios:

View Full Size

Namibia: Pension Expenditure, (in percent of GDP), 1995 to 2025

Citation: IMF Staff Country Reports 2006, 153; 10.5089/9781451828429.002.A003

Source: Staff calculations.

• Download Figure
• Download figure as PowerPoint slide

• Baseline Scenario (baseline GDP growth; UN population projections; pensions increase with inflation).

• Baseline With Output Shock (like the Baseline, but GDP growth is one standard deviation lower);

• Alternative Scenario 1 (low GDP growth, reflecting an unsuccessful implementation of MTP III; USCB population projections; pensions increase with inflation);

• Alternative Scenario 2 (like the Baseline, but with pensions rising by 1½ times the inflation rate);

• Alternative Scenario 3 (like Alternative Scenario 1, but with pensions rising by 1½ times the inflation rate).

The proportion of administrative costs³⁴ and the level of coverage³⁵ are assumed to be constant when computing the expenditure projections. In addition, the following details underlie the assumptions.

79. Number of pension recipients: The factual USCB and UN growth projections³⁶ for the number of old-age people are applied to the existing number of recipients to compute a series of old-age pensioners. While the UN projections are used in the Baseline Scenario (and Alternative Scenario 2) to represent a proxy for a successful implementation of the authorities’ plan to combat HIV/AIDS, the USCB projections are incorporated in Alternative Scenarios 1 and 3 to simulate a more pessimistic outcome of the fight against the epidemic.

80. Development of the level of the old-age pension: The Baseline Scenario (and Alternative Scenario 1) assume that the level of the old-age pension will increase with consumer price inflation. This is based on the historical observation that the change in the monthly pension tracked the inflation rate (Figure III.5); without having been formally indexed to it. The future consumer price levels used here to simulate a path of the pension grant are based on the projections in the accompanying Staff Report.³⁷ Alternative Scenarios 2 and 3 incorporate a pension path that is based on a growth rate of 1½ times the inflation rate to test for sensitivity.

View Full Size

Namibia: Consumer Price Inflation and Monthly Pension, 1994 – 2005

Citation: IMF Staff Country Reports 2006, 153; 10.5089/9781451828429.002.A003

Source: Staff calculations.

• Download Figure
• Download figure as PowerPoint slide

81. Growth of real GDP: The Baseline Scenario (and Alternative Scenario 2) incorporate the path of real GDP which is based on the baseline projections in the Staff Report.³⁸ Alternative Scenarios 1 and 3 assume a lower path of real GDP, which takes into account the growth effects of a failure to successfully implement the authorities’ strategy to combat HIV/AIDS (see Yackovlev, 2005; Box III.2). In addition, the Baseline With Output Shock incorporates a path of GDP subject to a permanent shock of one standard deviation of historic growth. The three different real GDP paths translate into different nominal GDP series—in order to compute a nominal expenditure to nominal GDP ratio—by using the same GDP deflator series.

82. The fiscal sustainability of the old-age pension is not endangered as different scenarios indicate a declining or constant cost in terms of GDP over the next 20 years (Figure III.4).

• The Baseline and the Alternative Scenario 2 incorporate a combination of Staff Report growth and UN population projections, assuming that the authorities’ strategy to combat HIV/AIDS is implemented. The Baseline Scenario yields a declining trend for the cost of the pension grant in terms of GDP, starting at 1.2 percent of GDP in 2005 and declining to 0.9 percent of GDP in 2025.³⁹ The difference in Alternative Scenario 2 (namely that the pension level is changed with 1½ times the inflation rate) results in a roughly constant cost for the simulation period.

• Alternative Scenarios 1 and 3 assume low GDP growth and USCB population dynamics, consistent with a more pessimistic HIV/AIDS development. Alternative Scenario 1 projects a declining expenditure to GDP ratio, while Alternative Scenario 3 simulates an average expenditure to GDP ratio of 1.2 percent. The difference is again due to the different adjustment mechanism in the pension level.

83. The old-age pension system does not generate unsustainable fiscal pressures. Using alternative population projection models does not yield a difference in the trend of costs. While the Baseline Scenario assumes an increasing number of pensioners, it also assumes relatively high GDP growth. In contrast, Alternative Scenario 1 is based on lower growth in the number of pensioners as well as lower GDP growth. Both scenarios project the same declining trend in costs. The adjustment process of the pension level proves to be more important. Adjusting the pension along with inflation ensures a declining trend in the expenditure to GDP ratio, while increasing the pension with 1½ times the inflation rate only keeps expenditure constant. An even higher adjustment will yield an increasing trend in the expenditure to GDP ratio once the combined growth of both grant recipients and the pension level outpace nominal GDP growth.

84. The largest danger to fiscal sustainability arises from overly generous adjustments to the level of the pension benefit. This is illustrated in Alternative Scenarios 2 and 3. The authorities should thus carefully reflect on the fiscal consequences of any adjustment. The pension expenditure to GDP ratio could act as a useful guide for policymakers. The simulations indicate the sustainability of an inflation-indexed pension path, which maintains the pension’s purchasing power. It might be an option to define a basket of goods and services that reflects the poor’s consumption decision more closely than the overall price index. The adjustment of the pension grant could then be guided by the change in this newly defined index.

85. Means-testing should only be introduced if the additional administrative costs do not outweigh the savings achieved through reducing the number of recipients. From a conceptual point of view, restricting the pension scheme to those in need can only be recommended. However, as pointed out earlier, it is not clear how the introduction of a means-test affects net costs. This procedure would reduce the number of recipients and at the same time increase costs to establish a sophisticated administrative system. Therefore, a more targeted approach depends on the possibility to implement a relatively simple and cost-effective means-test.

E. Conclusion

86. In sum, the non-contributory pension scheme is an important and reasonably efficient safety net tool to combat poverty. The universal pension system in Namibia has the following advantages: (i) it reaches old-age persons, who work(ed) in the informal economy and who live in rural areas; (ii) it does not crowd out inter-generational support; (iii) it has relatively low administrative costs; and (iv) it is fiscal sustainable.

87. The absence of unsustainable fiscal pressures is one of the key features of Namibia’s old-age pension system. Demographic developments are not expected to substantially worsen the ratio of pensioners to the working-age population. While HIV/AIDS reduces population growth, destroys human capital, lowers productivity growth, and exerts upward pressure on the fiscal deficit, it does not endanger the sustainability of the non-contributory pension system. The analysis in this chapter shows that the old-age pension system does not contribute to increasing fiscal deficits over time, provided that the authorities continue to show prudence in adjusting the level of the pension benefit over time. If this assumption holds true, the various scenarios in this chapter project a constant or declining trend of costs in terms of GDP over the next 20 years.

88. Conceptually, further efficiency gains could be made if means-testing was introduced. However, this procedure should only be adopted if the savings from restricting the pension to the needy is not overcompensated by the additional administrative costs necessary to enforce the means-test.