78501 Ethiopia Economic Update II Laying the Foundation for Achieving Middle Income Status June 2013 2nd Ethiopia Economic Update Laying the Foundations for Achieving Middle Income Status June 2013 the World Bank This Draft: 18 June 2013 Table of Contents Acknowledgements.................................................................................................................................. v List of Abbreviations............................................................................................................................vii Executive Summary................................................................................................................................. ix Recent Economic Developments...................................................................................................... 1 The Short View...........................................................................................................................................................1 Real Sector..............................................................................................................................................................1 Monetary Sector......................................................................................................................................................3 Financial Sector.......................................................................................................................................................5 Fiscal Sector............................................................................................................................................................6 External Sector........................................................................................................................................................8 The Long View.........................................................................................................................................................11 Middle Income Status by 2025: Is Ethiopia on Track?...........................................................................................11 Sources of Growth and Structural Change.............................................................................................................12 Public Investment: Ethiopia’s Growth Engine........................................................................................................14 Annex: Selected Economic Indicators.......................................................................................................................21 Savings............................................................................................................................................................. 25 Introduction.............................................................................................................................................................25 The Savings Rate in Ethiopia and the World.............................................................................................................26 Determinants of Savings Rates: Theory and Evidence...............................................................................................28 Income, Income Growth, and Terms of Trade.......................................................................................................30 Demographics, Urbanization, and Public Savings Rates........................................................................................32 Macroeconomic Stability and Financial Development...........................................................................................34 Recent Initiatives to Increase the Savings Rate in Ethiopia........................................................................................36 The Pension Scheme..............................................................................................................................................38 Expanding Bank Branches.....................................................................................................................................38 Micro Financing...................................................................................................................................................40 Diaspora Bonds.....................................................................................................................................................43 Housing Saving Scheme........................................................................................................................................43 Policy Recommendations..........................................................................................................................................44 Annex: Savings Rate.................................................................................................................................................47 Trade Logistics.......................................................................................................................................... 51 Introduction.............................................................................................................................................................51 Trade Logistics for Trade Development: Ethiopia’s Performance...............................................................................52 iv 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Diagnosing Ethiopia’s Trade Logistics Challenges.....................................................................................................54 Structure and Performance of the Transport and Logistics System.........................................................................55 Customs, Border Control and Other Trade Related Regulations...........................................................................61 Efficient Logistics and Distribution Services..........................................................................................................64 Policy Recommendations..........................................................................................................................................66 Annex: Trade Logistics..............................................................................................................................................69 References.................................................................................................................................................... 75 List of BOXES Box 1: Debt Sustainability and the IDA Non-Concessional Borrowing Policy................................................................9 Box 2: A Diagnostic Framework for Assessing Public Investment Management............................................................20 Box 3: Policy Recommendations on Savings.................................................................................................................45 Box 4: State Capitalism in East Asia – Bringing Competition to Markets.....................................................................57 Box 5: Informal Cross Border Trade (ICBT)................................................................................................................62 Box 6: Policy Recommendations on Trade Logistics.....................................................................................................68 List of FIGURES Figure 1: Economic Activity..........................................................................................................................................2 Figure 2: Monetary Sector.............................................................................................................................................4 Figure 3: Financial Sector..............................................................................................................................................7 Figure 4: External Sector.............................................................................................................................................10 Figure 5: Ethiopia: Economic Growth and its Sources................................................................................................13 Figure 6: Public and Private Investment Trends...........................................................................................................15 Figure 7: Public Investment: Sources and Financing...................................................................................................17 Figure 8: Private Sector Constraints in Ethiopia..........................................................................................................19 Figure 9: Infrastructure Project Costs and Time Overrun, Cost Baseline Sample.........................................................19 Figure 10: Savings and Investment in Ethiopia and the World (Percent of GDP)..........................................................27 Figure 11: GDS Rates in Ethiopia and Relevant Peer Groups (Percent of GDP)...........................................................29 Figure 12: Determinants of Savings: Persistence, GDP per Capita, Income, and Terms of Trade...................................31 Figure 13: Determinants of Savings: Demographics, Urbanization, and Public Savings.................................................33 Figure 14: Determinants of Savings: Real Interest and Inflation Rate............................................................................35 Figure 15: Determinants of Savings: Inflation Rates, Financial Sector Development.....................................................37 Figure 16: The Pension System and Savings in Ethiopia................................................................................................39 Figure 17: Ethiopia’s Trade Logistics Performance.........................................................................................................53 Figure 18: Policy Dimensions of Ethiopia’s Trade Logistics Challenges..........................................................................55 Figure 19: Moral, Professionalism and Competence of ERCA Staff...............................................................................63 Figure 20: Process Map for Border Clearance in Ethiopia (Illustrative).........................................................................69 List of TABLES Table 1: Federal Government Fiscal Performance.........................................................................................................8 Table 2: Ethiopia: Selected Economic Indicators (High Frequency)...........................................................................21 Table 3: Ethiopia: Selected Economic and Social Indicators (Annual Frequency).......................................................22 Table 4: National Savings vs. Savings from Purely Domestic Sources (% of GDP).....................................................28 Table of Contents v Table 5: Bank Branch Expansiona Data by Region in Ethiopia, 2006 to 2012...........................................................40 Table 6: The Development of Micro-Finance Institutions in Ethiopia, 2003 to 2012................................................41 Table 7: Saving and its Determinants........................................................................................................................48 Table 8: Probit Model Predicting the Probability of Having a Bank Account, Ueus 2006–2012.............................50 Table 9: Demand for Trucks for Recent Imports of Bulky Cargo, May 2013.............................................................60 Table 10: Supply Chain Sequence and Bottlenecks, Lldcs........................................................................................61 Acknowledgements T he World Bank greatly appreciates the close col- selected contributions of Francesco Strobbe (Financial laboration with the Government of Ethiopia Economist, AFTFE), Giulia Mascagni (Consultant, (the Ministry of Finance and Economic AFTP2), Ha Nguyen (Economist, DECMG), Maya Development, in particular) in the preparation of Eden (Economist, DECMG), and Susana Gable are this report. The report was prepared by a team led by also acknowledged. The report was prepared under Michael Geiger (Country Economist, AFTP2) and Lars the overall guidance of Pablo Fajnzylber (Sector Christian Moller (Lead Economist and Sector Leader, Manager, AFTP2) and Guang Zhe Chen (Country AFTPR) and included the following core members: Director, AFCE3). The peer reviewers were: Dino Abbi Kedir (Consultant, AFTP2), Toru Nishiuchi Merotti (Lead Economist, PREMED), Allen Dennis (Consultant, AFTP2), Mesfin Girma (Economist, (Senior Economist, DECPG), and Ankur Huria (PSD AFTP2) and Eyasu Tseahaye (Economist, AFTP2). The Specialist, Trade Logistics, IFC). List of Abbreviations AfDB African Development Bank GSP Generalized System of Preferences AGOA African Growth Opportunity Act ICBT Informal Cross-Border Trade ASYCUDA++ Automated System for Customs ICT Information and Communications Data Technology BOP Balance of Payments IFC International Finance Corporation CAD Current Account Deficit IMF International Monetary Fund CBE Commercial Bank of Ethiopia LPI Logistics Performance Index COMESA Common Markets for Eastern and MDRI Multilateral Debt Initiative Southern Africa MOFED Ministry of Finance and Economic CSP Country Strategy Paper Development DFID Department of International MoT Ministry of Transport Development (UK) MTS Multimodal Transport System DRC Democratic Republic of Congo NBE National Bank of Ethiopia EAC East African Community NEPAD New Partnership for Africa’s EDRI Ethiopian Development Research Development Institute NIT Net Income Transfers (sum of net EEPRI Ethiopian Economic Policy income from abroad and current Research Institute transfers) EMAA Ethiopian Maritime Affairs ODA Overseas Development Assistance Authority PRS Private sector saving EPA Economic Partnership Agreement SOE State Owned Enterprise ERCA Ethiopian Revenue and Custom SSA Sub-Saharan Africa Authority ToT Terms of Trade ESLSE Ethiopian Shipping and Logistics UNCTAD United Nations Conference on Enterprise Trade and Development eSW Electronic Single Window UNECA United Nations Economic FDI Foreign Direct Investment Commission for Africa FTA Free Trade Area UNIFM United Nations Development Fund GDP Gross Domestic Product for Women GDS Gross Domestic Savings WCO World Customs Organization GNI Gross National Income WEO World Economic Outlook GNS Gross National Savings WTO World Trade Organization GoE Government of Ethiopia Executive Summary O ver the past decade, Ethiopia has achieved high The Short View. Economic growth is driven economic growth, averaging 10.7 percent per by services and agriculture on the supply side and year. In 2012, Ethiopia was the 12th fastest by public investment on the demand side. Inflation growing economy in the World. If the country can has been brought down to single digits, as a result continue its historically impressive growth perfor- of tighter monetary policy and a drop in imported mance, it could potentially reach middle income status inflation. Financial intermediation and monetiza- by 2025. This, in turn, may require an adjustment in tion are declining, however, with important nega- economic policy to phase in the private sector as an tive implications for private sector development and additional engine of growth. Moreover, Ethiopia needs savings. The general government fiscal performance to make progress on two related important fronts: was prudent in the first half of 2012/13, although enhancing domestic savings and resolving the bottle- the consolidated public sector fiscal stance was necks of the trade logistics system. This 2nd Ethiopia expansionary as a result of State Owned Enterprise Economic Update, prepared in collaboration with the investment activities. The external current account Government of Ethiopia, offers policy guidance on balance is widening as import growth, driven by how to move forward.1 public capital imports, is outpacing sluggish export Chapter 1 discusses Ethiopia’s growth strategy, growth. This is, to some extent, explained by real which emphasizes a strong expansion of public invest- currency appreciation. ment. This model has delivered impressive results, The Long View. Projections indicate that Ethiopia although the underlying macro-policy mix highlights could potentially reach middle-income status by 2025 important challenges going forward, suggesting that if the historical growth momentum can be sustained. an adjustment to strategy may be warranted. One This, in turn, may require a change in economic strat- policy challenge relates to raising sufficient domestic egy and the way in which growth is achieved. The savings to finance one of the highest public investment current “big push” of public investment-led develop- rates in the world, as discussed in Chapter 2. Another ment has delivered very positive results. However, the challenge relates to strengthening the competitiveness development of a strong and vibrant private sector of the economy, to boost the lagging export perfor- would eventually be needed to sustain high growth, mance and attract foreign direct investment. Ethiopia’s as the experience of other high performing countries trade logistics system is a key constraint in this regard, (including those in East Asia) demonstrates. A gradual as highlighted in Chapter 3. The key messages of each phasing-in of the private sector in Ethiopia, therefore, of the three chapters are summarized as follows: offers improved prospects for achieving the country’s middle income aspirations. Recent Economic Developments The Ethiopian economy continues to perform well, although important structural challenges may need to 1 The First Economic Update Report for Ethiopia was published in De- be addressed to sustain this performance. cember 2012. xii 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Savings a plethora of interdependent factors. The implemen- tation of the recent multi-modal transport system is Increasing the domestic savings rate in Ethiopia is a key challenge. While the Government has started desirable, given the substantial investment propos- important analysis and implemented some changes, als embedded in the Growth and Transformation much more is needed to improve the system in a Plan (GTP), and the limits and risks associated with coordinated fashion. As a first step, the Government external sources of financing. Ethiopia’s savings rate is is encouraged to complete the development of a substantially lower than what would be expected for National Trade Logistics Strategy, which is currently a low-income sub-Saharan Africa country. Moreover, underway. Additional policy recommendations fall the savings rate has declined during the last decade into three broad areas: of very high economic growth. The empirical analysis of cross-country experiences presented in the chapter A. Improve the structure and performance of the point to four key determinants of savings: real per transport and logistics system: capita GDP (level and growth rate), demography, the  Strengthening the multimodal system degree of monetization, and macroeconomic stability. by encouraging increased competition. In the case of Ethiopia, low and declining savings are Addressing the infrastructure challenges in mainly attributed to the negative real interest rate and roads, trucks, and railways.Developing mul- demonetization. To increase the domestic savings rate, tiple trade corridors. Ethiopian policy makers should focus on a combina- B. Improving customs, border controls, and other tion of macroeconomic and financial sector measures: trade-related regulations: (1) facilitating a stable macroeconomic environment  Reforming customs procedures with new with positive real interest rates; (2) pursuing a strat- technologies and Public Private Partnerships. egy of monetization, while keeping inflation under Improving policy transparency, predictability control; (3) implementing policy measures to maxi- and communication. mize the potential of remittances: (4) continuing the C. Ensuring efficient logistics and distribution ongoing process of expanding bank branch networks, services particularly in rural areas; (5) facilitating the capacity  Training ERCA staff to improve efficiency and of micro finance institutions to access loanable funds; professionalism. (6) integrating informal savings schemes into formal  Equipping warehouses with modern technol- ones; (7) establishing mobile banking, drawing upon ogy and build more terminals.Implementing successful examples in neighboring countries, and; sound risk-based control systems with opti- (8) improving financial literacy. mal levels of inspections (including joint inspections). Trade Logistics The fiscal cost of implementing most of the trade The efficient functioning of the trade logistics system logistics recommendations is either low or moderate, in Ethiopia remains a major policy challenge. The except spending on infrastructure and trade corri- analysis sheds light on a highly complex system with dor developments, which would need to be weighed numerous actors, lack of clarity on regulations, and against other expenditure priorities. Recent Economic Developments 1 The Short View to grow faster than African countries such as Rwanda, Mozambique, Zambia, and Ghana, as well as China Real Sector and India (Figure 1.3). The sources of growth in Ethiopia are, how- The Ethiopian economy continued to expand ever, somewhat different from the rest of the at a rapid pace in 2011/12, registering a growth region. While Sub-Saharan Africa is also experienc- rate of 8.5 percent. Agriculture, industry, and services ing solid growth (4.4 percent in 2012), its expansion grew by 4.9 percent, 13.6 percent, and 11.1 percent, is driven by favorable commodity prices, public and respectively. However, given the relative size of each private investment, and robust private consumption. sector, expansion of the services and agricultural sec- In Ethiopia, meanwhile, soaring public investment tors explain most of this growth (57 and 26 percent, explains most of the 2011/12 growth (about two- respectively), while the contribution of industry was thirds), with private consumption explaining about relatively modest (16.7 percent). one-third (Figure 1.4).2 The relative slowdown in economic activity Leading economic activity indicators sug- compared to previous years is mainly explained gest slower growth for 2012/13. Crop production by lower crop production. Agriculture, which is forecast to grow by 4.8 percent in 2012/13, com- accounts for close to half of output, experienced pared to 7.4 percent in 2011/12. Since crop produc- a markedly lower growth rate in 2011/12 of 4.9 per- tion accounts for about a third of GDP:3 this would cent, compared to 9.0 percent in 2010/11. This was potentially slow down the 2012/13 growth rate explained by a decline in crop production growth by about 0.75 percent compared to the previous year. from 10.3 to 5.0 percent. Electricity generation, on the other hand, increased A difficult external environment also took its by 20 percent (y/y) in the first six months of 2012/13. toll on growth. Economic activity declined among Electricity sales to industries grew by 31 percent, sug- trading partners, including the EU, China, and neigh- gesting solid manufacturing activity, though the sector boring countries (Figure 1.1). International prices is small in total output (Figures 1.5 and 1.6). for coffee, Ethiopia’s biggest traditional export crop, External sector data reveal a mixed picture declined by about a third. Only gold prices remained regarding aggregate demand for 2012/13. Strong high and stable in the first half of the fiscal year, though capital imports (+45 percent) during the first seven prices have dropped recently (Figure 1.2). As a result, the positive growth impulse from net exports in the 2 This chapter relies substantially on demand side analysis of the national two previous years converted into a drag on growth accounts. Caution must be exercised in interpreting the results, owing in 2011/12. to the relatively lower quality of such estimates, compared to supply side estimates. This phenomenon is typical for sub-Saharan Africa. The Ethiopia was the 12th fastest growing economy authorities are aware of these limitations, and are collaborating with worldwide in 2012. In sub-Saharan Africa, only four development partners to improve the quality of national account statistics. 3 The agricultural sector accounts for 48.7 percent of GDP. This includes: other countries grew faster in 2012 (Sierra Leone, crop production (35.3 percent of GDP), animal farming and hunting Niger, Cote d’Ivoire, and Liberia). Ethiopia managed (9.8 percent of GDP), and forestry (3.6 percent of GDP). 2 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Figure 1: Economic Activity 1. International Commodity Prices 2. GDP Growth in China and the EU 700 2000 14% 1800 600 12% 1600 500 1400 10% 400 1200 8% 1000 6% 300 800 600 4% 200 400 2% 100 200 0% 0 0 –2% Jan-10 May-10 Sep-10 Jan-11 May-11 Sep-11 Jan-12 May-12 Sep-12 Jan-13 Q.1 Q.2 Q.3 Q.4 Q.1 Q.2 Q.3 Q.4 Q.1 Q.2 Q.3 Q.4 2010 2011 2012 Gold (left axis) Coffee, Arabica Crude Oil, Average China EU 27 3. GDP Growth in Selected Economies 4. Ethiopia: Demand Side Growth Decomposition 120% Private Private Sierra Leone Investment Investment Niger 100% Cote d’Ivoire Public 80% Public Investment Liberia Investment 60% Private Ethiopia Consumption Burkina Faso 40% Rwanda 20% Private Public China Consumption Consumption 0% Mozambique Public Net Export Zambia –20% Net Export Consumption Ghana –40% India –60% 0 2 4 6 8 10 12 14 16 18 2011–12 Average (2003–12) 5. Ethiopia: Crop Production 6. Ethiopia: Electricity Generation and Sales (GWH) 25% 220 700 200 600 20% 180 500 15% 160 400 10% 140 300 120 5% 100 200 0% 100 80 –5% 60 0 Jan-11 Apr-11 Jul-11 Oct-11 Jan-10 Apr-10 Jul-10 Oct-10 Jan-12 Apr-12 Jul-12 Oct-12 2007–08 2008–09 2009–10 2010–11 2011–12 2012–13 Cereal Pulses Oilseeds Grains Household Sales Industrial Sales Crude Production (GWH) (RHS) Source: 1.1: World Bank (2013), 1.2: Staff calculations based on MOFED data. 1.3. EUROSTAT. 1.4: World Bank. 1.5: CSA. 1.6. EEPCO. Note: Demand side decomposition uses GDP deflator for individual items. Note: Figure 1.1 prices are quoted in dollars/barrel (oil), cents/kg (coffee, Arabica), and US$ per troy ounce (gold). Recent Economic Developments 3 months of 2012/13 suggest that public investment means to achieve this. As a result, the growth in base continues to be the key growth engine. Private money has decelerated since June 2011 and contracted consumption imports, on the other hand, declined by 4.4 percent in 2011/12. In the first half of 2012/13, by 9 percent. Export growth was mute at 7 percent and however, base money growth increased by 17 percent. a widening trade balance could suggest an additional In addition, the Government has resumed borrowing drag on 2012/13 growth from net exports. from NBE in the current fiscal year. Changing bank reserve requirements can also Monetary Sector potentially undermine the effect of the develop- ment of base money. Successful targeting of broad Inflation is on a declining trend and has finally money (M2) through a base money nominal anchor returned to the single digit target. Since the peak depends on the steady relationship between base of 40.7 percent (y/y) in August 2011, headline infla- money and broad money. The NBE lowered the reserve tion dropped to 7.4 percent in March and 6.3 percent requirement ratio in January 2012 from 15 percent in May 2013—reaching a single digit level for the first to 10 percent, and in March 2013 lowered it fur- time since October 2010. The continuous decline ther to 5 percent, shifting the relationship between of headline inflation has largely been driven by the fast the two. The lowering of the reserve requirement decline in food price inflation. Non-food price infla- ratio in January 2012 increased the banks’ capacity tion, however, has also fallen noticeably by about five to extend credit and increased the money multiplier percentage points (Figure 2.1), which is an important (broad money divided by base money), thereby weak- break from past trends. ening the tightening effect of the base money contrac- International factors contributed to reduced tion in 2011/12 (Figure 2.5). inflationary pressure. A decomposition of inflation The recent reduction in the reserve requirement into tradable and non-tradable goods reveals that to 5 percent, in particular, may need to be accom- internationally traded goods (imported and exported panied by a change in the base money target commodities) in Addis Ababa exhibit a much faster to avoid potential unwarranted monetary stimulus. decline in inflation (Figure 2.2).4 Edible oil, cof- Given that banks are expected to deposit the additional fee beans, benzene, and chick peas are examples money in interest bearing (3 percent) certificates of tradable goods (Figure 2.3). Following the cur- of deposit, the measure is not expected to directly rency devaluations in 2009–10, tradable inflation affect the liquidity position of commercial banks. was generally higher than non-tradable inflation. However, the measure lowers the value of reserve Since February 2012, however, this trend reversed, money (the target of monetary policy) without chang- partly as the result of the slow pace of nominal cur- ing the liquidity in the banking system. As a result, rency depreciation. monetary stimulus could potentially be introduced Tighter monetary policy in 2011/12 contrib- if the base money target is kept unchanged, unless uted to the inflation decline, although the monetary base money supply growth is kept below the target. stance has loosened in 2012/13. Ethiopia is targeting Despite low base money growth, the expansion base money (or reserve money) as a nominal anchor of domestic credit kept broad money supply growth for monetary policy to control inflation. Accordingly, high. Broad money supply (M2) growth is declining broad money growth has been kept relatively low slightly compared to 2011/12, although it remains since the beginning of 2011/12. A temporary dis- continuation of direct financing of the budget by the 4 Owing to considerable data processing requirements, this point is illus- National Bank of Ethiopia (NBE) and sales of foreign trated using data for Addis Ababa only. Historically, there has been a very exchange (to reduce domestic liquidity) were the high correlation between Addis Ababa and other cities on these variables. 