WPS8280 Policy Research Working Paper 8280 Quasi-Fiscal Deficits in the Electricity Sector of the Middle East and North Africa Sources and Size Daniel Camos Antonio Estache Mohamad M. Hamid Energy and Extractives Global Practice Group December 2017 Policy Research Working Paper 8280 Abstract The annual electricity investments needed in the Middle gross domestic product (but goes down to 2.9 percent if East and North Africa region to keep up with demand Lebanon, Djibouti, Bahrain, and Jordan are excluded). have been estimated at about 3 percent of the region’s Only five economies have a quasi-fiscal deficit below 3 projected gross domestic product. However, in most econ- percent of gross domestic product (Algeria, Morocco, omies of the region, the ability to make those investments Tunisia, Qatar, and the West Bank), and hence would not is limited by fiscal and macroeconomic constraints. This be able to finance the average investment requirement paper demonstrates that the solution is readily available: through elimination of inefficiencies. For most economies, by improving the management and performance of the the main driver of the quasi-fiscal deficit is the underpric- region’s utilities, more than enough resources could be ing of electricity, which costs on average 3.2 percent of freed up to make the investments needed. The paper pres- gross domestic product (but 2.2 percent without Lebanon, ents the first evaluation of the size and composition of the Djibouti, Bahrain, and Jordan). Commercial inefficiency quasi-fiscal deficit associated with the management of the comes next, at an average cost of 0.6 percent of gross domes- electricity sector in 14 economies in the Middle East and tic product. Technical and labor inefficiencies represent, North Africa region. The estimations are for 2013. They respectively, 0.4 and 0.2 percent of gross domestic product. show that the average quasi-fiscal deficit is 4.4 percent of This paper is a product of the Energy and Extractives Global Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at dcamos@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Quasi-Fiscal Deficits in the Electricity Sector of the Middle East and North Africa: Sources and Size1 Daniel Camos, Antonio Estache, Mohamad M. Hamid JEL codes: H54:, H69, L32, L94, L98 Keywords: Quasi-fiscal deficit, electricity, utilities, Middle East and North Africa 1 This paper is a background note prepared to support the recent region-wide diagnostic conducted by the World Bank and published as Camos et al. (2018). However, none of the assessments conducted here should be credited or blamed to the World Bank. We are grateful to R. Bacon and V. Foster for useful comments and suggestions but the authors are solely responsible for any mistake or misinterpretation. 1. Introduction Despite its huge oil and gas reserves and its efforts to increase its reliance on renewables, the Middle East and North Africa region (MENA) may soon be unable to meet the electricity needs of its fast-growing population and business activities. In a region with a long tradition of generous subsidies in the sector, fiscal constraints are starting to become binding in many of the economies and the scope to continue funding these subsidies is quickly disappearing.2 The region has indeed started to find ways to cut public expenditures to address unsustainable fiscal deficits close to 10% of Gross Domestic Product (GDP) in 2015 and 2016. One of the effects of these adjustments is that MENA may not be able to allocate the estimated 3% of GDP needed annually over the next 30 years to cover the cost of annual electricity investments required to keep up with demand.3 If, and when, this happens, the current strong coverage rates and quality of service will drop, probably to the surprise of many in the region now used to overall good coverage rates. Turning on the light would no longer be a sure thing for many users. Part of the adjustments required can be managed by the sector itself so as to reduce the risks of investment rationing. As recognized already by many policy makers and utility managers of the region, there is a solid margin to improve the financing space of the sector within the sector itself by cutting the hidden costs linked to various sources of inefficiencies.4 These are seen as implicit subsidies to the sector’s producers, users and workers even if they do not usually appear in the budget. Their total cost is known as the quasi-fiscal deficit (QFD) among macroeconomists. It has already been assessed for other regions, but no estimations have so far been produced for MENA.5 To get a sense of the importance of the QFD at the economy level, the first step is thus to actually quantify them, which is the first purpose of this paper covering 14 economies of the region: Algeria, the Arab Republic of Egypt, Bahrain, Djibouti, Iraq, Jordan, Lebanon, Morocco, Oman, Qatar, the Republic of Yemen, Saudi Arabia, Tunisia, and the West Bank.6 The quantification requires a detailed diagnostic of the financial, technical, commercial, and labor-related inefficiencies.7 And this disaggregation allows, in turn, the assessment of the relative importance of the various inefficiency sources and of the specific areas on which reforms need to focus if the sector is to increase its ability to finance its investment needs on its own. This prioritization is the second main purpose of this paper. To report the results of the assessment, the paper is organized as follows. Section 2 explains the methodology adopted. Section 3 discusses the data and the assumptions which had to be made when data constraints were an issue. Section 4 discusses the results, including a diagnostic of the priority areas in each economy, and an estimation of the size of the effort required from economies to enable them to fulfill their desire of improving their ability to finance their investment needs. Section 5 offers some concluding comments on the main policy options hinted at by the results of the analysis to allow economies to improve their ability to finance their investment needs. 