World Bank Malaysia Hub Doing Business Research Notes No. 1/June 2017 Electricity Sector Constraints for Firms Across Economies: A Comparative Analysis Jean Arlet What are the main types of electricity sector constraints and where businesses self-report outages over a period of time. Doing Business, how do they vary across economies? How do power outages however, collects two indices directly from distribution utilities for each economy’s largest city:5 the system average interruption duration index and electricity tariffs impact firms’ demand for energy inputs? (SAIDI) and the system average interruption frequency index (SAIFI). SAIDI This note aims to provide insights into these questions using measures the average total duration (in hours) of outages, and SAIFI the recent Doing Business data from the World Bank. average number of outages, experienced by a customer over the course of a calendar year. SAIDI and SAIFI include all types of outages, including load More than one billion people do not have access to electricity1 and, accord- shedding or planned power cuts for maintenance. ing to the World Bank, an equal number receive electricity services that do not meet adequate reliability standards.2 A lack of electricity negatively Looking at SAIDI and SAIFI data across economies, several observations can impacts welfare by undermining areas such as education (Khandker and be made. First, the measures are highly correlated. Because SAIDI is a others 2014) and healthcare (Adair-Rohani and others 2013). Firm function of the number of service interruptions as well as the average performance is one area that is arguably the most affected by poor electric- disruption time, this is not surprising. Second, the data show a large variation ity services. Data from the World Bank Enterprise Surveys indicate that in outages from one economy to another. In 2015, customers in the main business owners in developing economies perceive a lack of reliable electric- cities of nearly 50 upper-middle-income and high-income economies ity supply as the biggest obstacle to the operation of their businesses, experienced less than one hour of blackouts—including Costa Rica, Germany behind only access to finance, the informal sector and political instability.3 and Singapore. By contrast, power was interrupted for more than 1,000 hours in 2015 in Iraq, Nigeria, Pakistan and South Sudan, among others. However, obstacles to getting electricity vary. For a newly incorporated Outages also fluctuate on a yearly basis; over a third of economies saw startup firm, for example, obtaining a new electricity connection may be outages increase or decrease by 30% or more in 2015. In Zambia, for difficult owing to a burdensome connection process. Or, once connected to example, SAIDI and SAIFI more than doubled compared to the previous year the grid, a business may face blackouts that force it to halt production or as insufficient rainfall resulted in low water reserves at hydropower dams resort to self-supply through generators, at a significant cost (Foster and (Mutale 2015). Steinbuks 2010). Finally, firm performance may be hindered in economies where electricity tariffs are high relative to income levels. The OECD high-income economies had the lowest duration and frequency of interruptions in 2015, occurring on average less than once a year per This note explores electricity sector constraints across 190 economies by customer. On the other hand, economies in Sub-Saharan Africa suffered the drawing on recent data from the Doing Business indicator set for getting most blackouts in 2015, averaging almost 741 hours over 253 interruptions electricity, which—in addition to measuring the process for obtaining an (Figures 1 and 2). South Asia and the Middle East and North Africa follow as electricity connection for a firm—now includes electricity tariffs and power the second and third regions, respectively, with the most power interrup- outages. Stylized facts are presented throughout. The data on tariffs and tions. Indeed, firms in the three regions mentioned above are most likely to outages are collected directly from utilities and do not rely on firm surveys own generators or report electricity as a major obstacle to doing business.6 (as commonly used in the literature). Further adding to existing research, Electricity shortages are so chronic in some economies—including Afghani- this note examines how power outages and electricity tariffs are related to stan, Guinea-Bissau and Sierra Leone—that utilities advise new