The Heavily Indebted Poor Countries Initiative: Does it Work?

(1)

The Heavily Indebted Poor Countries Initiative:

Does it Work?

Anieka van Leeuwarden

S1198424

Groningen, August 2007

Supervisor: Dr. ir. D.J. Bezemer

(2)

The Heavily Indebted Poor Countries Initiative: Does it Work?

Abstract

(3)

Abbreviations

AfDF African Development Bank

BWIs Bretton Woods Institutions

HIPCs Heavily Indebted Poor Countries

IMF International Monetary Fund

IDA International Development Association

MDGs Millennium Development Goals

MDRI Multilateral Debt Relief Initiative

NGOs Non-governmental Organizations

NPV Net Present Value

ODA Official Development Assistance

PRGF Poverty Reduction and Growth Facility

(5)

The debt problem of poor countries is nowadays a ‘hot’ topic and has attracted considerable public attention. Almost everyone has heard of the Jubilee 2000 campaign or the Make Poverty History campaign and not to mention the activities of Bono, the lead singer of U2, who made it his personal goal to attract the attention of the whole world to the poverty crisis. These actions and actions of other non-governmental organizations (NGOs) all fight for complete debt cancellation for Heavily Indebted Poor Countries (HIPCs) to make an end to extreme poverty. Under continuing pressure the World Bank and the IMF launched the HIPC Initiative in 1996 to cut back debt to sustainable debt levels and enhanced the initiative in 1999 to focus more on poverty reduction. Debt relief should reduce debt servicing costs and represent an increase in funds available to government which can be spent on poverty reducing expenditures.

Debt relief is not a new policy. Estimations suggest that, during the 1980s, the HIPCs as a group received US$6 billion in debt forgiveness and US$1 billion in debt-stock reduction. In 1990-96, these amounts jumped to about US$19 billion and US$6 billion, respectively (Development Committee, 1998). These debts were granted by Paris club creditors and reduced bilateral and commercial debt, but not multilateral debt, and the debt burdens of many low-income countries continued to grow, particularly in Sub-Saharan Africa (Addison et al, 2004). The HIPC Initiative is the first time that multilateral debt became eligible for debt relief.

(6)

and political history of a country will be of great influence together with poor economic management, weak governance, and external factors like war, climate problems and deterioration of terms of trade. Serious questions remain about the ability of poor countries, with weak institutions, to achieve a fast and effective shift of the resources into better services and infrastructure for the poor. Easterly (2001) even claims that debt forgiveness grants aid to those recipients that have been best proven their ability to misuse that aid. He states that the cause of the debt problem was not just bad luck for HIPCs. These countries were not any more likely than non HIPCs to be at war and there was no massive terms-of-trade deterioration for these countries during the past few years. It may be that countries that borrowed heavily did so because they were willing to mortgage the welfare of future generations to finance this generation’s (mainly government’s) standard of living. He refers to this as irresponsible borrowing and countries running their economies through irresponsible policies should not be rewarded by debt relief. Debt relief would lead in this case to wrong incentives for a continuing of those policies, to the benefit of the powerful ‘elite’ and to the detriment of the population as a whole.

The HIPC Initiative is designed to free up resources for an increase in poverty reducing expenditures like healthcare and education. These policy measures are undoubtedly important elements of any poverty reduction strategy but according to Nissanke and Ferrarini (2004) the unfounded expectation that poverty can be reduced by applying these measures only should not be encouraged, as poverty is the outcome of economic, social and political processes and their interactions. So can we really expect poverty reduction as a result of debt relief?

According to Bono and all the campaigning NGOs debt relief is the answer to end poverty but still there is a lot of criticism to the Initiative of the World Bank and the IMF which was merely a respond to their own request. Damaging conditions like liberalization and privatization are said to be undermining the ability of African governments to look after the welfare of their populations and the reform programs focus too narrowly on macroeconomic stabilization and not enough on poverty reduction.1 Another issue of worrying is that the funds donor countries devote to writing off poor countries debt will be counted as aid. Some of the ‘grants’ figures today already include official debt reduction which suggests that donors see grants and debt reduction to some extent as substitutes (Birdsall and others, 2004).

(7)

(8)

Chapter 2: The HIPC Initiative

(9)

The process of the Initiative

The first step towards qualifying for HIPC assistance is the entry requirement. The Initiative is open to all HIPCs that pursue or adopt programs of adjustment and reform supported by the IMF and IDA (World Bank and IMF, 1998). In addition, for a country to be eligible for the Initiative, it must meet the criteria that its end-December 2004 external public debt exceeds 150 percent of its exports. Figure 1 presents the process of the Initiative as fully described below.

Figure 1

Source: http://siteresources.worldbank.org/INTDEBTDEPT/Resources/review-HIPCProcess.pdf

(10)

emphasizing infrastructure programs in the areas of water, roads, electricity and telecommunications. Some strategies also propose: providing houses for the poor; and strengthening social safety nets to include food subsidies or other food security programs, social assistance programs, labour intensive public works, and food for work programs (IMF and IDA, 2001-a). Furthermore, countries should achieve satisfactory performance under the Poverty Reduction and Growth Facility (PRGF) of the IMF. The IMF established the PRGF to make the objectives of poverty reduction and growth more central to its lending operations. The targets and

policy conditions in a PRGF-supported program are drawn from the country’s PRSP2. When a

country reached its decision point, it may immediately begin receiving interim relief on its debt service. However, this debt relief is revocable if policy performance fails, in other words, countries should implement their PRSP and maintain macroeconomic stability.

In order to reach the completion point, countries must meet the following criteria; (a) establish a further track record of good performance under IMF-and IDA supported programs, (b) implement satisfactorily key reforms agreed at the decision point, (c) adopt and implemented the PRSP for at least one year.3 At the time of the completion point lenders are expected to provide the full debt relief committed at decision point. This debt relief is irrevocable. Since the beginning of the Initiative, 22 countries have been granted irrevocable debt relief which represent the amount of 30.208 billion U.S. dollars, in end-2005 NPV terms (IMF & IDA, 2006-b). Furthermore, 8 countries have reached their decision point (see table 1).

(11)

Table 1: Countries participating in the HIPC Initiative

* = decision point, # = completion point

2000 2001 2002 2003 2004 2005 2006

22 countries at completion point

Benin * july # march

Bolivia * feb # june

Burkina Faso * july # april

Cameroon * oct # may

Ethiopia * nov # april

Ghana * feb # july

Guyana * nov # dec

Honduras * july # april

Madagascar * dec # oct

Malawi * dec # end

Mali * sep # march

Mauritania * feb # june

Mozambique * april # sept

Nicaragua * dec # jan

Niger * dec # april

Rwanda * dec # april

São Tome

Principe * dec # end

Senegal * june # april

Sierra Leone # march # end

Tanzania * april # nov Uganda *feb # may

Zambia * dec # april

8 countries at decision point

Burundi * aug

Chad * may

Dem. Rep. Congo * july

Haiti * end

Rep. Congo * march

The Gambia * dec

Guinea * dec

Guinea-Bissau * dec

10 countries at pre-decision point

Central African Republic Comoros Côte d’Ivoire Eritrea Kyrgyz Republic Liberia Nepal Somalia Sudan Togo

(12)

The fiscal and monetary policy reform targets set for each individual country typically include; restructuring of government expenditure in favour of the social sectors; continued exchange and trade liberalization if needed; financial sector reforms; public enterprise privatization and the development of appropriate regulatory frameworks for private sector activity; and the removal of policy-induced distortions and inefficiencies in public utilities, agriculture and other productive sectors (Development Committee, 1998). These reforms are specified for each country individually.