2 See for instance, Fattouh and El-Katiri (2012) and Sdralevich et al. (2014). 3 See Ianchovichina et al. (2012). 4 Algeria, Oman, Qatar and Saudi Arabia have already started to address these issues in 2016 though energy price reforms and have, as a result, improved their fiscal situation quite significantly (IMF 2017). 5 See Petri et a. (2002), Saavalainen and ten Berge (2006) and Ebinger (2006) for Europe and Central Asia or Eberhard et al. (2011) and Trimble et al. (2016) for Sub-Saharan Africa. 6 A diagnostic at the utilities level is also available in Camos et al. (2018) mostly targeted at utility managers to allow them to get a monetary value associated with the inefficiencies they need to address at the level of their firm. 7 The QFD (or hidden-cost) approach has been used in numerous analyses as a powerful tool to communicate with policy makers. It also has been applied to other infrastructure sectors, notably water. For example, the methodology used for the utility QFD in this paper was largely inspired by Trimble and others (2016). 2  2. Defining the quasi-fiscal deficit Following the methodological insights provided by the earlier diagnostics conducted for Eastern Europe, Central Asia, and Sub-Saharan Africa, and in particular building on the approach presented in Trimble et al. (2016), the analysis focuses on the following sources of inefficiencies:  Financial inefficiency, usually labeled underpricing, is measured by the size of gap between the average end-user tariff (Te, expressed in $/kWh) and the cost-recovery tariff (Tc, expressed in $/kWh) weighted by the level of end-user consumption (Qe, expressed in kWh).  Technical inefficiency, usually labeled technical losses in the engineering literature, is measured by the relative difference between actual transmission and development (T&D) losses (lm) and those of an “ideal” utility T&D losses (ln) as documented by Prasad et al. (2009), valued at the cost recovery tariff (Tc) and weighted by the volume of end user consumption.  Commercial inefficiency focuses on revenue collection losses and is measured from the collection rates (Rct) estimated for each economy weighed by the theoretical revenue (Qe.Te) which is also the revenue billed.  Labor inefficiency, focuses on overstaffing and is estimated by comparing the number of customers (NC) per utility employee (NL) against an “efficient” or best practice customer per employee benchmark (BENL), weighted by the cost of labor per employee expressed in $ (CL). All four inefficiencies can be expressed in absolute monetary terms or as a percentage of GDP and adding up these valuations defines the QFD as seen in equation (1): 1 1 (1) Financial Technical Commercial Labor inefficiency inefficiency inefficiency inefficiency (overstaffing) (underpricing) (technical losses) (bill collection losses) 3. The data and the assumptions Technical efficiency is the only hidden cost relatively easily computed from available data simply because it builds on indicators commonly monitored in the sector by engineers. All of the other forms of inefficiencies demand some extra data work implying significantly more creativity than the simple formula just introduced implies. The task relies on multiple sources of information and a number of, sometimes strong, assumptions on certain of the variables. But none of these is very different from those made in earlier studies on other regions. This is one of the common consequences of the poor commitment to accountability in the sector, evidenced by the very narrow sets of data available from public sources on the sector. Most data come from the MENA Electricity Database recently produced by the World Bank Energy Global Practice, the World Development Indicators (WDIs), reports from the Arab Union of Electricity and the International Labor Organization (ILO). Table 1 explains how the various variables were approximated and Appendix 1 provides further details on data sources. The quantification of the financial, commercial and labor inefficiencies was particularly complex. Often, data for all 14 economies were not available in a single source, requiring further 3  collection, verification and occasionally assumptions. For some of the economies, data had to be collected at the utility level and information aggregated to produce economy-specific data. The main additional details on the way the data were generated is the main focus of the rest of this section. Table 1: Descriptions and assumptions of economy-level quasi-fiscal deficit components Variable Description and assumptions (Qe) End-user Calculated by multiplying the electric power consumption per capita by the total population of the economy for the consumption year 2013. (Te) Average Taken to be the average residential tariff for a consumption of 250 kWh/month for the year 2013. Values for all end-user tariff economies were calculated based upon the 2014 Arab Union of Electricity’s “Electricity Tariff in the Arab Countries.” In the case of Djibouti, calculations were based upon the official tariff document published by the country. (Tc) Cost- Estimated using the LCOE unit cost of energy per technology type ($/kWh) weighted according to the energy mix recovery tariff of each economy. Data is from WDI for the energy mix information, and an LCOE modeling tool(1) developed by rate ESMAP for most of the LCOE values. Since the unit cost of fuel and renewables used in the modeling tool did not reflect the current state of energy sources in the MENA region, values from Lazard’s LCOE Analysis 