4 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Figure 2: Monetary Sector 1. Inflation (year-on-year) 2. Inflation, Tradables and Non-tradables (Addis, y/y) 100% 50% 80% 40% 60% 30% 40% 20% 20% 0% 10% –20% 0% Jan-11 Jun-11 Nov-11 Feb-11 Jun-11 Oct-11 Feb-08 Jul-08 Dec-08 May-09 Oct-09 Mar-10 Aug-10 Apr-12 Sep-12 Feb-13 Feb-08 Jun-08 Oct-08 Feb-09 Jun-09 Oct-09 Feb-10 Jun-10 Oct-10 Feb-12 Jun-12 Oct-12 Feb-13 General (N) Food Non-Food General Tradables Non-Tradables 3. Prices of Selected Tradable (y-o-y, Addis Ababa) 4. Money Supply Growth (year-on-year) 160% 140% 140% 100% 120% 100% 60% 80% 60% 20% 40% 0% 20% 0% –20% –20% –40% –60% May-11 Oct-11 Jan-10 Apr-10 Jul-10 Oct-10 Jan-11 Apr-11 Jul-11 Oct-11 Jan-12 Apr-12 Jul-12 Oct-12 Jan-13 Jan-08 Jun-08 Nov-08 Apr-09 Sep-09 Feb-10 Jul-10 Dec-10 Mar-12 Aug-12 Jan-13 Edible Oil Coffee Beans Benzene Chickpeas Net Foreign Asset Domestic Credit Broad Money 5. Money Growth (year-on-year) 6. International Reserves 50% 4000 3.0 2.8 3500 40% 3000 2.5 30% 2.4 2500 20% 2000 2.0 2.0 1500 10% 1.6 1000 1.5 0% 500 –10% 1.2 0 1.0 May-11 Oct-11 Feb-11 Jul-11 Dec-11 Jun-08 Nov-08 Apr-09 Sep-09 Feb-10 Jul-10 Dec-10 Mar-12 Aug-12 Jan-09 Jun-09 Nov-09 Apr-10 Sep-10 May-12 Oct-12 Money Multiplier (left axis) Broad Money (M2) Base Money Foreign Reserve (mill $) Reserve in Months of Import (right axis) Source: 2.1 CSA. 2.2: Own calculations based on CSA data. 2.3: World Bank. 2.4–2.6: NBE. Recent Economic Developments 5 at high levels (29 percent in January 2013). This is rel- option, but preparations need to be made now. atively high compared to the money supply growth In order to establish inflation expectations as a tar- target of the Growth and Transformation Plan (GTP), get of monetary policy in Ethiopia, the institutional and is consistent with the real GDP growth and annual framework for the public to form inflation expecta- inflation targets, which is less than 20 percent.5 Broad tions would need to be strengthened. The NBE would money growth is driven by domestic credit growth need to start publishing credible short- and medium- (27 percent in January 2013). Meanwhile, net for- term inflation forecasts. In addition, other private eign assets declined by 15 percent over the same and public institutions, as well as academia and think period, reflecting the deficit in the overall balance tanks, would have to play their role to publish inde- of payments. pendent views on the future inflation outlook. At the In conclusion, while attainment of single digit same time, the central bank could continue its analysis inflation is a commendable achievement, there of monetary and credit developments, which then is a risk that this could become a short-lived expe- could be fed into the overall mechanisms to inform rience. Monetary policy has loosened over the past the public on the future outlook on inflation. six months, including: (a) stronger reserve money growth; (b) a return to direct central bank financing Financial Sector of budgetary outlays; and (c) the impact of changing reserve requirements on base money targeting. This, Financial intermediation is relatively low combined with an expansionary consolidated fiscal in Ethiopia and it is on a declining trend. Financial policy stance (discussed later), makes it challenging intermediation is a driving force for economic devel- to maintain single digit inflation. opment, but Ethiopia is falling behind its peers in this Going forward, the monetary authorities area (Figure 3.1). In 2011, credit to the private sec- may also wish to rethink their current monetary tor in Ethiopia was equivalent to about 14 percent policy strategy. The attainment of single digit level of GDP, compared to the regional average of 23 per- inflation (provided that it can be maintained) offers cent of GDP. Moreover, while the worldwide trend has an opportunity to rethink the current monetary policy been an increase in private sector credit, Ethiopia has framework. The introduction of the monetary base experienced a decline of about five percentage points as a nominal anchor was the right strategy at a critical since 2004. An expansion in credit to the private sector moment in time, when inflation was a serious macro- enables firms to invest in productive capacity, thereby economic challenge. The strategy has also strengthened laying the foundation for a sustainable growth path. the central bank’s reputation for being able to control The macroeconomic environment has con- inflation. The current tool kit, however, is essentially tributed to the declining depth of the financial limited to foreign exchange market interventions. sector. Historically high inflation resulted in steeply A new strategy could pursue a more indirect conduct negative real interest rates. Low nominal deposit rates, of monetary policy operations, in line with experiences which do not respond to inflation because of excess in modern central banks around the world. Open liquidity of commercial banks, erode the real value market operations based on an effective interbank of deposits, discourage savings, and dampen demand market, as well as short-term credit facilities to balance commercial bank balances with the central bank, are effective instruments seen elsewhere, but are largely 5 This comparison is based on the quantity theory of money absent in Ethiopia. (∆M=πtarget+∆Y–∆V) which states that for a constant velocity of money (∆V=0), money supply growth (∆M) at rates faster than real GDP growth The eventual introduction of an inflation- (∆Y) plus expected inflation rate (πtarget) will induce inflation above targeting regime is, in the long term, one possible the targeted inflation rate. 6 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status for broad money. Moreover, in April 2011 the mon- is much more concentrated than the SSA and Low etary authorities issued a directive requiring private Income Group averages (Figure 3.5). Finally, accord- banks to purchase NBE bills equivalent to 27 percent ing to NBE, the non-performing loan ratio is at very of any new loan disbursements.6 These bills have a low- low levels of 1.4 percent (Figure 3.6). interest earning of 3 percent and a maturity of 5 years. This directive, while not affecting the level of liquidity Fiscal Sector in the system, appears to have had an additional nega- tive impact on private banks’ intermediation activi- The General Government fiscal balance registered ties. In particular, it has encouraged banks’ purchases a surplus of Birr 11.3 billion during the first half of Government bonds (T-bills and NBE bills), while of 2012/13 as a result of revenue growth and reducing the growth rate of loans. Figure 3.2 compares spending restraint.7 By comparison, a fiscal sur- the 15 month period before and after the introduction plus of Birr 7 billion was registered during the same of the directive. It shows that the investment of private period of 2011/12. While caution should be exercised banks in Government bonds increased by 127 per- in interpreting fiscal results prior to the closure of the centage points, while loans increased by 20 percentage fiscal year, this performance suggests prudent fiscal points in comparison. management at the General Government level. Government-owned banks dominate Ethiopia’s Total revenues and grants are growing banking system. This makes Ethiopia an exception at a slightly slower pace than output in real terms. within sub-Saharan Africa and across the devel- Total revenues reached Birr 67.5 billion during oping world, where banking systems have much the first six months of 2012/13, which represents higher shares of private and foreign participation. a 23.5 percent increase in nominal terms, compared As of December 2012, public banks represented to the same period of the previous fiscal year. When over 70 percent of total banking sector assets and discounting for consumer price inflation, real revenues provided 77 percent of total loans to the economy, increased by only 5.4 percent. with the remaining 23 percent being divided among Increased government revenues are attrib- the 15 private banks. By comparing the 15 months uted mainly to higher tax collection. Tax revenues before and after April 2011, the percentage of new increased by 9.3 percent in the first half of 2012/13, loans from public banks to the private sector has as a result of solid performance in direct tax and been reduced by half, from 21 percent to 10 percent domestic indirect tax collection, while revenues from (Figure 3.3.) foreign trade were unchanged in real terms. Non-tax Despite the overall disintermediation trend, revenues, on the other hand, declined by 15 percent the Ethiopian financial sector continues to have in real terms. the potential to be a driver of growth. The banking General government exercised spending sector remains stable, well capitalized, and continues restraint as total expenditures increased by less to be highly profitable. Figure 3.4 shows that the than 1 percent in real terms. Total expenditures Ethiopian banking sector ranks higher than the SSA reached Birr 56.2 billion in the first half of FY12/13, average in terms of profitability, measured on the basis of Return on Equity (ROE). High profitability 6 This measure was introduced with the purpose of ensuring long term is also explained by limited competition. Although finance to the private sector. The proceeds of NBE bills are directed to the Development Bank of Ethiopia (DBE), which lends mainly to the private the total number of banks operating in Ethiopia has sector. Data constraints on DBE lending activities and loan term compo- increased from 11 in 2006 to 18 in 2012, the bank sition would have facilitated a more comprehensive analysis of whether the objective of the measure is being achieved. assets concentration index (focusing on the three 7 The results reported here refer to those of the General Government, biggest banks) shows that Ethiopia’s banking sector which exclude State Owned Enterprises. Recent Economic Developments 7 Figure 3: Financial Sector 1. Private Sector Credit (% of GDP) 2. Private Banks: Lending & Bond Investments 40 160% 35 140% 120% 30 100% 25 80% 20 60% 40% 15 20% 10 0% % Change in Total Lending % Change in Government 5 Bonds’ Investment 0 From Private Banks 2004 2005 2006 2007 2008 2009 2010 2011 Jan 2010–Mar 2011 (before introduction of the 27% rule) Kenya SSA Average LIC Average Tanzania Ethiopia Apr 2011–Jun 2012 (after introduction of the 27% rule) 3. Banking Sector Credit Growth 4. Profitability of the Banking Sector (ROE) 100% 50 90% 45 80% 40 70% 60% 35 50% 30 40% 25 30% 20 20% 10% 15 0% 10 All Public Private All Public Private Banks Banks Banks Banks Banks Banks 5 0 Jan 2010–Mar 2011 Apr 2011–Dec 2012 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 Lending to Private Sector Lending to Cooperatives Lending to Public Enterprises Ethiopia Kenya SSA Average 5. Bank Assets Concentration Index 6. Non-Performing Loans 100 16 90 14 80 70 12 60 10 50 40 8 30 6 20 10 4 0 2 2004 2005 2006 2007 2008 2009 2010 2011 0 SSA Average LIC Average Ethiopia 2006 2007 2008 2009 2010 2011 2012 Uganda Tanzania Kenya Ethiopia SSA Average LIC Average Source: 3.1: World Bank (FinStats 2013) & NBE; 3.2 and 3.3: Own calculations based on monetary survey; 3.4 and 3.5: World Bank (FinStats 2013); 3.6: World Bank (FinStats 2013) & NBE. 8 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status which was 18 percent higher than in the same period fiscal sector in Ethiopia is the lack of consolidated of FY11/12 in nominal terms. Given high inflation, public sector fiscal data, which also includes the however, this represents a 0.8 percent increase in total balance of State Owned Enterprises. In the absence spending in real terms. Capital expenditures increased of such numbers, the IMF and the World Bank, in the by 2.6 percent in real terms, while recurrent expendi- most recent Debt Sustainability Analysis (DSA, see tures declined marginally in real terms (Table 1 refers). Box 1) prepared estimates of the consolidated public Capital spending is expected to rise during the second sector fiscal balance. The results suggest a primary half of FY12/13. balance deficit averaging 5.4 percent of GDP in the The federal government did not make use period 2011 to 2013, compared to a central govern- of direct central bank advances during the first ment primary balance of 2 percent of GDP over half of the fiscal year. In July 2011, the central bank the same period. The difference reflects substantial ceased providing new credits to the government for investments by State Owned Enterprises as a part budget financing, as part of the Government’s efforts of the overall growth strategy of promoting public to reduce inflation. It is therefore encouraging that investment (as discussed in Section B). However, the Government continued this good practice during even this estimate may be on the low side when ana- the first semester of FY12/13. During the second half lyzed from the domestic and external financing side. of FY12/13, however, the Central Bank has advanced Improved data availability on State Owned Enterprise Birr 2 billion out of a planned Birr 9 billion. Going activity would go a long way in strengthening the fis- forward, this practice should be limited as much cal analysis. as possible in order to maintain the commendable results achieved in taming inflation. External Sector Despite Federal Government prudence, the consolidated public sector fiscal stance continues The trade deficit increased by 15 percent to be expansionary. A key challenge in analyzing the as import growth outpaced export growth . Table 1: Federal Government Fiscal Performance FY11/12 FY12/13 1.Semester 1.Semester Nominal Change (Birr Million) (Birr Million) (%) Real Change (%) Total revenues & grants 54,659 67,497 23.5 5.4 Tax revenues 42,102 53,930 28.1 9.3 Direct taxes 15,328 19,903 29.8 10.8 Domestic Indirect taxes 10,363 14,775 42.6 21.6 Foreign trade 16,411 19,253 17.3 0.1 Non-tax revenue 7,436 7,410 –0.3 –15.0 Grants 5,121 6157 20.2 2.6 Total expenditures 47,611 56,243 18.1 0.8 Recurrent 22,638 26,219 15.8 –1.2 Capital 24,973 30,024 20.2 2.6 Balance 7,048 11,254 — — Source: MOFED. Note: Average inflation was 17.2 percent between the two periods. Recent Economic Developments 9 Box 1: Debt Sustainability and the IDA Non-Concessional Borrowing Policy Ethiopia’s risk of external debt distress remains low, according to the most recent Joint IMF World Bank Debt Sustainability Analysis (DSA). In the latest DSA, published in October 2012, external debt burden indicators under the baseline and shock scenarios are projected to rise, but remain below the country-specific indicative thresholds. Under the baseline scenario, the present value of public and publically guaranteed external debt is projected to rise from 14 percent of GDP in in 2011/12 to a peak of 18 percent of GDP in 2016/17. Similarly, the present value of debt as a share of exports will peak at 107 percent in 2016/17. The debt service-to-exports ratio will peak at 9 percent in 2018/19 (see IMF Article IV Report for Ethiopia for more details). According to the 2012 DSA, Ethiopia’s overall public sector debt dynamics appear sustainable, although sensitive to several alternative scenarios. Public sector debt ratios are projected to rise over the medium term, suggesting that close monitoring of borrowing by public sector enterprises remains a necessity. The present value of the public debt to GDP ratio is projected to increase from 29 percent in 2012 to 35 percent in 2015. Alternative scenarios, including growth shocks or an unchanged fiscal balance, result in a rising debt trajectory. As a recipient of MDRI (Multilateral Debt Relief Initiative) support, the Government of Ethiopia has committed itself to prudent fiscal and debt management to avoid a renewed excessive buildup of external public debt. An important instrument in this respect is the IDA Non-Concessional Borrowing Policy, which applies to MDRI recipients, including Ethiopia. In April 2013, IDA authorized a US$1 billion ceiling for Ethiopia for FY13 and, in principle, a similar ceiling for FY14 and FY15. This implies that Ethiopia can borrow up to US$1 billion per year from other creditors, on non-concessional or commercial terms, as long as these loans are used to finance projects that are growth enhancing. According to the IDA Non- concessional Borrowing Policy, a loan counts at the point of signing the loan contract (irrespective of the disbursement profile). The decision was informed by the above-mentioned DSA analysis, which demonstrates that such a ceiling is consistent with the maintenance of low risk of external debt distress. During the first seven months of 2012/13, goods if export prices are quoted in dollars). The rising black exports increased by 7 percent while goods imports market premium is indicative of currency misalign- increased by 13 percent. Both items experienced ment (Figure 4.4). slower growth compared to the previous year, where A weak external environment also contrib- export growth reached 15 percent and import uted to poor export performance. As previously growth 33 percent. mentioned, economic activity in key export markets The deteriorating trade balance is partly is not conducive, as illustrated by low cut flower explained by the appreciation of the real effective demand in the EU. Declining international prices exchange rate. Although the Birr has nominally were also detrimental. depreciated by 0.4 percent per month over the past A widening trade balance adds to the struc- two years, domestic inflation has exceeded foreign tural current account deficit that Ethiopia is fac- inflation by a wide margin. As a result, the real effec- ing. Goods imports are about four times larger tive (i.e., trade weighted) exchange rate has appre- than goods exports, giving rise to a structural trade ciated by about 40 percent since September 2010. deficit. Although services exports are about the same To illustrate, a monthly nominal depreciation size as goods exports, a similar amount of services of 1.7 percent would have been necessary to maintain imports imply that the services balance is relatively the real effective exchange rate constant over the past small. As a result, the trade balance is the major two years. Real exchange rate appreciation encour- driver of the goods and services deficit. The latter, ages imports, as they become relatively cheaper, while in turn, is financed through a combination of private discouraging exports, which become relatively more transfers (remittances and NGO transfers), public expensive (or yield lower local currency revenues, transfers (official development assistance), FDI, and 10 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Figure 4: External Sector 1. Quarterly Balance of Payments (Percent of GDP) 2. Annual Balance of Payments (Percent of GDP) 15% 30 10% 5% 20 0% 10 –5% –10% 0 –15% –20% –10 –25% –20 –30% QII QIII QIV QI QII QIII QIV QI QII QIII QIV QI QII –30 2004 2005 2006 2007 2008 2009 2010 2011 2012 2009–10 2010–11 2011–12 2012–13 Loans FDI Public Transfers Private Transfers FDI Overall Balance CAB Trade Balance Service Balance Trade Balance CAB 3. Export Growth and Real Effective Exchange Rate 4. Exchange Rates: Black Market Premium (Pct.) 90 45 25% Major devalutions 5% 10% 5% 20% 100 40 35 20% 110 30 120 15% 25 130 20 140 10% 15 150 10 5% 160 5 170 0 0% Mar-11 Jul-11 Nov-11 Jul-03 Nov-03 Mar-04 Jul-04 Nov-04 Mar-05 Jul-05 Nov-05 Mar-06 Jul-06 Nov-06 Mar-07 Jul-07 Nov-07 Mar-08 Jul-08 Nov-08 Mar-09 Jul-09 Nov-09 Mar-10 Jul-10 Nov-10 Mar-12 Jul-12 Nov-12 Mar-13 Jun-11 Dec-11 Jun-06 Dec-06 Jun-07 Dec-07 Jun-08 Dec-08 Jun-09 Dec-09 Jun-10 Dec-10 Jun-12 Dec-12 180 –5 Export growth REER 5. Growth in Export Value by Commodity (percent) 6. Export Price and Quantity Effects, 7m of FY2013 Others –4% Leather 18% 24% –19% Leather –4% 14% 47% Live Animals –30% Live animals –20% 51% Pulses –5% 67% 75% Pulses 8% 1% Flower Flower –6% –13% 16% Chat –2% Chat 7% 10% 1% –1% 16% Gold Gold 17% 53% 0% Oilseeds 9% Oilseeds 49% –8% Coffee 13% Coffee –29% –7% 59% –25% –5% 15% 35% 55% 75% –40% –20% 0% 20% 40% 60% 80% 7 Months FY2013 7 Months FY2012 Unit Price Quantity Source: National Bank of Ethiopia and IMF (REER). Recent Economic Developments 11 external borrowing (Figure 4.2). Official development The Long View assistance (ODA) finances about half of the current account deficit, which averaged 10.8 percent of GDP Middle Income Status by 2025: Is Ethiopia in 2004–12 without ODA, and 5.2 percent of GDP on Track? when including ODA. A slower goods export during the first seven Ethiopia has experienced strong economic growth months of 2012/13 can be explained by a decline over the past decade. Economic growth aver- in most products, except coffee and chat. A product- aged 10.7 percent per year from 2003/04 to 2011/12,8 by-product analysis of volume and price effects yield compared to the regional average of 5.4 percent the following (products are ranked by their share of the (Figure 5.1). Since a rising population contributed export basket): to economic growth, output per capita growth more accurately reflects the underlying economic perfor-  Coffee experienced increased value growth as ris- mance. Still, at 8.3 percent per capita growth, Ethiopia ing volumes more than compensated for a sharp outperformed the average for sub-Saharan African drop in prices. countries (3.3 percent). Growth was induced through  Gold and chat experienced modest value growth, a mix of factors, including agricultural modernization, on account of higher volumes in the context the development of new export sectors, strong global of unchanged prices. commodity demand, and government-led develop-  Oil seeds exports were stagnant following strong ment investments. growth in 2011/12, as rising volumes compen- The overarching aspiration of the Government sated the price decline. of Ethiopia is to achieve middle-income status by 2020–23. This aspiration raises several ques- Imports of public capital goods explain most tions: Would a continuation of the historically high of the observed increase in goods imports. Eighty- growth performance be sufficient to reach this goal? seven percent of the increase in imports during the Does Ethiopia need more of the same type of growth, first seven months of the current fiscal year can or is a change in the sources of growth warranted? be explained by an increase in capital imports. Given A narrow interpretation of the middle-income that public investment is almost three times larger aspiration is to examine whether Ethiopia will pass than private investment, most of this can be attributed the relevant World Bank income threshold by 2025. to public investment activity. The import of consumer Currently, the World Bank classifies countries into goods, meanwhile, actually declined by 9 percent over four groups, depending on their Gross National the period. Income (GNI) per capita: low income (less than US$1,025), lower middle income (US$1,026–4,035), upper middle income (US$4,036–12,475), and high income (above US$12,475). With a GNI per capita of US$370 in 2011, Ethiopia was classified as a low- income country and one of the ten poorest countries 8 This report uses GDP growth rates as reported by the Government of Ethiopia (GoE). According to the IMF, however, alternative arms- length methodologies suggest that Ethiopian historical annual growth rates could be overestimated by as much as three percentage points in recent years. The IMF has indicated that the official methodology improved considerably in 2011/12. 12 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status in the world. The regional average is US$1,270 with a value added contribution of just 8 percent (US$920 excluding South Africa). (Figure 5.3). When examining the potential attainment In fact, the service sector is now larger than of middle-income status, it is important to realize agriculture in terms of value added. Over the past that the threshold increases over time. While the decade, the economy has been undergoing a shift current threshold is US$1,025, this has not always from agriculture production towards increased reli- been the case. In 1987, for instance, the threshold ance on services. Between 2003/04 and 2011/12, was US$480. This is mainly the result of the fact the value added contribution of agriculture fell that this is a nominal threshold, which is affected from 51 to 44 percent, while the services sector by inflation. By applying the historically observed increased from 38 to 45 percent. The value added annual nominal growth rate (3.2 percent) to the growth of the industrial sector remained relatively threshold, we derive an unofficial and approximate unchanged at 9–10 percent (Figure 5.4). expected threshold by 2025 of US$1,430. In sum, Public investment has become the growth engine Ethiopia would have to surpass a GNI per capita in recent years, according to a demand side analysis of US$1,500 by FY 2025 to be classified by the World of national accounts.10 Between 2003/04 and 2011/12, Bank as a lower middle-income country. 71 percent of output growth was explained by private If Ethiopia can repeat its recent historical consumption, while public investment contributed by growth performance of 10.7 percent per year, it 23 percent. Over the past three years, the public invest- would classify as a middle-income country by ment share in contributing to growth has increased 2023. This projection is based on a modified historical from 31 to 63 percent, however. In other words, almost scenario, calculated using the real GDP growth rate two-thirds of the 8.5 percent growth in 2011/12 can be achieved in the 2003/04–2011/12 period: 10.7 per- attributed to public investment (Figure 5.5). cent per year. Additional assumptions (not historically Consequently, the public investment share observed) include (annual rates): single digit inflation of GDP is increasing, while other demand compo- (9 percent), currency depreciation (7 percent), and a nents are declining. Between 2007/08 and 2011/12, population growth rate of 2.5 percent.9 Figure 5.2 public investment increased from 14 to 25 percent illustrates the projections. of GDP. Private consumption, on the other hand, To sustain such a high pace of economic declined from 85 to 77 percent of GDP. Public con- growth, it is also important to consider its sources. sumption fell from 11 to 7 percent of GDP; while pri- As discussed next, the sources of growth in Ethiopia vate investment fell from 13 to 10 percent of GDP, with have gradually shifted over the decade: (a) from agri- the residual being explained by net exports (Figure 5.6). culture to services, and (b) from private consumption In sum, Ethiopia has achieved high economic to public investment. growth in recent years through a “big push” in pub- lic investment, aimed at creating the conditions for Sources of Growth and Structural Change Ethiopia’s recent high growth has been the result 9 The historical values for the nominal exchange rate depreciation and inflation, 9.1 and 18.4 percent, respectively, were not considered realistic of an expansion of the services and agricultural for a forward projection, hence the modification. sectors. Fifty percent of the growth in value added 10 National account estimates in Ethiopia can be derived from either the supply side (by economic sectors) or the demand side (expenditure com- between 2003/04 and 2011/12 is attributed to the ponents). These two approaches are theoretically and methodologically services sector. The agriculture sector, meanwhile, consistent. All data used here are from the National Accounts Directorate of the Ministry of Finance and Public Credit. As explained in footnote 1, accounted for 42 percent of value added growth. some caution must be exercised in interpreting demand side statistics Industrial sector performance was relatively meager given the quality of data. Recent Economic Developments 13 Figure 5: Ethiopia: Economic Growth and its Sources 1. Economic Growth, 2004–12 2. GNI Per Capita Projection 14% 1600 12% 1400 10% 1200 8% 1000 6% 800 4% 600 2% 400 0% 2004 2005 2006 2007 2008 2009 2010 2011 2012 200 0 Ethiopia Real GDP Growth (annual %) 2012 2014 2016 2018 2020 2022 2024 Ethiopia Real GDP Per Capita Growth (annual %) SSA Real GDP Growth (annual %) Historical Growth Rate (10.7%) SSA Real GDP per Capita Growth (annual %) Lower Middle Income Threshold (projected) 3. Sectorial Contributions to Real GDP Growth 4. Sectorial Shares of GDP (Constant Prices) 14% 60% 12% 50% 10% 8% 40% 6% 30% 4% 20% 2% 0% 10% 2010–11 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2011–12 0% 2010–11 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2011–12 Others Pub Adm, Educ, Health Finance & Real Estate Trade & Transport Construction Manufacturing Agriculture Agriculture Services Industry 5. Demand Contributions to Real GDP Growth 6. Demand Shares of GDP (Constant Prices) 20% 30% 90% Private Consumption (% of GDP) 15% 25% 85% Percentage Points Percent of GDP 10% 20% 80% 5% 15% 75% 0% 10% 70% –5% 5% 65% –10% 0% 60% 2000–01 2010–11 2000–01 2010–11 2001–02 2002–03 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2011–12 1999–00 2001–02 2002–03 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2011–12 Gov. Cons Priv. Cons Public Inv. Public Consumption Private Investment Private Inv. Net Export GDP Growth Private Consumption (right axis) Public Investment Source: 5.1: MOFED and IMF. 5.2: Staff calculations based on WB data. 5.3–5.6 Staff calculations based on MOFED data. Note: The demand composition of GDP growth is based on staff estimates using the GDP deflator for individual demand components. Source: IMF, World Bank, MOFED. 14 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status self-sustained growth. The Government of Ethiopia has therefore reversed over the past 25 years. In 1987, is pursuing this strategy so that the private sector can private investment amounted to 16.1 percent of GDP, ultimately become the key driver of growth. In that while public investment equaled 11.7 percent of GDP. sense, public investment is intended to be an enabler By 2011, public investment was three times larger than of growth. For that reason, the next section takes private investment (Figure 6.5). a more detailed look at the long-term trends in public, The public investment strategy initially suc- private, and total investment over the past 25 years. ceeded in raising the total investment rate, but coin- The analysis, in turn, provides an appropriate con- cided with declining private investment in recent text for the more detailed analysis of savings in the years. In the decade from 1994 to 2004, total invest- next chapter. ment increased from 16.4 to 25.5 percent of GDP. This was driven by a public investment increase Public Investment: Ethiopia’s Growth Engine11 of 10.6 percent of GDP and a relatively moderate decline in private investment by 1.6 percent of GDP. Long-term trends Since 2004, however, total investment remained stag- nant. The decline in private investment intensified Ethiopia has the third highest public invest- between 2004 and 2011 with about 3 percent of GDP. ment rate in the world. In 2011, Ethiopia’s public Increases in public sector investment could only keep investment to GDP ratio was 18.6 percent. This the total investment rate constant, but not increase number was exceeded only by Turkmenistan (38.6 per- it any further (Figure 6.6). cent) and Equatorial Guinea (24.3 percent), as shown in Figure 6.1. On the other hand, the private invest- Types of public investment and sources ment rate in Ethiopia is the sixth lowest in the world of financing at 6.9 percent. Only Angola (2.8 percent), Azerbaijan (3.9 percent), Swaziland (5.2 percent), South Sudan The increasing public investment is financed through (6.6 percent), and Malawi (6.6 percent) had lower a combination of restraint in government consump- private investment rates in 2011. (Figure 6.2). In spite tion, as well as through borrowing. Public investment of this, Ethiopia registers a total investment rate which has essentially two sources: budgetary and non-budget- is higher than expected, given its level of development. ary, where the latter refers to investment of State Owned Total investment (public and private) equaled 25.5 per- Enterprises. An expansion of the budgetary component cent of GDP in 2011, as illustrated in Figure 6.3. was enabled through a shift in budgetary priorities, Encouragingly, this level of total investment is in line as well as aid and domestic/external borrowing. State with levels recommended by the Commission Owned Enterprises investments, in turn, benefitted on Growth and Development (2008) for a country from domestic and foreign borrowing. to achieve high and sustained economic growth. Initially, higher public investment was accom- The total investment rate increased by almost panied by increased government consumption, ten percent of GDP since 1987, mainly as a result but the latter has been decreasing since 2000. of an expansion in public investment. The total investment rate increased from 16.4 to 25.5 percent of GDP between 1987 and 2011 (Figure 6.4). Over 11 This report uses the national accounts definitions of the Ministry of Finance and Public Credit. Investment is defined as the sum of fixed this period, public investment rose by 14.2 per- capital formation and change in inventories. Public investment is the cent of GDP, while its private counterpart declined sum of investment by federal and regional governments as well as pub- lic enterprises. Private investment includes investment by households, by 4.8 percent of GDP. The relative importance of the private enterprises and non-profit institutions serving households. Total public and private components of total investment investment is the sum of private and public investment. Recent Economic Developments 15 Figure 6: Public and Private Investment Trends 1. Public Investment Rate by Income, 2011 2. Private Investment Rate by Income, 2011 40 50 40 30 30 Tanzania 20 Ethiopia India 20 Uganda Rwanda 10 Tanzania Rwanda 10 India Uganda Ethiopia 0 0 6 8 10 6 7 8 9 10 11 ln GNI Per Capita, PPP (current int. dollars) ln GNI Per Capita, PPP (current int. dollars) 3. Total Investment Rate by Income, 2011 4. Ethiopia: Total Investment, 1987–2011 60 30 50 25 China 20 40 Percent of GDP Tanzania Vietnam 15 Public Investment 30 Korea, Rep. Ethiopia India Singapore 10 Uganda 20 Rwanda Kenya 5 Private Investment 10 0 6 8 10 1991 2001 2011 1989 2009 1987 1993 1995 1997 1999 2003 2005 2007 ln GNI Per capita, PPP (current int. dollars) 5. Public and Private Investment, 1987–2011 6. Public and Private Investment 1987–2011 20 30 18 16 25 14 20 Percent of GDP Percent of GDP 12 10 15 8 6 10 4 5 2 0 0 1987 1994 2004 2011 1991 2001 2011 1987 1989 1993 1995 1997 1999 2003 2005 2007 2009 Private Investment Public Investment Private Investment Public Investment Source: 6.1–6.6. World Bank (WDI). 