DETERMINANTS OF EMERGING MARKET BOND SPREAD: EVIDENCE FROM TEN AFRICAN COUNTRIES ABSTRACT This paper investigates the determinants of bond market spreads over the period 1991-2012 in 10 African countries. The empirical findings reveal that the spread of sovereign bonds in African countries has increased from 2011 to 2012, and public debt to GDP ratio, political stability index and current account deficit are found to be the major determinants of spreads in most African countries. However, Development Financial Institutions (DFIs) such as the African Development Bank, World Bank and International Monetary Fund, should support African governments to effectively use debt for investment purposes not for consumption. Thus, DFIs should also support African countries to mobilize domestic resources that could leverage additional investment if spent in appropriate manner. Furthermore, this study can give benefit to policy makers, who could get a better overview and understanding on the determinants that can affect African countries cost of borrowing. Keywords: Sovereign bond spreads, Macroeconomics, Liquidity, Political stability.
TABLE OF CONTENTS Abstract.. 1 Table of contents... 2 List of Tables..... 3 List of Figures 3 1. Introduction.. 4 2. Stylized Facts... 5 3. Litterature Review... 8 4. Methodology..... 10 4.1 Data Description.... 10 4.2 Empirical Model.... 12 5. Empirical Findings..... 13 6. Conclusion.. 15 References.... 17 2
List of Tables Table 4.1 Definition and sources of variables... 11 Table 4.2 Descriptive statistics with other emerging markets... 12 Table 5.1 Spread regressions..... 14 List of Figures Figure 1.1 Evolution of the interest rates in USA, UK, Europe and Africa... 5 Figure 2.1 Size of sovereign bonds issued by African governments... 6 Figure 2.2 Yields on sovereign bonds issued by governments... 7 Figure 4.1 Evolution of sovereign bond spreads... 12 Figure 5.1 Evolution of the public debt to GDP ratio... 15 3
1. INTRODUCTION African countries have for many years depended on bank financing. As a result, African bond markets are still at a nascent stage (except for South Africa). However, to improve the intermediation of savings and the structure of the financial system, many African countries have taken decisive steps to promote the development of their bond markets. Africa has grown on average by 5.5% per year over the last decade (IMF, 2013). Thus, growth momentum and infrastructure, have prompted African governments to tap into international markets for sources of funds. Indeed, since 2011, USD 8.75 billion has been raised by Nigeria, Rwanda, Tanzania, Morocco, Zambia, South Africa, Namibia and Senegal on international markets. The latest case of sovereign bond is Rwanda with USD 400 million (IMF, 2013). Despite the rise, African governments have issued international sovereign bonds for a variety of reasons. These include: infrastructure spending, benchmarking, and debt restructuring. The global financial conditions-low interest rates in advanced economies (Figure 1.1) and portfolio diversification opportunities, have led investors to increase their appetites for African sovereign bonds. Therefore, this study seeks to investigate the determinants of sovereign international African bonds for the period 1991 to 2012. Such an exercise would allow African countries to pinpoint policies which can be adopted to ultimate the real world cost of borrowing. This study can give benefit to policy makers, who could get a better overview and understanding on the determinants that can affect African countries cost of borrowing. Thus Development Financial Institutions (DFIs) may apply this study as a baseline in implementing new policies regarding the allocations of resources to its regional member countries (RMCs). 4
Figure 1.1: Evolution of the Interest rates in USA, United Kingdom, Europe and Africa 12,00 10,00 8,00 6,00 4,00 2,00 0,00 Source: Thomson Routers (DataStream) USA UK France Japan Africa Note: Interest rates for Africa are the average of interest rates in Morocco, Nigeria, South Africa, Tunisia, Egypt and Cote d Ivoire. The structure of this paper is as follow: Section 2 analyses the stylized facts. Section 3 reviews the relevant literature. Section 4 describes the data and the methodology used in this paper. Section 5 assesses the main findings. Section 6 concludes. 2. STYLIZED FACTS Recently, such countries as Morocco, Nigeria, South Africa, Senegal, Zambia, Namibia and Rwanda have been able to raise funds in international bond markets (Figure 2.1). This sudden surge in borrowing by African countries is due to a variety of factors: Large borrowing needs: Africa has an estimated financing requirement to close its infrastructure deficit estimated at 93 billion annually until 2020 (AfDB, 2010). Sovereign bond proceeds could 5