2014 were used instead and adjusted to include T&D contribution to the unit cost by adding ¢ 3.2/kWh to the figures. (Lm) Technical The technical loss rate is defined as the electric power transmission and distribution losses (% of output) and was loss rate obtained from WDI database. WDI did not include data for West Bank (calculated alternatively as the average of the technical losses of West Bank distribution utilities in the MED) and Djibouti (value obtained as the grid losses from AEEP, 2013). (Ln) Normative The choice of 5% was done so as to have values of Ln below the region’s best-performing economies, namely loss rate Bahrain and Qatar with technical loss rates of 5.2% and 6%, respectively. (Rct) Collection The bill collection rate indicates the income effectively collected during the year by the utility in relation to the rate income billed. In the cases where a single utility existed (a VIU in the case of Algeria, for example), the collection rate of the economy was that of the utility. When more than one utility existed, the average value of the distribution utilities was used (in the case of Egypt, for example). (NC) Number of This figure was easily obtained for economies with a single VIU. For economies with several utilities, the presence customers of a regulator would allow for an aggregate official figure to be obtained from the regulator’s annual report. (connections) However, in the case of no regulator present, the sum of individual utility customers was calculated. (NE) Number of The number of full time equivalent (FTE) employees was used for all utilities, except for Oman, where the number employees of total (direct and indirect) employees was used. This is because several utilities in Oman have a very low number of FTE while the number of outsourced (or indirect) employees is high. (CL) Cost of The cost of labor is defined as the annual cost of personnel directly employed by the utility, and was sourced mainly labor from the financial statements of utilities. However, when this was not available, estimates were made to calculate a unit labor cost per employee, which was then multiplied by the number of employees present in the utilities for which labor cost data were not available. A calculated sum then allowed the economy-level aggregated estimate of the cost of labor to be obtained. Approximation had to be made for a number of economies. (BENL) Customer per employee is an indicator of performance with values commonly above 500 in the OECD economies. Benchmark The value of 413 used in this study was obtained using the same benchmark value for the number of customers per number of employees in low-income economies (Ebehard et al. 2008). customers per employee in LICs (1) A compilation of economic costs of more than 50 electricity generation and delivery technologies, the Model for Electricity Technology Assessment (META) was rolled out to the World Bank Group and selected partners and clients in June 2012. The modeling tool can be downloaded here: http://esmap.org/META. Notes: ESMAP = Energy Sector Management Assistance Program; FTE = full-time equivalent; ILO = International Labour Organization; kWh = kilowatt-hours; LCOE = levelized cost of electricity; LICs = low-income countries; MED = MENA Electricity Database; MENA = Middle East and North Africa; META =; T&D = transmission and distribution; VIU = vertically integrated utility WDI = World Development Indicators. 4  a. Methodology for estimating the cost-recovery tariff Cost-recovery tariffs were calculated using the basis of the economy’s fuel mix, and the levelized cost of electricity (LCOE) from different energy sources, as described by equation (2): % % . % . % % (2) where  “%Coal, Hydro, Natural Gas, Fuel and Renewables” represents the percentage share of each technology in generation; the share for each economy is reported in Table 2.  LCOEi represents the levelized cost of electricity (LCOE) for each generation technology expressed in US$ cents per kilowatt-hour; the costs are reported in Table 3. Table 2: Share of energy mixes used in the calculation of Tc (%) Economy Coal Hydro Natural gas Fuel Renewables Algeria 0 1 93 7 0 Bahrain 0 0 100 0 0 Djibouti 0 0 0 100 0 Egypt, Arab Rep. 0 8 77 15 1 Iraq 0 8 55 19 0 Jordan 0 0.3 25 74 0.1 Lebanon 0 7 0 93 0 Morocco 43 10 21 21 5 Oman 0 0 97 3 0 Qatar 0 0 100 0 0 Saudi Arabia 0 0 53 24 0 Tunisia 0 0.3 96 0.4 2 Yemen, Rep. 0 0 32 68 0 Israel* 54 0 42 36 1 Source: WDI.Note: * for the West Bank, all electricity is imported from Israel, therefore Israel’s LCOE is used for Tc. 5  Table 3: LCOE values used to calculate the cost-recovery tariffs and their sources Generation type LCOE (US$ cents) /kWh Source Coal 7.44 ESMAP META Model Hydro 2.86 ESMAP META Model Natural gas 8.12 ESMAP META Model Fuel 31.45 Average Lazard Renewables 6.9 Average Lazard Source: Author calculations based on ESMAP META Model and Lazard (2014). Note: (1) ESMAP = Energy Sector Management Assistance Program; (2010 as base year); META = Model for Electricity Technology Assessment; (2) The META does not cover T&D LCOE. In order to address this limitation, a value of US$ cents 3.2 per kilowatt-hour (kWh) was added to their LCOE estimations;8 (3), for renewables utility-sized photovoltaics (PV) and wind only are considered. b. Methodology for collection rates in Oman, Saudi Arabia, and Qatar The bill collection rate is defined by revenues collected divided by the billed amount. More specifically, the billed amount is defined as the income effectively collected from customers for energy consumption and related services/revenues related to energy consumption and services. The data were readily available for all economies except for Oman, Saudi Arabia and Qatar. For those economies, the collection rate was calculated from the annual reports and financial statements of the utilities. Since the annual reports do not provide a value for billed amounts, these were approximated as follows: 1. The income effectively collected is approximated by the annual sales of, or annual revenues from, electricity in the financial statement. 