16 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Between 1987 and 2000, the Government was able Policy options to increase both recurrent and capital spending as overall budgetary resources increased with higher International experience underscores the relevance foreign aid and taxes. Since 2000, however, the budget of emphasizing public investment as key element envelope has remained largely unchanged (as a share of a growth strategy. According to the Growth and of GDP). As a result, the continued increase in public Development Commission (2008: 36): “No coun- investment has been financed through lower govern- try has sustained rapid growth without also keeping ment consumption (including recurrent expenditures, up impressive rates of public investment—in infra- such as the public wage bill) (Figure 7.1). structure, education, and health. Far from crowding Budgetar y and non-budgetar y public out private investment, this spending crowds it in. investments are approximately similar in size. It paves the way for new industries to emerge and Between 2006 and 2011, budgetary public investment raises the return to any private venture that benefits and off-budgetary public investment through State from healthy, educated workers, passable roads, and Owned Enterprises each amounted to about 6 percent reliable electricity ... Unfortunately, we discovered, of GDP. In the decentralized fiscal system, budgetary infrastructure spending is widely neglected.”12 public investment is financed from both the central In order to achieve its impressively high public and regional levels of government. The central level investment rate, the Ethiopian Government has assumes the greater share of the investment (two- implemented a number of macroeconomic policies. thirds) and regional levels assume the remainder (one- These include: (a) low nominal interest rates to reduce third). (Figure 7.2). the cost of financing; (b) a gradual real appreciation Public investment is primarily financed from of the exchange rate to keep capital import costs domestic sources. About three quarters of public low; (c) channeling banking system credit and for- investment (budgetary and non-budgetary) is financed eign exchange primarily towards public enterprises; from domestic sources through the budget or the (d) restraining recurrent public expenditures to create domestic banking system with the remainder financed sufficient fiscal space to finance public capital outlays; through external borrowing (Figure 7.3). and (e) maintaining low external foreign exchange State Owned Enterprises are increasingly reserves to finance high yielding public investment. absorbing domestic banking sector credit. In the Policies that promote public investment six-month period from June 2011 to December 2011, in Ethiopia, meanwhile, also affect other policy 71 percent of new loans were directed towards public objectives. The same set of macroeconomic policies enterprises. This share increased to 89 percent during that are fostering high public investment and eco- the second half of 2012. A substantial share of the nomic growth may simultaneously make it difficult available foreign exchange is similarly diverted towards to achieve other policy goals. Whether the right bal- public investment. ance is currently being struck is ultimately a political Finally, it is important to emphasize gross decision. This analysis aims to highlight the impor- domestic savings, transfers, and external borrow- tant policy trade-offs associated with it. To illustrate: ing as sources of finance for public investment. (a) low nominal interest rates can be associated with Figure 7.5 illustrates this point for total investment. low levels of saving—thus limiting the foundation Given the size of public investment in total invest- for investment and sustained economic growth (see ment in the 2000s, it is clear that private and official transfers played an increasing role in financing public 12 The commission, headed by Nobel Laureate Michael Spence investment over this period (11.7 percent of GDP per (2008),based its recommendations on the experience of 13 high growth year, on average, in the 2000s). economies (9 from East Asia). Recent Economic Developments 17 Figure 7: Public Investment: Sources and Financing 1. Public Investment and Consumption, 87–11 2. SOE, Regional, Public, and Private Investment 20 30% 18 16 25% 14 Percent of GDP Percent of GDP 20% 12 10 15% 8 6 10% 4 5% 2 0 0% 2006/07 2007/08 2008/09 2009/10 2010/11 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Private Public (Central) Government Investment Government Consumption Public (Regional) State Owned Enterprises 3. Public Ivestment by Sources of Financing 4. Banking Sector Credit Growth 20% 100% 18% 90% 16% 80% Percent of GDP 14% 70% Percent of GDP 12% 60% 10% 50% 8% 40% 6% 30% 4% 20% 2% 10% 0% 0% 2007/08 2008/09 2009/10 2010/11 Jun 11–Dec 11 Dec 11–Jun 12 Jun 12–Dec 12 Domestic Foreign Public Enterprises Private 5. Financing of Total Investment (% of GDP) 6. Tax Revenue to GDP by Income Level 25% 22 Tax Revenue/GDP, percent, 2007–2011 20% 20 Malawi Kenya Senegal 18 15% Benin Zambia 16 10% Togo 14 Uganda DRC Tanzania Rwanda 5% 12 Ghana 0% 10 1990s 2000s Ethiopia 8 Gross Domestic Savings Private and Official Transfers 0 500 1,000 1,500 FDI External Borrowing GDP Per Capita (Current US$), Average 2007–2011 Source: 1.1 World Bank (WDI), 1.2: Staff calculations based on MOFED data. 1.3: MOFED. 1.4: IMF, 1.5. Staff calculations based on WDI data. 1.6: World Bank. 18 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status chapter 2); (b) a strong currency (in real terms) makes effect can hurt private businesses. A weaker private it difficult to maintain external competitiveness and sector, in turn, reduces the ultimate “crowding-in” thereby achieving high export growth, which is key effect that the Government is aiming to achieve to economic growth; (c) the imperative of priori- in the end. By adjusting existing policy levers accord- tizing some sectors and activities in terms of credit ingly, it should be possible to contribute to a stronger and foreign exchange has meant other sectors have private sector already today instead of waiting for less access to such resources, thereby constraining it to be ready when the public infrastructure is in place their development; (d) recurrent public expenditure tomorrow. This also has the advantage of making the restraint, including on the public wage bill, may have growth strategy more sustainable, as it facilitates export implications for public sector effectiveness; and (e) low competitiveness and enhances the savings base. Policy foreign exchange reserves increase external vulnerabil- options include increase credit and foreign exchange ity, as this policy reduces an important buffer against to the private sector, a faster pace of nominal cur- external shocks. rency depreciation, and a gradual rise in the nominal The current policy mix makes it difficult for pri- interest rate. vate investment, private consumption, and exports The development of a strong and vibrant pri- to flourish. The analysis presented here suggests that vate sector in Ethiopia will eventually be needed private sector-led activities appear to be constrained to sustain high growth. In addition to the change by the policy choices favoring public investment. This in the policy mix discussed above, this would also is illustrated by the low levels of private investment, require an improvement in the general business envi- slow private consumption growth, and declining ronment (see Figure 8), especially in the area of trade export growth observed in recent years. logistics (discussed in Chapter 3). It could also While the ultimate objective is for high public include industrial policy, as already practiced by the investment to “crowd-in” the private sector, some Government of Ethiopia, including the development “crowding-out” may inevitably occur in the pro- of industrial zones. cess. In the case of Ethiopia, public investment proj- A stronger private sector would also expand ects “crowd-in” private activity in two ways. In the the taxable base and therefore provide addi- short term, private contractors of public investment tional resources to finance public investment. projects may benefit from this additional activity. The tax to GDP ratio in Ethiopia is one of the low- This largely benefits the private, non-tradable sector. est in the sub-Saharan Africa region (Figure 7.6). In the medium to long term (following the comple- According to IMF data, general government rev- tion of the project), the private sector will benefit enue (excluding grants), increased from 12 per- from the existence of the enhanced public infrastruc- cent of GDP in 2008/09 to 14.2 percent of GDP ture. On the other hand, in the process of provid- in 2011/12 compared with the low-income regional ing additional public investment, the private sector average of 18.9 percent in 2012. There is considerable in Ethiopia is crowded out through the credit and potential for Ethiopia policy makers to increase the foreign exchange allocation systems. tax base in the medium term, because of a thriving To nurture private activities, Ethiopia may private sector. need to reduce the pace of building public infra- The jobs agenda offers an additional motivation structure investments. Ideally, Ethiopian policy for policy makers to strengthen the private sector. makers should maximize the “crowding-in” effect and Between 2 and 2.5 million young people are enter- minimize the “crowding-out” effect that high public ing the labor market every year. This contrasts with investment can have on the private sector. If public the 2.5 million people in paid employment in 2011, investment proceeds too rapidly, the crowding out of which only 1.5 million were employed in the Recent Economic Developments 19 Figure 8: Private Sector Constraints in Ethiopia 1. World Bank Enterprise Survey (2011) 2. WB Survey of Chinese Investors in Ethiopia (2012) Access to Finance 33 Customs and Trade Regulation 84 Tax Administration 53 Access to Land 23 Access to Finance 40 Electricity 12 Tax Rates 35 Macroeconomic 35 Tax Administration 7 Labor Regulations 32 Tax Rates 7 Electricity 21 Inadequately Educated 18 Customs and Trade 5 Access to Land 17 Practices of the Informal 5 Transportation 15 Corruption 3 Corruption 11 Business Licensing and Permits 10 Transportation 2 Crime, Theft And Disorder 6 Inadequately Educated 2 Competitors in the Informal Sector 5 Political Instability 1 0 10 20 30 40 Source: World Bank Enterprise Survey (2011) and World Bank Survey of Chinese Investors in Ethiopia (2012). Note: Based on interviews with private sector enterprises. The numbers indicate the share of respondents identifying the problem. formal private sector. While new entrants to the labor Figure 9: Infrastructure Project Costs and market have to find some kind of job to sustain their Time Overrun, Cost Baseline Sample livelihoods, typically in the informal sector, the policy Ethiopia challenge is to generate as many high-quality jobs Tanzania as possible. Given the typical budgetary constraints Malawi faced by the Government of Ethiopia, the private sec- tor plays the most important role in addressing this Bangladesh policy challenge. Zambia Finally, in addition to ensuring a better balance Vietnam between public and private sector development, Philippines it is critical to ensure that public investment man- Guatemala agement is of high quality. Since weakness in public UK investment management can negate the core argu- ment that impressive rates of public investment are 0% 20% 40% 60% 80% 100% 120% 140% necessary for a country to sustain rapid economic Time Over-Run Cost Over-Run growth, attention to the processes that govern proj- Source: World Bank (forthcoming). ect selection and management is critical (see Box 2). Figure 9 illustrates some of the challenges often faced by countries, in terms of infrastructure project costs and time overrun, both of which appear relatively high in Ethiopia, in the sample of countries for which such data was available. 20 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Box 2: A Diagnostic Framework for Assessing Public Investment Management If well-managed, public investment can be a critical driver for growth. On the other hand, there are important risks to be aware of in the public investment management process. International experience offers many examples of low efficiency of public investment in a number of dimensions, including: (a) poor project selection, including wasteful “white elephant” projects, (b) delays in design and completion of projects, (c) corrupt procurement practices, (d) cost over-runs, (e) incomplete projects, and (f), failure to operate and maintain assets effectively. In this context, World Bank (2010) identifies eight key “must have” features of a well-functioning public investment system along with some diagnostic questions for evaluating efficiency. A summary is provided below to give a flavor of the issues, although readers are encouraged to consult the source for a comprehensive treatment of the topic. 1. Investment guidance, project development, and preliminary screening • Is there well-publicized strategic guidance for public investment decisions? • Is there a well-established screening process? How many projects are rejected? 2. Formal project appraisal • Does a formal process exist? If yes, is it mandatory? • What proportion of projects is formally appraised for costs and benefits? 3. Independent review of appraisal • Are reviews undertaken independently? What is the quality of such appraisals? 4. Project Selection and budgeting • Are donor-financed projects subject to the same rules as government projects? • Are appraisals screened by an external agency? • Is final project selection part of the budget process? 5. Project implementation • What is the completion rate of the public investment program? This is defined as the annual public investment budget divided by the estimated cost to complete the current investment program. • Do ministries undertake procurement plans in line with good practice (e.g., competitive tenders)? 6. Project adjustment • Has the government undertaken a rationalization of its public investment management program in the recent past? Did the rationalization improve prioritization? Did it result in cancelation or closure of the public investment program? 7. Facility operation • Are projects commissioned to private contractors and, if so, are contracts awarded using competitive bidding? Are international firms permitted to bid? 8. Project Evaluation • Is the actual Net Present Value (NPV) of completed projects measured? Is a project end evaluation undertaken to review the nature of differences relative to the estimated NPV at appraisal? Source: Rajaram, et. al (2010). Recent Economic Developments 21 Annex: Selected Economic Indicators Table 2: Ethiopia: Selected Economic Indicators (High Frequency) Sep-12 Oct-12 Nov-12 Dec-12 Jan-13 Feb-13 Mar-13 Apr-13 Inflation (Year-on-Year): % 18.9 15.9 15.7 12.9 12.5 10.3 7.6 6.1 Food 17.6 13.2 13.4 11.8 11.3 7.9 5.2 1.6 Non-Food 21.0 20.4 19.4 14.7 14.4 14.3 10.5 11.5 Inflation in AA (Year-on-Year): % 18.3 15.3 15.4 12.4 12.4 12.3 Traded Goods 6.8 4.9 5.1 2.7 3.6 1.7 Non-Traded 22.5 19.1 19.2 16.0 15.6 16.2 Monetary Growth (Year-on-Year): % M2 27.4 30.5 32.3 29.1 28.7 Domestic credit 36.8 41.3 35.7 27.4 29.3 Net Foreign Assets –29.2 –25.6 –19.4 –15.1 –15.4 Reserve Money –2.7 12.2 10.7 16.9 23.8 Gross reserves (Mill. $) 2,315 2,404 2,465 2,564 2,461 2,725 Exchange Rate Exchange rate (Birr/$) 17.9 18.0 18.1 18.1 18.2 18.3 18.4 18.5 Real Exchange rate 135.0 133.9 134.8 133.7 133.8 133.6 135.5 Black market premium (%) 3.4 2.6 3.3 5.0 7.4 8.6 Trade Deficit: Cumulative 32.4 36.2 23.8 17.9 15.1 (Growth: y-on-y) (%) Export –7.3 0.2 3.8 5.0 7.1 Import 19.3 25.0 18.2 14.5 13.0 International Prices Crude oil, average ($/bbl) 106.3 103.4 101.2 101.2 105.1 107.6 102.5 Coffee, arabica (cents/kg) 394.6 382.1 352.5 336.7 346.8 329.5 330.2 Gold ($/troy oz.) 1,745 1,747 1,722 1,685 1,672 1,628 1,593 World Growth (quarterly: y-o-y) % China 7.4 7.9 Euro area –0.6 –0.9 US 2.6 1.6 OECD-Total 1.2 0.7 22 Table 3: Ethiopia: Selected Economic and Social Indicators (Annual Frequency)a Avgerage 2003/04– Fiscal year ending July 7 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2010–11 2011–12p 2011/12 Income and Economic Growth GDP growth at factor cost (annual %) 11.7 12.6 11.5 11.8 11.2 10.0 10.6 11.4 8.5 11.0 GDP growth at constant market price 13.7 11.7 10.8 11.4 10.8 8.7 12.7 11.2 8.5 10.7 (annual %) Atlas GNI per capita, US$ 140 160 190 230 280 330 360 370 400 273 GDP per capita, PPP (current interna- 562 636 711 798 883 951 1035 1116 n/a 836 tional $) Private Consumption, nominal (an- 7.2 33.8 25.9 26.9 51.8 35.7 15.3 28.6 41.8 32.1 nual %) Gross Fixed Investment (% of GDP) 29.2 26.1 27.7 24.3 24.6 25.0 27.2 27.9 34.6 27.4 Money and Prices Inflation, consumer prices (annual %, 1.7 13.0 11.6 15.1 55.3 2.7 7.3 38.1 20.8 18.4 end of year) Inflation, consumer prices (annual %, 8.6 6.8 12.3 15.8 25.3 36.4 2.8 18.1 33.4 17.7 period average) Treasury bill rate (91-days maturity, 0.5 0.1 0.0 0.8 0.6 0.9 0.9 1.3 1.9 0.8 annual average) Nominal Exchange Rate (End of 8.6 8.7 8.7 9.0 9.6 11.3 13.5 16.9 17.8 10.8 period) 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Real Exchange Rate Index 94.3 100.0 107.2 111.2 136.7 126.0 107.7 113.2 134.0 112.0 (2005=100) Fiscal Revenue (% of GDP) 16.1 14.6 14.8 12.7 12.0 12.0 14.1 13.5 14.6 13.8 Expenditure (% of GDP) 23.7 23.3 22.3 20.7 18.9 17.2 18.6 18.4 17.6 20.1 Current (% of GDP) 13.8 12.4 11.6 10.0 9.2 8.1 8.4 8.0 7.3 9.9 Capital (% of GDP) 9.5 10.7 10.7 10.7 9.7 9.1 10.3 10.4 10.3 10.1 Overall Fiscal Balance including –3.0 –4.4 –3.9 –3.1 –2.9 –0.9 –1.3 –1.6 –1.2 –2.5 grants (% of GDP) (continued on next page) Table 3: Ethiopia: Selected Economic and Social Indicators (Annual Frequency)a (continued) Avgerage 2003/04– Fiscal year ending July 7 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2010–11 2011–12p 2011/12 Overall Fiscal Balance excluding –7.6 –8.7 –7.4 –8.0 –6.9 –5.2 –4.6 –4.9 –3.1 –6.3 grants (% of GDP) Primary Fiscal Balance including –1.7 –3.4 –3.1 –2.4 –2.4 –0.6 –0.9 –1.3 –0.9 –1.9 grants (% of GDP) Total Public Debt (% of GDP) 105.6 78.9 66.8 43.9 38.5 35.5 39.0 37.4 34.2 53.3 External Accounts Export growth (%, yoy) 24.4 41.1 18.1 18.7 23.1 –1.0 38.3 37.1 14.8 23.0 Import growth (%, yoy) 39.4 40.4 26.4 11.6 32.8 13.4 7.7 –0.9 34.0 19.9 Merchandise exports (in % of GDP) 6.1 7.0 6.7 6.2 5.5 4.6 6.8 8.7 7.4 6.5 of which coffee exports (in % of GDP) 2.3 2.8 2.4 2.2 2.0 1.2 1.8 2.7 2.0 2.2 Merchandise imports (in % of GDP) 26.2 30.0 30.8 26.6 25.8 24.4 28.5 26.3 25.9 27.3 Services, net (in % of GDP) 3.1 2.3 1.0 0.8 0.5 1.3 1.8 2.4 0.4 1.6 Private transfers, net (in % of GDP) 7.8 8.4 8.2 8.8 9.0 8.5 9.3 8.7 7.6 8.6 Current account balance before grant –9.8 –12.6 –15.1 –10.7 –10.6 –10.1 11.0 –6.6 –10.7 –10.8 (in % of GDP) Current account balance after grant –4.1 –6.1 –9.3 –4.5 –5.6 –5.1 –4.4 –0.7 –6.6 –5.0 (in % of GDP) Foreign Direct Investment (in % of 1.5 1.2 2.4 2.5 3.1 2.8 3.3 4.0 2.5 2.6 GDP) External debt, total (in % of GDP) 74.5 49.7 37.8 12.0 10.5 13.7 19.3 24.9 20.8 30.3 External debt, total (% of GDP) 73.3 48.9 37.3 11.8 10.4 13.5 18.1 22.0 19.2 28.3 Multilateral debt (% of total external 63.5 82.7 81.1 51.6 55.7 46.7 48.6 46.0 45.1 57.9 debt) Debt service ratio (% of goods and 11.1 9.1 8.4 3.6 2.9 2.4 3.2 4.3 6.8 5.7 NFS) (continued on next page) Recent Economic Developments 23 24 Table 3: Ethiopia: Selected Economic and Social Indicators (Annual Frequency)a (continued) Avgerage 2003/04– Fiscal year ending July 7 2003–04 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 2010–11 2011–12p 2011/12 Population, Employment and Poverty Population, total (millions) 72.5 74.3 76.0 77.7 79.4 81.2 83.0 84.7 86.4 79.5 Unemployment Rate (urban except 5.4 17.0 20.4 18.9 18.0 17.5 16.2 2004/05) Poverty headcount ratio at national 38.7 29.6 34.2 poverty line (% of population) Poverty headcount ratio at $1.25 a 39.0 30.7 34.8 day (PPP) (% of population) Poverty headcount ratio at $2 a day 77.6 66.0 71.8 (PPP) (% of population) Inequality – Gini Coefficient 0.300 0.298 0.30 Population Growh (annual %) 2.5 2.4 2.3 2.3 2.2 2.2 2.2 2.2 2.0 2.3 Life Expectancy 54.4 55.2 56.0 56.7 57.5 58.1 58.7 59.2 n/a 55.5 Other: GDP (current LCU, billions) 85 105 130 169 244 330 377 506 737 298 Nominal GDP (current US$, billions) 10 12 15 19 26 32 29 31 43 24 b Doing Business (rank) 101 97 102 116 107 104 111 105 c Human Development index ranking 170 170 169 169 171 157 174 173 169 a 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status This report uses the national accounts definitions of the Ministry of Finance and Public Credit. Investment is defined as the sum of fixed capital formation and change in inventories. Public investment is the sum of investment by federal and regional governments, as well as public enterprises. Private investment includes investment by households, private enterprises, and non-profit institutions serving households. Total investment is the sum of private and public investment. b This indicator is ranked out of 175 countries in 2007, 178 in 2008, 181 in 2009 and 183 in 2010 and 2011. c The HDI ranking in 2001 is in relation to 175 countries; from 2005 to 2008, to 177; in 2009, to 181; in 2010, to 169 countries; in 2011, to 187 and in 2012 186 countries. Savings 13 2 Introduction X is exports, M is imports; Yf  is net factor income from abroad, and, Tf is net foreign private and offi- Ethiopia’s domestic savings rate is low compared cial transfers. to the fast pace of capital accumulation observed The relationship between Gross Domestic between 2004 and 2011. Ethiopia has been expe- Savings (GDS) and Gross National Savings riencing single-digit domestic saving rates while (GNS) can then be written as: economic growth was in double digits, supported by investment rates beyond 25 percent of GDP. GDS = GDP – (C + G) Consequently, Ethiopia is confronted with a persistent GNS = GNDI – (C + G) = GDS + Yf + Tf and wide domestic saving and investment gap, which has been financed by external sources. As discussed Hence, Gross Domestic Savings (GDS) and Gross in the previous chapter, the Government of Ethiopia National Savings (GNS) differ substantially if a coun- has very ambitious public investment plans. Given the try has large current transfers in the form of public current levels of domestic and external savings, how- (e.g., official aid) and private transfers (e.g., remit- ever, it may be difficult to finance this investment plan. tances) from abroad, such as in the case of Ethiopia. The risks associated with external sources The effort to mobilize domestic savings is a sub- of financing offers an additional motivation for set of a broader set of efforts to mobilize domestic increasing domestic savings. Those risks include that resources, which include: taxation, increasing access external transfers—private and official—may decline to financial services, and deepening financial sec- in the future. Moreover, resorting to external bor- tors. Countries pursue the latter at various stages rowing, in order to finance high levels of investment, of their development with the aim of supporting will result in a gradual build-up of debt which could long-term growth (Mavrotas: 2009). Domestic savings eventually weaken Ethiopia’s current record of low mobilization is a complement to domestic resources levels of debt distress. mobilization, but it is not a substitute for it. Alternative The two most common definitions for savings sources of financing and investment, including taxa- refer to Gross Domestic Savings (GDS) and Gross tion, FDI, aid, and remittances remain equally impor- National Savings (GNS). These concepts are derived tant and should be pursued as simultaneous options. from the national accounts. Gross Domestic Product This chapter examines the trends in the Gross (GDP) and Gross National Disposable Income Domestic Savings (GDS) rate over the past three (GNDI) can be expressed as: decades. It sheds light on the causes of the low rate observed and discusses possible measures to increase it. GDP = C + I + G (X – M) Cross-country comparisons are made to distill useful GNDI = GDP + Yf + Tf 13 In preparing this report, MOFED requested the Bank to include suc- cessful examples from other countries that could be used in Ethiopia. Where C is private consumption, I is investment (gross Hence, this chapter tries, wherever possible, to make reference to other capital formation), G is government consumption, country experiences. 26 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status lessons for Ethiopia. The chapter also discusses recent Rising private savings have driven increases government measures to boost the saving rate. in GDS in countries such as China and Vietnam. The chapter is structured as follows: Section Figure 10.4 compares Ethiopia’s savings rate with B provides an overview of trends in saving and sub-Saharan Africa (SSA), China, and Vietnam over investment in Ethiopia and the World, including the past three decade. In SSA, the ratio of public sav- a decomposition of saving into its private and pub- ing to GDP is stable across decades and constitutes lic components. Section C establishes the deter- major portion of gross domestic savings. In China and minants of savings in theory and practice. Section Vietnam, on other hand, the share of public saving D describes current government policies in the area in gross domestic savings is small and increases in gross of saving mobilization. Finally, Section E offers pol- domestic savings are mostly attributable to increases icy recommendations. in private saving. When remittances are excluded, Ethiopian The Savings Rate in Ethiopia and the households dissave15 in the aggregate. Table 4 gives World an overview of the contribution to Gross Domestic Savings by economic units, based on work car- The remarkable recent growth performance was ried out by the Ethiopian Development Research supported by robust investment—but not matched Institute (EDRI, 2010). On average, rural and urban by similarly high savings rates. As illustrated households had a negative domestic savings rate of in Figure 10.1, the gap between Gross Domestic Savings –4.5 percentage points of GDP in 2006–09. The lead- (GDS) and the investment rate widened over the past ing domestic contributors to savings are State Owned three decades. Investment rose from 15.7 percent Enterprises, central government and private enter- of GDP in the 1980s to 23 percent in the 2000s, while prises, with a combined effect of ten percent of GDP Gross Domestic Savings declined from 10.5 percent over this period. The government saving increased to 6.1 percent of GDP over the same period. Recent over time through a combination of factors, including revisions in the national accounts of Ethiopia indicate improved tax revenue collection. a growing savings rate again over the past years.14 Public and private enterprises account for Investment financing has shifted gradually a third of gross national savings and general gov- away from gross domestic savings towards net ernment for a quarter. The remainder is explained income transfers, foreign direct investment, and by the remittances received by households (37 per- external borrowing. In the 1980s, gross domestic cent) and private enterprises (see Figure 10.5). savings mostly financed investment. In the 2000s, This is not an unusual pattern, and somewhat an expansion of investment was made possible similar to what was observed in East Asia. In China, by an increase in net income transfers and a larger enterprise savings (including SOEs) rose sharply current account deficit (financed, in turn, by FDI and from 35 to 50 percent of Gross National Savings external borrowing). Figure 10.2 refers. in 1992–2008 (Chamon, et al., 2011). Similar fig- A large drop has driven the decline in Gross ures for both government savings and corporate Domestic Savings (GDS) rates in Ethiopia in both private and public savings. The private component of gross domestic savings declined by more than 14 The Government of Ethiopia has recently carried out a revision of its national account methodology, including the introduction of a new base nine percentage points in the 1980s to less than five year. This resulted in slightly higher saving and investment rates for the percentage points in the 2000s. Public savings fell two years. Using the new savings and investment data, and applying it pro-rata over the past decade leads to only minor discrepancies between from 4.2 to 1.4 percent of GDP over the same period the new and the old series. (Figure 10.3). 15 If consumption is greater than income, dissaving takes place. Savings 27 Figure 10: Savings and Investment in Ethiopia and the World (Percent of GDP) 1. Savings and Investment 2. Investment and Financing 25% 25% 20% 20% 15% 15% 10% 10% 5% 5% 0% 0% 80s 90s 00s 80s 90s 00s GDS Rate GNS Rate Investment Rate GDS NIT CAD 3. Public and Private Savings in Ethiopia 4. Public and Private Savings in the World 14% 50% 12% 40% 10% 30% 20% 8% 10% 6% 0% 4% –10% late 80s 90s 00s 80s 90s 00s 80s 90s 00s 80s 90s 00s 2% 0% late 80s 90s 00s Ethiopia China Vietnam SSA Public Private Public Private 5. GNS by Economic Units (2006–09 Average) Others 3% Private enterprises Rural HH 16% 21% Urban HH 16% General Gov't 27% SOEs 17% Source: World Bank staff calculations, based on World Development Indicators (WDI) and EDRI (2010) in the case of 9.5. Note: NIT = Net Income Transfers (i.e., a summation of net income from abroad and current transfers). CAD = Current Account Deficit. Note: Late 80s refers to the pe- riod 1986–1989. 