At the start of the HIPC Initiative in 1996, a sunset clause was introduced to restrict the eligibility under the Initiative to countries that had started qualifying IMF-and IDA-supported programs within a two-year time period. The aims of this clause were to prevent the Initiative from becoming permanent, minimize the moral hazard problem of excessive borrowing in anticipation of debt relief, and to encourage early adoption of reforms. This sunset clause was extended (for the fourth time) eventually to the end of December 2006. As this date passed, the final list of countries consists of the 40 listed in table 1, which all meet the income and indebtedness criteria of the Initiative. However, not all of these countries have adopted programs of adjustments and reforms yet.

(13)

HIPC Initiative for the 40 HIPCs is estimated (in August 2006) at US$ 63.2 billion in end-2005 NPV terms. In addition, the cost of MDRI debt relief to these countries is estimated at US$ 24.9 billion in end-2005 NPV terms (IMF and IDA, 2006-b).

Table 2: Proportion of debt due to the IMF, IDA and AfDF.

Public and publicly guaranteed (PPG) External Debt, in current million U.S. dollars

2005 of which owned by the creditors of the MDRI

Benin 1520.5 1150.2 75.7% Burkina Faso 1921.1 1403.9 73.1% Bolivia 5114.0 2079.0 40.7% Ghana 6073.8 5400.1 88.9% Mali 3440.0 2182.0 63.4% Tanzania 7114.8 4783.8 67.2% Uganda 4422.0 3802.0 86.0%

Source: IMF Staff Assessments of Qualification for the Multilateral Debt Relief Initiative, December 8, 2005.

Progress of the Initiative

At the beginning of 2007, 22 countries have reached their completion point. This means that all of these countries have received or are receiving irrevocably the debt relief committed at decision point. From the 8 decision point countries Burundi, Haïti and the Republic of Congo have only reached this point recently. The other five decision point countries (Chad, Dem. Rep. of Congo, Gambia, Guinea, and Guinea-Bissau) still did not reach completion point because they have been experiencing difficulties in the implementation of their macroeconomic programs and haven’t implemented their PRSP for at least one year.

(14)

Table 1 shows that the speed at which countries move through the HIPC process differs over countries. Weak budget execution and poor policy implementation have often slowed the achievement of the agreed triggers and preparing fully participatory PRSPs has taken longer than expected, given that many countries lack the institutional and human resource capacity needed to prepare such documents. Furthermore, progress in several pre-decision point countries has been hindered by internal conflict, governance issues, or protracted arrears. For example, Côte d’Ívoire reached its decision point under the original HIPC Initiative already in 1998, but the security and political stability deteriorated soon afterwards which pulled the country back to pre-decision point again.

The impact of the Initiative

A recent study by the World Bank’s Independent Evaluation Group, “Debt relief for the Poorest:

An evaluation update of the HIPC Initiative (2006)” assesses the impact of the Initiative.

First, to free up resources for poverty reducing expenditure, it is necessary that the debt relief given is additional to other aid transfers. This additionality is not easy to measure. Creditors use a variety of measures to account for debt relief which results in unreliable data. And more fundamentally, assessing whether debt relief is additional requires assumptions about the resources that would have been transferred in the absence of debt relief. Several studies up to 2000 found that debt relief has not been additional but the latest assessment found that net resource transfers to HIPCs doubled from US$8.8 billion in 1999 to US$17.7 billion in 2004, while transfers to other developing countries grew by only one-third.

Second, to achieve the objectives of the Initiative, it is necessary that all creditors will participate and deliver the promised level of relief. The Initiative remains voluntary in nature and full creditor participation has therefore been difficult to achieve. The main architects of the Initiative, the World Bank, the IMF and the Paris Club, have committed their full share of relief, which

counts for 64% of the total estimated costs4, but participation of non-Paris Club members and

commercial creditors has been limited.

4_{40.7 out of 63.2 billion U.S. dollars, in end-2005 NPV terms, for all 40 participating countries.}

(15)

Third, an important objective of the Initiative is to reduce poverty by freeing up resources for higher social spending. It is clear that the reporting of such spending has improved in participating countries since this is part of the PRSPs written and implemented during the process. But several problems arise concerning this reporting. Social spending is defined differently across countries and over the years countries may have included more expenditures as ‘poverty reducing’ which could bias the increase in these expenditures. Furthermore, data is still very limited. Figure 2 shows the change in poverty reducing expenditures as a positive result from the HIPC Initiative, but clearly this should interpreted with care due to a change in reporting and more important, this figure does not show what would have happened without the Initiative and what happened in non-HIPC countries. In the following chapters this issue will be further discussed. Another remarkable feature of this figure is the chosen sample of countries that had reached decision point as of March 2006. It is obviously not possible to measure the result after HIPC in 2005 while some of the sample countries have not even received a dollar debt relief until 2006. The latest assessment concludes that the PRSPs and the Initiative have focussed heavily on the increase in social spending and maybe not enough on a balanced approach to growth and poverty reduction.

Figure 2

(16)

Fourth, the Initiative is supposed to provide a sustainable debt position for participating countries. The Initiative has contributed to reducing the debt stocks of completion point countries but this does not ensure long term sustainability. Figure 3 shows the reduction in debt service and debt stock over the years of HIPC which indicate an obvious decrease but when this is compared to, for instance, the actual debt service paid in 1999 (US$ 3437.6 mill.) and in 2005 (US$ 3425.0 mill.), not found in this table, there appears no significant difference5. Some of the reasons for this phenomenon are described in the review of literature. Furthermore, at completion point, HIPCs’ NPV of debt-to-exports ratios were, on average, 38.9 percentage points higher than expected at decision point. This was caused by exogenous changes in discount and exchange rates and by some new borrowing. In eight completion point countries debt ratios once again exceed HIPC thresholds. It can be concluded that debt reduction alone cannot ensure debt sustainability but this has to be accompanied by other efforts to improve repayment capacity. This is further discussed in chapter 3.

Figure 3

Source: http://go.worldbank.org/XLE8KKLEX0

(17)

Chapter 3: Review of the literature

In this chapter, an overview is given of several subjects regarding the assessment of the impact of participation in the HIPC Initiative and the reduction of debt and debt service on economic growth and poverty reducing expenditures. This overview gives insight into how economic growth and poverty are affected by debt and what the role of social spending is.