2. The income not collected is approximated by the receivables from customers, as stated in the financial report. 3. The billed amount is therefore the sum of what was not collected (the receivables) and what was actually collected (the sales revenue reflected in the financial report). 4. The collection rate = Sales revenue / (sales revenue + receivables from customers). 5. If the economy has several utilities, steps 1 to 4 above were applied to each utility and the average of all utilities was taken to be the economy collection rate. Appendix 2 reports the detailed computations from the raw data. c. Estimating labor costs with data limitations The cost of labor defined as the annual cost of personnel directly employed by a utility was collected from the financial statements of utilities. However, when this was not available, estimates were made to calculate a unit labor cost per employee from the partially available data found in the MED (if several utilities were present in the economy) or the average cost data available from the ILO as in the case of the Republic of Yemen. To get to the total cost, this average unit cost was then multiplied by the total number of FTE. The utilities' specific data were then aggregated at the economy level. For the Arab Republic of Egypt, Jordan, Morocco, Oman, and the Republic of Yemen, labor costs were unavailable for several 8 This is based on a recent tariff study for Morocco (see Macroconsulting, 2013) which found a transmission LCOE of about US$ cents 0.08 per kWh and a distribution LCOE of US$ cents 0.24 per kWh. This is also aligned with figures reported in IEA, 2014. 6  utilities and were obtained based on calculations making use of an estimated unit labor cost, as described next. Appendix 3 reports the detailed computations. Arab Republic of Egypt Egypt counts 12 utilities (including generation, distribution, and transmission). The cost of labor for all utilities was available except for the important Hydro Power Plants Electricity Production Company. To calculate the total cost of labor, including that of the Hydro Power Plants Electricity Production Company and EEHC, a unit average cost per employee was produced from the data for the utilities with labor costs and number of employees available. This unit cost was then multiplied by the number of employees to obtain the value for the total labor cost for Egypt. Jordan For Jordan, of 10 utilities, the number of employees was available for nine, and labor costs for eight. Data from the report of the Jordanian regulator, the Energy and Minerals Regulatory Commission (EMRC), were used for the utility with the missing number of employees, Qatrana Electric Power Company, (QEPCO). Since the Amman Asia utility was not operational in the year of study (2013), it was omitted from the computations. From then on, the methodology used in the case of Egypt was applied to calculate the total cost of labor for the nine utilities in Jordan. Morocco In the case of Morocco, the number of employees of all the utilities was available from public sources, but not the labor costs per utility. This demanded some extrapolations. Essentially, they consisted in applying the average unit labor cost observed to the utilities for which these costs were not observed and multiplying them by the observed FTE. The details are provided in Appendix 3 Oman For Oman, the challenge was that utilities often have a larger number of outsourced employees than full- time employees. For consistency, the total number of employees is included in the labor cost estimates. Twelve utilities had data for both the total number of employees and the labor costs. A unit cost of labor was calculated from these 12 utilities. The total number of employees for the 12 utilities was then used to compute the total cost of labor obtained for these 12 utilities. From an aggregate value for the total direct and indirect employees in 2013 obtained from the Authority for Electricity Regulation (AER) annual report for 2014, the unaccounted-for number of workers was identified. Combined with an assessment of the average labor cost, it allowed an estimation of the total labor cost for the economy. Republic of Yemen The cost of labor for the Public Electricity Corporation (PEC), the Yemeni public VIU, was unavailable. Data on the number of employees were obtained. An estimate of the cost of labor was done using Republic of Yemen–specific average values from the ILO for the main professional categories. These were then used to produce an average unit cost used in turn to produce the total labor cost. 7  4. The economy-specific estimations of QFD levels and composition Table 4 brings together the results of the estimations made for every source of inefficiency for each of the 14 economies covered by the sample as per the approaches and assumptions discussed in section 3. To make comparisons across economies easier, the estimations are normalized to GDP. The cost of the sum of the various sources is then reported as a share of GDP and in absolute value. The following are the main insights unveiled by the quantification of the burden imposed by hidden costs on the sector in MENA. Table 4: Drivers of QFD in MENA, 2013 (except as noted) Cost of the sources of inefficiencies expressed as a share of GDP Absolute (%) QFD as QFD Commercial share of Economy Financial Technical Labor value inefficiency GDP inefficiency inefficiency inefficiency ($ (Collection (%) (underpricing) (T&D losses) (Overstaffing) million) losses) Lebanon 8.20 0.41 