28 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Table 4: National Savings vs. Savings from Purely Domestic Sources (% of GDP) Saving from Domestic Source Only Gross National Saving 2006 2007 2008 2009 Average 2006 2007 2008 2009 Average Rural HH 0.2 –0.2 –0.6 –1.3 –0.5 5.2 4.9 4.4 3.6 4.5 Urban HH –4.3 –3.6 –3.9 –4.0 –4.0 3.5 3.9 3.3 3.1 3.5 SOEs 2.9 3.9 3.9 2.9 3.4 3.3 4.2 4.2 3.0 3.7 General Gov’t 3.2 2.8 2.8 3.8 3.2 4.0 5.2 5.1 6.7 5.3 Private enterprises 3.3 3.6 3.8 3.0 3.4 3.1 3.5 3.7 2.9 3.3 Quasi Govt 0.0 0.4 0.3 0.3 0.3 0.0 0.4 0.3 0.3 0.3 Oils stab. Fund –1.4 –0.5 –2.0 –0.2 –1.0 –1.4 –0.5 –2.0 –0.2 –1.0 Source: Staff calculations based on EDRI (2010). profits were observable in South Korea (henceforth initial rate of 35.4 percent in the 1980s and continu- referred to as Korea between 2000 and 2008. In the ing rise to 41.2 percent in the 1990s and 45.8 per- Philippines (2000–08), a declining household savings cent in the 2000s (Figure 11.2). Vietnam and Korea rate was fully compensated by increases in corporate followed similar paths: for Vietnam, a substantial and government contributions, in order to keep the increase in saving rate from an initial rate of 4.4 per- gross national savings constant at just below 20 per- cent in the 1980s to 28.3 percent in the 2000s; and for cent of GDP (Prasad, 2011). Korea, from 8.6 percent in the 1960s to 30.9 percent The trends in savings in Ethiopia are distinct in the 1980s (Figure 11.3). from most other countries. Middle and low-income Kenya and Ethiopia followed a similar declin- countries raised their saving rate in the 1990s and ing trend in GDS rates over the past three decades. the 2000s (Figure 11.1). Most middle-income Figure 11.4 shows trends of saving rates in neighboring countries achieved positive per capita GDP growth East African countries. Kenya follows a similar trend in the 1990s, whereas all low-income countries, to Ethiopia, lowering its saving rate from 18.3 per- including Ethiopia, experienced more rapid and cent in the 1980s to 8.7 percent in the 2000s. Unlike substantial economic growth in the 2000s. However, Ethiopia and Kenya, Tanzania shows a reversion of the Ethiopia is one of few countries in low-income SSA saving rate from 7.2 percent in the 1990s to 14.6 per- that reduced its gross domestic savings rate during its cent in the 2000s and Uganda shows an increase in the rapid growth period of the last decade. saving rate from 2.3 percent in the 1980s to 9.5 per- Savings rates in East Asia increased to much cent in the 2000s. higher levels over the past three decades whereas in East Africa they remained stagnant. Saving rates Determinants of Savings Rates: Theory for NIEs and ASEAN increased substantially both and Evidence16 from 15 percent of GDP in the 1960s to 35 percent and 30 percent in the 1980s, respectively, following There is a common division of determinants of sav- similar paths of their GDP. China and Vietnam started ings into policy and non-policy determinants. from a level of initial per capita real GDP in the 1980s, which was similar to Ethiopia’s, while Korea had 16 The result of an empirical analysis that supports the theoretical con- siderations is shown in Annex Savings Rate 1. There, the analysis follows a similar level of per capita income in the 1960s. Loayza, et al. (2000) by employing Generalized Method of Moments China’s saving rate is remarkable in terms of its high (GMM) estimators applied to dynamic models using panel data. These Savings 29 Figure 11: GDS Rates in Ethiopia and Relevant Peer Groups (Percent of GDP) 1. Selected Income Groups in SSA 2. Emerging Markets and East Africa 12% 40% 35% 10% 30% 8% 25% 6% 20% 15% 4% 10% 2% 5% 0% 0% Ethiopia Middle Income Low Income Fragile Ethiopia NIEs ASEAN East (80s, 90s, (60s, 70s, (60s, 70s, (80s, 90s, 80s 90s 00s 00s) 80s) 80s) 00s) 3. China, Vietnam, and Korea 4. Selected East African Countries 50% 20% 45% 40% 15% 35% 30% 25% 10% 20% 15% 5% 10% 5% 0% 0% Ethiopia NIEs ASEAN East Ethiopia Kenya Tanzania Uganda (80s, 90s, (80s, 90s, (80s, 90s, (60s, 70s, 00s) 00s) 00s) 80s) 80s 90s 00s Source: World Bank staff calculations, based on World Development Indicators (WDI). Note: NIEs are Hong Kong, Korea, Singapore and Taiwan. ASEAN includes Indonesia, Malaysia and Thailand. East Africa includes Kenya, Tanzania and Uganda. Non-policy determinants are: persistence of saving The saving decision can be regarded as intertem- rate, the level and growth of GDP, and demographic poral allocation of consumption to maximize util- changes. Policy determinants, in turn, include: finan- ity. The theories of saving and consumption are based cial liberalization and development, government on the life-cycle approach (Modigliani, 1970) and the saving, and macroeconomic stability. This section permanent income hypothesis (Friedman, 1957). The provides a detailed analysis from both macro and life-cycle model argues that households seek to smooth micro perspectives using panel and country-specific out consumption over time because of precautionary econometric methods. The methodology used relies motives. The permanent income hypothesis predicts on the insights from Hevia, Ikeda and Loayza (2010). that households smooth out the consumption path based on anticipation on future income. Both the precautionary motive for saving and the anticipation estimators allow the control of unobserved country-specific effects and potential endogeneity between savings rates and other macroeco- of future income determine saving behavior through nomic variables. a number of channels, as discussed below. 30 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status This section uses scatter plots to contextualize Income, Income Growth, and Terms of Trade the Ethiopian experience with other country expe- riences.17 To capture growth-saving episodes in each Saving rates increase as countries grow richer.19 country, periods analyzed for each country correspond The impact of higher income on savings also depends to those when the country experienced rapid economic on the level of income (Loayza et al., 2000). Changes growth. For Ethiopia, SSA, China, and Vietnam, the in real per capita income in developing countries have periods analyzed are from 1981 to 2011 whereas for NIEs larger impact on saving than in advanced countries. and ASEAN the periods are from 1961 to 1989 during This implies that policy actions targeted to increase which the GDS rate rose sharply, exceeding or reaching income through, for instance, improving total factor close to 40 percent of GDP. In subsequent graphs, the productivity, are indirect but effective ways to raise following labeling is used: Ethiopia (red diamonds); the saving rate. Figure 12.2 plots the savings rate and low-income SSA (green triangles); and Asia (blue log of real per capita GDP across countries, showing circles). Regression lines for low-income SSA (light strong positive correlation for both Asia and SSA. blue) and for Asia (red). The relationship between saving and income It takes time to change savings behavior. Lagged growth is unclear in the empirical literature. Some values of saving rate remain strongly significant even show that there is causality between saving and growth after controlling for a large set of policy and non-policy (i.e., that higher savings induce more growth). Others determinants. As a result, a change in any determi- try to examine the question of how income growth nants of savings requires a number of years before affects saving. Aghion, et al. (2009) find that a higher it has an impact. The long-run response is approxi- saving rate predicts higher income growth in devel- mately twice as large as the short-run response (Loayza oping countries. Loayza. et al. (2000) find that one et al., 2000). percentage point increase in income growth increases Ethiopia’s saving rate of 5.5 percent of GDP Private Sector Savings (PRS) rate by roughly the same in the 2000s is approximately five percentage points amount. Rodrik (2000) concludes that a permanent lower than the one predicted for low-income SSA increase in the saving rate induces a temporal increase countries. In Figure 12.1, the horizontal axis displays in output growth, whereas a permanent increase the average saving rate for a given decade T. The ver- in income growth is followed by a permanent increase tical axis displays the average saving rate in the next in saving rate. decade (T+1). In order to compare Ethiopia with the Economic theory does not offer conclusive relevant countries at the appropriate stage of devel- results either. The life-cycle model predicts that aggre- opment, the decade T differs by country of region. gate savings will increase in response to an increase In particular, T covers the 1990s for Ethiopia, SSA, in income growth, through an increase in the saving China, and Vietnam. For NIEs and ASEAN countries, of active workers relative to the dissaving of retirees. it covers the 1960s. Figure 11.1 is used to illustrate The consumption habit theory reinforces this view, whether the savings rate rises or declines over time indicating that household consumption adjusts slowly in different countries or comparator groups. If a coun- try is located above the 45-degree line, it implies that 17 Asian countries considered are those in NIEs, ASEAN, China, its savings rate increased, and vice versa. All Asian and Vietnam. countries and most low-income SSA countries lie 18 For instance, saving rates for Korea and Indonesia in the 1960s are 8.6 percent and 6.2 percent, respectively, which are lower than above the 45-degree line, indicating that their sav- Ethiopia’s saving rate in the 1990s (9.6 percent). In the 1970s, saving ings rate increased over time.18 Ethiopia lies below rates for Korea and Indonesia were 22.1 and 25 percent, respectively. 19 Notable examples in the literature are: Edwards (1996), Dayal-Gulati the low-income SSA regression line, indicating that and Thimann (1997), Loayza, et al. (2000), Metin-Ozcan and Ozcan savings are on the decline. (2004), Chaturvedi, et al (2009). Savings 31 Figure 12: Determinants of Savings: Persistence, GDP per Capita, Income, and Terms of Trade 1. Persistence of Saving Rate 2. Saving and Per Capita GDP 50% 50% Saving Rate in Dedade T+1 40% 40% Saving Rate (% of GDP) 30% 30% (% of GDP) 20% 20% 10% 0% 10% –10% 0% –20% –10% –10% 10% 30% 50% 4 6 8 10 Saving Rate in Decade T (% of GDP) Log of Real Per Capita GDP Ethiopia Ethiopia Note: T = 1990s for Ethiopia, SSA, China and Vietnam and T = 1960s for other countries. 3. Saving and Income Growth 4. Saving and Terms of Trade 50% 50% Saving Rate (% of GDP) Saving Rate (% of GDP) 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% –10% –10% –5% 0% 5% 10% 15% 0 0.5 1 1.5 2 Real Per Capita GDP Growth Rate Terms of Trade (2000=1) Ethiopia Ethiopia Source: Staff calculations based on World Development Indicators (WDI). Note: The horizontal axis (period T) contains the average saving rate over the period 1990–1999 for Ethiopia, SSA, China, and Vietnam and the period 1960–1969 for NIEs and ASEAN countries. The vertical axis (period T+1) is the average saving rate in the next decade (i.e., the pe- riod 2000–2009 for Ethiopia, SSA, China, and Vietnam and the period 1970–1979 for NIEs and ASEAN countries). to increase in income.20 Therefore, increase in income growth because returns to capital diminish and saving leads to increase in saving. Contrarily, the permanent itself does not affect total factor productivity (e.g., the income hypothesis predicts that households dissave long-term determinant of economic growth). against anticipated future higher income. Ethiopia is one of few countries in the world The argument on the causality of saving and that achieved higher income while lowering its sav- growth has profound policy implications. If saving ing rate. Figure 12.3 plots saving and income growth, causes growth, saving-enhancing policies are likely showing rapidly growing Asia with a positive correla- to induce growth whereas if the direction is opposite, tion, while low-income SSA exhibits no correlation. such policies may fail to promote permanent growth. In Asia, a one percentage point increase in per capita These policies may promote growth in the short run by fueling investment through saving. In the long term, however, they may fail to realize permanent 20 See Carroll and Weil (1994) and Carroll, et al. (2000). 32 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status GDP is associated with a 2.5 percentage point increase point increase in the youth dependency ratio is asso- in the saving rate. Economic growth positively affects ciated with five percentage point decline in domestic saving rates in SSA. In both regions, the income saving rate, on average. Figure 13.1 shows a strong level positively affects the savings. Ethiopia and East negative correlation between the savings rate and the African countries are exceptions, possibly because the youth dependency ratio for Asia and no correlation for income level is so low that it is hard to increase sav- SSA. In all decades, Ethiopia’s saving rates lie around ings. To illustrate, the income levels in Ethiopia and the low-income SSA regression line with youth depen- East African countries in the 2000s are lower than the dency ratio ranging from 85 percent in the 1980s level in ASEAN in 1960s. to 86 percent in the 2000s. Therefore, an increase Most empirical studies find that improvements in youth dependency ratio by one percentage point in the terms of trade have a positive impact on the in Ethiopia is negligible, compared to the average saving rate. Improvement in the terms of trade decline in youth dependency ratio by 27 percentage increases real income through increase in purchas- points in Asia. Ethiopia has not experienced a demo- ing power. Therefore, improvement in the terms graphic transition, and is unlike emerging Asia during of trade is predicted to have positive impact on saving the period of rapid economic growth. As Ethiopia goes as income does. Figure 12.4, however, does not show through the demographic transition in the coming such a positive correlation between the terms of trade decades, savings are expected to increase. and saving rate for any of the regions analyzed. This There is a positive correlation between the may reflect the existence of omitted variables that saving rate and the old dependency ratio for Asia, correlate with both. The multivariate model pre- and negative correlation for low-income SSA and sented in Annex Savings Rate 1, suggests the strong Ethiopia. This finding contradicts the life-cycle theory importance of terms of trade: a one percentage point (Figure 13.2). The strong positive correlation for Asia improvement in the terms of trade raising domestic may be explained by the existence of a third variable: saving rate by 0.08 percentage point. the level of income. This is because life expectancy increases with the level of income, and income is posi- Demographics, Urbanization, and Public tively correlated with the saving rate. The old depen- Savings Rates dency ratio is low for Ethiopia and Asia. Urbanization is also considered as a proxy The life-cycle theory predicts that saving follows of uncertainty and demographics. The reason for a hump-shaped pattern.. In other words, the young this is that rural income is substantially more volatile workers dissave against their future income, the than urban income, leading rural households to save middle-aged workers save for their retirement and a large portion of their income in case of poor har- the elderly dissave upon their retirement. Therefore, vest for precautionary motive. Therefore, in the more demographic changes have a significant impact urbanized society, households tend to save less because on household saving patterns. Both the microeco- their income is more foreseeable. Figure 13.3 shows nomic and macroeconomic literature confirms that a positive correlation between the saving rate and the an increase in youth and old dependency ratios tend urbanization ratio. As in the case for the old depen- to lower the gross domestic and private saving rates. dency ratio, the level of income may be a possible Muhleisen (1996) claims that the age dependency ratio, sum of aged and young, dominates the behavior of the private sector saving rate.21 21 Empirical estimates suggest that a one percentage point increase in aged dependency ratio lowers the gross domestic savings rate by 0.5 to 1.6 per- Demographic effects have yet to unfold and cent whereas an increase in the youth dependency ratio reduces savings affect savings in Ethiopia. In Asia, a ten percentage by 0 to 0.7 percent. Savings 33 Figure 13: Determinants of Savings: Demographics, Urbanization, and Public Savings 1. Saving and Youth Dependency Ratio 2. Saving and Old Dependency Ratio 50% 50% 40% 40% Saving Rate (% of GDP) Saving Rate (% of GDP) 30% 30% 20% 20% 10% 10% 0% 0% –10% –10% –30% 50% 70% 90% 4% 6% 8% 10% 12% Youth Dependency Ratio Old Dependency Ratio Ethiopia Ethiopia 3. Saving and Urbanization 4. Gross Domestic and Public Saving Rates Gross Domestic Saving Rate (% of GDP) 50% 50% 40% 40% Saving Rate (% of GDP) 30% 30% 20% 20% 10% 10% 0% 0% –10% –10% 0% 20% 40% 60% 80% 100% –10% –5% 0% 5% 10% 15% Urbanization Rate Public Saving Rate (% of GDP) Ethiopia Ethiopia Source: World Bank staff calculations, based on World Development Indicators (WDI). Note: The horizontal axis (period T) contains the average saving rate over the period 1990–1999 for Ethiopia, SSA, China, and Vietnam and the period 1960–1969 for NIEs and ASEAN countries. The vertical axis (period T+1) is the average saving rate in the next decade (i.e., the pe- riod 2000–2009 for Ethiopia, SSA, China, and Vietnam and the period 1970–1979 for NIEs and ASEAN countries). explanation. Income level is strongly correlated with Ricardian equivalence holds only partially.23 However, the degree of urbanization and income level again estimates on the degree of increase in domestic sav- correlates with saving rate. ing vary across studies. Loayza, et al. (2000) report Public saving shows a negative correlation that a one percentage point increase in public saving for low-income SSA and a positive correlation leads to a decline of 0.3 percentage points in the pri- for Ethiopia (Figure 13.4). The permanent income vate savings rate in the short run and 0.6 percentage hypothesis predicts that changes in the timing of taxa- points in the long run. Alternatively, one percentage tion do not affect household consumption behavior. point increase in public saving leads to 0.7 percentage Given a sequence of public expenditures, it does not points increase in gross national savings in the short matter when the government raises taxes to finance run and 0.4 percentage points in the long run. This these expenditures. The theory predicts that changes in public savings fully crowds out private sector sav- 22 Note that if the change in public saving is due to changes in public expenditures, the change has impact on domestic saving. ings, leaving domestic saving unchanged (Ricardian 23 Refer Edwards (1996), Dalai-Gulati and Thimann (1997), Loayza, equivalence).22 Most empirical studies show that the et al. (2000) and IMF (2005). 34 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status suggests that public saving can be an effective policy In Ethiopia, there is no evidence for overcon- instrument to raise gross national savings. sumption by high-income group during periods of inflationary pressure. One hypothesis for this Macroeconomic Stability and Financial analysis is that the low savings rate may be the result Development of higher income groups not increasing their sav- ing behaviors the way one would typically expect. Real interest rates—or alternatively the inflation Part of the reason, as the original hypothesis states, rate under a given nominal interest rate—affects may be that there are insufficient alternative mon- saving by affecting the rate of intertemporal sub- etary saving instruments available that would allow stitution of consumption. Balassa (1986) claims them to save. As a consequence the analysis tried that a stable and substantial level of domestic saving to find evidence that higher income groups tend to can be achieved by keeping real interest rate stable “over-consume” in Ethiopia as compared to other and sufficiently high. Alternatively, Balassa proposes East African countries. However, the analysis carried that low and stable inflation, supported by a small out based on household data from Ethiopia, Kenya, budget deficit, encourages saving. In theory, how- Uganda, and Tanzania did not provide evidence for ever, the net impact of the real interest rate on saving such overconsumption, which may explain the low is ambiguous due to the income and substitution savings rate in Ethiopia. effects moving in opposite directions. In a situation The theory of precautionary saving predicts of rising real interest rates, the income effect could that an increase in economic uncertainty, often lead to a decline in saving given the increase in interest measured by inflation rate, leads to higher savings income (the income effect). Contrarily, the substitu- rates in the absence of complete financial markets. tion effect could lead to increases in saving because The logic behind this is that risk-averse households, saving is more attractive by increasing future con- which cannot perfectly insure risks, increase buffer sav- sumption.24 Figure 14.1 shows a negative correlation ings to avoid large adjustment in consumption in the for Asia, where decade-average real interest rate never face of uncertain events. However, the inflation rate falls below zero. This suggests that the income effect also reflects macroeconomic stability, which indicates dominates the substitution effect (i.e., that higher real that high inflation rate reduces saving. This is observed interest rates reduce savings). in Ethiopia. In fact, the empirical literature disagrees Negative real interest rates lead the saving rate on the effect and extent of impact of inflation on the to deviate from the equilibrium saving rate. This saving rate, reflecting the contrasting views above.25 could partly explain why the gross domestic savings rate declined in Ethiopia at a time when the country 24 McKinnon (1973) and Shaw (1973) discuss that the real interest rates experienced persistent high inflation—and negative capture the extent of financial repression under which real interest rates are artificially kept low by government. Financial liberalization raises real interest rates—over many years in the 2000s real interest rate and provides opportunities for saving and borrowing. (Figures 14.2, 14.3, and 14.4). Empirically, for coun- McKinnon and Shaw claim that higher real interest rates raise saving for low-income financially repressed countries due to improved sav- tries with low real interest rates, the substitution effect ing rate and opportunity. Empirical studies disagree about the impact dominates the income effect (i.e., higher real interest of real interest rates. Loayza, et al. (2000) finds negative impact whereas Edwards (1996), Masson et al. (1998) and IMF (2005) find no impact rates increase savings). For countries with low real and Masson, et al. (1995) shows the impact is positive. interest rates, the opposite is the case. Therefore, real 25 Loayza, et al. (2000) find a positive and significant effect of inflation on savings. Chaturvedi, et al (2009) also reports a positive and significant interest rates and saving rate are related in a hump- effect of inflation on savings in South-East Asian and South Asian countries, shaped (inverted-U shaped) form. This implies that as shown by Ferrucci and Miralles (2007) for a large group of emerging market economies. Other studies, however, find that the inflation rate is not higher real interest rates can be conducive to savings, significantly associated with private savings (e.g., Edwards, 1996) or even as long as the real interest rate is not too high. negatively associated with it (e.g., Dayal-Gulati and Thimann, 1997). Savings 35 Figure 14: Determinants of Savings: Real Interest and Inflation Rate 1. Saving and Real Interest Rate 2. Real Interest Rates in Emerging Countries 50% 12% Saving Rate (% of GDP) 40% 10% 30% 8% 20% 6% 10% 4% 0% 2% –10% –10% 0% 10% 20% 30% 0% Real Interest Rate –2% Ethiopia NIEs ASEAN East Ethiopia (90s, 00s) (70s, 80s) (70s, 80s) (90s, 00s) 3. Saving and Real Interest Rates in Ethiopia 4. Saving and Inflation Rates in Ethiopia 20% 20% 18% 18% 16% 16% Saving Rate (% of GDP) Saving Rate (% of GDP) 14% 14% 12% 12% 10% 10% 8% 8% 6% 6% 4% 4% 2% 2% 0% 0% –20% –10% 0% 10% 20% –10% 20% 30% 50% Real Interest Rate Inflation Rate Source: World Bank staff calculations, based on World Development Indicators (WDI). Figure 15.1 shows that inflation has a marginally development on savings rate is hump-shaped and finds negative correlation with saving for Asia, whereas such an empirical relationship. In order to capture the it is positively correlated for SSA. nonlinear relationship between savings rate and finan- The impact of financial development on saving cial development, the empirical analysis below incor- is complex. It can be divided into direct, short-run porates the square of financial development indicators. impact and indirect, long-run impact. The direct, Figure 15.2 shows no correlation between the savings short-run impact, which is predicted as negative, rate and private credit flow for Asia, and a positive cor- is further decomposed into price and quantity chan- relation for low-income SSA. Private credit flows are nels. Through the price channel, higher interest strongly associated with income level, which negates rates lead to income and substitution effects, which the predicted negative correlation between saving and affect savings in the opposite directions. Contrarily, private credit flow. through the quantity channel, Jappelli and Pagano The indirect impact of financial develop- (1994) find that elimination of credit ceiling at both ment—which is predicted as positive—is real- corporate and consumer levels has a negative impact ized through improved financial intermediation. on saving. Wang, Xu, and Xu (2011) provide theoreti- Countries with efficient financial intermediation cal analysis that shows the impact of financial sector can enhance domestic resource mobilization, leading 36 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status to income growth. As described above, income growth affects savings. (4) Macroeconomic stability, measured is associated with higher saving. Therefore, in the end, by developments in real interest rates (or alternatively financial development and saving rate is expected inflation rates) affects savings through its impact to show a positive correlation.26 Empirical studies on the rate of intertemporal substitution of consump- analyzing the long-term impact of financial develop- tion. Negative real interest rates observed in Ethiopia ment use the stock of private credit or broad money over much of the 2000s can partly explain the decline (M2) as proxies for financial depth. Figure 15.3 shows in savings during that time. a positive correlation between the savings rate and M2. Ethiopia is demonetizing, while its peers are Recent Initiatives to Increase the doing the opposite, and this negatively affects sav- Savings Rate in Ethiopia ings. In phases with high inflation rates, as the nomi- nal GDP becomes large, the ratio M2/nominal GDP The Government of Ethiopia has taken various declines and a country demonetizes.27 Additionally, measures to increase savings. These measures range high inflation increases the velocity of money and from schemes to increase household savings through hence M2/GDP declines even further. The decade- improving financial sector accessibility, to attracting averages in Figure 15.4 initially lead to the conclusion funding from the large Ethiopian diaspora, to main- that the degree of monetization in Ethiopia increased taining macroeconomic stability. Concretely, five main over the three decades under observation. However, measures are in place:28 a closer look at the annual figures in Figure 15.5 reveals that it has been continuously falling from 44.3 percent  Pension scheme: The government has established in 2003 to 22.9 percent in 2009 in Ethiopia. Such a pension scheme for employees of private com- a demonetization is observed only for Ethiopia. The panies and NGOs. decline in the degree of monetization seems a key  Expanding bank branches : In an attempt factor in explaining the low saving rate in Ethiopia. to increase access to financial ser vices In sum, this section identifies four key deter- from 20 to 67 percent at the end of the GTP minates of savings: real per capita GDP and its period, the commercial banks continue their growth rate, the old dependency ratio, the degree expansion of the branch network. The number of monetization and macroeconomic stability. of branches for all banks increased from 681 at the Their impact on the actual developments of the sav- end of FY10 to 1,286 at the end FY12 (an increase ings rate in Ethiopia varies: (1) The level and growth of 89 percent). rate of income (real GDP) is a key determinant  Micro-financing: Microfinance institutions are of savings worldwide. However, Ethiopia and East being promoted and their clients and deposits are African countries are exceptions, possibly because the increasing over time. More than 2.6 million clients income level is so low that it is hard to increase savings. (2) Demographic changes have a significant impact on household saving patterns and this can be observed 26 See Levine, et al. (2005) and Levine (2000). in Asia. In Ethiopia, however, demographic effects 27 Other monetary variables (measured in percent to GDP), such as money in circulation and monetary base, follow a similar declining have yet to unfold. (3) The effects of the degree trend since 2005/06. of monetization and financial development on sav- 28 There is also a planned housing saving scheme, which has yet to be im- plemented. The scheme aims to encourage low- and middle-income ings are complex; however, in the long run, financial earners to deposit money from their income for a given period with the depth, measured through the ratio of M2 to GDP, aim to reach sufficient savings to allow for buying a home. By the time the savings will have reached 40 percent of the housing cost, the savers will are expected to show a positive correlation to savings. be entitled to a house where the remaining cost of the house be covered However, Ethiopia is demonetizing and this negatively through long term, possibly subsidized loans. Savings 37 Figure 15: Determinants of Savings: Inflation Rates, Financial Sector Development 1. Saving and Inflation Rates 2. Saving and Private Capital Flow 50% 50% Saving Rate (% of GDP) Saving Rate (% of GDP) 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% –10% –10% –10% 10% 30% –5% 0% 5% 10% 15% Inflation Rate Private Capital Flows (% of GDP) Ethiopia Ethiopia 3. Saving and Monetization 4. M2 in Emerging Countries (% of GDP) 50% 80% 70% Saving Rate (% of GDP) 40% Saving Rate (% of GDP) 30% 60% 50% 20% 40% 10% 30% 0% 20% –10% 10% –10% 40% 90% 140% 0% M2/GDP Ethiopia NIEs ASEAN East (80s, 90s, (60s, 70s, (60s, 70s, (80s, 90s, Ethiopia 00s) 80s) 80s) 00s) 5. M2 in Ethiopia and East Africa (% of GDP) 45% 40% 35% 30% 25% 20% 15% 1988 1990 1992 1984 1986 1988 2000 2002 2004 2006 2008 Ethiopia East Africa Source: World Bank staff calculations, based on World Development Indicators (WDI). 