customers firm demand for energy inputs. with moderate electricity needs to purchase their own generators instead of connecting to the grid (Geginat 2009). Electricity reliability varies considerably across economies and regions Power outages are associated with an economy’s income Infrastructure is one of the main pillars of competitiveness.4 While it encom- level—and economies that do not monitor outages tend to passes many types of facilities and systems, an economy’s electricity supply have more of them is one of the main determinants of firm productivity (Escribano and others 2009). Furthermore, productivity is boosted by reliable electricity services An economy’s income level is associated with its infrastructure development (Fedderke and Bogetic 2006; Kirubi and others 2009; Grimm and others (Calderon and others 2014). This holds true when infrastructure develop- 2012). A weak power infrastructure, however, can act as a drag on economic ment is proxied by service reliability (Figure 3). A firm operating in a growth. In Sub-Saharan Africa, economic growth is constrained by about two low-income economy in 2015 faced nearly 400 power cuts on average, while percentage points by a weak power infrastructure (Andersen and Dalgaard a firm in a high-income economy experienced about one such power cut. 2012). Moreover, the 42 economies where no SAIDI/SAIFI data are available have twice as many power outages on average7 and significantly worse service To assess the impact of electricity infrastructure on firm performance, most reliability.8 This is consistent with research suggesting that economies with studies use proxy measures of power outages. Some studies employ meteo- poor service reliability often do not record or disclose data on the rological satellite data on lighting density, while others use firm-level data performance of public infrastructure (Alcott and others 2014). Figure 1. Power outages have the longest duration in Figure 2. Power outages occur most frequently in Sub-Saharan Africa and South Asia Sub-Saharan Africa and South Asia 741 253 Average frequency of power outages in Average duration of power outages in 217 99 2015 (hours) 160 70 2015 15 12 27 26 7 12 1 1 Sub-Saharan South Asia Middle East & East Asia Latin America Europe & OECD high Sub-Saharan South Asia Middle East & Latin America East Asia Europe & OECD high Africa North Africa & Pacific & Caribbean Central Asia income Africa North Africa & Caribbean & Pacific Central Asia income Source: Doing Business database. Source: Doing Business database. Note: The sample includes 142 economies. It excludes economies where SAIDI data Note: The sample includes 142 economies. It excludes economies where SAIFI data are not available for 2015. Scale break is used on the average outage duration for are not available for 2015. Scale break is used on the average outage frequency for Sub-Saharan Africa to illustrate an outlying value. Sub-Saharan Africa to illustrate an outlying value. Affiliation: Doing Business, World Bank Group. Acknowledgement: Hulya Ulku contributed to this research note. Objective and disclaimer: Doing Business Research Notes present short analytical studies on business environment. The notes carry the name of the author and should be cited accordingly. The findings, interpretations, and conclusions are entirely those of the author. They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the governments they represent. Global Knowledge & Research Hub in Malaysia Electricity Sector Constraints for Firms Across Economies: A Comparative Analysis Figure 3. High-income economies have less outages in total duration 12 CHE NOR QAT LUX 11 AUS GNI per capita (logatirhm) DNK SWE USA SGP SMR IEU RL IS N L LD CAN AUT FIN D GB AREB EL R K JPNW FRA HK T G NZL ISR KORESP ITA R² = 0.5033 CYP 10 TWN BRN S PRT CZE VN BHR SVK SAU GRC EST MLT TTO PRI OMN URY LTU LVA BRBARG CHL SYC RUS KAZ POL HU N HRV ATG PAN PLW MEX MYS TUR BRA GAB 9 BLR LCA DMA AZE GR D CHN ROM BG R COL SUR THA ECU I R N PER DOM IRQ SR JAMM BKD JOR BIH DZA FJI B LZ TON PRY G TUEO ARM N SLV WSM MNG KSV ALB GUY LKA IDN G TM PHL EGY 8 MAR B T MDAN VUT UKR BOL WBG CPV HND SWZ NGA UZB SLB VNM PNG LAOSDNNIC IND B N ZMKE PAK CIV CMR 7 MRT KHM TZA ZWE SSD UGA BFA COM ERI GIN 6 ZAR LBR NER 5 0 1 2 3 4 5 6 7 8 9 SAIDI for 2015 (logarithm) Source: Doing Business database; World Development Indicators database (http://data.worldbank.org/data- catalog/world-development-indicators), World Bank. Note: The figure shows the average total duration of power outages