0.21 0.03 8.9 3,826 Djibouti 0.98 1.08 5.24 0.88 8.2 101 Bahrain 7.86 0.02 0.02 0.13 8.0 2,640 Jordan 5.96 0.84 0.75 0.21 7.8 2,608 Egypt, Arab Rep. 5.61 0.42 0.06 0.28 6.4 18,219 Saudi Arabia 4.81 0.11 0.17 0.07 5.2 38,467 Yemen, Rep. 3.16 0.81 0.08 0.11 4.2 1,494 Iraq 2.44 0.83 0.13 0.21 3.6 7,888 Oman 2.70 0.22 0.18 0.10 3.2 2,496 Algeria 1.46 0.37 0.10 0.32 2.3 4,720 Qatar 1.47 0.02 0.10 0.01 1.6 3,224 Tunisia 0.34 0.39 0.54 0.15 1.4 655 Morocco 0.65 0.33 0.20 -0.21 1.0 948 West Bank -0.84 0.30 0.30 0.13 -0.1 -13 Average 3.2 0.4 0.6 0.2 4.4 Average without Bahrain, Djibouti, 2.2 0.4 0.2 0.1 2.9 Jordan and Lebanon Source: Authors’ calculations. Note: The year is 2013 for all except the following: 2012 for Lebanon, Iraq, Morocco, and the West Bank; and 2011 for Djibouti. This variation reflects data availability. GDP = gross domestic product; MENA = Middle East and North Africa; QFD = quasi-fiscal deficits; T&D = transmission and distribution. The first insight emerging from the table is that the hidden costs of financial inefficiency drive the high QFD values in MENA whether outliers (Lebanon, Djibouti, Bahrain and Jordan) are included or not. As seen in Figure 1, in most economies the difference between cost recovery and actual tariffs is often quite significant. It leads to a financial inefficiency averaging 3.2% of GDP in the region. Overall this is more than the sector’s investment requirements for the region on an annual basis. Ignoring the outliers in terms of underpricing (Lebanon, Bahrain, Jordan and Egypt) however, leads to an average of 1.7% which implies that improving cost recovery alone will not be enough for many of the economies. 8  But underpricing is an issue in most economies. In 8 of the 14 economies, it represents more than three- quarters of the QFD and in 11, it represents at least two-thirds. Lebanon and Bahrain suffered from a particularly strong underpricing issue in 2012. It is unfortunately hard to distinguish between underpricing linked to electricity subsidies to users and subsidies to producers for fuels used to generate electricity. This is because the cost-recovery tariff used to estimate the economy-level QFD is based on levelized energy costs, computed as weighted averages of each economy’s energy mix, to which a factor was added to account for transmission and distribution (T&D) costs. Djibouti and the West Bank are notable exceptions to the trend of underpricing as a driving force of the QFD. This is because they have high average end-user tariffs: $0.31 per kilowatt-hour (kWh) and $0.16 per kWh, respectively. Note also, that the negative values for underpricing in the West Bank simply mean that the West Bank’s cost-recovery tariff is smaller than the average end-user tariff (based on the energy mix of Israel, given that the West Bank imports all of its electricity from there). Figure 1: Comparing average end-user and cost-recovery tariffs in MENA, 2013 (or most recent year with data, 2009–12) 40 US¢ / kWh 30 20 10 0 Te ‐ average end‐user tariff Tc ‐ cost recovery tariff Source: Authors’ calculations. The second most important QFD driver is commercial inefficiency. It averages 0.6% of GDP (although, without Djibouti the average drops to 0.2% of GDP). In relative terms, it is the main driver of QFD in Djibouti, Tunisia and the West Bank. In absolute terms, it is a significant problem in Djibouti (5.24% of GDP), Jordan (0.75), Tunisia (0.54) and the West Bank (0.3). In most of the other economies, collection rates are reasonably high and poor collections do not represent a major issue in absolute or relative terms. It is worth mentioning however, that, somewhat counterintuitively, high financial efficiency (i.e. when cost recovery is high, it does not mean that collection is bad) is not strongly correlated with high commercial efficiency. The correlation coefficient is -0.23. Technical inefficiencies come third in relative importance as a driver of QFD with an average size of 0.4% of GDP. In absolute terms, it is an above average problem for Djibouti, Jordan, the Republic of Yemen, and Iraq, and to a lesser extent Algeria and Lebanon. In relative terms, it is a notable issue in an important part of some economies’ QFDs: they represent more than one-fifth of the total QFDs in Morocco, the West Bank, Tunisia, Iraq, and the Republic of Yemen. It is useful to note that there seems to be a strong correlation (0.61) between technical inefficiency and labor inefficiency. Both of these dimensions are, to some extent management issues but may also reflect the age of the technology in the economies characterized by the joint presence of these two forms of inefficiencies. 9  Labor inefficiencies are the weakest driver of QFD in absolute and often in relative terms. On average, it represents about 0.2% of GDP but it is higher than average in absolute terms in Djibouti, Egypt, and Algeria. The low average impact in the region partially reflects the low average labor cost in MENA which tends to reduce the impact of overstaffing when this takes place. It represents between 10 and 15 percent of the QFDs in Algeria, Tunisia, and Djibouti. In short, the QFD’s share of GDP is relatively small in Maghreb economies, and large in some Mashreq and Gulf Cooperation Council (GCC) economies.9 Addressing this type of inefficiency may be a delicate act for governments, since it often implies reducing the size of state-owned enterprises (SOEs). Providing public jobs—and subsidized basic services—has been part of the social contract in the region for the past several decades, in exchange for social stability. The aggregation of these different components provides a clear picture of the high total costs of management and policy weaknesses in the region. The average QFD is 4.4% of GDP. Excluding the four main outliers (Bahrain, Djibouti, Jordan and Lebanon) brings down the average to 2.9% of GDP, which is right below the average value of the estimated annual investment requirement in the sector to allow the economies to meet the future consumption needs. Figure 2 illustrates the wide range of experiences visually. A more detailed look at table 4 and figure 2 shows that 9 of the 14 MENA economies studied have a QFD above 3% of their GDP. In other words, these economies of the region have enough margin to increase their financing space by simply reducing their inefficiencies. The margin is particularly strong for Lebanon, Djibouti, Bahrain, and Jordan with a QFD between 8 and 9 percent of GDP in 2012-13. Only five economies have a QFD below 3 percent of GDP (West Bank, Morocco, Tunisia, Qatar, and Algeria). For those economies, cutting inefficiencies will help but not be enough to cover the investment needs. Figure 2: QFD (% of GDP) in MENA, 2013 10% 9% 8% 7% 6% 5% 4% 3% 2% 1% 0% ‐1% Source: Authors’ calculations. Note that by international standards, the margin for action is quite strong as MENA’s QFD tends to be high. For Sub-Saharan Africa (SSA), Trimble et al. (2016) find values ranging from -0.3% to 6% of GDP for an average of 0.8% of GDP vs 4% for MENA. In other words, MENA’s utilities have more hidden 9 Note that the negative values for overstaffing in Morocco simply mean that Morocco’s ratio of customers to employees is better than the efficiency benchmark (413:1) used here. 10  costs than SSA’s. Another difference is that the MENA QFD appears to be driven mostly by financial inefficiency while for SSA, technical and commercial inefficiencies play the largest role. To complete the snapshot, Table 4 also shows that, in absolute terms, the highest QFDs are to be found in Saudi Arabia ($38 billion), Egypt ($18 billion), and Iraq ($8 billion), and the lowest in the West Bank (with a negative QFD of $13 million), Djibouti ($101 million, despite having the second-highest QFD when expressed as a percentage of GDP), and Tunisia ($655 million). These values strongly correlate to the size of the economy and to the consumption levels of its population. 5. Concluding comments This paper suffers from several data constraints, which demanded some creativity to be able to come up with decent approximations of the values for key variables. Despite these data issues, it seems quite reasonable to argue that the analysis conducted here provides enough evidence of the existence of an important QFD problem in MENA and on its sources in 2013. The key to its reduction and to increasing space to finance investment from the available resources within the sector itself, resides, in fixing the significant underpricing problem characterizing the region. The user and producer subsidies, still widespread in the region, not only distort price signals and hence production and consumption patterns, but also decrease the region’s odds of achieving its investment needs. Therefore, tariffs need fixing indeed. But fixing tariffs does not simply mean increasing tariffs. Tariff structures often also need to change particularly in a context of social tensions linked to the limited capacity to pay of many families. Tariff reforms in most economies of the region could help improve the political viability of efforts to increase cost-recovery rates. However, to achieve improved cost recovery, subsidy cuts, and better targeting, there has to be a political will to assess the current design of electricity tariffs and its incidence in the various economies of the region. Focusing on prices alone would be a mistake, as prices are not the only problem for the region and for many economies, this will not suffice. MENA also has some margin to increase its financing space by addressing the other components of the QFD, i.e. T&D losses, collection losses, and overstaffing, which add up to as much as 1 percent of GDP in some economies. The economy-specific diagnostics reported here show that the actual priorities are different across economies. Moreover, it is quite likely that, within economies also, there is some scope for differentiation, as many of the economies contain several utilities with very different constraints. Yet assessing these constraints is very specific to each utility and this would get into much deeper details than what this paper allows.10 At the economy level, the policies to address the broad economy-specific issues and the priorities are all relatively straightforward now that the relative importance of each source of inefficiency has been identified. The real issue for the region is that most of the solutions are politically sensitive. Tariff increases, improvements in revenue collection efforts and reductions in overstaffing are not easy to sell in the current social and political context. But there is enough margin to be fair and more efficient in the region such as to move in the right direction at the financial level, while also addressing social concerns. This should ease the political tensions even if, as always, such reforms would imply some losers likely trying to slow down the efforts made. The next analytical step for the region to prepare the implementation of the reforms may require a more detailed look at the winners and losers of the various policy options 10 Camos et al. (2018) actually report the outcome of this assessment of the drivers of the QFD at the utilities level. 11  needed to cut the QFD. Without this additional information, progress may continue to be slow, at least in some of the economies of the region where private interests continue to dominate the public interest. 12  References Arab Union of Electricity. 2014. Electricity Tariff in the Arab Countries. Statistical bulletin, Arab Union of Electricity, Amman. AEEP (Africa-EU Energy Partnership). 