be crucial to financing infrastructure projects-such as electricity generation and distribution, roads, airports, ports and railroads, which often require resources that exceed aid flows and domestic savings. Low borrowing costs: In several cases, many African countries have been able to issue sovereign bonds at lower interest rates than troubled European economies such as Greece and Ukraine, Latin American countries such as Venezuela and Argentina (Figure 2.2). However, political instability in some African countries such as in Egypt has increased the yield of sovereign bonds. Figure 2.1: Size of sovereign bonds issued by African governments Rwanda (2013) Namibia (2011) South Africa (2012) South Africa (2011) Nigeria (2013) Nigeria (2011) Morocco (2012) Senegal (2011) Zambia (2012) $0 $400 $800 $1 200 $1 600 $2 000 $2 400 Size of offering, million dollars Source: International Monetary Fund, Bloomberg Note: Year in parentheses is the settlement date 6
Figure 2.2: Yields on sovereign bonds issued by governments Greece Ukraine Kazakhstan Russia Philippines Malaysia China EMBIG Asia EMBIG Africa EMBIG Latin Venezuela Argentina Brazil Egypt Morocco Nigeria Cote d'ivoire South Africa 1 3 5 7 9 11 13 15 Sovereign bond yields, percent Source: International Monetary Fund, DataStream Note: Bond yields as of August 23, 2013, in the secondary market. EMBIG is the JP Morgan Emerging Market Bond Index Global However, the climate for African bonds issuance may become less favourable in the near future. There is a possible end to the U.S. Federal Reserve s bond-buying program in 2014, and therefore an increase of the U.S. Treasury yields could increase the cost of borrowing, and the risk appetite for foreign investors. As a result, African governments will have to compete with other issuers for funding. 7
3. LITTERATURE REVIEW Since the resolution of the 1982 crisis, with the issue of the popular Brady Bonds through the Brady plan, bonds have become the most common source of funding for emerging market sovereigns. However, many researchers have investigated the determinants of emerging market bond spreads in Eastern and Central Europe, Latin America and Asia, while few studies have been done for Africa. Many studies have examined whether debt variables, GDP growth, reserves, interest rates and external debts, play an important role in explaining sovereign bond spreads (see, for example, Eichengreen and Mody, 1998; Min 1998, Bellas, Papaioamou and Petrova, 2010). They find that, external debts, GDP growth, reserves and interest rates are all significantly correlated with sovereign bond spreads. Ferrucci (2003) extended these studies by identifying short-and longterm determinants of sovereign bond spreads. He concludes that the degree of openness, the ratio of amortizations to reserves, the external debt to GDP ratio, and the ratio of current account to GDP are all significantly correlated with sovereign bond spreads. He also found that global liquidity conditions and U.S. equity prices are also correlated with sovereign bond spreads. Hartelius and Kristian (2008), by using a set of country-specific and common external explanatory variables, estimated emerging markets sovereign bond spreads. He found that low volatility environment of global financial markets played a role in explaining the sovereign market bond spreads since 2003. In addition, Luengnaruemitchai and Schadler (2007) used a similar model in order to see whether market participants underestimated or not the riskiness of holding sovereign bonds issued by Central and Eastern European Countries (CEECs). Their empirical findings revealed that, before the global financial crisis, market participants were requiring lower yields to hold CEECs sovereign bonds in their portfolio. 8
