CHAPTER ONE INTRODUCTION
Background of the Study
Following the Euro Area Sovereign Debt crisis and the 2008 financial crisis, the discussion on the debt-to-GDP and economic growth relationship has been recurrent in the literature (Chudik, Mohaddes, Pesaran & Raissi, 2017). Another call for the discussion of the debt-to-GDP relationship in recent times is the groundbreaking findings of Reinhart and Rogoff (2010), “Growth in time of Debt”. The discussion has touched well on developed countries (see Reinhart & Rogoff, 2010; Eberhardt & Presbitero, 2015; Herdon, Ash & Pollin, 2014) and some developing countries (see Chudik et al., 2017; Pattillo, Poirson & Ricci, 2011). A recent study was done by narrowing developing countries to look at Africa in context due to the heterogeneity in the growth and debt relationship (Ndoricimpa, 2017). Debt is an inevitable source of finance, needed to complement the national savings of developing countries, geared towards the countries’ growth and development (Siddique, Selvanathan & Selvanathan, 2016).
Studies have shown that debt thresholds are different across countries based on varying literature findings provided by these studies. The findings of Reinhart and Rogoff (2010) showed that advanced countries with debt-to-GDP fraction above 90 percent tend to realise negative effects on growth. Developing countries have a different case, where debt-to-GDP of about 35 to 45 percent has a detrimental effect on the growth of the economy (Pattillo et al., 2011). These ratios are sensitive to the variables introduced in the growth model during the estimation (Ndoricimpa, 2017). The level of debt a country has can be a propeller of growth or a detriment to the economy’s growth. The issue of the debt and growth nexus is to ascertain a policy measure for countries to
sustain their growth levels with their level of debt stock in order to avoid unexpected financial crises especially when the high debt levels are funded with short-term borrowing (Reinhart & Rogoff, 2010).
Figure 1.1 shows the trend of average debt-to-GDP ratios from 1960 to 2015 for all the African countries being studied in this thesis. The markers represent the mean debt-to-GDP for the African countries within the sample. Public debt here represents gross government debt which will also represent debt as used throughout this thesis. The cubic nature of Figure 1.1 represents history repetition. In 1960, the average debt-to-GDP for African countries was 44.43%, which realised a decrease for about a decade followed by a rise for about two decades before the presence of the Highly Indebted Poor Country (HIPC) initiative by the World Bank, the IMF, and some bilateral and commercial creditors (IMF, 2017).
Data source: IMF Historical Public Debt Database (Abbas et al., 2010)
The HIPC initiative was targeted at reducing the indebted countries’ debt to levels that can be sustained in order to avoid debt rescheduling (Chowdhury, 2004; Claessens & Detragiache, Kanbur & Wickham, 1996). From Figure 1.1, average debt levels in Africa have started rising after the relief introduced by the HIPC initiative in the early 2000s. The question academicians and policymakers are asking is the long-term sustainability of debt by the HIPC initiative (Chowdhury, 2004).
Data source: IMF Historical Public Debt Database (Abbas et al., 2010)
The public debt trend for Economic Community of West African States (ECOWAS) and Common Markets for Eastern and Southern Africa (COMESA) as depicted in Figure 1.2 and Figure 1.3 respectively shows similar behaviour to that of all the countries within the sample.
Data source: IMF Historical Public Debt Database (Abbas et al., 2010)
This study tries at analysing the effect of debt threshold that could significantly reduce the growth of an economy within Africa, using new methodologies. To put in a different perspective, whether debt build-up is a potential to slow economic growth. Although debt at high levels enhances economic progress in the short run by increasing the capital base of an economy thereby increasing productivity, it also has a crowding out effect in the long run on spending and reduces output (Chudik et al., 2017; Pattillo, Poirson & Ricci, 2004). The problem persists in developing countries as high debts tend to cripple growth due to the huge debt services that face developing countries (Pattillo et al., 2011). Since the mid-1980s, high levels of debt have restricted the growth of developing countries especially those in Africa (Fosu, 1999).