Debt and growth

One of the motivations for debt relief stems from the presumed impact of a heavy debt burden on per capita income growth. In the second half of the 1990s, policy makers recognized that earlier debt programs did not provide sufficient results and a new program was necessary since very high levels of debt were seen as a contributing factor to the extreme poverty of low income countries. This gave cause for creating the HIPC Initiative. The channel through which debt affects growth is a considerable subject of research over the last years. In 1992, Eaton suggested that ‘reasonable’ levels of borrowing by a developing country are likely to enhance its economic growth. Developing countries have small stocks of capital and are likely to have investment opportunities with high rates of return. As long as the borrowed funds are used for productive investment and the countries do not suffer from macroeconomic instability, policies that distort economic incentives, or external shocks, economic growth should increase and allow for timely debt repayments.

(18)

efficiently used borrowed funds and the focus on long term results for the entire population. Furthermore, they say that external debt service can affect growth by crowding out private investment or altering the composition of public spending. Higher debt service can raise budget deficits and reduce public savings which in turn may raise interest rates or crowd out credit available for private investment, and so discourage economic growth. Public spending may also be limited so the amount of resources available for infrastructure and human capital decreases which has again negative effects on growth.

Pattillo and others (2004) argue that large debt stocks may have negative effects on economic performance especially in HIPC countries. This is because of the uncertainty of debt relief, debt rescheduling and additional lending. It is not clear what portion of the debt will actually be serviced with the country’s own resources. This uncertainty may reduce the level for long term investment and investment that does take place may be poorly allocated to activities with quick returns. Iyoha (1999) studied the impact of external debt on economic growth in sub-Saharan Africa for the period 1971-1994 and her research confirms that an excessively high stock of external debt depresses investment and lowers the rate of economic growth. Furthermore, Clements and others (2003) expect that the reduction in the stock of external debt projected for HIPCs would directly increase per capita income growth by about 1 percentage point per annum. Their expectations are based on research on low-income countries with data covering the period 1970-1999. They found that stock of debt does not appear to depress public investment but a reduction in debt service does provide an indirect boost to growth through their effects on public investment. Hussain and Gunter (2005) find that HIPC debt relief has boosted economic growth in 18 HIPCs by 2.9% and that much of the positive impact has been eroded due to recent deteriorations in terms of trade and growth has been reduced to an average of 0.9%. However, this research is based on the negative effect of debt service on economic growth and by simply reversing the sign of the values will give them the benefits of HIPC debt relief. So they did not actually find that the HIPC Initiative will boost economic growth but a high debt service depresses economic growth, as we found earlier.

(19)

the negative effects of debt are taken away and a country is able to service its debts again. For developing countries it is important that debt will be reduced to sustainable levels before economic growth can resume.

Debt sustainability

The sustainable level for debt according to the World Bank6, is an external public debt of not

more than 150 percent of exports (or in some cases 250 percent of fiscal revenue7). Pattillo and others (2004) found indeed that by reducing the net present value of debt to 150 percent of exports, debt reduction would eliminate the negative effect of debt on growth. However, this is not the growth maximizing level of debt. By setting this level as sustainable, the HIPC Initiative reduces the level of indebtedness just to a level where a new increase in debt (which may be likely, given the level of development and poverty of these countries) would have a negative impact on growth.

Noreena Hertz (2004) does not agree with the way debt sustainability is calculated on the basis of exports and not on what a country could reasonably afford to pay. ‘Given that external shocks to world commodity prices were not built in to World Bank and IMF calculations, and given that most of the countries on the list of HIPCs continue to rely on a small list of volatile and vulnerable export commodities, the export projections used to make the HIPC calculations were completely overstated’. Also Hussain and Gunter (2005) say that external shocks are an important factor why HIPCs will not reach long-term sustainability. While the HIPC Initiative has led to marked reduction in debt indicators in recent years, the pattern of trade for participating countries (especially in Africa) and its specialization in primary product exports call into question if this reduction is sustainable. Falling prices of primary exports have large negative effects on external indebtedness through increasing the domestic currency equivalents of such debts, as well as indirectly through increasing debt service ratios. Currency depreciation encourages domestic supply of the same primary commodities and reduces the price of primary commodities even further so that exchange earnings may not improve, leading to further indebtedness. Second, unless debt relief is used effectively to boost exports relative to imports or attract more private

6_{Source: http://web.worldbank.org/WBSITE/EXTERNAL/NEWS/.html}

7_{As long as a country has both (a) an export-to-GDP ratio of at least 30 percent and (b) a government}

(20)

capital, the debt problem will re-emerge if countries attempt to grow at rates higher than its balance equilibrium growth rate (current export earnings equal imports and debt repayments). More debt is needed to finance imports and debt repayments. Easterly (1999) found in his research on HIPCs in the years before the HIPC Initiative (1989-1997) that new borrowing in these countries was correlated with debt relief so that debt ratios actually got worse. In other words, new borrowing was the highest in countries that got the most debt relief.

HIPC countries encounter debt problems for many different reasons, at very different levels and with very different consequences for their economies and the welfare of the poor. According to Hjertholm (2003), considering this diversity of debt problems and economic circumstances, it would not be justified to expect that the sustainable level for debt ratios would be the same for each country. This results in allocating available funding inconsistent with individual country needs. Some low-income debtor countries could receive manageable debt positions; in other cases the sustainability targets may not be low enough.

The Independent Evaluation group of the World Bank (2006) indicated that in at least 11

post-completion point countries8, out of 22, the key indicator of external debt sustainability has

deteriorated since completion point. In 8 of these countries9, the ratios once again exceed HIPC thresholds. Changes in discount and exchange rates caused an increase in debt ratios and the effect of improved exports has been offset by new borrowing. This result, together with he reviewed literature, leads to the conclusion that debt reduction alone is not a sufficient instrument to affect the multiple drivers of debt sustainability. Sustained improvements in export diversification, fiscal management, the terms of new financing, and public debt management are also needed, measures that fall outside the ambit of the HIPC Initiative.

Debt and social spending

The HIPC initiative requires that the participating countries adopt a poverty reduction strategy which includes that debt service savings due to the HIPC initiative are directed to pro-poor social sector spending, mostly targeting primary education and health services for the poor. This

8_{These countries are: Uganda, Bolivia, Tanzania, Burkina Faso, Mauritania, Benin, Niger, Nicaragua, Guyana,}

Mozambique, and Ethiopia.

(21)

increase in social sector spending is expected to reduce poverty (in the long-term) and to increase economic growth. The IMF and IDA (2001-a) indicate that public spending is a critical instrument for poverty reduction in developing countries. However, to have a significant impact on poverty, it must be budgeted and disbursed for activities that help the poor expand their access to resources and their income-earning potential.