38 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status of MFIs contributed 4.5 percent to domestic sav- workers, diplomatic, and international organization ings in the country in 2012. staff—to register to the private pension scheme. The  Diaspora bonds: The purpose of these bonds total number of employees covered with the new has been to raise savings domestically and abroad scheme reached about 360,000 at the end-Decem- to finance investments of EEPCO, in general, and ber 2012. The employee contribution rates started the construction of the Grand Renaissance Dam, at five percent of gross salaries in 2012, and will reach in particular. seven percent in FY14. The employers’ contribution  Housing saving scheme: A new scheme is being should reach up to 11 percent of gross salaries by FY14, implemented in the city of Addis Ababa since similar to the scheme for civil servants. In the first year June 2013. of its operation, the private employees’ pension scheme managed to collect about 0.5 billion Birr (0.3 per- The Pension Scheme cent of gross national savings); ongoing collection in FY13 points to an increased collection level in FY13. The share of the pension fund to domestic savings is small (net of pension payments), but the contri- Expanding Bank Branches bution to gross national saving is increasing. The current pension system in Ethiopia was established Ethiopia’s financial intermediation is small relative in 1963 with the aim of providing pension social secu- to other African countries. In 2011, private sector rity to public servants and state owned enterprises. Yet, credit in Ethiopia was equivalent to about 14 percent even today, the overwhelming majority of the Ethiopian of GDP compared to the regional average of 23 percent population has no coverage of the social security system of GDP. Likewise, the banking value added in share of any kind. Nevertheless, for those who have access, of GDP is only 1.7 percent. It is therefore far lower their contributions are on the rise. The targets set forth than, for instance, in South Africa (18.7 percent) and by government are to increase the contributions of civil Kenya (7.2 percent). Ethiopia’s economy is character- servants from five to seven percent of employees’ gross ized as a large agrarian society with poor infrastructure salaries over three years (FY12–14) and the govern- and relatively weak financial system coverage. Under ment contribution from seven to 11 percent employees’ these circumstances, the expansion of the banking gross salaries, respectively. For the subset of military sector to rural areas might be costly. On the other and policy services employees, the target is to increase hand, the deposit mobilization of the banking sec- the employee contribution to seven percent and the tor has been increasing over time. The average saving employer contribution to 25 percent of employees’ deposit by individuals in the formal banking system has gross salaries. The share of the pension fund to saving increased from 27 Birr in 2004 to 316 Birr in 2012. Yet, is small (net of pension payments), but the contribu- the share of per capita saving deposits that is channeled tion to gross national savings is increasing (Figure 16). into the formal banking sector is small, and accounted However, the relative share of the pension fund in gross for only to about 4.5 percent in 2012 (in relation domestic saving is declining because the former is not to a per capita income of US$380, or about 7000 Birr). growing as fast as the latter. The past two years saw a substantial increase The newly established Private Organization in the number of bank branches in Ethiopia. Employee Pension Scheme is a potential source for In an effort to increase access to financial services increased domestic savings in the economy. Large from 20 to 67 percent at the end of the GTP period, portions of the urban labor force are engaged in the pri- the commercial banks continued their drive to open vate sector. A new law introduced in FY12 requires any significant numbers of branches across the country. private employees—with the exception of domestic As an effect, the number of commercial bank branches Savings 39 Figure 16: The Pension System and Savings in Ethiopia Net Pension Fund Share in Gross National Saving, % Contribution to Pension Fund, Million Birr 1.6 5000 1.4 4500 550 4000 1.2 3500 1.0 3000 0.8 2500 586 0.6 2000 3990 1500 0.4 2400 1000 1743 1700 2070 0.2 1598 500 0 0 2008 2009 2010 2011 2012 2008 2009 2010 2011 2012 2013 6mon Public Private Pension Fund Public Pension Scheme Private Pension Scheme Quarterly Performance of Private Employees Pension Fund, Million Birr 350.0 18000 300.0 16000 14000 250.0 12000 200.0 10000 150.0 8000 6000 100.0 4000 50.0 2000 0.0 0 Q1 Q2 Q3 Q4 Q1 Q2 2012 2013 Pension Contribution No of Institutions (right axis) Source: Pension and social security agency and private employees’ social security agency. in operation increased by 90 percent in the last two on the savings behavior in Ethiopia. Based on nation- years, from 680 in 2010 to 1,289 in 2012 (Table 5). ally representative data collected via the Urban The majority of the increase comes from the state- Employment and Unemployment Survey (UEUS)29 owned Commercial Bank of Ethiopia (CBE), which an econometric analysis provides insights into the increased its branch network by 167 percent during question of the likelihood of saving in banks (Annex the first two years of the GTP. Private banks show Savings Rate 2).30 The survey data consists of ques- a slower increase in their bank branch network, which tions related to bank saving, individual characteristics is a reflection of their smaller size and ability; the latter is constraint given the 27 percent rule that limits the 29 UEUS is designed to provide data on employment and unemployment availability of funds to onlend and hence the ability at national and regional urban levels. It follows a household approach and covers all urban areas of the country except nomadic areas of the 3 zones to expand. CBE also promotes deposit mobilization of Afar and 6 zones of Somalie regions. through the provision of lottery coupons for additional 30 The UEUS data used in the analysis was collected in 2006, 2009 (Due to quality concerns, the multivariate analysis excludes the 2009 sample), new saving made by depositors and for those who 2010, 2011, and 2012. It also covers more than 70,000 individuals opened new accounts (with a fixed minimum amount). except in 2006 where there are about 60,000 respondents. Since there is no information on the amount of saving the outcome variable is dis- Looking at micro data indicates that the crete which takes a value of one when an individual has a bank account expanded bank branch network has had an effect and zero otherwise. 40 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Table 5: Bank Branch Expansiona Data by Region in Ethiopia, 2006 to 2012 Region 2006 2010 2011 2012 Addis Ababa 148 265 352 430 Tigray 26 43 76 94 Afar 3 5 5 9 Amhara 69 88 132 190 Oromia 115 182 262 354 Somalie 7 10 16 24 Beneshangul 1 2 4 7 SNNPR 38 59 95 143 Gambela 1 2 4 6 Harari 4 10 10 12 Diredawa 9 14 17 20 Total 421 682 973 1289 Source: National Bank of Ethiopia (NBE). a The number of bank branches includes is given as the sum of all banks in a given region. Hence, it includes both public (e.g., Commercial Bank of Ethiopia) and private banks (e.g., Dashen Bank). of savers (e.g., age, gender, marital status, education), environment accompanied it. In situations where the and employment status. The analysis is possible overall environment is right for savings, (i.e., the infla- through matching the UEUS data with data on the tion rate is low, the real interest rate is moderate, and number of bank branches in each of the 11 regions,31 the dependency rate is not changing over time) then which is provided by the National Bank of Ethiopia. the improved access through an increase in the branch The branch expansion in all regions of Ethiopia network can unfold its full effects to increase sav- has led to a significant increase in the likelihood ing. A similar idea for getting the environment right of saving in banks by individuals. The analysis shows comes from reviews of compulsory saving schemes. the increasing importance of bank deposits as finan- For instance, Husain (1995) and Faruqee and cial instruments used by individuals. Expanding Husain (1998) find that compulsory saving schemes bank branches is therefore a major saving-supporting in Malaysia and Singapore are offset by a reduction policy measure in Ethiopia. This is because the num- in voluntary savings and are insignificant in terms ber of branches has a positive and significant impact of impact on private saving. Rather, they find that on improving the access of individuals to formal demographic change and per capita income account for saving opportunities. King and Levin (1993) show strong saving performance in Malaysia and Singapore. that bank expansion and economic growth are posi- tively correlated in cross-country data while Burgess Micro Financing and Pande (2005) show that branch expansion in to rural, unbanked locations in India significantly Microfinance is a dynamically developing sector reduced poverty. in the financial industry in Ethiopia. As per the International experience indicates that expand- ing the bank network will maximize effects 31 These are Addis Ababa, Tigray, Afar, Amhara, Oromia, Somalie, Ben- if an overall saving-conducive macroeconomic eshangul, SNNPR, Gambela, Harari and Diredawa. Savings 41 end of 2012, there were 31 licensed microfinance and 9.46 ATMs per 100,000 adults. Access to finance institutions (MFIs) operating in Ethiopia. Their also remains a major constraint for enterprises. deposits amounted to 5.5 billion Birr and represented The recently published data of the 2012 Ethiopia the savings of 2.6 million clients (Table 6). In 2012, Enterprise Survey confirms that access to finance MFI savings constituted around 4.5 percent of GDS. remains a major constraint for enterprises. This is per- Some MFIs are sizeable financial institutions in their ceived as the main business environment constraint own rights and bigger than some of the commercial by both small (38 percent) and medium (29.5 per- banks. The sector is highly concentrated, with the five cent) enterprises in Ethiopia, compared to an SSA largest MFIs (owned by regional governments and average of 21.2 and 15.2 percent respectively. The same operating in different regions without competing with survey indicates that almost 93 percent of small enter- each other) corresponding to 89 percent of total sec- prises and over 95 percent of medium enterprises have tor assets and 83 percent of total borrowers. The pre- either a checking or a savings account (a percentage dominant loan methodology is group loans, but some higher than the respective SSA averages) but only three MFIs have started to offer individual loans and this percent of small enterprises and 23 percent of medium development could be expanded to support the devel- have a loan or a line of credit. These low percent- opment of micro and small enterprises. Strengthening ages can be explained by (among other factors) the the microfinance policy framework and supervision extremely high value of collateral needed for a loan, through NBE and the inclusion of large MFIs in key corresponding to 249.3 percent (253.5 percent) of the strategic NBE projects, such as the payment system loan amount for small (and medium) enterprises, and the credit information center, are key priorities against a SSA average of 160 percent. going forward. MFIs provide financial intermediation mostly However, access to financial ser vices for communities with low access to formal finan- remains highly limited all over Ethiopia with cial institutions. Low access is often the case due only 1.97 commercial bank branches and 0.33 ATM to the absence of banking institutions, limited capac- per 100,000 adults. This is compared to, for instance ity of beneficiaries, and lack of provision of small Kenya where there are 5.17 commercial branches loans without collateral. Thus, MFIs serve as a social Table 6: The Development of Micro-Finance Institutions in Ethiopia, 2003 to 2012 MFI Savings as % MFI Savings as % Year Number of MFI Number of Clients Total Saving in ETB of GDP of GDS 2003 22 746,136 323,503,677 0.45 6.3 2004 22 622,650 411,234,819 0.48 3.3 2005 26 939,585 583,664,099 0.56 6.2 2006 27 1,299,896 799,356,324 0.62 7.9 2007 28 1,700,396 1,172,879,769 0.69 5.8 2008 27 2,172,823 1,489,128,630 0.61 7.0 2009 30 2,197,688 2,223,443,931 0.67 7.3 2010 30 2,325,914 2,555,729,721 0.68 7.7 2011 31 2,480,810 3,696,016,796 0.73 5.7 2012 31 2,637,625 5,474,346,625 0.74 4.5 Source: National Bank of Ethiopia (NBE); and Adapted from AEMFI (2012) Performance Analysis Report, Table 3; p. 8. 42 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status function by filling the gap where formal financial For example, a group of 10 individuals may contribute institutions failed to reach the majority poor people. US$100 per month for 10 months. MFIs in Ethiopia initially started as semi-financial The ROSCA fund is collected each month and institutions with the support of government and given to one member either by a random draw32 donors, but they have since transformed into deposit or via a bidding process. After having received the taking institutions. Integrating them fully to the for- ROSCA pot when it is his or her turn, the winning mal intermediaries such as banks will greatly help the member continues his or her monthly contribution government’s plan to mobilize savings for investment until the last member of the ROSCA receives the pot purposes. As a policy, the current developmental phase at the end of the ten months. After one ROSCA cycle of MFIs in Ethiopia requires intervention of develop- is completed, the members decide to continue saving ment financing agencies such as the Development by starting another cycle with a similar or different Bank of Ethiopia (AEMFI, 2012). level of contribution. Alternatively, they may disband Expanding the activities of MFIs is one of the the ROSCA and stop the informal saving altogether. key instruments to support the poor and vulner- ROSCAs mobilize large sums of money, which can able and to meet the saving mobilization objectives be used by participants for a variety of purposes such of the five-year GTP. However, the regulations are as insurance against shocks, to buy durables, to start rather restrictive and thereby prohibit growth to fully up a business, to expand a business, and to buy com- meet demand and to allow the provision of flexible and mercial vehicles. responsive services to clients where they are needed. ROSCAs satisfy the financial needs of the They could also be provided with extra loanable funds majority of the Ethiopian population and have by government since current loan volumes are insuf- relative merits compared to banks. The funds con- ficient to meet demand (Amha, 2012). tributed to ROSCAs can be instantly drawn on, and In addition to participating in MFIs, many the institutions are useful for information sharing individuals in Ethiopia save either in Savings and and social interaction. The winner of the ROSCA pot Credit Cooperatives (SACCOs) or via Rotating has a possibility repaying the loan to other members Savings and Credit Associations (ROSCAs), of ROSCAs over an extended period. Members can which are often preferred to formal saving join ROSCAs with limited or no collateral, mem- institutions. Contrary to ROSCA, which are bership entails low transaction costs, and there are informal groups, SACCOs are usually supervised no restrictions on what members spend their money by the Federal Cooperative Agency (or its regional on. In ROSCAs, there are no form filling procedures, branches); informal surveys suggest that SACCOs excessive waiting time for withdrawals and loan hold around 13 million ETB in their accounts approval delays. Since membership is usually not through 450 rural cooperatives and 30,000 mem- anonymous and is based on close social networks, bers. The members of ROSCA are usually neigh- there is trust among participants and few reported bors in a given community, friends, and employees defaults. The characteristics of members participating of a given organization. In a ROSCA, these members in them are significant determinants of participation come together and make regular contributions to what in these informal schemes (Kedir and Ibrahim, 2011; is referred to as the ROSCA pot or fund. Members Kedir, et al 2011). Since some of the big ROSCAs meet in cycles on a daily, weekly or monthly basis. The already keep their deposits in banks, further effort fund is collected by a ROSCA judge who is selected by members as a responsible person to organize the 32 Random ROSCAs are the most common in Ethiopia. Sometimes funds ROSCA meetings and keeping records. The ROSCA are also released based on the financial need of a given member and the fund is given to one member in each of meeting cycle. judge decides to allocate the funds to the person in need. Savings 43 to encourage them to use formal banking facilities savings from the diaspora. Sales were slow during will support the existing strong saving mobilization the first months of offering despite the efforts of the effort of the Ethiopian government. Commercial Bank of Ethiopia, and the embassies and consulates, to sell them. Some risks that the diaspora Diaspora Bonds perceived were (Plaza, 2011): perceptions on the pay- ment ability of EEPCO on its future earnings from The government introduced two sets of diaspora the operations of the hydroelectric power; lack of trust bonds over the past 3 years. The first one was called the in the government as a guarantor; and political risks. “Millennium Corporate Bond” in 2009, and was used The issuance of the “Renaissance Dam Bond” was for raising funds to finance hydropower dam plans for designed to overcome some of the weaknesses observed the Ethiopian Electric Power Corporation (EEPCO). in the first diaspora bond. The government took con- The second one, which consolidated the “Millennium siderable additional marketing and awareness-raising Corporate Bond” was the “Renaissance Dam Bond” campaigns to encourage the diaspora to buy it. The launched in 2011 to finance the building of the “Grand minimum denominations were lowered to US$50, Renaissance Dam.” The idea behind the diaspora bonds so more Ethiopians could have access. Additionally, is simple and has been tried out in countries like Israel the bond can be transferred to up to three people and and India. With those bonds, which are marketed it can also be used as collateral in Ethiopia. Finally, the in developed countries to reach the emigrated dias- Commercial Bank of Ethiopia committed to cover any pora, poor countries try to create a formal instrument remittance fees associated with the purchase of these for emigrants to invest back in their country of origin. bonds (Plaza, 2011). Still, at the end of FY2011/12 the Advantages of reaching out to a diaspora are that they consolidated diaspora bonds in EEPCO’s books are often patriotic and hence benevolent, they are reached a mere US$888,000. A domestic component patient, and they are less prone to follow irrational to the “Renaissance Bond,” however, is reported to have market sentiments (The Economist, 2011). around US$500 million in pledges and US$200 mil- India and Israel combined have both raised lion in actual collection by April 2013 according to the about US$40 billion by reaching out to the wealth Council for the Coordination of Public Participation of their diaspora communities. Diaspora bond on the Construction of the Grand Ethiopian issuance by the Development Corporation for Israel Renaissance Dam. (DCI) has been a recurrent feature of that nation’s annual foreign funding program, raising well over Housing Saving Scheme33 US$25 billion since 1951. The State Bank of India (SBI) has issued diaspora bonds on just three occa- The Government started to implement a new hous- sions: in 1991, following the balance of payments ing scheme in June 2013, in the city of Addis Ababa, crisis; in 1998, after the country conducted nuclear through the Commercial Bank of Ethiopia. The tests; and in 2000. The SBI has raised US$11.3 billion. scheme aims to encourage low- and middle-income Jewish diaspora investors paid a steep price premium earners to deposit money for a given period, with the (perhaps better characterized as a large patriotic yield goal of reaching sufficient savings to allow for buying discount) when buying DCI bonds. Indians living a home. By the time the savings will have reached abroad purchased SBI bonds when ordinary sources a predefined percent of the estimated housing cost, of funding for India had all but vanished (Kethkar the savers will be entitled to a house. and Ratha, 2009). “Millennium Corporate Bond” issuance has 33 At the time of report, preparation of the scheme was too new to fa- not met the expectations thus far in attracting cilitate an evaluation. 44 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status The new scheme allows for three main choices rates are key determinants of savings in the World for the individual saver: the “10/90”, the “20/80” and in Ethiopia. and the “40/60” options. Accordingly, an individual  Second, pursuing a strategy of monetization, saves ten percent (or 20 percent or 40 percent) of the while keeping inflation under control could housing cost in closed saving accounts at Commercial be an efficient means to increase the saving Bank of Ethiopia branches. Upon eligibility, which rate. Ethiopia is the only country in sub-Saharan varies according to each option, the Government facil- Africa that experienced a major decline in the itates long term loans to cover the remaining balance: degree of monetization in the late 2000s (cap- tured by M2/GDP). Monetization would need  The 10/90 scheme targets low income groups to be accompanied by appropriate financial where the potential house owners will save a fixed deepening with a vibrant banking sector that sup- monthly amount of up to ten percent of the ports both public and private real sectors of the estimated house price. After three years, the economy. More competition in the financial sector Government’s long-term loan will kick in for the could increase saving products through innovation remaining 90 percent. and lead to a larger customer base—and savings.  The 20/80 scheme targets lower middle-  Third, the importance of official transfers and income households.34 Savers will deposit a pre- remittances for government and households defined amount for five years until the amount to save offers the opportunity to implement reaches 20 percent of the estimated total housing policy measures that help channel and formal- cost before the long-term loan is triggered. ize those transfers into the economy. While the  The 40/60 scheme targets the upper middle- reliance on external funding causes some concern income group of the population. Registered as to the dependency of the saving rate in Ethiopia individuals are expected to save for five years on variables outside the control of policymakers, before the long-term loan is arranged for 60 per- three concrete steps could help formalize these cent of the estimated housing cost (the quality transfers (RMA, 2010): of houses is supposed to be best in this scheme,  Establishing formal savings products for due to better building material). An alternative remittances. Ethiopia’s banks and other RSPs approach in this scheme is that individuals can have not yet made much concrete progress pay upfront a 40 percent down payment and in using remittance transfers as an entry point the housing agency will arrange for loan for the for formal financial products. This is the case remaining balance. partly because of the relatively underdevel- oped state of Ethiopia’s financial system. Policy Recommendations Allowing banks to better serve remittance receivers in Ethiopia would help banks to for- The following recommendations emerge from the malize remittances. analysis of this chapter:  Increasing transparency of remittances- related fees. Making information on remittance  First, establishing a stable macroeconomic environment with low, but positive real inter- 34 This scheme complements a mechanism (i.e., the raffle for condo- est rates and low levels of inflation, is among minium houses). There, however, the winners of the raffle are expected the main policy tools to increase the saving rate to settle 20 percent of the cost of the house up front at the time of the hand over the house. The new scheme is expected to use a similar raffle in Ethiopia. The analysis showed that macroeco- mechanism to ration the availability of the housing option. Savers cannot nomic variables such as real interest and inflation withdraw money from accounts once deposited. Savings 45 credible deposit guarantees to further strengthen Box 3: Policy Recommendations the growth trend in bank saving. One important on Savings consideration is also to focus in some areas of the Policy Recommendation country where the saving culture is not strong. For Area instance, the Southern part of Ethiopia is mainly Macro- 1. Macroeconomic stability with positive rich in cash crops. Therefore, saving in formal economic real interest rates. institutions is weak compared to areas in the policy 2. Pursue a strategy of monetization Northern part of the country where households Financial 3. Facilitate remittance transfers, including face years of droughts and other calamities, which sector establishing formal savings products, measures increasing transparency, and reduc- caused them to be risk averse and develop a thrift ing costs. culture. Policy making that recognizes this het- 4. Continue bank branch network expan- erogeneity could be more flexible in incentivizing sion. 5. Enhance Micro Finance Institution saving mobilization where the impact promises (MFI) capacity. highest returns. 6. Integrate informal savings schemes  Fifth, the capacity of MFIs to access loanable with formal. 7. Develop mobile banking. funds should be facilitated with government 8. Improve financial literacy. support. MFIs are entrusted with the responsibil- ity of being the key institutions that implement the MSE Development Strategy of Ethiopia, employ- ment creation, and poverty reduction. So far, they transfer fees publicly available to both poten- have raised large sums of savings especially in the tial senders and receivers enhances market last two years, but this covers only half of their transparency among the market’s Remittances loanable funds and additional support could Service Providers (RSPs) and can further help to further MFI penetration in Ethiopia. increase the remittance inflows sent through Strengthening the microfinance policy framework formal channels. and supervision through NBE and the inclusion  Reducing cost through new technology. The of large MFIs in key strategic NBE projects, such development of new technologies and prod- as the payment system and the credit information ucts for the delivery of cross-border remit- center are key priorities going forward. tance inflows, such as mobile money transfers  Sixth, given the great importance of informal and card-based technologies, could further saving, efforts must be channeled into inte- reduce the cost of remittance transfers and grating the informal and formal sectors. There further boost competitiveness among RSPs is limited integration between the informal and in Ethiopia’s remittances services market. formal financial sectors. An attempt to link the  Fourth, the branch network expansion policy two could potentially lead to getting the saving should be continued and furthered with more mobilization into the open at an integrated mar- efforts targeted also to reach the rural areas ket price (Senbet, 2012). Policy can also focus where saving is mainly in kind. Due consider- on encouraging Rotating Savings and Credit ation is needed to facilitate bank branch expan- Associations (ROSCAs) and community saving sion also of private commercial banks. Since trust cooperatives to hold their saving deposits in banks and risk aversion is one of the key factors leading to formalize and secure otherwise informal sav- households to save in informal schemes, formal ings. Policies could aim to link informal schemes saving institutions such as banks can provide that operate in rural areas with excess liquidity 46 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status to more stable and regulated deposit taking MFIs, mobile operators help to push the price down which are in need of deposits to expand. for the services.  Seventh, establishing mobile banking draw-  Finally, to benefit from branch expansion and ing upon successful examples in neighboring use of money transfer technologies, individu- countries. To illustrate, “Branchless Banking” als’ financial literacy levels need improvement. is practiced in Kenya via M-PESA, payment A recent cross-country study showed that a num- method using mobile phones. Tanzania will ber of key ingredients are commonly seen in suc- soon launch a similar approach, called Z-PESA. cessful financial literacy programs and it includes Ethiopia is a latecomer to permit mobile bank- the following recommendations to improve the ing in Africa. With most of the population financial education for consumers: (1) start with “unbanked” (in 85 million total population), awareness building rather than with traditional mobile banking is a lucrative venture. Approval numeracy skills; (2) leverage social networks and of the currently considered legislation for mobile peer effects; and (3) identify and target vulner- banking in Ethiopia could help kick start the able populations. Women, youth, the elderly, and development of the sector. In addition, the mobile those with lower incomes and educational attain- banking business model in other countries ben- ment are less likely to be financially literate and efited from deregulation of mobile money services targeting those promises the most success (World (e.g., Kenya, Tanzania, and Uganda). Multiple Bank, 2012e). Savings 47 Annex: Savings Rate literature, where the lagged private savings rate has a coefficient of 0.59 (Loayza, et al., 2000). Annex Savings Rate 1  Income Level And Growth: The estimation reveals that the level of income is positive and Empirical analysis of the determinants statistically significant—weakly for SSA and of savings strongly for Asia only for PRS. The coefficient on the growth rate of income is positive and sig- This annex provides an empirical analysis of the nificant for SSA. This indicates that there is a vir- saving determinants for SSA countries and emerg- tuous circle from higher growth to higher saving. ing Asian countries. The analysis follows Loayza, Previous literature incorporating instrumental et al. (2000) by employing Generalized Method variable method and various causality tests also of Moments (GMM) estimators applied to dynamic confirm the virtuous circle (Loayza, et al., 2000). models, using panel data. These estimators allow  Macroeconomic Stability: The coefficients control of unobserved country-specific effects and on squared real interest rate are negative and sig- potential endogeneity between savings rates and other nificant for all specifications, suggesting a nonlin- macroeconomic variables.35 ear, hump-shaped relationship with saving rates. The reduced form estimating equation is given by: In other words, the real interest rate has a posi- tive impact on domestic savings up to certain st = ast–1 + θ’Xt + ϕ + εt,(1) threshold, but its impact turns negative as real interest rate rises beyond such a threshold. This Where is savings rate; denotes a set of determinants indicates that the substitution effect overwhelms potentially affecting saving rate; represents a set the income effect up to the threshold while the of unobserved, time-invariant, country-specific effect; income effect dominates the substitution effect and is the error term. All the determinants analyzed beyond the threshold.37 in the previous section enter linearly, and squared real  Financial Development: Coefficients on M2/ interest rate, and broad money enter to capture pos- GDP ratio are insignificant for SSA but signifi- sible nonlinear relationship in the above equation.36 cant for Asia, indicating financial development Table 7 shows the estimation results for GDS leads to higher saving in Asia. However, this does and PRS, respectively. GMM estimation results are not necessarily mean that monetization does not satisfactory and broadly consistent with the determi- promote saving in SSA. As IMF (2010b) points nants highlighted in the main text and the literature. out, monetization and financial deepening are Key insights on the determinants of the saving rate the dynamos for economic growth. Furthermore, are as follows:  Persistence: The presence of inertia in saving 35 Interested readers may refer to Hauk and Wacziarg (2009) which is evident as the coefficient of the lagged saving overcomes issues of endogeneity, small sample and measurement er- rors by simulating data from a true economic growth model. This said, rate is positive and statistically significant across they show that both fixed effect and Arellano-Bond GMM models lead all samples. The coefficients of the lagged GDS to similar results as simple OLS estimation. 