2013. “Country Power Market Brief: Djibouti.” European Union Energy Initiative (EUEI). http://www.euei- pdf.org/sites/default/files/field_publication_file/AEEP_Djibouti_Country_market_brief_EN.pdf. AER (Authority for Electricity Regulation). Annual Report 2013. Oman: AER. AUE. 2014. “Electricity Tariff in the Arab Countries.” AUE, Amman, Jordan. Camos, Daniel, R. Bacon, A. Estache and Mohamad M. Hamid (2018). Shedding Light on Electricity Utilities in the Middle East and North Africa: Insights from a Performance Diagnostic. Directions in Development. World Bank. Washington, DC Eberhard, A., O. Rosnes, M. Shkaratan and H. Vennemo. 2011. “Africa’s Power Infrastructure. Investment, Integration, Efficiency.” Directions in Development report, No: 61390, World Bank, Washington, DC. https://openknowledge.worldbank.org/bitstream/handle/10986/2290/613090PUB0Afri158344 B09780821384558.pdf?sequence=1. Ebinger, J. O. 2006. “Measuring Financial Performance in Infrastructure: An Application to Europe and Central Asia.” Policy Research Working Paper 3992, World Bank, Washington, DC. EEHC (Egyptian Electricity Holding Company). Annual Report 2014. Egypt: EEHC. ESMAP (Energy Sector Management Assistance Program). ESMAP Meta Model. http://esmap.org/META Fattouh, B., and L. El-Katiri. 2012. “Energy Subsidies in the Arab World.” Arab Human Development Report Research Paper Series, United Nations Development Programme, New York. Ianchovichina, E., A. Estache, R. Foucart, G. Garsous, and T. Yepes. 2012. “Job Creation through Infrastructure Investment in the Middle East and North Africa.” Policy Research Working Paper No. 6164, World Bank, Washington, DC. International Energy Agency (IEA), 2014. Electricity Transmission and Distribution. Energy Technology Sstems Analysis Programme, Technology Brief. https://iea-etsap.org/E-TechDS/PDF/E12_el-t&d_KV_Apr2014_GSOK.pdf International Labor Organization, ILOStat, consulted on 20 January 2017. http://www.ilo.org/ilostat/faces/oracle/webcenter/portalapp/pagehierarchy/Page3.jspx?MBI_ID= 6 International Monetary Fund (IMF). 2017., Middle East and Central Asia: Regional Outlook Reflecting Global Developments, April, available at https://www.imf.org/en/Publications/REO/MECA/Issues/2017/04/18/mreo0517 Iraq Energy Institute. 2015. “Iraq’s Biggest Power Threat.” Iraq Energy Forum, August 6, 2015, Iraq Energy Institute, Baghdad. http://iraqenergy.org/home/articles_details.php?id=5. Jeddah Chamber of Commerce. 2015. Sectorial Report on Saudi Arabia—Electricity July 2015. Jeddah: Jeddah Economic Gateway. http://www.jeg.org.sa/data/modules/contents/uploads/infopdf/2832.pdf. KAHRAMAA (Qatar General Electricity and Water Corporation). 2013. Sustainability Report 2013. Qatar: KAHRAMAA. Lazard. 2014. “LCOE Analysis.” https://www.lazard.com/media/1777/levelized_cost_of_energy_- _version_80.pdf 13  L’Algérie profonde/Ouest. “Plus de 184 millions de DA de pertes pour la Sonelgaz.” http://www.liberte- algerie.com/ouest/plus-de-184-millions-de-da-de-pertes-pour-la-sonelgaz-228183/print/1. Lebanon Ministry of Environment and United Nations Development Programme (UNDP). http://climatechange.moe.gov.lb/viewfile.aspx?id=64. Macroconsulting, 2013. Morocco electricity sector tariff study. NEPCO (National Electric Power Company). Annual Report 2013. Jordan: NEPCO. Petri, M., G. Taube, and A. Tsyvinski, 2002. “Energy Sector Quasi‐Fiscal Activities in the Countries of the Former Soviet Union.” IMF Working Paper WP/02/60, International Monetary Fund, Washington, DC. https://www.imf.org/external/pubs/ft/wp/2002/wp0260.pdf. Prasad, T. V. S. N., M. Shkaratan, A. K. Izaguirre, J. Helleranta, S. Rahman, and S. Bergman. 2009. Monitoring Performance of Electric Utilities: Indicators Benchmarking in Sub-Saharan Africa. Washington, DC: World Bank. https://www.esmap.org/sites/esmap.org/files/P099234_AFR_Monitoring%20Performance%20of %20Electric%20Utilities_Tallapragada_0.pdf. Saavalainen, T. and J. ten Berge. 2006. “Quasi Fiscal Deficits and Energy Conditionality in Selected CIS Countries.” IMF Working Paper WP/06/43, International Monetary Fund, Washington, DC. www.imf.org/external/pubs/ft/wp/2006/wp0643.pdf. Sdralevich, C, R. Sab, Y. Zouhar, and G. Albertin. 2014. Subsidy Reform in the Middle East and North Africa: Recent Progress and Challenges Ahead. Washington, DC: IMF. Trimble, C., M. Kojima, I. P. Arroyo, and F. Mohammadzadeh. 2016. “Financial Viability of Electricity Sectors in Sub-Saharan Africa: Quasi-Fiscal Deficits and Hidden Costs.” Policy Research Working Paper 7788, World Bank, Washington, DC. World Bank. 2017. “MENA Electricity Database.” World Bank, Washington, DC. World Bank 2014b. West Bank and Gaza—Assessment and Action Plan to Improve Payment for Electricity Services in the Palestinian Territories: Study on Electricity Sector Contribution to Net Lending. Report No: ACS9393. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/120271468317065014/pdf/ACS93930WP0P1469990 Box385388B00OUO090.pdf. World Bank. 2009. Energy Efficiency Study in Lebanon—Final Report. Washington, DC: World Bank. http://climatechange.moe.gov.lb/viewfile.aspx?id=205. World Bank. 2013. World Development Indicators (database). World Bank, Washington, DC. https://data.worldbank.org/data-catalog/world-development-indicators . 