over the course of 2015 for each customer served. There is a negative correlation coefficient of -0.71 between the natural logarithm of outages and the logarithm of GNI per capita. The relationship is significant at the 1% level. The sample includes 142 economies. It excludes economies where SAIDI are not available. Electricity tariffs have decreased in the past three years average at the global level. Interestingly, the Middle East and North Africa was the only region where average commercial tariffs increased as several Business performance is sensitive to the cost of indirect inputs (Eifert and economies in the Gulf lowered subsidies on domestic electricity prices due others 2008) and energy bills can constitute up to 30% of operating costs for to falling revenues from fossil fuel exports. In Kuwait, for example, electricity an average company (Jewell 2006). A limited but growing body of research tariffs for commercial clients were raised for the first time in 30 years, with examines the impact of electricity tariffs on firm behavior. Abeberese (2016), tariffs increasing over tenfold.11 for example, shows that manufacturing firms respond to exogenous tariff increases by reducing electricity consumption and switching to less energy- intensive industries. Boonzaaier and others (2015) find that electricity Electricity tariffs are likely to be lower in resource-rich econo- demand in South Africa is becoming more elastic in the wake of tariff surges mies and that eroding profits may lead to a change in investment decisions. Resource-rich economies offer the lowest commercial electricity tariffs, Doing Business measures electricity tariffs based on a set of assumptions and while economies with low natural resource endowments—and relying on a hypothetical monthly consumption. On the basis of the assumptions about power generation using imported oil and gas—have the highest tariffs. The monthly consumption, a monthly bill for a commercial warehouse in the latter group includes several small island economies12 such as the Solomon largest business city of the economy is computed. For comparability Islands (96 cents/kWh), Kiribati (46 cents) and Antigua and Barbuda (44 purposes, the price of electricity is measured in cents per kilowatt-hour cents). These economies rely almost entirely on imported fossil fuels for (kWh). The data are collected from utilities and regulatory agencies and power generation, making it expensive to produce electricity—and exposing checked against sample bills sent by private sector professionals. them to oil price shocks. Economies with the lowest tariffs also tend to Electricity tariffs per kWh varied significantly in 2016, from less than 1 cent generate electricity through fossil fuel sources but they are able to subsidize in Kuwait to 96 cents in the Solomon Islands.9 The global average was 18 domestic commercial tariffs thanks, in part, to oil and gas export revenues. cents per kWh in 2016. The data show that, unlike outages, end-user tariffs These economies include Algeria (3 cents/kWh), Bahrain (5 cents) and Qatar are not associated with the income-level of economies.10 At the regional (5 cents). Fuel exports as a percentage of total merchandise exports can be level (Figure 4), electricity tariffs are the lowest on average in the Middle East used as a proxy for natural resource endowment (Asiedu and others 2013). and North Africa (11 cents) as well as Europe and Central Asia (11 cents). This measure shows a correlation with electricity tariffs; the relationship is They are highest in East Asia and the Pacific (24 cents). Electricity tariffs also even more pronounced for economies that export a large share of fuels fluctuate from year to year. In 2016, for example, tariffs decreased by 7% on (Figure 5). Figure 4. Electricity tariffs have been declining since 2014 35 30 Electricity tariffs (cents per kWh) 25 20 15 10 5 0 East Asia & Latin America Sub-Saharan South Asia OECD high Europe & Middle East & Pacific & Caribbean Africa income Central Asia North Africa 2014 2015 2016 Source: Doing Business database. Note: The sample includes 188 economies. República Bolivariana de Venezuela is excluded as it is an outlier. Somalia is also excluded as no data for 2014 are available. 