To explain issues regarding the yield spreads, Min (1998) used several variables such as liquiditysolvency, macroeconomic, external shock and debt-related, and with the Mexican crisis in 1995, he used dummies to capture the effect of the crisis on bond spreads. His empirical findings, revealed that macroeconomic, debt-related and liquidity-solvency, influenced the volatility of bond spreads. Thus, the external shock variables (oil prices and US 3-month T-bill rate) have no power on bond spreads. Kamin and Kleist (1999) extended this study by introducing credit ratings as an explanatory variable. They found that, credit ratings played a significant role in explaining the evolution of emerging market bond spreads, and the dummy for the Mexican crisis had a significant power on bond spreads. Nazim and Salvatore (2013), by using smooth transition regression model with a panel of 26 emerging market economies investigated the impact of debt sustainability on emerging market bond spreads. They found that debt sustainability is a major determinant of bond spreads. Thus, they concluded that market interest rates react more to debt sustainability concerns in a country with a high level of debt compared to a country with a low level of debt. In this study, however we use the Emerging market bond index global (EMBIG) as dependent variable to measure the spread of the African sovereign bonds. Our independent variables follow Min (1998) and Nazim and Salvatore (2013), and are used together to investigate the relationships among them and sovereign bond spreads. Therefore, these variables are: (1) Macroeconomic variables: CPI inflation, Nominal exchange rate; (2) Liquidity variables: Current account to GDP, Reserves to GDP, Gross international reserves to GDP, Public debt to GDP, Real GDP growth, Real export growth and Real import growth; (3) Political fragility: Political stability. 9
4. METHODOLOGY 4.1 Data Description Building on the IMF paper but focusing on Africa, this study investigates the determinants of sovereign bond spreads in 10 African countries for the period 1991-2012. These countries are a subset of all the African countries included in the JP Morgan EMBIG index: Morocco, Nigeria, South Africa, Egypt, Tunisia, Cote d Ivoire, Ghana, Namibia, Gabon and Senegal. However, to be included in the EMBIG index, countries have to satisfy one of the following criteria: (i) be classified as low or middle per capita income by the World Bank; (ii) regardless of their World Bank-defined income, have restructured external or local debt in the past 10 years. This study also compare the African sovereign bond spreads with other emerging markets. Definition and sources of the variables used are reported in Table 4.1. Summary statistics and a graphical representation are however, provided in Table 4.2 and Figure 4.1. From the figure, we noticed that sovereign bond spreads is higher in Latin America than in Europe, Africa and Asia. The figure also shows that from 2010, African sovereign bond spread has increased, which is definitely due to some troubles experienced in such countries as Cote d Ivoire, Morocco, Tunisia, Egypt and Nigeria. Table 4.2 furthermore shows that the mean value and the volatility of sovereign bond spreads in Asia are relatively low. 10
Table 4.1: Definition and sources of variables Variable Name Source Sign Unit Definition Dependent Variable EMBIG Spread DataStream Basis Points Independent Variables Macroeconomic Variables Inflation AfDB (Dataportal) + Percentage Nominal Exchange Rate AfDB (Dataportal) - Percentage Liquidity Variables International reserves to GDP AfDB (Dataportal) - Percentage Total reserves to GDP World Bank - Percentage Current Account deficit World Bank + Percentage Public debt to GDP AfDB (Dataportal) + Percentage EMBIG spread is the defined as the yield difference over U.S. Treasuries. Higher yield spreads are associated with weak liquidity position, such as high debt to GDP ratio, low international reserves to GDP ratio, low export growth rate, high import growth rate and high public debt ratio. As measured by the consumer price index reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services. The higher the inflation rate, the higher the spread. It is the annual average based on monthly averages (local currency units relative to the U.S. dollar). An appreciation of the local currency against the U.S. dollar could lower the spread. Is the sum of the country's foreign exchange, its reserves position in the IMF, its monetary gold reserves, and the United States dollar value of SDR (Special Drawing Rights) holdings by its monetary authorities. The higher the ratio, the lower the spread. Comprise holdings of