Gupta and others (2001) provide evidence that the poor are more favorably affected by public spending on healthcare than the non-poor and the relationship between public spending on healthcare and the status of the poor is stronger among low-income countries than in other countries. This indicates that there may be higher returns to health spending in low-income countries, when compared to other countries. Heltberg and others (2004) measure a progressive distribution in public services in Mozambique. Public services seem to reduce inequality in welfare, which means that there are possibilities for directing welfare benefits to poor people through well designed public interventions in the health area, education and road infrastructure. In Mozambique it is likely that debt relief will have significant poverty-reducing effects by helping poor people access public services. In contrast, Castro-Leal and others (1999) found that public spending on education and healthcare in several African countries favor not the poor but those who are better off. However, they acknowledge that although the effectiveness of public social is not optimal, these basic services are still essential in any effort to combat poverty. Immediate evidence for sub-Saharan Africa over the last twenty years suggests direct relationships between generally improving social indicators in the region and rising social spending levels. Lopes (2002) analyzes trends in social indicators in sub-Saharan Africa and their correlation with government social spending. Research shows that social government spending both per capita and as a percentage of GDP to be of relevance to social outcomes.

(22)

service’s crowding out of governments’ social spending. They argue that high debt service directly reduces government budgetary allocations on health, education, social safety nets, and water and sanitation, in part because governments find it politically easier to cut back spending in such sectors because the poor are not usually organized to have a voice in such decisions. This effect should be decreasing since the poor are more and more represented by NGOs and monitors of the Millennium Development Goals although this monitoring is not well organized yet.

(23)

been meeting their full debt service obligations before the Enhanced HIPC Initiative. When actual debt service paid was small or negligible due to accumulation of external arrears, HIPC assistance may be associated with increased debt service payments resulting from the regularization of relations with creditors.

Conclusion

As Eaton (1992) sums the requirements for leading a ‘reasonable amount’ of borrowed funds to enhance economic growth, this literature shows that most of those requirements do not hold for developing countries; funds borrowed are not used for productive investment, countries do face external shocks, HIPC brings policies that distort economic incentives, like investment opportunities, and macroeconomic instability is not an exception. So from this point it is clear that a high stock of debt does not stimulate growth and that countries need debt service relief in order to reach this growth. However, it is not clear if debt relief provided by the HIPC Initiative is given on the right terms and conditions to lead to economic growth. Earlier research indicates many hiccups for this process. In the empirical research following in the next chapters, a comparison with non-HIPCs will be made to see whether participating countries receive a higher growth from debt relief than non-HIPCs, and also to find if debt (service) relief is the most important variable of influence on economic growth or could it be simply the fact that countries participate in the Initiative.

(24)

Chapter 4: Methodology

Much research has been performed on specific conditions and regulations of the Initiative, why these could be good or bad for participating countries, but there is no clear empirical research if the HIPC Initiative actually meets it objectives. Evaluation reports show a decrease in debt burdens due to the Initiative but do not show the linkage between these changes in debt related to the objectives of the Initiative. Furthermore, former research merely focuses on HIPCs solely, or at developing countries in general. This means that the question what would happen without the Initiative is marginally assessed. In this research the objective of promoting growth by the Initiative and the objective of increasing social spending are being assessed. This leads to the following research questions:

Does the HIPC Initiative promote economic growth for participating countries?

And

Does the HIPC Initiative promote higher social spending for participating countries?

In order to answer these questions, the influence of participation in the HIPC Initiative will be measured on GDP and social spending, together with debt reduction and debt service reduction over the period of the Initiative. A distinction will be made between countries participating in the Initiative and countries which are not. The model used in this research is referred to as the linear least-square method. The method of least squares minimizes the total squared errors of the estimations. Multivariate regressions will be done merely in SPSS and for some additive results in Eviews.

Dataset

A major struggle in doing research on economic and social indicators in less developed countries is the scarcity of useful and consistent data. Data series used in this research were taken from the World Bank’s World Development Indicators (WDI) 2006, The United Nations Statistics Division10, and the statistical update 2006 of the IMF11.

(25)

The dataset used in this research should at least include all participating countries in the HIPC Initiative. These 40 countries are listed in table 1, as published by the World Bank. To measure the significant influence of the HIPC Initiative, a comparison should be made so countries which are not participating should be included as well. To cover this, the list of Least Developed Countries 2006, put together by the United Nations is used as a basis for the used sample, supplemented by the HIPC countries not covered by this list. The countries Afghanistan and Tuvalu are excluded from the list since there is no data available on these countries. This results in a dataset of 57 countries, see appendix 1.

Computing variables

HIPC dummy variable

To determine the influence of participation in the HIPC Initiative, it should be decided which countries to include in the variable HIPC. Difficulties arise since most countries reached decision point and completion point on different dates so they are all in different phases of the Initiative. Table 1 indicates in which stage of the Initiative the countries are placed at this moment. The dummy-variable HIPC-1 includes all countries participating in the Initiative (40). All these countries passed the entry requirements and started the phase of restructuring in compliance with IMF conditionalities so any effect due to the Initiative could be measured on economic growth and social spending. However, this variable also includes countries which did not reach decision or completion point yet or which reached these points only recently (and there is no data available afterwards), so it is not possible to really measure what has happened after provided debt (service) relief for this whole group. Another variable for HIPC is constructed which consist of only those countries which passed at least the first period of the Initiative and reached decision point in the years 2000 and 2001. This results in the variable HIPC-2 which consists of 24 out of the 40 eligible HIPCs. In the future it would be better to assign the countries which passed the whole program to a different dummy variable to understand the consequences of each period of the Initiative separately. However, now this would result in only a few cases.

Dependent variables and Independent variables

(26)

changes started in this period, growth and debt figures used in these regressions are calculated as the percentage difference between the average before this period and the average after this period. For the period before HIPC, an average is taken of the years 1990 until 1999. The period of 2002-2004 is used to calculate the averages after HIPC. The period after HIPC consists of a considerable shorter period than before due to data limitations. Unfortunately the results could be influenced due to heavy fluctuations in this shorter period. To overcome the problem of fluctuations over the years in the early period, this period is taken considerably longer. The dependent, independent and control variables are calculated using the following formula, when used in the regressions for economic growth. See appendix 2 for an overview of the individual variable calculations.

ΔY = [(Ya – Yb) / Yb] * 100

Where ΔYis the growth rate for the variable, Ya is the average level of the variable in de period

after HIPC (2002-2004) and Yb is the average level in the period before HIPC (1990-1999).

The data series used for the hypotheses concerning social spending are only available for the year 1998 and later. This makes it impossible to measure a difference over a distinctive period before and after. In this case growth of the variables is reproduced by the average of the annual percentage change taken over the years of HIPC. In general these are the years 1998-1999 until 2004-2005. The dependent, independent and control variables are calculated using the following formula when used in the regressions for social spending.

ΔY annual= 1 100 * ] / ) [( 2004 / 2003 1999 / 1998 1 − −

∑

+ n Y Y Yt t t

Where ΔY annual is theaverage annual growth of the variable, Y,t+1 is the level of the variable in

the year of observation plus one, Y,t is the level of the variable in the year of observation and n

(27)

Models and Hypotheses

Economic growth

In the first two models, the influence of whether or not a country participates in the Initiative will be tested on economic growth together with the influence of change in debt figures. Both debt stock and debt service are included since these are the pillars of the Initiative and the best measurable changes the Initiative directly influences. HIPC-1 and HIPC-2 do not complement each other so these dummy variables are used in different formulas.