36 The analysis drops the inflation rate because of its high correlation and PRS rates for SSA are 0.63 and 0.50 respec- with real interest rate. tively, implying that the factors affecting saving 37 Loayza, et al. (2000) also find a similar negative impact of real interest rate on private savings. In general, the results regarding real interest rates rate have 2.70 and 2.01 times larger long-term and savings in developing countries should be taken as given, because impact than their short-term impact. This out- the real interest rate measure may reflect more the action of nominal interest rate controls and financial repression than the intertemporal rate come is consistent with findings by previous of substitution of consumers. 48 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Table 7: Saving and its Determinants Gross Domestic Savings Private Savings Saving Determinants SSA Asia SSA Asia Lagged saving rate 0.625 *** 0.503 *** 0.503 *** 0.330*** (14.28) (7.11) (12.18) (5.70) Real per capita GDP 0.053 0.013 0.111* 0.053*** (0.96) (0.75) (1.78) (3.10) Per capita GDP growth 0.310*** 0.066 0.226*** –0.048 (4.63) (1.48) (3.11) (–1.01) Real interest rate –0.112*** 0.025 –0.055 0.009 (–2.92) (0.75) (–1.36) (0.25) Squared real interest rate –0.296*** –0.560** –0.224*** –0.742*** (–3.42) (–2.22) (–2.65) (–2.58) M2/GDP –0.181 –0.062 ** 0.130 –0.080*** (–1.38) (–2.41) (0.88) (–2.94) Squared M2/GDP 0.088 0.011* –0.114 0.016*** (0.76) (1.77) (–0.99) (2.54) Terms of trade 0.094*** 0.074** 0.060*** 0.039 (4.23) (2.41) (2.79) (1.27) Urbanization ratio 0.173 –0.094 0.047 –0.171** (1.08) (–1.21) (0.26) (–2.23) Old dependency ratio –0.133 –0.016 –2.134* –0.684** (–0.12) (–0.05) (–1.74) (–1.98) Youth dependency ratio 0.117 –0.232*** 0.128 –0.274***  (1.52) (–3.57) (1.49) (3.51) Public saving — — –0.760 *** –0.668*** — — (–9.86) (–8.34) Number of observations 460 195 394 183 Number of countries 19 8 19 8 Note: figures in brackets are t-statistics. *, ** and *** indicate statistical significance at 10%, 5% and 1% level, respectively. Savings 49 as examined above, higher income significantly countries. A rise in the old dependency ratio con- raises saving. Therefore, one should be careful tributes to dampen the positive impact of increases to not underestimate the impact of monetization in per capita income on the saving rate. on saving via economic growth channel.  Public Saving: The impact of public saving  Terms of Trade: The coefficient on terms of trade on PRS is strongly negative across country groups, is positive and significant with a one percentage confirming that public saving crowds out PRS. point improvement in the terms of trade raising Specifically, for SSA, the private sector reduces domestic saving rate by 0.08 percentage point. its saving rate by 0.76 percentage points for each The effect of temporal improvement in the terms percentage point increase in the public saving rate of trade on saving is important for SSA countries within the same year the policy change occurs. that export limited variety of primary commodi- An important issue, especially regarding on policy ties sold in highly volatile market. implications, is if the coefficient is statistically  Demographics: The coefficient on old depen- different from –1. One can reject the hypothesis dency ratio is significant with signs for PRS con- that public saving fully crowds out private saving sistent with prediction of the life-cycle theory. in the short run, indicating that the government The magnitude of coefficients tends to be larger can raise the GDS rate by increasing public sav- for old dependency ratio than youth dependency ing. However, for sustainable country-led eco- ratio as previous literature finds. Loayza, et al. nomic growth in Ethiopia, public saving may (2000) argue that this reflects the fact that the not be an appropriate policy tool given public labor force effectively includes a non-negligi- saving has a negligible impact on GDS rate in the ble proportion of the young workers in many long run. Annex Savings Rate 2 Table 8: Probit Model Predicting the Probability of Having a Bank Account, UEUS 2006–2012 Variable Variable 2006 2010 2011 2012 No of Bank Branches 0.0006 *** 0.001*** 0.001*** 0.001*** (0.0001) (0.000) (0.000) (0.000) Female –0.180*** –0.086*** –0.092*** –0.086*** (0.021) (0.018) (0.017) (0.017) Age 0.020*** 0.004*** 0.004*** 0.005*** (0.001) (0.001) (0.001) (0.001) No schooling –0.215*** –0.262*** 0.034 0.082 (0.062) (0.074) (0.070) (0.063) Primary 0.220*** 0.103*** 0.127*** 0.105*** (0.044) (0.037) (0.035) (0.035) Secondary 0.080*** 0.186*** 0.311*** 0.302*** (0.029) (0.025) (0.027) (0.028) Married –0.135*** –0.058*** –0.075*** –0.031 (0.028) (0.022) (0.022) (0.021) Divorced –0.356*** –0.207*** –0.153*** –0.225*** (0.042) (0.034) (0.032) (0.033) Wage employed 0.342*** 0.973*** 1.087*** 1.049*** (0.029) (0.028) (0.028) (0.027) Self employed 0.032 0.854*** 0.945*** 0.810*** (0.030) (0.032) (0.031) (0.029) Employer 0.445*** 1.497*** 1.679*** 1.781*** (0.121) (0.106) (0.093) (0.081) Unemployed –2.538*** –0.930*** –0.841*** –0.810*** (0.915) (0.035) (0.032) (0.030) Pseudo R-squared 0.21 0.25 0.26 0.25 No of observations 60,282 74,777 72,697 70973 Notes: ***, **, *= significant at 1%, 5%, and 10% levels. In recent years, the government has supported the expansion of branch networks of Commercial Bank of Ethiopia (CBE). This led CBE to triple its branch network in less than two years and more than doubled its deposits (Amha and Alemu, 2012). A similar analysis was also undertaken that included an additional variable on income. In the data set used, however, income is individual- based and not household based. Still, an analysis was carried out to test the influence of such an income variable, and it was found that the vari- able is insignificant. Much more, the number of branches variable is still statistically significant and positive. In addition, the coefficient on variable “branches” is very small compared to others such as employed, self-employed and employer dummies. Since the latter may be proxies for income levels (in monetary form), they may influence the overall result in favor of the significance of the branch network extension. Trade Logistics38 3 Introduction Agreement negotiations, have the potential to trig- ger improvements in the area, as could regional fora, Trade logistics efficiency plays an important role such as COMESA and bilateral trade agreements with in competitiveness and trade facilitation. This Ethiopia’s main trading partners. is because it can lead to important reductions in the The issue of trade logistics is particularly perti- time and costs of trading. A challenging task lies nent in the case of Ethiopia. This is so because there ahead for trade logistics in sub-Saharan Africa (SSA), is consensus among most observers that the current as summarized by the following fact: “Shipping system is not functioning optimally. The substantial a car from China to Tanzania costs US$4,000, but delays in shipping containers are one of many indica- getting it from there to nearby Uganda can cost tors hereof. Global indicators of logistics performance another US$5,000” (The Economist, 2013). The also give cause for concern: Ethiopia’s relative rank- emphasis of trade policy reforms across the regions ing in the World Bank Logistics Performance Index is often placed on trade liberalization through tariff dropped from 123 in 2010 to 141 in 2012 (out reductions to enhance international integration and of 155 countries surveyed, Figure 17.1). The recent strengthen competitiveness. While this generally introduction of the multimodal system, in particular, leads to reductions in time to trade and lower costs, has stimulated substantial public debate. and thereby promotes exports, more needs to be done This chapter addresses some of the following in lowering non-tariff barriers. Hence, the focus questions: What is Ethiopia’s trade logistics perfor- on trade logistics. mance compared to its peers? How has this changed Advances in trade logistics to facilitate trade over time? Why is trade logistics such a big challenge supported the dramatic transformation of suc- in Ethiopia? What are the implications of being land- cessful emerging economies, such as China and locked? Is the multimodal system a part of the problem India. According to a comprehensive global survey or the solution? What can Ethiopia learn from other of international freight forwarders, India and China countries? In addressing the problem, where should undertook aggressive reforms over the past decade policy makers start? to improve their logistics performance in support The chapter is structured as follows: Section of economic transformation (World Bank, 2012). B offers an overview of Ethiopia’s trade logistics perfor- At the same time, many countries in sub-Saharan mance compared to relevant peers. Section C presents Africa (SSA) have experienced limited improvement a diagnostic of the Ethiopian trade logistics system in their logistics performance and this ultimately ham- by analyzing three policy dimensions: (a) structure pers economic development. Trade logistics are also and performance of the transport and trade logistics instrumental for attracting Foreign Direct Investment (FDI), especially of the export-seeking kind that 38 In preparing this report, MOFED requested the Bank that this analysis is so important for small economies (Delvin and cover issues and policy recommendations along the whole supply chain and problems related to policy implementation of trade logistics. The Yee: 2005). Ongoing discussions at the World Trade team was also asked to place due consideration on specific issues arising Organization (WTO), as part of the Trade Facilitation from Ethiopia’s status as a landlocked country. 52 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status system; (b) customs, border control and other trade The key drivers of high trade costs are related related regulations; and (c) logistics and distribution to inland transportation and handling, and docu- services. Finally, Section D derives the policy recom- ment preparation.39 Figure 17.4 breaks down the mendations arising from the analysis. cost to export a container by cost category. Ethiopia shares the high cost of inland transportation with other Trade Logistics for Trade Development: landlocked countries (Rwanda, Uganda, and Zambia). Ethiopia’s Performance In the apparel sector, for instance, higher inland trans- port costs adds more than a two percent production Poor trade logistics penalize importing and cost penalty and a ten-day delay, due to longer dis- exporting firms. To illustrate, poor trade logis- tances, inadequate transport infrastructure, and a lack tics can add about ten percent to production cost of competition in the trucking industry (World Bank, in light manufacturing in East Africa (World Bank, 2012c). The cost of document preparation is also 2012c). It can also cause long and uncertain delays, an additional cost penalty on exporters. Here Ethiopia which are not acceptable to most global buyers, stands out, also in relation to other landlocked coun- especially in time-sensitive industries. As a result, tries. The cost of apparels increases by an estimated two production is often confined to small market niches. percent as a result, according to World Bank (2012c). For instance, Ethiopia exports only small volumes In Ethiopia, commercial banks charge three percent of low-value apparel products even though the of the value of the shipment on imports and a two Free-On-Board (FOB) price for an Ethiopian polo percent advisory fee on exports (compared with less shirt (US$3.20) is more than 40 percent lower than than one percent in China). the FOB price of an equivalent Chinese polo shirt Trade logistic costs are further increased by the (US$5.50). The higher Chinese price results from high cost of obtaining foreign exchange and the cost higher-quality shirts and the premium that global for shipping to and from Africa. The National Bank buyers put on China’s ability to offer greater choice, of Ethiopia charges a 1.5 percent foreign exchange bigger volumes, and shorter, more reliable delivery commission fee on the dollars needed to import the times (World Bank, 2012c). inputs. Waiting for the National Bank’s authorization Trade costs and transit times are very high can take up to six months when foreign exchange in Ethiopia. Importing a container in Ethiopia is scarce. In fact, the time needed for approvals and rather than in Tanzania adds an additional cost verification for foreign exchange cascades through of US$1,095 for an Ethiopian importer (Figure 17.2). the whole logistics chain. It starts at the beginning This difference can go up to US$2,060 when where pre-import permits require foreign exchange- comparing Ethiopia with Vietnam or be as low related documents. Moreover, it goes down the as US$310 compared to Kenya. On the other hand, chain to post clearance where there is a need to close importing a container into Uganda is even more the documentation loop with the national bank. expensive than in Ethiopia, as both are landlocked As such, foreign exchange regulations add even more countries. Similar relations are observable when looking at the cost of exports. Moreover, the tran- 39 Documents considered in this category of the Doing Business Report sit time taken to import and export is about twice for Ethiopia include: Bank permit, bill of lading, commercial invoice, as long for Ethiopia as it is for China, Vietnam, and customs-export declaration, export permit, health and fumigation cer- tificate, packing list. It also includes, consistent with standard procedures Kenya (Figure 17.3). Comparing Ethiopia with other in the Doing Business analysis, the time and cost related to obtaining landlocked countries such as Rwanda, Uganda, and a letter of credit through the local financial system. For more details, please refer to the Ethiopia overview on the Doing Business Report website: Zambia, however, shows that Ethiopia’s relative per- http://doingbusiness.org/data/exploreeconomies/ethiopia#trading- formance is better. across-borders. Trade Logistics 53 Figure 17: Ethiopia’s Trade Logistics Performance 1. Ethiopia’s Logistics Performance Index (LPI) over Time 2. Cost to Handle a 20-foot Container 80 3.0 Rwanda 2.9 90 Zambia 2.8 Uganda 100 2.7 Ethiopia 110 2.6 SSA 2.5 Kenya 120 2.4 Tanz 130 2.3 Korea 2.2 China 140 Vietnam 2.1 150 2.0 0 2000 4000 6000 2007 2010 2012 US$ LPI Rank (left axis) LPI Score (right axis) Cost to Export Cost to Import 3. Time to Import and Export (days) 4. Cost to Export a 20-foot Container by Category Zambia Rwanda Ethiopia Uganda SSA Zambia Uganda Ethiopia Rwanda Kenya Tanzania Tanzania Kenya China Korea Vietnam China Korea Vietnam 0 20 40 60 0 1000 2000 3000 4000 Days Documents Preparation Customs Clearance and Technical Control Time to Export Time to Import Ports and Terminal Handling Inland Transportation and Handling 5. Logistics Performance Index, 2012 6. Ethiopia: Trading Across Borders, 2009–13 LPI 3000 50 5 2500 48 4 Timeliness Customs 46 3 2000 2 44 1500 1 42 Tracking and Infrastructure 1000 40 Tracing 500 38 0 36 Logistics Competence International Shipments 2009 2010 2011 2012 2013 Ethiopia 2012 Kenya 2012 Tanzania 2012 Rwanda 2012 Time to Import (days) Cost to Export China 2012 Korea, Rep. 2012 Vietnam 2012 Cost to Import Time to Export (days) Source: World Bank Doing Business Report, Trading Across Borders (2012), World Bank (2012), World Bank Logistics Performance Index (2012). Note: Logistics Performance Scale from 1 (low) to 5 (high score). Data for Uganda and Zambia are not available for 2012. 54 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status complexity to documents and trade procedures caus- Looking at costs and time, Ethiopian trade ing further delays in the system. At the same time, logistics are stagnant. Figure 17.6 provides a sum- it costs 60 percent more to ship to the United States mary of trade costs and the number of days required from Djibouti than from China, and about the same for trading in Ethiopia.41 While there are slight to ship to Europe, despite the much greater distance improvements in the time to export since 2010, the from China (World Bank, 2012c). Reasons include time to import is getting longer since 2009 and with- the relative low traffic in Djibouti, as well as the struc- out any change for the last four years. In terms of costs, ture of exports and imports. A domination of heavy there are modest savings in exporting costs while the imports and largely low-scale, small exports compli- cost of importing increased since 2009 owing to the cate logistic matters. Therefore, fully loaded ships with positive correlation between trade costs and delays. heavy goods for Ethiopia into Djibouti are not neces- Overall, the situation is best described as stagnant. sarily at full capacity when they leave the port again. A low level of general infrastructure devel- The overall logistics performance is low opment, which hampers Ethiopia’s overall com- in Ethiopia compared to relevant peers. The World petitiveness amplifies the low trade logistics. Bank Logistics Performance Index (LPI) provides In the 2012/13 World Economic Forum (WEF) World a comprehensive measure of the state of trade logis- Competitiveness Report, Ethiopia’s infrastructure tics in a country facilitates comparisons with other development is ranked 119 out of 144, far behind countries.40 Ethiopia ranks at the lower end of the Vietnam and China (95 and 48, respectively), but rela- surveyed countries (141 out of 155 countries in 2012). tively better than Tanzania and Uganda (132 and 133, It is not only below the average for sub-Saharan Africa respectively). in all the six key dimensions of logistics performance measured, but also lags behind in the direct com- Diagnosing Ethiopia’s Trade Logistics parison with neighboring Kenya and Tanzania; yet Challenges it is only slightly lower performing as landlocked Rwanda (Figures 17.5). Ethiopia is generally consid- Ethiopia’s trade logistics challenges can be grouped ered “logistics unfriendly.” The difference from indi- into three major policy areas: (a) Improving the vidual countries in East Asia such as China, Korea, structure and performance of the transport and and Vietnam, as well as the grouping of East Asia and logistics system; (b) strengthening customs, border Pacific, is even starker. control and other trade related regulations; and The logistics system in Ethiopia is not improv- (c) enhancing the efficiency of logistics and distribu- ing over time. In fact, recent studies show even tion services. Policy area (point a) addresses issues a worsening situation relative to other countries. related to “inland transportation and handling” According to the LPI, the ranking of Ethiopia shows while (point c) captures “document preparation” (see a relative deterioration between 2007 and 2012. Figure 17.4). Figure 18 illustrates these three policy In the first LPI of 2007, Ethiopia still ranked 104 out dimensions and the rest of the chapter is structured of 150 economies; in 2010, it ranked 123 out accordingly.42 of 155 economies; and in the most recent version in 2012, Ethiopia reached only 141 out of 155. The situation gets even more worrisome if one includes 40 For a description of the LPI, please refer to the Logistics Performance Index Website. in the consideration Djibouti, Ethiopia’s only gateway 41 World Bank (2013) Doing Business Report, Washington D.C. to the world in terms of land transportation. Djibouti 42 The analysis in this chapter is delimited to land and sea trade logistics as these are the modes for the highest volume imports and exports. With ranked 145, 126, and 154, respectively for 2007, the introduction of the MTS system both modes are also the source 2010, and 2012. of the largest challenges. Trade Logistics 55 Figure 18: Policy Dimensions of Ethiopia’s contract from a place in one country at which the goods Trade Logistics Challenges are taken in charge by the multimodal transport opera- a. Structure and performance of the transport logistic system tor to a place designated for delivery situated in a dif- Lack of competition in the multimodal system (MTS) ferent country.”43 According to the UN Multimodal Lack of capacity and coordination in the MTS Convention, the Multimodal Transport System Lack of physical infrastructure Challenges of landlockedness (MTS) is the usage of multiple modes of transpor- tation for the delivery of goods in a single contract Trade costs in Ethiopia are with a carrier for it to assume all responsibilities for very high and time required the transportation of cargo between two countries. to trade is long, logistics performance is low and stagnant MTS is a system of transporting goods from source country directly to inland dry ports or buyers’ ware- houses instead of delivering only to border posts. The multimodal scheme has a number of advantages. b. Customs, border control c. Lack of efficient logistics It eliminates delay and dwell time in the port, mini- and other related regulations and distribution services mizes the risk of damage or theft, negates the need Source: Staff elaboration. to pay storage fees for the sale of goods at auction due to delays, and avoids numerous nonphysical barriers to movement in transit to inland dry ports (Gelalcha, Structure and Performance of the Transport 2009). If implemented appropriately, it leads to low and Logistics System costs of operations because it uses the most economi- cal combination of available transport systems, with The Multimodal Transport System (MTS) the least possible environmental damage. The MTS will lead to efficient trade logistics if two issues are The Government of Ethiopia has identified addressed. The first issue is operational, and is directly the logistics system as a priority strategic issue linked to the transport policies and infrastructure to improve the competitiveness of the economy. of a given country and the region it is located. The A first major step was taken in November 2011 with second issue has to do with documents and execution the introduction of a multimodal transportation sys- of responsibilities by various stakeholders44 responsible tem. As a result, the Ethiopian Shipping and Logistics for the movement of goods.45 Enterprise (ESLSE) was established by combining The efficient usage of a multimodal transport operations of the three major State-Owned Enterprises system depends on the actual implementation (SOEs) of Ethiopian Shipping Lines, Ethiopian within the country context. In Ethiopia, there are Maritime and Transit Services, and Ethiopian Dry mixed experiences so far. On the one hand, the recent Port Service Enterprise. The new system was enacted Ethio-Djibouti Multi Modal Transit and Transport to enable importers to save time and cost giving the Agreement (EDMTA) is one of the few function- mandate to one agent to take charge of transiting and ing accords in Africa. As per the agreement it would transporting shipments until delivery to the importer. The multimode system is a well-recognized 43 Article 1 of the United Convention on International Multimodal international concept that can potentially lead Transport of Goods. MTS is also referred to as “door-to-door” and “in- termodal” transport system. MTS has gained attention in the last 10 years to substantial improvements in trade logistics. among private operators as well as governments. International multimodal transport is defined as the 44 These include the owners of the merchandise (shippers), freight- forwarders, carriers, insurers and terminal operators. “carriage of goods by at least two different modes 45 Zuidwijk (2003) “A multimodal transport perspective” provides more of transport on the basis of a multimodal transport information on MTS. 56 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status be possible to achieve excellent transport and trade implementation challenges.46 Despite its potential facilitation. On the other hand, the low cost argument to streamline the logistics chain, the multimodal in favor of MTS has not been realized due to a series scheme has not led to better logistics management of implementation problems. The sole MTS opera- in Ethiopia. Implementation of the system may have tor, ESLSE, is currently overwhelmed with the task been over ambitious. For instance, ESLSE under- at hand. Lack of adequate transport infrastructure, took a number of restructuring activities in paral- unclear tariff structure, ad-hoc transport policies, and lel to the management of the new logistics services the severe clearance bureaucracy are all detrimental system, which led to capacity constraints. Moreover, to the MTS concept. There is a serious congestion the system is very complex and may require a multi- problem in the dry ports which has, in turn, resulted stakeholder effort to improve management, however in substantial operating costs for ESLSE. Supporters private public partnerships are currently limited to the of the new system argued before its implementa- unimodal framework. The pilot implementation of the tion that there would be savings on scarce foreign MTS, which focused only on 8 percent of government currency, warehousing costs, and port fees; this was cargo, was prematurely scaled up to full operational also the case in the initial piloting, where the MTS mode. Better sequencing of implementation might was only obligatory for government cargo. However, have led to better results. these benefits are yet to be realized for the large scale A gradual introduction of more market-based implementation that was rolled out to cover all cargo implementation of the MTS would potentially shortly thereafter. In fact, there are signs that delays be relevant in Ethiopia. In many African countries and costs have increased over the implementation freight rates are relatively high, due to anti-competitive period, rather than decreased. cargo reservation policies, and Ethiopia is no excep- tion (Teravanithorn and Raballand, 2008). The mul- Lack of competition, capacity and timodal scheme has protectionist tendencies because coordination other operators are not allowed to take part—except ESLSE. However, according to research on African Competition in the Ethiopian trade logistics system countries, such a monopolistic approach to logistics was absent even prior to the introduction of the management hampers trade (Mbuli, 2013). ESLSE multimodal concept. When importers financed enjoys monopolistic privileges, which are reflected operations through one of the state banks, the bank in its pricing behaviors. Effectively importers have required the importer to use either Ethiopian Shipping become price takers and ESLSE, as a multimodal Line (ESL) or Ethiopian Airlines to transport cargo operator, has been assigned responsibility for bringing into Ethiopia. As such, even in in the old system, ESL cargo from most ports of origin, via Djibouti, to the was the de facto default carrier from over 30 destina- dry ports and temporary storage facilities. There are tions. To overcome capacity constraints, ESL often currently no competitors for this single contract ser- contracted service to other shipping lines and charged vice offered by ESLSE and mandated for use through a mark-up per container on imports. This is still the the Ministry’s Directive (IFC, 2012). Experiences case, even though the recent restructuring increased from other countries show, however, that the gradual the fleet size and reduced the mark-up. A similar sys- creation of competition in markets is key for economic tem, which increases trade cost, also impacts air cargo on the Ethiopian Airlines side. The multimodal scheme is operated solely 46 Previously, the Ethiopian Shipping Lines Share Company (ESLSC), the Maritime and Transit and Services Enterprise (MTSE) as well as the by the Ethiopian Shipping and Logistics Services Dry Port Service Enterprise (DPSE) were the principal service providers Enterprise (ESLSE), which is facing substantial in the maritime sub-sector. Trade Logistics 57 Box 4: State Capitalism in East Asia – Bringing Competition to Markets The rise of many of the East Asian countries has led to a new model of capitalism that is often labeled “state capitalism” (Economist, 2012). The development evolution from (often) planned economies to market economies that was largely gradual in East Asia brought about a plethora of combinations between the two theoretical extremes of pure planning and pure market control (Pyle, 1997; Feltenstein et al., 1998; De Brauw et al., 2002; and So, 2003). “State capitalism” is not a new concept, however, and neither is it confined to experiences in East Asia. In fact, various forms of “state capitalism” have existed ever since capitalism came into existence. In reality every rising power has relied on the state to kick start growth, or at least to protect fragile industries. Even Britain, the crucible of free-trade thinking, created a giant national champion in the form of the East India Company (Economist, 2012). East Asia’s new “state capitalism” has its origins in Singapore. “Singapore’s experience illustrates an approach to economic planning which admits possibilities other than just “the market” or “the plan”, and shows that this is not a polarized debate” (Huff, 1995). In other words, Singapore followed both a strategy of opening up to market forces and pursuing government- led industrial policies. This created a system where state owned enterprises operated under market forces and hence could realize profits and losses; it is noteworthy that the model implied that unsuccessful companies on the market would be dissolved (Huff, 1995; and Chia, 2005). China’s reforms closely followed the Singapore model, which was built on the ideas of establishing competitive forces in the market, bringing foreign companies into the country, and using special economic zones to take advantage of the then new phenomenon of globalization. With this approach China embraced “corporatism” that required state-owned enterprises to be modeled and managed according to private corporation standards and to operate in a market environment. Over the three-decade reform period, Chinese SOEs started to face intense competition and efficiency pressures. Contributing factors were not only the rise of the growing domestic private sector, but to a large extent the opening up and invitation to foreign-owned enterprises to enter the Chinese market and compete. The gradual approach was preserved, and opening up after the WTO accession was phased. While key sectors gradually transformed, the trend of introducing more competition in the Chinese market remained one of the key success factors over the reform period (Ralston et al., 2006; and Economist, 2012). development; in fact, competition is so important that the globe, and Ethiopia has relevant experience with the actual ownership—domestic or foreign,47 state- management contracts from other sectors, for instance owned or private—may not matter (see Box 4 on East in telecoms, that could be applied. ESLSE’s current Asia’s experience with state capitalism that introduced considerations for a management contract with a for- competition in former monopoly markets). eign operator are therefore a welcome development. While introducing more competition is rel- (3) The lack of competition in inland transportation evant across the MTS, implementation could is probably the most pressing issue in the current situ- focus around three relevant subsystems. (1) For ation, being where the majority of delays and costs international shipping into Djibouti some lessons arise. Allowing additional operators to take part in the learned from air transportation and the management MTS, described in the preceding paragraph, could of Ethiopian Airlines. Ethiopian Airlines is state help alleviate the pressing challenges in this sub-sector. owned, but operates on a fully commercial basis and A limited understanding of, and the capacity faces competition from international airlines that of the Djibouti port and the dry port(s) to effi- serve Addis Ababa—strongest in the passenger seg- ciently work, the multimodal system aggravate ment, but also in cargo shipments. This is in contrast to international shipping into Djibouti, where choices 47 The Investment Law and the Investment Incentives and Investment Areas are more constrained, largely due to strict regula- Regulation, both of 2012, provide a general framework for investments tions that require ESLSE as operator for government in Ethiopia. While the trade logistics area is not specified as one of the permissible areas for foreign investors, the regulation states that foreign procured imports. (2) Port and terminal services are investors may be allowed to invest in areas other than the ones specified being managed by professional companies around (paragraph 4.2). 