monetary gold, special drawings rights, special drawing rights, reserves of IMF members held by the IMF, and holdings of foreign exchange under the control of monetary authorities. The higher the ratio, the lower the spread. Is the sum of net exports of goods and services, net primary income, and net secondary income, expressed as percentage of GDP. An increase of this ratio, increases the spread. It includes money owed by the government to creditors within the country (domestic or internal debt) as well as to international creditors (foreign or external debt). The higher this ratio, the higher the spread. Real GDP growth AfDB (Dataportal) - Percentage Is the annual growth rate of GDP at market prices. The higher this ratio, the lower the spread. Real Export growth AfDB (Dataportal) - Percentage Real Import growth AfDB (Dataportal) + Percentage Political Fragility Political stability Center for Systemic Peace - Index Is the growth rate of exports of goods and services in volume term. An increase of this ratio, lowers the spread Is the growth rate of imports of goods and services in volume term. The higher this ratio, the lower the spread This indicator is rated on a four-point fragility scale: 0 "no fragility", 1 "low fragility", 2 "medium fragility", 3 "high fragility". Higher political instability lead to a substantial widening of sovereign bond spreads. 11
Table 4.2: Descriptive statistics with other Emerging markets Africa Asia Europe Latin America Mean 446.09 252.20 543.79 480.37 Standard 206.02 103.20 529.61 203.89 Deviation EMBIG spread Minimum 183.16 65.32 122.84 159.91 Maximum 1247.41 879.80 2512.22 1138.79 N 3914 4957 3914 3914 Source: DataStream Note: Data are in Daily basis Figure 4.1: Evolution of Sovereign bond spreads 3,000 2,500 2,000 1,500 1,000 500 0 92 94 96 98 00 02 04 06 08 10 12 ASIA LATIN EUROPE AFRICA Source: DataStream Note: Spreads are the EMBIG spread for Asia, Europe, Latin America and Africa 4.2 Empirical Model Following Min (1998), and Nazim and Salvatore (2013), two regression models have been specified: Model 1 is the linear regression model and Model 2 represents the fixed effect regression model. 12
Model 1: Model 2: Where, is the EMBIG spread for the country i at year t represents independent variables (IV) is the coefficient for the IVs is the error term = represents all unobserved factors that vary across units but are constant over time represents all unobserved factors that vary across units and time 5. EMPIRICAL FINDINGS In this section, we summarize the regression coefficients of the two models, as described in section 4. Table 5.1 shows that public debt to GDP ratio and the political stability index are significant and have the largest impact on sovereign bond spreads: a 1 percentage point increase in public debt to GDP ratio increase the spread by 64.1 and 64.9 percentage points in model 1 and model 2, respectively. Thus, a 1 percentage increase in political stability index, lower the spread by 21.9 in model 1, and 47.4 in model 2. In addition, the current account deficit ratio and real GDP growth are also significant and have the expected signs in both models. On the other hand, real import growth has the least impact on sovereign bond spreads. The R- square is 0.731191 in model 1 and 0.761888 in model 2, suggesting that more than 70 percent of the variation of the spread is explained by public debt to GDP ratio, political stability index, current account deficit ratio and the real GDP growth. Africa remains broadly positive, and growth is expected to be 5.4 percent in 2013 (IMF, 2013). However, despite some political troubles such as the Arabic spring revolution in 2010 experienced 13
in Tunisia, Egypt and Morocco, many African countries have implemented stronger policy frameworks, and reduced their public debt to GDP ratio (Figure 5.1). Overall, the findings are consistent with previous studies such as Min (1998), Nazim and Salvatore (2013), Eichengreen and Mody (1998), Bellas, Papaiamou and Petrova (2010), who found public debt ratio, political stability index, current account deficit and real GDP growth significantly correlated with sovereign bond spreads. Table 5.1: Spread regressions Model 1 Model 2 Coefficients Inflation 4.041785 1.76E+00 (5.17574) (6.83663) Exchange rate -1.345307 0.740527 (0.829249) (1.253179) Political Stability -21.94638*** -47.39634*** (53.91376) (60.10196) Current account deficit 16.17791*** 23.61676*** (6.069355) (7.859463) Reserves 3.3694 5.4373 (1.0126) (3.7856) International reserves -11.1097-20.4438 (9.7212) (11.1442) Public debt 64.14058*** 64.89664*** (6.837931) (7.997012) Real growth -25.2097* -37.6029** (15.14873) (17.61077) Real Export Growth -5.037896-2.104159 (4.121428) (4.725669) Real Import Growth 6.588214* 9.187881* (3.932468) (4.66394) R 2 0.731191 0.761888 F-statistic 24.75304*** 8.769549*** N 102 102 Notes: ***, **, * indicate significance at the 1%, 5%, and 10% levels, respectively. Standard errors are in parentheses. The dependent variable is the EMBIG spread (Emerging Bond Market Index Global). 14