ΔGDPj = α0 + α1 HIPC-1 + α2 ΔD + α3 ΔD*HIPC-1+ α4 ΔDS + α5 ΔDS*HIPC-1 + α6 C + ε (1)

ΔGDPj = α0 + α7 HIPC-2 + α8 ΔD + α9 ΔD*HIPC-2+ α10 ΔDS+ α11 ΔDS*HIPC-2 + α12 C + ε (2)

Where:

• ΔGDPj represents the percentage change in GDP level between the period before HIPC

and after HIPC for every country j.

• HIPC-1 is a dummy value with value 1 if a country participates in the Initiative.

• HIPC-2 is a dummy value with value 1 if a country participates in the Initiative and has reached decision point in 2000 or 2001.

• ΔD and ΔDS represent the percentage change in debt (service) between the period before HIPC and after HIPC.

• C represents a set of control variables.

• The α’s are the estimated parameters in the regression equation (regression coefficients). α0 is the intercept and the other α’s are the slopes of the regression line.

• ε represents a common error term with mean zero.

(28)

This leads to the following hypotheses:

Hypothesis 1: Participation in the HIPC Initiative has a positive influence on the growth in GDP

within a country.

Hypothesis 2: A decrease in external debt has a positive influence on the change in GDP and

within HIPC countries this effect is larger.

Hypothesis 3: A decrease in debt service has a positive influence on the change in GDP and

within HIPC countries this effect is larger. Social spending

The third objective of the HIPC Initiative is to reduce poverty by freeing up resources for higher social spending. These resources are referred to as ‘Poverty Reducing Expenditures’. The definition of poverty-reducing spending will necessarily be country specific. It is important for countries to recognize that different types of public spending can contribute to poverty reduction, whether through their impact on the provision of direct services to the poor or through their impact on overall economic growth and security. Certain outlays (such as those for primary education and basic health care) tend to be counted as poverty reducing in all countries. However, others (e.g., for infrastructure and housing) may be specific to some countries (IMF and IDA, 2001-a). Unfortunately data on Poverty Reducing Expenditures is hardly available except for countries participating in HIPC. Since the objective of this research is to compare HIPCs with non-HIPCs, data referring to countries not participating in HIPC should be included as well. This is not possible for regressions concerning Poverty Reducing Expenditures so the model concerning this dependent variable doe not include the independent variable HIPC. An alternative for poverty reducing expenditures is a combination of health and education expenditures. As described above, these figures are the main part, and always included in the definition of ‘Poverty Reducing Expenditures’. Unfortunately data availability on education expenditures is marginal and would reduce our dataset considerably. On contrary, health expenditure data is available and will be used as an alternative for poverty reducing expenditures in model 4 and 5. The next model (3) is used to test the influence of changes in debt figures on poverty reducing expenditures.

(29)

Where:

• ΔPRE represents the average annual percentage change in Poverty Reducing Expenditures within a country, over the period 1999-2005.

• ΔD-annual and ΔDS-annual represent the average annual percentage change in debt (service) within a country, over the period 1998-2004.

• The α’s are the estimated parameters in the regression equation (regression coefficients). α0 is the intercept and α1, α2 & α3 are the slopes of the regression line.

• ε represents a common error term with mean zero. This leads to the following hypotheses:

Hypothesis 4: A decrease in external debt has a positive influence on the change in Poverty

Reducing Expenditures.

Hypothesis 5: A decrease in total debt service has a positive influence on the change in Poverty

Reducing Expenditures.

In the following models (4 & 5), health expenditures will be used as an alternative for poverty reducing expenditures. Since data of this dependent variable is sufficient for the whole sample, the independent dummy variable HIPC can be included again. This results in the same models as presented for economic growth except that the variables are calculated differently due to data limitations over the years, as described earlier. The influence of whether or not a country participates in the Initiative will be tested on health expenditures together with the influence of change in debt figures.

ΔHE = α0 + α1 HIPC-1 + α1 ΔD-annual + α2 ΔD-annual*HIPC-1 + α3 ΔDS –annual + α4 ΔDS

-annual*HIPC-1 + α5 C + ε (4)

ΔHE = α0 + α6 HIPC-2 + α7 ΔD-annual + α8 ΔD-annual*HIPC-2 + α9 ΔDS –annual + α10ΔDS

(30)

Where:

• ΔHE represents the average annual percentage change in Health Expenditures within a country, over the period 1998-2003.

• HIPC-1 is a dummy value with value 1 if a country participates in the Initiative.

• HIPC-2 is a dummy value with value 1 if a country participates in the Initiative and has reached decision point in 2000 or 2001.

• ΔD-annual and ΔDS-annual represent the average annual percentage change in debt (service) within a country, over the period 1998-2004.

• The α’s are the estimated parameters in the regression equation (regression coefficients). α0 is the intercept and the other α’s are the slopes of the regression line.

• ε represents a common error term with mean zero.

These models are based on the assumption that by participating in the HIPC Initiative, a country is better able to convert given reduction in debt (service) into an increase in health expenditures. So in the case of debt (service) reduction, the coefficients are expected to be negative, which results in a positive growth in health expenditures. When the coefficients of the interaction variable (ΔD-annual*HIPC) or (ΔDS-annual*HIPC) are also negative, they intensify the effect of the debt (service) reduction. This means that the change in debt figures has a greater impact in HIPCs than in non-HIPCs. The literature study indicates several reasons why it is not likely for this to happen, which will be addressed in the discussion of the results. For now, this research focuses on the assessment of the Initiative’s objectives and according to the objective of freeing up resources for social spending by providing debt (service) relief, models 4 and 5 should show the same signs for the coefficients of the variables involving debt.

This leads to the following hypotheses:

Hypothesis 6: Participation in the HIPC Initiative has a positive influence on the change in

Health Expenditures within a country.

Hypothesis 7: A decrease in external debt has a positive influence on the change in Health

(31)

Hypothesis 8: A decrease in debt service has a positive influence on the change in Health

Expenditures and within HIPC countries this effect is larger.

Control variables

Since the dependent variables, the indicators of economic growth and social spending, can be explained by more than just the HIPC Initiative and debt relief, several control variables should be included. These control variables are expected to contribute to the variance in the value of the dependent variables and are measured in the same way as the dependent and independent variables as mentioned.

In the model to assess economic growth, 4 control variables are used. First the dummy variable

Africa, with value 1 if a country is located in Africa, is included. The fact that location could be

an important factor leading to economic growth is subject of numerous studies. This can be a result of regional spill over effects, regional trade, regional policies and even climate. The HIPC Initiative and Africa are often mentioned in one sentence but 10 out of the 40 HIPCs are not located in Africa. This is reason enough to enclose the influence of location into the analysis. The second control variable to include is ODA. Official development assistance is a source of funding for social expenditures or public investment which both can lead to economic growth. Third,

export is included as a control variable. International trade is one of the most powerful forces

affecting the process of economic development and growth (Perkins et. al. 2001). And forth control variable for economic growth is primary education enrollment. Primary education is universal accepted factor leading to economic growth12.