58 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status implementation limitations. Congestion problems add to the complexity of the situation. One recent and consignment tracking complications at dry ports example of such a coordination issue is the case of the are among the most serious issues in implement- Ethiopian Maritime Affairs Authority (EMAA), which ing the MTS. Towards the end of December 2012, prepared a directive, for the approval by the Ministry ESLSE made repeated requests to traders to collect of Trade, to penalize traders that do not collect their their containers on time. Non-collection is an increas- items within 45 days. However, the recent offer ing problem, and possibly related to issues such by ESLSE for traders to collect their goods for free does as tracking problems, financial constraints, and the not tally with this tendency. Hence, no one knows incentives of traders to use dry ports as a means which directive is valid at a given point in time. This of storage. Another important complication is the creates confusion among traders and their relationship failure to track consignments. In fact, the congestion with the various logistics agencies. It also indicates the problem developed in such a serious way that ESLSE potential arbitrariness of handling the trade logistics decided to exempt its clients from paying demurrage problem and coordination failure among ESLSE, costs, including provision of credit and containers, EMAA, and MoT. for those with severe financial and transportation problems. Waiving fees and providing credit for Lack of physical infrastructure importers of containerized cargo is only temporary, and not a long-term solution to the logistics problem. As discussed in Chapter 1, the Government is mak- In a similar vein, the build-up of emergency storage ing substantial progress in providing physical infra- sites cannot be more than a temporary solution. structure. The area of trade logistics is an illustrative Trade logistics in Ethiopia are complicated example of the need for this public investment, par- by a plethora of actors. There are, for instance, ticularly in roads and railways. Many initiatives are over 25 stakeholders involved in the trade logistics under way but could need a final push for completion. supply chain in Ethiopia (IFC, 2012). The large For instance the road from Addis Ababa to Nazareth number of stakeholders leads to confusion about is almost completed. This route will be useful to reduce who deals with what and poses a coordination the congestion at Modjo dry port which is very close challenge. It may not be apparent for traders who to Nazareth. Another project underway is the rail- to approach and what procedures to follow, since way line on the Djibouti-Dire Dawa corridor. Now there are frequent and unpredictable changes in rules it is under renovation and this should be fast tracked and regulations. For instance, inspection is frequent to circumvent the current congestion problem. For and highly susceptible to rent seeking. Valuation high volume goods, rail transport can be low cost of imported cargo is contentious.48 Often inspection and more effective than land transport, and could is handled by more than one agency, and this obvi- be given similar priority as road construction proj- ously contributes to delays. The lack of coordination ects, especially to improve transport from Djibouti in inspection is also due to non-use of risk manage- to Addis Ababa. ment for border clearance. While ERCA use some risk management for clearance, other border control 48 While using the “transaction value” is recommended by the WTO—the agencies do not. As such, this leads to agencies making agreement stipulates that customs valuation shall, except in specified independent control decisions, including for inspec- circumstances, be based on the actual price of the goods to be valued, which is generally shown on the invoice—Ethiopia is not a member tions. Simultaneous or joint inspections could help of WTO. As such, Ethiopia is under no obligation to use “transaction overcome this problem. value”, but the proposed updated customs code is supposed to conform to international best practice. Moving swiftly to implementation of the Coordination issues and ad hoc implementa- newly proposed code would help overcome inconsistencies in the current tion of regulations to solve the current challenges valuation approach. Trade Logistics 59 Lack of sufficient road connectivity across the and state-owned, than the required weight of goods border is a particular concern for cross-border to be lifted. Table 9 quantifies the excess demand for transportation. A specific concern is the situation transporting trucks currently at Djibouti port. The of the roads from Addis Ababa to Djibouti. There are available number of trucks range from 800 to 1,500, two roads to Djibouti: Addis Ababa-Awash-Djibouti significantly below the total required number of vehi- and Addis Ababa-Dredawa-Dewole-Djibouti. The cles (i.e., 13,055) for transporting the existing and first route is fully paved. However, on the second expected bulky cargo.49 one, the part that connects Dredawa to Djibouti via The majority of heavy imports are destined Dewole is unpaved and in need of upgrading. This to the various infrastructure projects of the govern- results in hazardous conditions on this major route, ment. However, the existing trucks that are old and especially for the bulky project imports that are light in weight do not match the heavy and bulky transported. Furthering joint investment planning cargo. As a consequence the existing trucks are over- between Ethiopia and Djibouti could help to address loaded, which leads to the destruction of roads and these issues in developing cross-border road links. bridges. The solution is, in addition to strengthen- Additionally, the government could consider open- ing enforcement of axle load limits, to upgrade the ing up new routes with potential to reduce the dis- capacity of the road system and bridges; but more tance to Djibouti (e.g., Meison-Djibouti). Opening importantly to facilitate the expansion and construc- this route could reduce the distance from Addis tion of a functioning railway system, to provide to Djibouti by more than 100km. an alternative to road transport. In the meantime, the The existing railway network between Ethiopia government could introduce initiatives to increase the and Djibouti is outdated and slow. Narrow tracks number of trucks with the appropriate specifications and steep grades accommodate only small trains, required for both the cargo and the roads. which slow down transportation times on the railway. While this is partly attributed to the extremely rugged One Stop Border Posts topography of the line, the situation has been aggra- The concept of OSBPs between two countries vated through years of negligence. Poor maintenance is linked to developing procedures that allow all exit led to the deteriorating tracks that to a large extent can and entry formalities to be performed in sequential only be replaced but not be repaired anymore. To add order in one facility, enabling clients to stop once to these challenges, the axle load of the railway line at the border in either direction. The concept entails is limited and a number of bridges need rehabilitation the performance of border clearance extraterritorially, (Ranganathan and Foster, 2011; Aschenaki, 2004). through the creation of a designated common control The inadequate quantity, capacity, and quality zone where border agencies share facilities with the of trucks, particularly at Djibouti port, are a key aim of eliminating duplication of procedures. The challenge: There is just a fraction of the trucks essence of establishing OSBP facilities at borders needed available. The structure of Ethiopian imports is to streamline border agencies work, enhancing has changed over time, from finished and light weight effective agency collaboration, sharing of facilities, products to a combination of light finished prod- and generally putting in place mechanisms for joint ucts and project-bound bulky cargo. This structural change has not been accommodated by a change in the 49 The finding of a lack of trucks is in contrast to statistics of the Ethio- design of the logistics system, however. The upgrad- pian Road Transport Authority that show that overall there is a sufficient ing of the road network needs to go hand in hand number of trucks available in the country (more than 30,000 vehicles for the transport of dry cargo of various sizes). If so, the heart of the problem with the acquisition of trucks of the right length and may well be in the management and deployment of the existing fleet, quality. There are much fewer trucks, both private rather than the number of trucks and vehicles. 60 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Table 9: Demand for Trucks for Recent Imports of Bulky Cargo, May 2013 Bulky Cargo Being Unloaded Type of Cargo Quantity Importer Demand for Trucks Coal 64,833,335 Ethiopian Petroleum Enterprise 1,621 Steel 94,618,406 Various importers 2,365 Fertilizer 16,802,000 Agricultural Input Supply Corporation 420 Petroleum coke 2,509,000 NORC AGRO & IND 63 Petroleum coke 6,265,000 NOC 157 Wheat 38,867,900 World Food Program 972 Wheat 62,629,400 Ethiopian Grain Trade Enterprise 1,566 Wheat 36,869,000 Ministry of Agriculture 922 Wheat 156,000 Catholic Relief Society 4 Total 323,550,041 8,090 Bulky Cargo on vessels that arrived and are queuing to be unloaded Wheat 48,600,000 Ethiopian Grain Trade Enterprise 1,215 Bulky Cargo on vessels soon to arrive at Djibouti port with expected date in parentheses Fertilizer (6 to 9 May) 50,000,000 Agricultural Input Supply Corporation 1,250 Wheat (May 7) 33,000,000 World Food Program 825 Fertilizer (May 12) 13,500,000 Agricultural Input Supply Corporation 337 Fertilizer (May 17) 13,500,000 Agricultural Input Supply Corporation 338 Wheat 40,000,000 Ethiopian Grain Trade Enterprise 1,000 Grand Total 522,150,041 13,055 Note: As of May 7 2013, there were 9,306 containers at Djibouti port out of which 7,059 weighed 20 ton each while the remaining 2,247 weighed 40 ton each. These require about 5,747 trucks that can load 40 ton each given the maximum load allowed by law in Ethiopian roads. customs controls/clearance and inspections, and stop- projects is that they are driven by local stakeholders ping only once when crossing into a partner state and through the National OSBP Committee and Steering once when exiting. Committees. These are elements Ethiopia could Several African countries are making remark- potentially replicate to initiate a bilateral agreement able progress in building One Stop Border Posts with Djibouti and other potential transit neighbors. (OSBPs) to facilitate trade intra-Africa and with the rest of the world. Progress is facilitated by member- Challenges of being landlocked ship of a regional integration bloc with an operational free trade area (FTA). Tanzania agreed to implement Ethiopia is a landlocked country and therefore the OSBP facilities in its borders with Kenya, Uganda, needs to critically rely on trade and border logis- and Burundi, within the framework of the East African tics to develop a thriving and diverse export sector. Cooperation (EAC), and with Zambia on the basis A landlocked economy faces a number of specific chal- of a bilateral agreement. OSBPs reduce trade costs and lenges. The most important ones include higher trade transit times for persons and goods by at least 30 per- costs, increased delays due to extended supply chains, cent. One of the unique features of the Tanzanian expensive and unreliable access corridors with multiple Trade Logistics 61 clearances, vulnerability to instability of their transit In fact, Ethiopia is the only landlocked African coun- neighbors, and cumbersome administrative practices try that depends on one single corridor (Arvis et al, (e.g., high transit and custom charges). Interestingly, 2011). This provides unnecessary risk to the trading these challenges are related more to operational companies, domestic and foreign, and adds to the inefficiencies due to cross-border relationships than overall impediments of engaging into trade business to infrastructure capacity per se (Arvis, et. al. 2011). out of Ethiopia. Table 10 details some of the key challenges for LLDCs in their extended supply chain. Customs, Border Control and other Trade The cost of being landlocked is exacerbated Related Regulations because goods have to be shipped through transit countries. Higher transaction costs are generated The other key components of Ethiopia’s trade logis- because importing countries (e.g., Ethiopia) are at the tics problem are linked to customs, border control mercy of the administrative practices and politi- and other trade related regulations. New regulations cal stability of the transit country (e.g., Djibouti). should be considered to increase transparency and help The associated fees include the costs for documents, facilitate custom processes. Where new regulations administrative fees for customs clearance and techni- are drafted (e.g., in the case of the customs code), cal control, terminal handling charges, and inland implementation of the new law is critical as it would transport. Not surprisingly, export costs and import allow for many trade logistics procedures and prac- are ranked among the highest for most landlocked tices to be modernized. Trade transactions are getting countries, as previously illustrated in Figure 16. increasingly difficult to manage due to border related For Ethiopia, relying only on one trade cor- costs (e.g., policies, language, currency, information, ridor makes the management of the political security, working time differences between Djibouti economy of logistics particularly vulnerable to the and Ethiopia), and logistics costs (e.g., transport time, relationship with the partner country Djibouti. freight transport, and pre-shipment inspection). Table 10: Supply Chain Sequence and Bottlenecks, LLDCs Step Participants Typical Issues Unloading; declaration and initia- Port authority, terminal operator, Excessive time to clear transit, sometimes longer than for tion of transit forwarder, customs local clearance, cumbersome transit declaration Loading on truck Forwarder, trucking company, Overregulation of trucking industry, Inadequate market handling company structure, Formal or informal queuing system Loading on multimodal facility Terminal operator, transport Lack of coordination between port and transport opera- or railway company tors, Inefficient transport operators and long lead time Control en route Road Agencies; customs Convoys, Multiplication of checkpoints, formal or not, and payments; Weak link or congestion on the corridor Border crossing Forwarder’s agent or broker, Duplication of controls on each side of the border, Wait- Customs, Other border agencies, ing time, Inconsistency of procedures, Fragmentation Road agencies of brokerage services across borders Transit in destination country and Forwarder, Customs No continuity of procedures or portability of documents, final clearance Improper use of ICT for transit, Lack of capacity (IT risk management) for proper clearance at destination, Inef- ficient discharge of transit increases the cost of guarantee Source: Adapted from Arvis et al (2011) Table 2.2, page 19. 62 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Box 5: Informal Cross Border Trade (ICBT) Informal Cross Border Trade (ICBT) is legal, but unregulated or informal trade. ICBT is the least researched area, mainly due to the lack of reliable data. Surveys suggest that informal cross border trade represents a significant proportion of regional cross border trade in Africa. Ethiopia engages in informal trading with neighboring countries such as Sudan, Somalia, and Kenya, involving the exchange of goods such as live animals, oilseeds, and small manufactured items. Women play a key role in informal trade. For example, it is estimated that 70 percent of informal cross-border trade in southern Africa is by women (UNIFEM, 2009; Afrika and Ajumbo, 2012). ICBT is mainly conducted by individual traders and micro, small, and medium-sized enterprises, and often consists of small consignments (Lesser and Moise-Leeman, 2009). Despite the evident contribution of women in informal cross-border trade to the economy, mainstream trade policies and institutions tend to neglect them (Masinjila, 2010; UNIFEM, 2010; Ndaiye, 2010). The implication of ignoring this segment of the trading population is significant. It means that policy decisions are being made in response to incomplete data and information, and those informal businesses may be “missed” in policy decisions and not offered the support they need to expand. Ethiopia is a major participant in ICBT. The country loses revenue given the high incidence of ICBT. COMESA (of which Ethiopia is a member) estimates that US$ 35 million worth of goods are traded informally annually. This is in contrast to the limited formal trading which is US$ 19.2 million per year (Njiwa et al., 2010). A large number of cross border traders resort to ICBT due to various fees and regulations in formal crossings. Ethiopia is a source of non-processed tradable goods for Kenya and South Sudan. In Eastern Africa, Ethiopia and all other countries in the region (except South Sudan) are also involved in re-exports. By learning from other countries, Ethiopia can lower the incidence of ICBT by reducing documentary and other trade logistics requirements in the formal sector. A cross-border case study between Kenya and Tanzania indicates that the documentary requirements in the formal sector push traders to go informal (e.g., case of the Arusha-Namanga Border). Encouraging informal traders to come to the formal route by addressing the trade logistics problems of the formal sector is essential because ICBT can have detrimental consequences on health and safety of consumers (e.g., through smuggling expired food and medical supplies via informal crossings without checks). Therefore, Ethiopia can improve border infrastructure such as cross border storage facilities, stalls in key border markets, and check points, and also prioritize a gender based approach to ICBT to address some of the problems that affect women (Afrika and Ajumbo, 2012). A starting point for addressing the issue could be introducing the COMESA simplified trader regime (STR) for small traders (see COMESA STR Website). There are up to 103 procedures and docu- can contribute to a formalization of trade, with all its ment requirements at different stages of the chain positive implications (see Box 5). to obtain border clearance. These are illustrated in the Customs clearance is an important part Annex to this chapter.50 There are also many official of the issues observed in excessive documenta- fees due to the number of agencies involved. For tion and complicated trade logistics. There are instance, although the trader or the clearing agent can at times 21 documents required for custom clearance, submit the custom declaration electronically through which is a huge transaction cost for the business com- ASYCUDA++51 in a pilot initiative, the Ethiopian munity (IFC, 2012). ERCA is clearly aware of issues Revenue and Customs Authority (ERCA) requires in the customs area with a recent reform study,52 which its submission in hard copy along with supporting is important to effectively address shortcomings. For documents. This requirement undermines the benefits of electronic automation. In fact, ERCA is preparing to invest significant resources to automate trade transac- 50 The illustrative process map reflects both observed procedures (final clearance and technical control) and procedures recorded on the basis tions. Two large projects are planned: On the one hand, of interviews in April 2012. Some of the steps may have changed an upgrade of the existing customs processing system; since then. 51 The Automated System for Customs Data is a computerized system and on the other hand the introduction of an electronic designed by the United Nations Conference on Trade and Development Single Window. Successful implementation of these (UNCTAD) to administer a country’s customs. 52 ERCA (2013) A study about improving/reforming custom clearance projects is likely to impact the trade transactions sys- procedures, Ethiopian Revenue and Customs Authority (in Amharic), tem positively. Furthermore, simplification of processes Addis Ababa, Ethiopia. Trade Logistics 63 instance, the clearance of imported goods proceeds Figure 19: Moral, Professionalism and after completion of transit, and this leads to exces- Competence of ERCA Staff sive storage of goods in warehouses. Also, not allow- 60 ing subsequent imports without finishing clearance 52 50 48 of previous imported items severely limits the activi- 45 41 40 40 ties of importers, especially productive importers that 37 contribute to the growth of the economy. According 30 26 27 to ERCA, many stakeholders in the import clearance 23 20 16 system (i.e., traders, as well as public officials) are 15 11 found to be ill disciplined and in fact prone to the 10 temptations of rent-seeking/corruption53 (Figure 19). 0 Corruption Staff Accountability Genuine and The agency’s reform study also indicates that there Competence Polite Staff is lack of staff with sufficient experience in custom pro- Yes Partially No cedures, and those who are in office have a low capac- Source: ERCA’s Customer Satisfaction Survey 2012/13. Note: the ity for decision making. Often, ERCA’s senior staff bars indicate the percentage of survey respondents who respond members pass matters to their superiors even if the to the question. For instance, 41 percent of respondents indicated that ERCA staff are genuine and polite. issue at hand is trivial, and this problem is prevalent at each level of the decision-making hierarchy. The lack of decision-making initiative is even more prevalent in the Djibouti custom office, which leads to delays ERCA has given special attention to custom in releasing imported items that have completed all clearance for importers and exporters that are the required documentation and payments. working on areas of industrial and economic The trade supply chain is burdened with the development. There are a number of planned initia- prevalence of hardcopy documents for informa- tives for this specific group of customers. (1) They tion exchange both between the private sector will be served in separate “custom clearance units.” and government actors and between government (2) They will get clearance prior to the arrival of the actors themselves. Overall, there is no legal frame- goods (i.e., pre-arrival clearance). However, there work in place to recognize documents exchanged should be some caution depending on the nature electronically in relation to e-commerce, e-signatures, of imports (e.g., chemical products). (3) They will and e-payments. In fact, current Ethiopian laws benefit from the use of transaction value to value their require paper receipts to be issued for all transactions, imports. If there are valuation disagreements, they although the Government is reviewing legislation will be resolved by letter guarantees instead of hold- and assessing, among other things, the recognition ing goods as hostage; (4) They can have their items of electronic format. Most stakeholders in the trade inspected at the gate of the factories/establishments; transaction process also lack automation systems, and (5) They will be given full custom clearance on the where they exist (Customs, ESLSE) their implemen- industrial zones and at factor sites; (6) They will ben- tation levels could be further improved. Documents, efit from a “Simplified Clearance” procedure; and forms, laws, and regulations are not always available (7) Though it is risky, they will benefit from differed online while data sharing electronically is completely payment. The special treatment for strategic sector lacking between agencies. Modernization efforts are producers may be appropriate in the context of the underway, as ERCA plans to upgrade its customs processing system and implement an electronic Single 53 Recent reports about ERCA officials being accused of corruption al- Window (eSW). legations in May 2013 support this point. 64 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status developmental state economic strategy pursued by the sites, dry ports, and any other location identified and Government. However, it is not clear how the import- approved by ERCA. Having different sites is one of the ers and exporters are selected. Other institutions can many measures already taken by ERCA to improve the also enjoy some of the privileges listed above. For transit system in Ethiopia. ERCA also designed a sys- instance, any government entity, other organizations tem whereby items are processed only with insurance with a permit for a “Simplified Clearance” procedure, guarantees, without the requirement of tax payment, and users of the multimodal scheme are allowed and reduced the number of inspection posts along the to benefit from pre-arrival clearance. transit route. Follow-up and implementation is needed While ERCA has accepted risk management for plans to reduce or eliminate road blocks. Equally as a key element of its strategy and a structure has some road checkpoints could be allowed to work been put in place to enable its adoption, implemen- for 24 hours per day in order to speed up the movement tation of risk management still needs considerable of cargo from Djibouti to dry ports. Based on the Ethio- strengthening. Other border control agencies are Djibouti transit protocol agreement, ERCA designed still in infancy in their understanding or use of risk a separate transit arrangement for goods that are trans- management as a tool to facilitate and strengthen ported via the multimodal as well as the unimodal their operations. As such, cargo undergoes multiple schemes. There is a plan to conduct joint inspection inspections by different agencies, at different points of goods at the Djibouti-Ethiopia border, with reduced of time, through the supply chain. This has a bearing frequency and without a need for repeated inspection on inspection levels and on the speed, cost, and effi- both by Djibouti and Ethiopian authorities. The imple- ciency of clearance of cargo—thereby slowing down mentation of this plan depends on a joint agreement the trade logistics system. The impact on both the between the two countries. business community and on government resources Inspection of goods in warehouses is not sup- is costly and detracts from the need to facilitate ported by appropriate technology. The warehouses legitimate trade while concentrating enforcement do not have the necessary modern facilities that sup- resources on non-compliant businesses. Scaling up the port efficient service provision. Further, the working usage of “trusted” economic operators, which appears procedures in place in each warehouse are not appro- to be part of the currently discussed yet not yet passed priate to the type of warehouse. ERCA believes that new customs law, could potentially help better the there are no competent warehouse operators of the current situation in the short-run. scale required. On the other hand, inspection capaci- ties are also very low (ERCA 2013). Inspection institu- Efficient Logistics and Distribution Services tions often fail to assign the right personnel to match the required inspection activity in warehouses. And Encouraging measures have been taken by ERCA, but in the most extreme cases, inspection authorities either there are outstanding challenges related to incom- send inspectors with long delays or fail to send even plete transit activities, and insufficient alignment one inspector. There is also concern about the lack to international working procedures. In addition, they of acceptable floor surface (e.g., for food items and are not supported by modern technology. There are four medicine) when goods are unstuffed and inspected transit types: inward transit, through transit, interior out of containers in dusty dry ports such as Modjo transit, and outward transit. ERCA designed separate and Bekelcha. There are four steps and documentary procedures for the four transit types, with a reduced requirements to put goods in warehouses and five number of inspection posts. Ethiopia has a number steps and documentary requirements to get goods out of transit completion sites such as custom posts, per- of warehouses, all of which contribute to delays. The mitted warehouses, industrial zone(s), permitted project fundamental recent changes in warehousing are related Trade Logistics 65 to allocation of goods to different types of warehouses, and procedures, but this could be rectified if commit- and designing a different system of airport warehous- ted efforts were to be made in the short run. Transit ing of goods in recognition of the unique feature operators fail to provide information about custom of air transport. declaration, and do not respond to questionnaires sent The lack of technology or investment in ware- to them by ERCA. Operators responsible for check- housing and related logistics facilities is also due ing the weight of vehicles are often slow, and delay to the lack of adequate private participation in this the movement of goods. Informal evidence suggests space. For instance, in neighboring countries such that there is also widespread corruption, allowing the as Kenya there has been increasing investment in value movement of vehicles loaded beyond the legally per- added logistics services—that are steadily improving mitted cargo weight. the state of logistics. Taking advantages of the positive Limited dialogue between government agen- and innovative forces of broad private participation cies involved in trade logistics and private actors in the logistics arena of Ethiopia could enhance the dampens the prospects for rapid diagnosis of prob- system at large. lems or development of solutions to issues at hand There are particular problems associated with for the trading community. While the Government establishing efficient linkages of other public introduced some discussion fora,55 a lack of clarity and agencies working with ERCA and also challenges predictability on rules, regulations, and policy changes in relation to operations. These include problems is still prevalent in the trading community. Increasing with freight forwarders, warehouse operators, transit substantial dialogue between the parties involved operators, inspectors, and organizations responsible could lead to great advances. In relation to open and for granting permits. Freight forwarders often fail candid engagement with the private sector, it is impor- to disclose the required information about items tant for ERCA to assess the level of satisfaction of its and passengers (i.e., “cargo passenger manifest”). clients with the service it is providing. In line with this Warehouse operators perform their duties without approach, ERCA conducted a customer satisfaction the support of modern technology. They lack equip- survey from July 2012 to January 2013. “Develop the ment for moving goods (e.g., high-tech cranes) and Tax Payers Awareness and Customer Care Attitude of the hence delay movement of goods and containers. Here Staff” is one of the priorities listed under the “Way there is scope for foreign or local private participants forward” agenda of ERCA. to provide the necessary warehousing equipment and devices for moving goods. There is insufficient automated and technologi- 54 ERCA (2013). Customer Satisfaction Survey report, 2nd draft (unpub- lished), Addis Ababa, Ethiopia. cally supported exchange of information among 55 Current fora for dialogue include: (1) The Ethiopia Public Private the various agencies and duplication is common. Consultative Forum (EPPCF), chaired by the Ministry of Trade in col- laboration with the Chamber of Commerce. This initiative started The recent customer satisfaction survey report, pro- in February 2012 and is established to allow the private sector to engage duced by ERCA,54 shows that there is an absence with the government on urgent trade logistics issues. According to the Government, the forum since then has been institutionalized on a regular of a service provision mentality or proper customer schedule. (2) The Customs and Logistics Subcommittee, chaired by the handling spirit. An example of unnecessary duplica- Director General of ERCA. The committee, set up two years ago, deals with customs and logistic issues and is empowered to pass decisions tion of efforts is linked with the inspection of one item to improve the situation in these areas. There is a steering committee by more than one agency. Transit operators generally (SC) and a technical committee under this mechanism. The SC consists of the following members: Ministers/State Ministers of Finance and lack knowledge about customs rules and procedures, Economic Development, Industry, Transport, Trade, and Agriculture. and fail to settle the required payment with agents The technical committee includes members from priority industry sectors like textiles, leather, and others. In-depth reviews and evaluations may in Djibouti. To make matters worse, ERCA itself fails be required to assess the effectiveness of the two fora so to enhance the to inform the trading community of the current rules future dialogue through these channels. 