2 Doing Business Research Notes No.1 Figure 5. Low electricity tariffs are associated with high natural resource endowment 50 Electricity tariffs per kWh (cents) MDV JAM WSM 40 VC B T HS NIC S YC CP V GUY JPN BRB 30 SLE BFA DEU PRT MUS MD IN LT CRI DOM PA FJ ILN ITA CYP NER IRL KA R² = 0.1885 JOR BE N BLZ ESP GNQ SEN AUS CMR URYTGO S PAK L VGTM 20 RW KHMPHL U MKD AAFG LVA GA BEL LTU CZE USAHRV GCHE IN TZATHAMEX FIN SGP MYS BW LUX A H SVK MDG HKG UN MWICHN ISRNZLPO F T R URD ALN KGBR PER BLR GRC BOL COL BIH SW E NPL C B UKR AF DMAZW I A U T RE SVN CIV CAN RUS MDA BRA 10 CHL YE M ISLAR GROM TUN KORA BGR LB ZAF EG Y NOR KAZ AR MGEO ECU AZE IRQ MNG PRY MOZ SAU OMN BRN ETH ZMB BHR AGO KGZ QAT DZA KWT 0 0 10 20 30 40 50 60 70 80 90 100 Fuel exports (% of merchandise exports) Source: Doing Business database (tariffs are for calendar year 2015); World Bank staff estimates using data from the UN Comtrade database for 2016 (https://comtrade.un.org/), United Nations. Note: The figure compares the share of fuel exports as a percentage of merchandise exports with electricity tariffs. The sample includes 127 economies for which data are available. The correlation coefficient between tariffs and fuel exports is -0.43. The relationship is significant at the 1% level. A burdensome connection process is associated with utility Power outages are correlated with a difficult connection corruption process, while electricity tariffs are not associated with either The performance of infrastructure services is associated with the quality and measure efficiency of regulatory institutions (Kirkpatrick and others 2002; Cubbin and Historical Doing Business data show that power outages are negatively Stern 2006; Andres and others 2008). Doing Business measures the process associated with the efficiency of grid connections.14 This is consistent with of getting electricity based on the time, cost and interactions required to Geginat and Ramalho (2015), who suggest that utilities ensuring better obtain a new connection to the grid. customer service are more likely to ensure better service reliability. On the The process of getting an electricity connection varies significantly across other hand, electricity tariffs are not associated with the number or duration economies and regions (Geginat and Ramalho 2015). For example, in South of power outages.15 Electricity tariffs are subject to myriad forces beyond the Asia it takes on average 138 days to obtain a new connection—the most of control of the utilities and, as such, they do not always act as efficient price any region—compared to an average of 66 days in Latin America and the signals (Viera and others 2016). On the demand side, market size and Caribbean. The number of procedures varies widely: in the United Arab exogenous factors such as weather conditions have an impact on tariffs. On Emirates it takes three procedures to obtain a new connection compared to the supply side, tariffs are impacted by an economy’s natural resource nine procedures in Nigeria. Furthermore, it takes significantly longer to get endowment as well as the regulations it has in place. Governments may connected to the electrical grid in economies where a higher number of choose to subsidize tariffs to make electricity consumption affordable for the procedures are required. On average, it takes a customer 67 days to get population through public fund transfers or cross-subsidization between connected to the grid in economies where three procedures are required; consumer groups. Given these factors, it is not surprising that end-user this number rises to 245 days in economies where nine procedures are tariffs do not reflect an economy’s ability to meet peak demand or carry out required. connections in an efficient manner. Comparing estimates on the procedures, time and cost to connect,13 it can Electricity sector constraints are negatively associated with be observed that the higher the income group, the easier the connection process. The OECD high-income economies, for example, have the simplest measures of energy demand connection process. Another finding is that utility corruption is associated Power outages, electricity tariffs and burdensome connection procedures with more complex connection processes; there is a higher likelihood of can impact firms to varying degrees. Therefore, it is interesting to see how utility corruption in economies where there is more interaction between the these variables are associated to firms’ demand for energy inputs. The utility and customers, both in terms of time and procedures (Figure 6). International Energy Agency (IEA) provides data on electricity consumption Figure 6. The more complex the connection process, the higher the likelihood of utility corruption 8 7 N ZL FIN NOR LUX DNKSGP R² = 0.3977 IRL ISL CHEARE HKG CAN NLD