The model for poverty reducing expenditures includes 4 control variables. The first is committed

debt relief, which is included by the assumption that higher future cash flows could influences the

decisions taken on current expenditures. The other three variables, ODA, exports, and GDP are all expected to influence funds available for public spending. The model for health expenditures includes the same control variables except for committed debt relief since this data is only available for the limited sample used for poverty reducing expenditures. The other difference is that health expenditures also includes the variable Africa, which does not contribute to poverty

(32)

reducing expenditures since only 4 countries from the used limited dataset are not African countries.

Data analysis

The linear least-square method, used in this research, is based on some assumptions concerning the used variables:

Collinearity

The linear least-square method requires the absence of collinearity. Collinearity, the situation where two or more of the independent variables (including the control variables) are highly correlated, can have damaging effects on multiple regression. It makes it risky to interpret the coefficients as an indicator of the relative importance of predictor variables (Cooper and Schindler, 2000). The variance inflation factor (VIF) will reflect the presence or absence of multicollinearity. This is a measure of the effect of the other independent variables on a regression coefficient. The statistic is equal to 1 divided by the tolerance. The tolerance shows the percentage of each variable that is not related to the other predictors. Lower tolerances, which indicate overlap among the predictors, results in higher VIFs. In this research, most reported VIF values are around one, which means that almost nothing of the variation is related to other predictors.

Normality

(33)

probability than 0.05, the null hypothesis of normality can not be rejected (Hill et. al. 2001). The figures for JB are given in the tables for regression results. Since all the found statistics are well below the critical value, it can be assumed that the data is normally distributed and suitable for the regression model.

Descriptive statistics

In appendix 3 an overview is given of the descriptive statistics of all the variables used in this research. Some of these descriptives indicate a very high standard deviation which can be caused by the presence of outliers in the variable. In the regressions, the casewise diagnostics function of SPSS identifies the outliers where the residuals are outside 3 standard deviations from the mean of the predicted values of the dependent variables. Outliers were deleted from the sample for each dependant variable separately. In the following tables of descriptive statistics the outliers were identified by means of box plots (appendix 4). Cases with values more than 3 times the interquartile range from the middle 50 % of observations were identified as outliers. This resulted in only very little losses on the dataset (see appendix 3).

In table 3 and 4 the dataset was divided according to the two HIPC dummy variables to make a comparison between HIPCs and non-HIPCs. The variables analyzed refer to the same variables used in the regressions from which the precise calculation is given in appendix 2. Note that all these figures indicate a growth over a certain period of time and not the actual values.

Table 3: Descriptives separated by HIPC-1(=all participating countries).

ΔGDP

Δ

GDP-annual Δ Debt

Δ

Debt-annual Δ Debt service

(34)

Table 4: Descriptives separated by HIPC-2 (=countries with decision point in 2000 or 2001).

ΔGDP

Δ

GDP-annual Δ Debt

Δ

Debt-annual Δ Debt service

Δ Debt service -annual Δ Health Expenditures N 24 24 24 24 24 24 24 HIPC-2 Mean 7.24 3.14 15.71* 1.33* -2.78* 4.10* 4.20 St Dev 23.26 4.30 29.80 3.99 52.70 12.63 4.70 non N 32 32 29 29 28 28 32 HIPC-2 Mean 15.25 4.30 37.74* 4.16* 28.83* 17.20* 3.90 St Dev 27.76 5.25 46.58 5.30 72.76 17.79 6.57

An Independent Samples t-test in SPSS measures whether the means of the variables are statistically significant at a 10% level. The results of these tests are given in appendix 5. Whether the differences between the means of HIPCs and non-HIPCs are statistically significant is indicated by *.

With the analysis of these tables, a few relationships should be taken into account. Following the objectives of the Initiative, the HIPCs are expected to experience a positive economic growth together with a decrease in debt and debt service figures and an increase in health expenditures. To show that these developments are due to participation in the Initiative, these relationships should be stronger in HIPCs than in non-HIPCs. On the other hand, literature addresses the reasons why a decrease in debt service is not to be expected and that the lack of created fiscal space will lower economic growth and does not lead to an increase in health expenditures.

Economic growth shows in both tables a higher increase in non-HIPCs, although this difference is not significant for HIPC-2. This result could be caused by the selection of the HIPCs since a lack of growth is often part of the origin of their debt problem. However, according to the Initiative, the countries which are further in process of the Initiative (especially HIPC-2) should starting to experience a higher growth, which is not happening. The question is if this will happen at all, considering the critics on the design of the Initiative, as discussed earlier.

(35)

but this does not give any guarantee that a sustainable debt position will be reached, not as long as countries can borrow unlimited new funds.

The figures for debt service growth are expected, again according to the objectives of the Initiative, to show a negative growth, which they do for ‘∆ debt service’. Debt service annual, however, still indicates a growth in debt service over the years 1998-2004. Furthermore, since HIPC-2 consists of countries which all received debt service relief, it could be expected that the decrease in growth is larger in this group than the group of HIPC-1, in which not all countries received debt relief yet, but the contrary is true. As mentioned in the literature review, this is caused by debt service payments in arrears or falling donor contributions.

(36)

Chapter 5: Results

As mentioned in the previous chapter, regression analysis will be used to see whether any positive results concerning the objectives of the HIPC Initiative can be found. The first table will present the results found for the objective of increasing economic growth and the two following tables will present the results concerning social spending, represented by poverty reducing expenditures and health expenditures. The regression results are presented with and without control variables. At the bottom of each table, besides the Jarque-Bera and the number of observations (N), the regression statistics R-square and F are given. R-square represents the amount of variance explained by the predictors in the regression and F is calculated by dividing the mean square of the regression by the mean square of the residual. If this statistic is significant, the null-hypothesis that all the coefficients are equal to zero can be rejected.

The VIF value, which reflects the multicollinearity between the independent variables, is given for each variable in appendix 6. This table is an exact copy of the regression table for the particular dependent variable only then with just the VIF values. The values are all very low which means that only a very low percentage of the variation in each variable is related to other predictors. Between some variables there appeared to be a very high multicollinearity so the form of regressions had to be adjusted.

Economic Growth

Casewise diagnostics in SPSS indicates Equatorial Guinea (case 18) as an outlier. The residual of this outlier differed from the prediction by 6,953 standard deviations. This case will be deleted from the sample, when used for GDP.

In the regressions with independent variable HIPC-1, the variables service and

debt-service*HIPC-1 both have a VIF value of more than 40. This means that those two variables can

(37)

reason, regression 3 is only done with the variables debt and debt-service, and not the interaction variables.