66 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Policy Recommendations by multiple banks. Enhanced competition should be supported by measures to increase the capacity The following policy recommendations emerge of key players and their understanding of the mul- from the analysis: timodal system; this applies in particular to the management and operation of Djibouti port and 1. Develop a national trade logistics strategy. the dry port(s). On a general level, there is an urgent need for the 3. Address the infrastructure challenges: roads, Government to develop such a strategy to facilitate trucks, and railways. The Government is keenly all efforts in the area of trade logistics. This study aware of physical infrastructure challenges, has shown the complexity of the current trade and is making good progress in this area. The system in Ethiopia, the numerous actors involved, nearly completed Addis Ababa—Nazareth road manual processing systems, lack of clarity on regu- is expected to reduce congestion at the Modjo dry lations, and a plethora of other interdependent port. An upgrade of the Dire Dawa—Dewle road factors. In this context, the lack of a national to a higher standard could reduce the transport trade logistics strategy to provide an overarching distance by about 100km compared to current framework is acutely felt by all stakeholders. The alternatives. The acquisition of specialized trucks Government is therefore encouraged to prioritize for bulky imports is also a priority identified in the the development of such a strategy, work on which report. Finally, it should be emphasized that for has already started in collaboration with UNDP. some goods railway is a more cost effective means 2. Strengthen the multimodal system by encour- of transport. In light hereof, the Government aging increased competition. The efficiency should give this transport option due attention. of the multimodal transport system could In fact, the Government has already put due be enhanced by encouraging the participation emphasis on railway development, with the cur- of other operators, including the private sec- rent plan of building a new link between Addis tor. More competition and additional operators Ababa and Djibouti. Other railway projects are in a market-based system are needed along the in the pipeline. Still, some fast-tracking of current whole multimodal transport system, but imple- renovation work, for instance on the Djibouti- mentation could focus on relevant subsystems Dire Dawa railway line could reduce congestion such as international shipping into Djibouti, port in the short-run. and terminal services, and inland transportation. 4. Develop multiple trade corridors. Potential In fact, the lack of competition is felt most acutely alternatives to the current exclusive reliance in the latter, where the majority of delays and on Djibouti could be ports such as Berbera, costs arise. To illustrate, the existing private sector Mombasa, Port Sudan, and Lamu. Detailed fea- operators are not providing the trucks required for sibility studies would be needed, however. Many import cargo. With enhanced competition, indi- other landlocked countries have more than one vidual actors would have better incentives to make exit route for their trade. To illustrate, Uganda, such investments. There are several examples Rwanda, Burundi, and Zambia have 2–4 transit in Ethiopia, where the introduction of increased neighbors, which lead to the use of multiple trade competition has improved sector performance. corridors linked to ports. Multiple corridors would For instance, the introduction of private banks diversify Ethiopia’s options and thus improve its in Ethiopia improved the performance of the negotiating power with transit corridor(s). They government owned banks, due to the competitive would also introduce competition among transit spirit generated to provide good financial services neighbors, which could keep costs down; it would Trade Logistics 67 also provide some level of insulation against social, actual content. For instance, import valuation political, and economic ups and downs of tran- is a fairly complex issue, and training for ERCA sit neighbors. Caveats to using multiple corri- is best provided through international expertise. dors may relate to multiple infrastructure costs Support could possibly be mobilized through and diseconomies of scale in logistics services. International Organizations like the World Bank The involvement of PPPs would help address Group, WCO, and through visits to best practice these caveats. countries, as efforts intensify to promote South- 5. Reform customs procedures with new tech- South exchange. Other options include using the nologies and Public Private Partnerships. private sector for training purposes. The private The proposed ERCA custom clearance reform, sector could for instance provide training to ERCA including the implementation of an electronic staff at low cost on other key issues of concern such Single Window, is key in this respect. These as customer care and handling. There should also modernization efforts need to be coordinated be effective legal instruments to make corrupt staff nationally to ensure that the government adopts accountable for their actions. a coherent e-government strategy as part of the 8. Equip warehouses with modern technology and larger national logistics strategy agenda. For ICT build more terminals. It is necessary to buy and transformation of ERCA, private public partner- provide more port and warehouse management ships can be vital by introducing improvements equipment. One of the problems in Djibouti and in the custom clearance, and by learning from hinterland dry ports (e.g., Modjo and Kaliti) is the other African countries. Ghana, for instance, lack of high tech cranes. There are only a few tech- made effective use of PPP initiatives for the trans- nically advanced cranes and other logistics equip- formation of its custom authority (Calvin 2013). ment in the country. If one or more fails to work, Before considering the implementation of PPPs, excessive delays can occur. Increasing private par- however, it is important to review all concerned ticipation in the area of warehousing and related business processes, in order to be able to map out logistics facilities would support modernization clear requirements to address the issues at hand. and the introduction of innovative service prod- 6. Improve transparency and the predictability ucts and management techniques. There are not of regulations and communicate better with enough terminals in Modjo. The three terminals the business community. There is a need for under construction need to be finalized before improvements in administrative procedures the start of the rainy season to avoid damage of ERCA and the manner in which changes and further delays in logistics. The same is true in rules and regulations are communicated to the to accommodating loading trucks in Djibouti that business community in Ethiopia. Arbitrary and are fully packed in a dusty terminal. ad hoc changes in rules and procedures, lack 9. Implement sound risk-based control systems of predictability, and lack of transparency should at the agency and inter-agency level to facili- not be norms. This recommendation applies also tate trade and ensure effective safety controls to the various directives coming from the Ministry with optimal levels of inspections (including of Transport (MoT). Furthermore, it is feasible joint inspections). Inspection is frequent and to make use of the already existing mobile phone highly susceptible to rent seeking. Valuation technology to communicate with traders as a pos- of imported cargo is contentious. Often inspec- sible solution to tracking consignments. tion is handled by more than one agency and this 7. Train ERCA staff to improve their efficiency and obviously contributes to delays. The lack of coor- professionalism. Training options depend on the dination in inspection is also due to non-use 68 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Box 6: Policy Recommendations on Trade Logistics Policy Area Recommendation General 1. Develop a National Trade Logistics Strategy Structure and 2. Strengthen the multimodal system by encouraging increased competition performance of the 3. Address the infrastructure challenges: roads, trucks, and railways transport and logistics 4. Develop multiple trade corridors system Customs, border 5. Reform customs procedures with new technologies and Public Private Partnerships controls and other 6. Improve transparency, predictability of regulations and communicate better with the business trade-related community regulations Efficient logistics and 7. Train ERCA staff to improve efficiency and professionalism distribution services 8. Equip warehouses with modern technology and build more terminals 9. Implement sound, risk-based control systems, with optimal levels of inspections (including joint inspections) of risk management for border clearance. While and risk-based control systems (#9). In principle, ERCA use some risk management for clearance, introducing competition (#2) could also be rela- other border control agencies do not. As such, tively fiscally neutral, because higher revenues this leads to agencies making independent of another SOE or taxation of potential private control decisions including for inspections. operators could compensate for lower ESLSE Simultaneous or joint inspections could help revenues; PPPs could also play an important part overcome this problem. in this. A second set of recommendations would 10. Finally, when considering policy recommen- have some fiscal cost, including: reforming cus- dations, it is important to consider the fiscal toms procedures (#5), training ERCA staff (#7), cost of their implementation. Although no for- and equipping warehouses with modern technol- mal costing exercise was undertaken as a part ogy (#8). Finally, there are a couple policy options of the study, it is still possible to roughly classify with substantial cost implications: addressing policy recommendations in terms of their costs. infrastructure challenges (#3) and developing A good number of the recommendations listed multiple trade corridors (#4). Policy recommenda- in Box 6 would be relatively cheap to implement, tions with a fiscal cost would need to be consid- including: National Trade Logistics Strategy (#1), ered in the context of competing spending needs transparency and predictability of policy (#6), and priorities. Trade Logistics 69 Annex: Trade Logistics Figure 20: Process Map for Border Clearance in Ethiopia (Illustrative) Phase I: Application to Import (FMHACA Example) Certificate of quality and Importer advised. 2 safety from exporting End of process country No 1 5 Importer fills 4 Importer wishes to 6 imports product application and FMHACA officer Application regulated by submits with reviews application approved? FMHACA i.e. food supporting documents Yes Supporting documents submitted: FMHACA officer (i) Cover letter; (ii) Proforma Invoice; 7 approves 3 (iii) Insurance; (iv) Application for forex; PO/proforma invoice (v) Application for LC; (vi) Import License Phase II: Application for Forex and Letter of Credit Importer advised. End of process 9 No 10 8 Importer submits Bank refers to importers 11 registered account no. issued documents to the Application Bank officer reviews commercial bank for application and by National Bank if Ethiopia approved? (NBE) to ensure trader is not forex permit with checks forex position supporting documents delinquent 12 Yes Forex permit issued Supporting documents submitted: (i) Import License; (ii) Proforma Invoice; (iii) TIN Certificate; (iv) Application for 8 letter of credit; (v) Insurance document; (vi) Document confirming National Bank Account Number 13 15 16 ‘Chamberized Invoice’ Importer submits Bank officer reviews Application from supplier country documents to the application and confirms 17 approved? commercial bank for LC margin position 18 Supporting documents submitted: Yes No LC issued Importer advised. (i) Trade License; (ii) ‘Chamberized End of process 14 Invoice’ (approved by FMHACA); (iii) TIN No.; (iv) Application for LC; and (v) Insurance 70 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Phase IIIA: Vessel Arrival, Cargo Discharge and Temporary Storage Vessel arrives at port Djibouti control authorities Container Cargo unloaded of Djibouti and berths board ship to perform control from vessel and moved to 24 21 procedures Dolareh terminal 19 Parallel mode 23 Manifest validated 20 22 22 Shipping Agent receives Shipping Agent submits Containers stored in terminal manifest from Shipping manifest to Djibouti Customs pending clearance Line and Port Authorities (8 days w/o charge) Djibouti clearance process Phase IIIB: Release of Original Documents from the Bank Importer notified of Importer goes to commercial Bank officer releases bank and requests for release Bank officer issues ship arrival originals when received release order of original documents 26 27 28 13 Phase IIIC: Transit Clearance Process in Ethiopia Note: This process Declaration rejected and is described for Invoice, transportation documents, returned to agent commercial cargo insurance documents, forex at Kality Customs documents, permits and certificates, 30 House at Addis trade and other licenses No Ababa Clearing agent self-assess 35 Importer chooses Importer provides and fills up declaration at clearing agent and Kality OR premises in ASYCUDA Declaration documentation to validated signs authorization clearing agent ASYCUDA. Prints and signs validates data papers 3 copies 31 32 34 33 Yes Clearing agent Importer presents transit Bank issues Cash provides draft transit documents and makes Payment Order (CPO) declaration to importer payment to bank 36 37 38 39 Importer returns documentation including CPO to clearing agent Trade Logistics 71 Phase IIIC: Transit Clearance Process in Ethiopia (continued) Clearing agent presents 3 copies of Clearing agent Transit officer opens documents to cashier’s office payment receipt submits all documents manifest and creates 45 at transit shed at Kality received by agent to transit at shed dispatch order 44 41 42 (I) Draft Transit Declaration; (II) Transit (I) Transit Declaration; (II) Transit Application; (III) Bill of Lading; Application; (III) Bill of Lading; (IV) Commercial Invoice; (V) Packing List; 43 43 (IV) Commercial Invoice; (V) Packing List; (VI) Forex Application/LC; (VII) Payment (VI) Forex Application; (VII) Payment Request; (VIII) ERCA Certified Signature Request; (VIII) ERCA Certified Signature copy of FF; (IX) CPO copy of FF; (IX) Payment Receipts 49 Transit Officer stamps transit Clearing Agent couriers all 46 declaration and returns all necessary documents to documents Djibouti agent 52 (I) Original Transit (I) (Separate) Transit Declaration to Transit Officer emails Declaration; (II) Bill of Djibouti Customs; (II) Guarantee for dispatch orders to 48 Lading; (III) Packing List duty; (III) Original Transit Djibouti (IV) Commercial Invoice Declaration; (IV) Bill of Lading; (V) Packing List (VI) Commercial 47 Invoice Phase IV: Djibouti Clearance Process 54 Shipping Agent Djibouti Clearing Agent Djibouti Clearing Agent submits documents to Djibouti customs 50 makes payment to issues delivery order release issued (DO) Djibouti Customs who shipping agent conduct control activities 51 53 56 58 57 59 Djibouti Clearing Agent Djibouti Clearing Agent submits documents to Ethiopian Customs submits DO and other Djibouti Clearing Agent Ethiopian Customs at issues waybill documents to Port makes payment to Djibouti Authority/Dolareh terminal Port/terminal authority Freight Truck is loaded and Forwarder Gate pass 60 ready to leave for issued (I) Ethiopian Transit Declaration; organizes Addis Ababa (II) Original Transit Declaration; 55 transport (III) Bill of Lading; (IV) Packing 62 List; (V) Commercial Invoice 61 Note: This is not an account of the new intermodal process 72 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Phase V: Inland Transportation 63 65 66 68 70 Truck enters At checkpoint 1: At checkpoint 2: At checkpoint 3: At checkpoint 4: Ethiopia and stops Border control Border control Border control Border control at checkpoint 1 agencies inspect agencies inspect agencies inspect agencies inspect document and seals documents and seals documents and seals documents and seals (I) Transit Declaration; (I) Transit Declaration; (I) Transit Declaration; (I) Transit Declaration; (II) Bill of Lading; (II) Bill of Lading; (II) Bill of Lading; (II) Bill of Lading; (III) Packing List; (III) Packing List; (III) Packing List; (III) Packing List; (IV) Invoice (IV) Invoice (IV) Invoice (IV) Invoice 64 65 67 69 Phase VI: Truck Arrival at Comet Warehouse 72 73 74 75 Truck arrives at Customs Inspector Inspector stamps Transit Officer Reconciliation of Comet inspects seals and arrival seal on closes transit performed Warehouse documents documents manifest with Djibouti team 71 Cargo is unloaded from truck into Comet 76 warehouse Phase VII: Border Clearance at Kality Customs 78 79 80 Clearing Agent Clearing Agent brings documents receives documents to Facevet Customs Officer. Assessment from transporter and Documents listed are those notice issued importer submitted mandatorily/observed as being submitted Trade Logistics 73 Phase VII: Border Clearance at Kality Customs (continued) 84 85 Physical Inspection Inspection report Valuation assessed Red Officer performs Inspection No created. Penalties 100 and documents inspection and notes passed? and fees assessed returned to findings (on back of verification office declaration) Yes Either release order issued Documents OR serious cases referred to 99 transferred to Kality investigations Valuation Officers No 86 97 97 98 95 Yellow Document Scrutiny Documents Documents registered Officer places Officer checks Documents transferred to in ASYCUDA and query with agent? documents ok? No verification office officer assigned Yes Yes 87 No Technical Documents and payment slip Clearing Agent makes control Release order issued brought to verification office necessary payments required? 103 102 101 Yes 89 91 90 Importer goes back to Importer notified or arrival FMHACA for approval to FMHACA Officer Inspection of goods at Kality/Comet import and submit reviews application required supporting documents 88 Note: FMHACA technical control assumed Supporting documents submitted: (I) Airway FMHACA Inspector Bill/Bill of Lading; (II) Invoice; (III) Packing 92 checks cargo List; (IV) Application/Letter to request permission to import; (V) Certificate of Origin; and (VI) Certificate of Analysis No Cargo ok for Importer advised. 93 release? End of process Yes FMHACA Officer issues technical control release/certificate 94 References AEMFI (2012) Ethiopian Microfinance Institutions Carroll, Christopher, Judy Overland and David Performance Analysis Report, Bulletin 8, Addis Weil (2000) “Saving and Growth with Habit Ababa, Ethiopia. Formation,” American Economic Review, 90, 3, Afrika, J-G. and G. Ajumbo (2012) Informal Cross pp. 341–55. Border Trade in Africa: Implications and Policy Chamon, Marcos, Kai Liu and Eswar Prasad Recommendations, African Development Bank, (2011) The puzzle of China’s rising household Chief Economist Complex, Africa Economic Brief, saving rate, http://prasad.dyson.cornell.edu/ doc/ Vol. 3(10): 2012. ChinaSavingsVoxEU.pdf. Amha, W. (2012) The Development of Deposit-Taking Chaturvedi, Vaibhav, Brajesh Kumar and Ravindra Microfinance Institutions (MFIs) in Ethiopia: H. Dholakia (2009) “Inter-relationship Between Performance, Growth Challenges and the Way Economic Growth, Savings, and Inflation Forward, Occasional Paper no. 32, AEMFI, Addis in Asia,” Journal of International Economic Ababa, Ethiopia. Studies, 23, pp. 1–22. Amha, W. and T. Alemu (2012) Household Saving Commission on Growth and Development, 2008, Behaviors and saving Mobilization in Ethiopia, ‘The Growth Report – Strategies for Sustained mimeo, AEMFI, Addis Ababa, Ethiopia Growth and Inclusive Development’, World Aschenaki, B. (2004) Transport Costs in Ethiopia: Bank, Washington, DC. An impediment to Exports? A background Dayal-Gulati, Anuradha and Christian Thimann study for the World Bank’s FY04 Country (1997) “Saving in Southeast Asia and Latin America Economic Memorandum for Ethiopia, Addis Compared – Searching for Policy Lessons,” IMF Ababa, Ethiopia. Working Papers 97/110, International Monetary Balassa, Bela (1986). “Toward Renewed Economic Fund (Washington, DC). Growth in Latin America” Institute of International Edwards, Sebastian (1996) “Why Are Latin Economics (Washington, DC). America’s Saving Rates So Low? An International Burgess, R. and R. Pande (2005) Do Rural Banks Comparative Analysis,” Journal of Development Matter? Evidence from the Indian Social Banking Economics, 51(1):5–44. Experiment, American Economic Review, Vol. ERCA (2012/13) Customer Satisfaction Survey, 2nd 95(3): 780–795. draft, January 2012, Addis Ababa, Ethiopia. Calvin, M. (2013) ICTs for Regional Trade and Faruqee, Hamid and Aasim Husain (1998) “Saving Integration in Africa, Chapter 5, Assessing Trends in Southeast Asia: a Cross-Country Regional Integration in Africa (ARIA) VI report, Analysis,” Asian Economic Journal, 12(3): UNECA Addis Ababa, Ethiopia. 195–217. Carroll, Christopher and David Weil (1994) “Saving Ferrucci, Gianluigi and Cesar Miralles-Cabrera and Growth: A Reinterpretation,” Carnegie- (2007) “Saving Behaviour and Global Imbalances: Rochester Conference Series on Public Policy, The Role of Emerging Market Economies,” ECB 40, pp. 133–92. Working Papers No. 822, European Central Bank. 76 2nd Ethiopia Economic Update – Laying the Foundations for Achieving Middle Income Status Friedman, Milton (1957) “A Theory of Consumption Levine, Ross (2005) “Finance and Growth: Theory Function,” National Bureau of Economic and Evidence,” In Handbook of Economic Research (Cambridge, MA). Growth. Phillipe Aghion and Steven Durlauf, eds. Gelalcha, G. (2009) Public and Private Partnerships The Netherlands: Elsevier Science. for the Development of Infrastructure to Facilitate Levine, Ross, Norman Loayza, and Thorsten Beck Trade and Transport, Initiatives and mea- (2000) “Financial Intermediation and Growth: sures implemented in Ethiopia, Incentive Causality and Causes,” Journal of Monetary and Facilitation Team, Ministry of Trade and Economics, 46, pp. 31–77. Industry, Ethiopia. Loayza Norman, Klaus Schmidt-Hebbel, and Luis Hauk, William and Romain Wacziarg (2009) “A Servén (2000) “What Drives Private Saving Monte Carlo Study of Growth Regressions,” Across the World?” The Review of Economics and Journal of Economic Growth, 14(2), 103–147. Statistics, LXXXII, 2, pp. 165–181. Hevia, Constantino, Yuki Ikeda and Norman Masinjila, M. (2010) Gender Dimensions of Cross- Loayza (2010) “Saving in Egypt: Trends and Border Trade in the East African Community: Explanations,” mimeo, World Bank(Washington, Kenya/Uganda and Rwanda/Burindi Borders, DC). Africa Trade Policy Centre Gender and Trade Husain, Aasim (1995) “Determinants of Private Policy Briefs, No. 1, February. Saving,” In Singapore: a Case Study in Rapid Ma v ro t a s , G . ( 2 0 0 9 ) Do m e s t i c Re s o u rc e De ve l o p m e n t . Ke n n e t h Be rc u s o n , e d . Mobilization and Financial Development: International Monetary Fund. Looking at the MDGs from a Different Angle? IFC (2012) Ethiopia Trade Logistics Scoping World Institute of Development Economics Memo, Ethiopia Investment Climate Project, Research (WIDER) Angle newsletter, January, Washington, DC. WIDER, Helsinki. International Monetary Fund. (2010a). “Regional Mbuli, E. (2013), Harmonization of Trade Facilitation Economic Outlook: sub-Saharan Africa” Measures, Chapter 4, Assessing Regional (Washington, DC). Integration in Africa (ARIA) VI report, UNECA International Monetary Fund (2010b). “2010 Article Addis Ababa, Ethiopia. IV Consultation and First Review of the McKinnon, Ronald (1973). “Money and Capital Arrangement under the Exogenous Shocks in Economic Development” Brookings Institution Facility”, (Washington, DC). (Washington, DC) Kedir, A. and G. Ibrahim (2011) ROSCAs in Urban Metin-Ozcan, Kivilcim, and Yusuf Ziya Ozcan Ethiopia: Are the Characteristics of the Institutions (2004) “Determinants of Private Savings in the More Important than those of members? Journal Middle East and North Africa,” In Money and of Development Studies, Vol. 47(7): 998–1016. Finance in the Middle East: Missed Opportunities Kedir, A., R. Disney and I. Dasgupta (2011) Why Use or Future Prospects? Research in Middle East ROSCAs When You can Use Banks? Theory and Economics, 6, pp. 95–113. Evidence from Ethiopia, IZA Discussion paper Modigliani, Franco (1970) “The Life-Cycle Hypothesis Series, DP NO. 5767. of Saving and Intercountry Differences in the Lesser, C. and E. Moise-Leeman (2009) Informal Saving Ratio,” In W.A. Eltis, M.F.G. Scott, and Cross-Border Trade and Trade Facilitation Reform J.N. Wolfe, Eds., Induction, Trade, and Growth: in Sub-Saharan Africa, OECD Trade Policy Essays in Honor of Sir Royd Harrod. Oxford: Working Paper, No. 86, Paris. Clarendon Press. References 77 Muhleisen, Michael (1996). “India—Policies Saving Rates: A Hump-Shaped Relationship.” to Increase Domestic Saving.” International mimeo., Hong Kong University of Science and Monetary Fund, (Washington, DC). Technology, Hong Kong. Plaza, Sonia (2011) Ethiopia’s new diaspora bond: World Bank (2011) Light Manufacturing in Africa, will it be successful this time? World Bank. People Creating Jobs and Prosperity by Fostering Move: A blog about migration, remittances, Competition: with a Case Study of Ethiopia, and development. Volume 1. Prasad (2011). Rebalancing growth in Asia: World Bank (2012a) Global Trade Logistics http://onlinelibrary.wiley.com/doi/10.1111/ Performance Slows Down Amid Recession and j.1468–2362.2011.01276.x/abstract Major Events, Press Release No:2012/446/ Rajaram, A., T. Minh Le, N. Biletska, and J. Brumby, PREM, Washington, DC. 2010, ‘A Diagnostic Framework for Assessing World Bank (2012b) Chinese FDI in Ethiopia: Public Investment Management.” World Bank A World Bank Survey, Africa Region, World Bank, Policy Research Paper 5397. Washington, DC. Washington, DC. Ranganathan, R. and V. Foster (2011) East Africa’s World Bank (2012c) Development Economics, Infrastructure, a continental perspective, World Operations and Strategy; Africa Finance & Private Bank Research Paper 5844, Washington, DC. Sector Development, Light Manufacturing Senbet, L. (2012) African Financial Systems, the in Africa: Creating Jobs and Prosperity by Fostering Oxford Companion to the Economics of Africa, Competition. Volume I, The World Bank Group, pp. 365–372, Oxford University Press. Washington, DC. Shaw, Edward (1973) “Financial Deepening World Bank (2012d) African can Help Feed Africa: in Economic Development.” Oxford University Removing Barriers to Regional Trade in Food Press (New York, NY). Staples, Report 66500-AFR, World Bank, Teravaninthorn, S. and G. Raballand (2008) Transport Washington, DC. prices and costs in Africa: A review of the World Bank (2012e) Financial Literacy around the international corridors. World Bank directions World – An Overview of the Evidence with in Development, Washington, DC. Practical Suggestions for the Way Forward, Policy Wang, Pengfei, Lifang Xu, and Zhiwei Xu Research Working Paper 6107. World Bank, (2011) “Financial Development and Aggregate Washington, DC.