14  Appendix 1: Sources of data used for the economy-level quasi-fiscal deficit calculations Economy Qe: End-user Tc: Cost- Te: Avg. End- Lm: Technical loss Number of Number of Cost of laborb Rct: Collection GDP consumption recovery user tariff ratesa customers employees (FTE) Rates (kWh) tariff (connections) Algeria MED WDI MED MED MED Online Bahrain WDI MED Online MED MED Djibouti MED Onlinec MED MED MED MED Egypt, Arab WDI WDI EEHC Annual Report EEHC Annual Estimation MED (average) Rep. of 2014 Report 2014 Iraq MED Calculations WDI MED Onlined MED Online (World (WDI; Bank)e Jordan WDI ESMAP Arab Union of WDI NEPCO Annual NEPCO Annual Estimation MED MED (average) META Electricity Report 2013 Report 2013 Model; (2014), Lazard’s Electricity Lebanon WDI LCOE Tariff in the WDI MED MED MED Onlinef WDI Morocco WDI Analysis, Arab Countries WDI Estimation MED ONEE contact Oman WDI 2014) WDI AER Annual Report AER Annual Report Estimation MED Estimated 2013 2013 Qatar WDI WDI KAHRAMAA KAHRAMAA KAHRAMAA MED Sustainability Report Sustainability Annual Report 2014 2013 Report 2013 Saudi Arabia WDI WDI MED MED MED SEC statistics 2000 to 2014 Tunisia WDI WDI MED Data from utility West Bank MEDg MED (average) MED MED (average) Yemen, Rep. WDI WDI MEDh Estimated MED Source: Author’s calculations. Note: AER = Authority for Electricity Regulation; EEHC = Egyptian Electricity Holding Company; ESMAP = Energy Sector Management Assistance Program; FTE = full-time equivalent employee; GDP = gross domestic product; kWh = kilowatt-hours; LCOE = levelized cost of electricity; MED = MENA Electricity Database; META = Model for Electricity Technology Assessment; NEPCO = National Electric Power Company; ONEE = Office National de l’Eau et l’Electricité; SEC = Saudi Electricity Company; WDI = World Development Indicators. a WDI technical losses (distribution and transmission losses). b Refer to appendix C for calculation details. c AEEP d Iraq Energy Institute 2015. e World Bank 2016a. f Lebanon Ministry of Environment and UNDP g Calculated as the sum of energy volume billed (from MED) for the three distribution utilities in the West Bank (TUBAS, JDECO, and NEDCO). h Used 2012 value in the case of the Republic of Yemen due to lack of data for 2013. Appendix 2: Data and sources used for calculating collection rates Economy Oman Oman Oman Utility name Muscat Electricity Majan Electricity Company Mazoon Electricity Distribution Company Distribution Company Source of data Annual report 2013 Annual report 2013 Annual report 2013 Amounts due from private 33,562,000 17,357,000 20,344,000 customers in Omani Rials (RO) Amounts due from government 13,610,000 6,029,000 5,776,000 customers (RO) Electricity sales to private 98,814,000 79,265,000 67,567,000 customers (RO) Electricity sales to government 37,479,000 10,221,000 18,815,000 customers (RO) Collection rate (%) 74 79 77 Economy Saudi Arabia Economy Qatar Utility name Saudi Electricity Company Utility name Kahramaa (SEC) Source of data SEC publication: electric Source of data Kahramaa Annual Report data 2000–14 2013 Receivables form Saudi riyal (SRl) Accounts receivable Qatari riyal (QR) customers and revenues 18,452,000,000 585,434,000 accrued net Total electricity sales SRl 32,878,000,000 Revenues from sale of QR 1,553,741,000 electricity Collection rate (%) 64 Collection rate 73 16  Appendix 3: Approximation of labor costs in economies with data gaps Egypt Formula Description Value A Number of employees without the Hydro Power Plants Electricity Production Company and without EEHC 172,733 B Cost of labor in all utilities except the EEHC and Hydro Power Plants Electricity Production Company $1,359,678,577 C=B÷A Unit cost of labor $7,872 D Number of employees in the EEHC 3,586 E Number of employees in the Hydro Power Plants Electricity Production Company 3,038 F = (D+E) × C Cost of employees in the EEHC and Hydro Power Plants Electricity Production Company $52,141,228 G=F+B Total estimated cost of labor including EEHC and Hydro Power Plants Electricity Production Company $1,411,819,806 Source: MENA Electricity Database and Authors’ calculations. Note: EEHC = Egyptian Electricity Holding Company. Jordan Utility No. employees (A) Labor costs in $ (B) 1 AES Levant Holding B.V. 47 Not available 2 Amman East Power Plant (AES) 51 3,248,314 3 Central Electricity Generating Company 1,037 18,788,759 4 Electricity Distribution Company 1,320 19,813,536 5 Irbid District Electricity Company 1,088 16,270,190 6 Jordan Electric Power Company 2,602 86,150,700 7 National Electric Power Company 1,373 22,166,850 8 Qatrana Electric Power Company 78 Not available 9 Samra Electric Power Generation Company 345 6,096,730 Total number of employees used in computation 7941 C =A2+A3+A4+A5+A6+A7+A9 Total labor cost used in computation $175,535,079 D = B2+B3+B4+B5+B6+B7+B9 Unit labor cost: E=D/C $22,075 Estimated labor costs for Qatrana: F=A8xE $1,721,819 Estimated labor costs for AES Levant: G = A1xE $1,037,506 Final total labor cost estimation for Jordan $ 175,294,404 H=D+F+G Source: Authors’ calculations based on MENA Electricity Database. Morocco Utility No. employees (A) Labor costs in $ (B) 1 AMENDIS Tanger 401 25,306,122 2 AMENDIS Tetouan 468 25,772,595 3 LYDEC 1,432 92,912,657 4 ONEE 8,796 252,453,751 5 RADEEL 134 6 REDAL 511 44,702,600 7 Regie de Kenitra 196 8 Regie de Marrakech 370 8,355,024 9 Regie de Meknes 208 10 RADEEJ 188 4,131,731 11 Regie de Fes 439 12 Regie de Safi 118 Total number of employees available: C = SUM (A1 to A12) 13,261 Total labor costs available $453,634,480 D = SUM (B1 to B12) Unit labor cost: E = D/C $34,208 Estimated labor cost in non-available utilities: F = E x (A5+A9+A11+A12) $37,457,940 Total cost of labor for all utilities: G=D+F $491,092,421 Source: Authors’ calculations based on MENA Electricity Database. Yemen Position Monthly salary in YRls (Yemeni Riyals) Managers 30,290 Clerical support workers 42,591 Technicians and associate professionals 69,439 Average monthly earning calculated 47,440 Average annual cost in U.S. dollars per employee (assuming salary paid for 12 months; $2,797 and using an exchange rate of $1 = 203.4 Yemeni riyals (corresponding to January 1, 2013) Number of employees in PEC 18126 Total estimated salary bill in U.S. dollars (cost of labor) $50,706,483 Source: Authors’ calculations based on MENA Electricity Database and ILO data 18