Table 5: Regression results for Economic Growth, using dependent variable ΔGDP

ΔGDP HIPC-1 HIPC-2 1. 2. 3. 4. 5. 6. 7. Constant 27.999***** 18.445**** 26.386***** 22.787**** 15.255***** 12.767**** 7.823 (6.024) (8.414) (7.185) (12.885) (4.586) (5.616) (13.274) Predictors HIPC-1 -22.650***** -12.319* -21.638***** -14.058 (7.128) (9.346) (7.813) (11.758) HIPC-2 -8.012 -4.611 5.615 (7.005) (8.427) (10.140) Debt 0.143*** 0.065 -0.017 0.102*** 0.001 (0.074) (0.052) (0.087) (0.057) (0.094) Debt*HIPC -0.167*** -0.162 (0.085) (0.087) (0.206) (0.225) Debt service -0.036* 0.011 -0.052*** 0.024 (0.027) (0.044) (0.030) (0.050) Debt service *HIPC 0.046 -0.147 (0.116) (0.149) Africa -15.708** -24.812**** (10.584) (9.440) ODA -0.089 -0.214 (0.203) (0.234) Exports 0.121 0.202**** (0.094) (0.096) Primary education 0.181 0.114 (0.168) (0.190) R square 0.158 0.218 0.188 0.238 0.024 0.094 0.234 F 10.096***** 4.639***** 3.868**** 1.559* 1.308 0.996 1.118 Jarque-Bera 0.023 0.067 0.100 0.179 0.163 0.565 0.585 N 56 54 54 43 56 54 43 Standard errors are given in parentheses.

*p<0,2 **p<0,15 ***p<0,1 ****p<0,05 *****p<0,01

In the first three regressions, there is a significant negative relationship between HIPC-1 and

∆GDP. This corresponds with the found descriptive statistics where ∆GDP in non-HIPCs

(38)

negatively influenced by just being an eligible country for the HIPC Initiative and not only because of the associated debt relief (note that HIPC-1 consists of all eligible countries from which at least 10 did not receive any relief yet). One reason for this was already mentioned; one of the origins of the debt problem for HIPCs is a very low economic growth, so these results can be caused by selection. However, an additional reason why HIPCs did not get a chance to speed up their economic growth is because their finances are directed by the IMF and World Bank to economic reforms and if possible, to social spending.

In regression 2, debt has positive influence on ∆GDP, but for HIPCs this effect is reversed since the coefficient of debt*HIPC is more negative than the coefficient of debt is positive. Also regression 6 indicates a positive influence of debt on ∆GDP. This result is partly surprising since earlier research found that a higher stock of debt leads to a lower economic growth due to depressing investments and disincentive of governments to strengthen their financial position. These earlier conclusions seem to be right for countries participating in the Initiative which corresponds with Pattillo and others (2004), who argue that the uncertainty of debt relief, debt rescheduling and additional lending reduces the level of investment. In other words, investors and governments value the Initiative as a high risk program which does not guarantee an increase in economic growth, not even on the long term.

In regression 3, debt-service indicates a negative significant relationship which implies that a reduction in debt service leads to an economic growth. This was also found in the literature and can be explained by the effect on public investment and social spending. The same result was found in regression 6 for HIPC-2, but both effects are overshadowed when control variables are entered.

(39)

exports on economic growth. This relationship is not a new one to find but it is however

important in this setting. The HIPC Initiative is designed to enhance economic growth, reduce debt burdens, and increase poverty, and the only focus on exports is for the critical level of debt sustainability. Why is the efficiency and composition of exports not a pillar of the Initiative? An increase of steady exports could reduce trade deficit and lead to macroeconomic stability. Further, it enables countries to service their debts without cutting social spending. And if a country can service its debts, the negative effects of the ‘debt overhang’ can be diminished and this may increase foreign investment.

Poverty Reducing Expenditures

(40)

Table 6: Regression results for Poverty Reducing Expenditures, using dependent variable ΔPRE. ∆PRE 1. 2. Constant 18.509***** 14.601***** (2,373) (3.951) Predictors Debt-annual -0.157 0.020 (0.571) (0.481) Debt service-annual 0.094***** 0.133***** (0.014) (0.028) Control variables Committed 0.003 (0.002) ODA-annual -0.353*** (0.184) Exports-annual 0.312** (0.203) GDP-annual 1.534***** (0.350) R square 0.655 0.857 F 23.776***** 20.978***** Jarque-Bera 0.339 1.192 N 29 28

Standard errors are given in parentheses.

*p<0,2 **p<0,15 ***p<0,1 ****<0,05 *****<0,01

In both regressions of table 6, debt service-annual shows to have a positive significance influence on ∆PRE which implies that an increase in debt service leads to an increase in poverty reducing expenditures. This contradicts with the thoughts of the Initiative that a reduction in debt service should lead to free up funds available for social spending. The explanation can be found in the literature. In these regressions, the sample consisted of only HIPCs. All these countries wrote a PRSP and were more or less obliged to increase their poverty reducing expenditures. Furthermore, the literature indicated already that it is not necessary logical that HIPCs experience a decrease in debt service over the years participating in the Initiative. So it could be that the relationship presented in the regressions is only there because all countries are HIPCs and being HIPC implies an increase in poverty reducing expenditures and an increase in debt service. To find out if this theory holds, the regression should be done in a dataset including non-HIPCs, which is not possible due to unavailable data on poverty reducing expenditures.

(41)

higher GDP growth and a higher exports growth can lead to more funds available for poverty reducing expenditures. It is surprising to find that ODA shows a negative significant result. This implies that an increase in ODA leads to a decrease in poverty reducing expenditures while a country should have more funds available for these expenditures. Since ODA is often bound to a lot of constraints and obligations, this may cause a decrease in available funds free to spent by the government and so a decrease in poverty reducing expenditures.

Health Expenditures

Casewise diagnostics in SPSS indicates Myanmar (case 39) as an outlier. The residuals of this outlier differed from the prediction by 3,895 standard variations. This case will be deleted from the sample, when used for health expenditures.

In the regressions with independent variable HIPC-1, the variables service annual and

debt-service annual*HIPC-1 both have a VIF value of more than 850. This means that those two

(42)

Table 7: Regression results for Health Expenditures, using dependent variable ∆Health.

∆Health HIPC-1 HIPC-2

1. 2. 3. 4. 5. 6. 7. 8. Constant 6.579***** 7.759***** 1.972 7.440***** 2.264 3.898***** 3.815***** -0.346 (1.404) (1.918) (1.972) (1.665) (1.745) (1.034) (1.399) (1.484) Predictors HIPC-1 -3.575**** -4.325**** -0.735 -4.064**** -0.851 (1.661) (2.136) (2.017) (1.766) (1.729) HIPC-2 0.297 1.083 3.929***** (1.579) (1.897) (1.451) Debt-annual -0.236* -0.146 -0.196*** -0.166*** -0.086 -0.029 (0.178) (0.152) (0.116) (0.098) (0.138) (0.104) Debt-annual*HIPC 0.067 -0.130 -0.096 -0.237 (0.235) (0.195) (0.329) (0.231) Debt service-annual 0.003 0.007 0.003 0.012*** (0.006) (0.006) (0.006) (0.006) Debt service-annual*HIPC -0.115 -0.142**** (0.094) (0.068) Control variables Africa 0.784 0.736 -0.024 (1.570) (1.548) (1.362) ODA-annual -0.006 -0.035 -0.033 (0.033) (0.044) (0.040) Exports-annual -0.148 -0.129 -0.079 (0.081) (0.081) (0.077) GDP-annual 0.692***** 0.698***** 0.785***** (0.137) (0.136) (0.121) R square 0.079 0.112 0.423 0.115 0.437 0.001 0.054 0.540 F 4.633**** 2.109** 4.614***** 2.167** 4.870***** 0.035 0.545 5.483***** Jarque-Bera 1.048 5.039 2.506 5.155 1.033 0.824 1.507 1.488 N 56 54 52 54 52 56 54 52

Standard errors are given in parentheses.