Problem Statement
The IMF and the World Bank in a joint programme, started the HIPC initiative from 1996 to help countries with unsustainable debt levels (IMF, 2017). The initiative was to help these countries reduce their debt burden and to reduce poverty. Out of 36 countries that have benefited from the programme, 30 are African countries with 3 additional African countries said to be eligible for the HIPC initiative assistance. This is alarming and suggests that debt levels held by African countries are not well managed and policies need to be put in place to control this high debt levels. Debt reliefs from the IMF and World Bank cannot be a sustainable solution to the high debt levels that African countries assume (Arnone, Bandiera & Presbitero, 2008).
The aim of the IMF and World Bank HIPC initiative was achieved in mitigating the debt levels of poorly performing economies in order to help them raise their poverty reduction expenditure as shown in Figure 1.2. Nevertheless, these countries, mostly African countries were not able to achieve a greater portion of the targets of the Millennium Development Goals (MDGs) (IMF, 2014).
Source: Adapted from IMF (2014) with t representing HIPC completion point
Even after the intervention of the IMF and World Bank in their HIPC initiative to help mitigate the debt burden of most African countries, some of these countries are still carrying high levels of debt as at 2015 (IMF, 2017). This is because the debt relief from the IMF and World Bank is not a long-term debt sustainability measure and that having a threshold policy will guide African countries to achieve their Sustainable Development Goals (SDGs) without relying on debt relief packages.
In their paper, Panizza and Presbitero (2013) posited that the public debt and economic growth relationship is heterogeneous across countries. While some countries have experienced slow growth with relatively low levels of debt during debt distress, some other countries have been able to grow amidst high levels of debt. This makes it important to consider the analysis of the threshold based on heterogeneity across countries (Chudik et al., 2017).
Reinhart and Rogoff (2010) estimated the non-linear relationship that exists between debt and growth by ignoring the issue of heterogeneity across countries, the growth feedback to debt, and without much attention to the possible existence of error cross-sectional dependencies within the sample (Chudik et al., 2017). Due to the magnification of spillover effects from one country to the other during a financial crisis, the analysis of the debt threshold effect on growth must consider error cross-sectional dependencies (Chudik et al., 2017). Reinhart and Rogoff (2010) employed an exogenous approach to establishing the relationship by using means and median. Recent studies have used robust endogenous econometric methods to analyse the public debt and economic growth relationship (Minea & Parent, 2012; Eberhardt & Presbitero, 2015; Ndoricimpa, 2017; Chudik et al., 2017).
The study, therefore, adopts the panel Autoregressive Distributed Lag (ARDL) model by Chudik et al. (2017) to estimate the non-linear relationship that exists between public debt and economic growth because it is robust to providing feedback effects from economic growth to public debt. The model can, therefore, handle the possible reverse causality that may be present between the two variables under study. Additionally, this model allows for slope heterogeneity across countries which helps to control for the country-specific factors within the sample of countries for the study (Chudik et al., 2017). Another interesting feature of the model is its ability to capture both short- run and long-run dynamics based on the number of lags specified. For this study, the model uses a maximum of three (3) lags for all variables and across countries which is enough to capture the dynamics of the short and long-term periods based on the persistence of economic growth rates (Chudik, Mohaddes, Pesaran & Raissi, 2013)
Subsequently, the ARDL is augmented with the cross-sectional averages of all the regressors to control for possible cross-sectional error dependence. The CS-ARDL model helps to reduce the
biases of the estimations of the ARDL that may be due to the cross-sectional error dependence resulting from financial integration and shocks from international activities such as oil price fluctuations. This study adds to the existing literature on public debt and economic growth by providing additional evidence on the debt and growth nexus within the African context. This is based on new methodologies by considering dynamics, heterogeneity across countries, and cross-sectional dependencies among countries. The study seeks to add new evidence on the relationship between public debt and growth in the African region, by considering the impact of the persistent growth of public debt on economic growth. Although the problem of debt threshold effects has been addressed in the literature, not all methodologies have been exploited on the relationship. Ndoricimpa (2017) pointed out that the relationship is sensitive to methodological approaches which creates room for a further probe with developing methodologies capable of establishing the relationship with the least bias.