60 Figure 5.1: Evolution of the public debt to GDP ratio 50 40 30 20 10 0 92 94 96 98 00 02 04 06 08 10 12 MOROCCO NIGERIA SOUTHAFRICA EGYPT TUNISIA IVOIRE Source: African Development Bank 6. CONCLUSION This study was set out to investigate the determinants of bond market spreads in 10 African countries: Morocco, Nigeria, South Africa, Egypt, Cote d Ivoire, Tunisia, Ghana, Namibia, Gabon and Senegal. The empirical findings reveal that public debt to GDP ratio, political stability index, real GDP growth and current account deficit are the major determinants of sovereign bond spreads. Thus, real import growth is found to have the least impact on spread. Moreover, African sovereign bond spreads has increased since 2011, due to political troubles experienced in some countries. However, macroeconomic variables are all found to be insignificant and have no impact on sovereign bond spreads. 15
To prevent debt crises in African countries, Development Financial Institutions (DFIs) such as the African Development Bank, World Bank and International Monetary Fund should support African government to effectively use debt, i.e. for investment purposes such as infrastructure, education and health, not for consumption because issuing sovereign bonds for consumption purposes can create difficulties when the debt obligation falls since it does not yield future economic benefits. Thus, DFIs should also support African countries to mobilize domestic resources that could leverage additional investment if spent in an appropriate manner. Furthermore, African governments should issue bonds in local currencies and issue bonds to their own diaspora known as Diaspora bonds rather than issuing sovereign bonds. 16
REFERENCES Bellas, D., Papaionnou, G. and Petrova, I. (2010). Determinants of Emerging Market Sovereign Bond Spreads: Fundamentals vs. Financial Stress. Working paper No. 10/281, International Monetary Fund Eichengreen, B. and Mody, A. (1998). What Explains Changing Spreads on Emerging-Market Debt: Fundamentals or Market Sentiment? Working Paper No. 6408, National Bureau of Economic Research Ferruci, G. (2003). Empirical Determinants of Emerging Market Economies Sovereign Bond Spreads. Working paper No. 205, Bank of England Gersovitz, M. (1985). International Debt and the Developing Countries. Working paper No. 268, World Bank Hartelius, J. (2008). Emerging Market Spread Compression: Is it Real or Is it Liquidity? Working paper No. 08/10, International Monetary Fund Jahjah, S. and Yue, V. (2004). Exchange Rate Policy and Sovereign Bond Spreads in Developing Countries. Working paper No. 04/10, International Monetary Fund Jennifer, M. (2008). Mapping of Current Ongoing Initiatives related to Bond Market Development in Africa. Working paper, African Development Bank Kamin, S. and Kleist, V. (1999). The Evolution and Determination of Emerging Market Credit Spreads in the 1990s. Working paper No. 653, Federal Reserve Luengnaruemitchai, P. and Schadler, S. (2007). Do Economist and Financial Markets Perspectives on the New Members of the EU Differ? Working paper No. 07/65, International Monetary Fund Mc. Donald, C. (1982). Debt Capacity and Developing Country Borrowing. Working paper No. 69, International Monetary Fund Min, H. (1998). Determinants of Emerging Market Bond Spread: Do Economic Fundamentals Matter? Working paper No. 1899, World Bank Mu, Y., Phelps, P. and Stotsky, G. (2013). Bond Markets in Africa. Working paper No. 13/12, International Monetary Fund Nazim, B. and Salvatore, D. (2013). The Impact of Debt Sustainability and the Level of Debt on Emerging Markets Spreads. Working paper No. 13/93, International Monetary Fund Sachs, J. (1981). The Current Account and the Macroeconomic Adjustment in the 1970s. Brookings papers on Economic Activities, 1, 201-268 17