*p<0,2 **p<0,15 ***p<0,1 ****<0,05 *****<0,01

(43)

increasing social spending, at least for one year before completion point but it can be questioned if the financial resources were provided by the Initiative or through new borrowing, and furthermore, if this increase is sustainable and can lead to economic growth in the long term. Regression 2, 4, and 5 indicate a significant negative relationship between debt-annual and

∆Health which implies that a reduction in debt should increase growth in health expenditures, but

the assumption that this effect will especially occur in HIPCs can not be confirmed. In the literature the relationship between debt and social spending is more related to the reduction of debt service in stead of debt stock but regression 8 indicates the opposite. Debt service-annual shows a significant positive relationship with ∆Health but within the framework of HIPC (debt

service-annual*HIPC) this relationship is significant negative and has even more impact

(coefficient is larger). This would imply that the countries within the classification of HIPC-2 are better able to direct a reduction in debt relief to an increase in health expenditures, which corresponds with the objective of the Initiative. However, it should be noted that debt

service-annual did not show a reduction in the descriptives but only less growth than in non-HIPCs.

(44)

Chapter 6: Discussion & conclusion

In this thesis, the objectives of the HIPC Initiative are playing the central theme and research is based on the question if the Initiative can actually meet its objectives.

The Initiative is designed in 1996 to promote growth in HIPCs, provide debt sustainability, and to free up resources for social spending. Now that the 10th anniversary of the Initiative has taken place, many review reports have been written to evaluate its impact. In these reports, it is clear that the status and progress of individual countries differ enormously and it is therefore difficult to answer the question if the Initiative in general works. This research attempts to answer this question based on literature review and statistical research.

Earlier research indicates that debt and growth are indeed related which makes the first objective, to promote growth by reducing debt, a logical one. However, to promote growth, debt should not exceed a certain level and should be sustainable. Research have found that the sustainable level given by the World Bank, 150% of exports, could indeed eliminate the negative effects of debt on growth, but than debt should stay constant at this level. It appears that simply reducing a country’s debt level is not enough to achieve sustainability. Low income countries are very sensitive to external shocks and prices of primary commodities which can heavily influence the level of debt and the ability to repay. It seems that the HIPC Initiative does not focus enough on the individual debt situation of participating countries to encounter these problems. A large part of completion point countries face debt ratios that once again exceed HIPC thresholds. Furthermore, literature shows that a decrease in debt burden is beneficial for social spending and poverty. However, a decrease in debt burden by the Initiative should free up funds available for this social spending. Many studies found that this is not necessarily the case in HIPCs.

(45)

also seem to be able to convert debt service reduction into growth in health expenditures. Since the period assessed in this research does not provide a distinctive period after the Initiative, it can not be concluded if these positive effects of the Initiative are sustainable or that they imply just the obligation to increase health expenditure in order to receive debt reduction.

This leads to the conclusion and answer of the first research question that participation in the HIPC Initiative does not have any positive influence on economic growth, not directly and not by means of converting debt reduction into growth. Especially in the beginning of the process, the conditionalities attached to the Initiative cause a decrease in economic growth. The influence of participation in the Initiative on social spending is more difficult to assess. It seems that participation in the early years depresses funds available for social spending but for countries in the final period of the Initiative this effect is less. However, economic growth seems to be a much better variable leading to increasing social spending so it might be better to focus on ways to increase growth (other than debt relief) to reach the objective to reduce poverty by an increase in social spending.

Limitations

First, an important limitation to this study was the availability of sufficient data over a longer period of time. Due to missing data on some variables the research models had to be adapted and, as mentioned before, the definitions of some variables changed over time which biased the growth figures. And given the short period of time after the Initiative, it is not possible to present good figures of the result measured after the Initiative.

Second, for the classification of participation in the HIPC Initiative, two measures are being used. However, both ways are not perfect to analyze the process of each individual country but considered best for the amount of available information and data. When periods before and after decision point and completion point are taken for each individual country, this would probably lead to better results. Changes in variables could be measured also after for instance the start of IMF implications.

(46)

implementation of changes will not happen overnight but it might take several years to give a result. Again, this was not possible due to the short period after the Initiative.

Fourth, as a measure of social spending, health expenditures should be extended by education and other social spending. Furthermore, poverty reducing expenditures should be interpreted with care since these are defined differently across countries and over the years countries may have included more expenditures, which could bias the increase in these expenditures.

Recommendations

(47)

References

Addison, T., Hansen, H., Tarp, F. (2004), “Debt Relief for Poor Countries, The Introduction”

Debt relief for poor countries, United Nations University WIDER, Palgrave Macmillan.

Birdsall, N., Claessens, S., Diwan, I. (2004), “Policy Selecticity Forgone: Debt and Donor Behaviour in Africa” Debt relief for poor countries (ch. 3), United Nations University WIDER, Palgrave Macmillan.

Castro-Leal, F., Dayton, J., Demery, L., Mehra, K. (1999), “Public Social Spending in Africa: Do the Poor Benefit?”, The World Bank research observer, vol 14, no. 1.

Clements, Benedict, Rina Bhattacharya and Toan Quoc Nguyen (2003), “External Debt, Public Investment, and Growth in Low-Income Countries”, IMF working paper No. 03/249

(Washington DC: International Monetary Fund)

Cooper, D. and Schindler, P. (2000), Business Research Methods, McGraw-Hill Irwin, seventh edition, ISBN 0-07-231451-6.

Easterly, William (1999), “How did Highly Indebted Poor Countries Become Highly Indebted? reviewing two decades of debt relief”, policy research working paper 2225, November 1999, The World Bank.

Easterly, William (2001), The Elusive Quest for Growth, Economists’ Adventures and

Misadventures in the Tropics, The MIT Press, Cambridge, Massachusetts/London, England ISBN

0-262-05065-X.

Eaton, Jonathan (1992), “Sovereign Debt: a Primer”, policy research working papers, WPS 855, The World Bank.

The Heavily Indebted Poor Countries Initiative: Does it Work?

The Heavily Indebted Poor Countries Initiative:

Does it Work?

Anieka van Leeuwarden

S1198424

Groningen, August 2007

Supervisor: Dr. ir. D.J. Bezemer

The Heavily Indebted Poor Countries Initiative: Does it Work?

Abstract

Table of contents

Abbreviations

Chapter 2: The HIPC Initiative

Chapter 3: Review of the literature

Chapter 4: Methodology

∑

Chapter 5: Results

Chapter 6: Discussion & conclusion

References