EST AUS QAT GBRJPN_Osak 6 BEL GEO MKD SWE WEF utility bribe score CHLL RWA URY VA AUT SAU MNG L TU PRT ARM SV N TU DE WN CYP OMNB HR POL GTM BTN USA_Losa HUN ECU ISR FRA MYS MUS CZE 5 CIV ESP BG RB ZM AZE LKA MAR BW GRC ITA AR PE CRI TURPAN PHL KOR MNE LAO SWAZLB S YC SRB COL HRV JOR SLV GM JAMB KAZ MDA DOM IRN HND SVK NIC NAM INDTHA TUN TJK ZWE CPV CHN MLT EGY VNM ZAF 4 LBR GAB MWI SEN KHM UKR MOZ DZA MEX KWT ARG NPL ETH GHA RUC SR M PRY ID N KEN TTO BRA MDG BDI BIH BOL LBN VEN NGA UGA PAK_Laho KGZ LSO TZA 3 TCDSLE ROM GUY MLI MMR HTI BEN BGD GIN 2 MRT 1 0 20 30 40 50 60 70 80 90 100 Distance to frontier score for efficiency of connection process Source: Doing Business 2017 data; World Economic Forum 2015. Note: The sample includes 140 economies for which data are available. The correlation coefficient between the distance to frontier score for the ease of connecting (DB2017) and the WEF’s Global Competitiveness Index (GCI) score for frequency of bribe payments for calendar year 2015 is 0.63. This relationship is significant at the 1% level. The GCI utility bribe index is based on survey results from businesses which are asked how common it is to make undocumented bribes to public utilities. The index is scored from 1 to 7; a score of 1=very common and a score of 7=never occurs. 3 Electricity Sector Constraints for Firms Across Economies: A Comparative Analysis economy’s natural resource endowment, while burdensome electricity Table 1. Robust regression analysis of the percentage of firms connections are associated with utility corruption. owning a generator (city level) Consistent with the existing research, the data reveals that electricity sector constraints impact firm behavior in terms of demand for energy inputs. One (i) Generator (i) Generator major question, however, remains to be explored: how is the performance of ownership (%) ownership (%) with firms impacted by specific electricity sector constraints, namely (i) power income control outages, (ii) electricity tariffs and (iii) the connection process? This is will be Outages (log of SAIDI) 8.383*** 8.099*** explored in an upcoming policy note. (1.021) (1.254) Electricity Tariffs (cents per kWh) 0.618*** 0.610*** NOTES (0.13) (0.13) 1 IEA 2016. Ease of connecting (DTF) -0.260** -0.241* 2 World Bank Group 2015. 3 According to World Bank Enterprise Surveys data, over 11% of business owners in (0.13) (0.12) developing economies perceive a lack of reliable electricity supply as their biggest Income (log of GNI) - (1.19) obstacle, behind access to finance (15%), the informal sector (12%) and political instability (12%). For more, see the website at http://www.enterprisesurveys.org. - (2.01) 4 According to the World Economic Forum (WEF), infrastructure is one of four pillars that are Observations 96 96 “basic requirements” for global competitiveness. 5 Economies where two cities are measured are Bangladesh, Brazil, China, India, Indonesia, R-squared 0.661 0.662 Japan, Mexico, Nigeria, Pakistan, the Russian Federation and the United States. 6 According to World Bank Enterprise Surveys, electricity is perceived as major constraint to Standard errors in parentheses doing business by 39% of firms in Sub-Saharan Africa, 41% in the Middle East and North *** p<0.01, ** p<0.05, * p<0.1 Africa and 46% in South Asia. These represent the highest share of all regions. 7 Controlling for sample size and the city considered, and for 35 economies where no SAIFI Source: Doing Business database; Enterprise Surveys database (http://www.enter data are available, there is an average of 131 outages a year, according to data from the prisesurveys.org), World Bank. World Bank Enterprise Surveys. For a sample of 103 economies where SAIFI is available, on Note: Sample excludes all World Bank Enterprise Surveys data prior to 2010. average 53 outages a year are reported. 8 Controlling for sample size, and for 24 economies where no SAIFI data are available, the average World Economic Forum 2015 Global Competitiveness Index (GCI) score on the per capita. A regression analysis of the IEA measure of firm consumption quality of electricity supply is 2.4. In contrast, for a sample of 116 economies where SAIFI against the Doing Business electricity infrastructure measures shows that data are available, the score is 4.9. both power outages and electricity tariffs are negatively associated with 9 República Bolivariana de Venezuela is excluded from the sample. 10 The correlation coefficient between tariffs and GNI is 0.12. The relationship is not consumption.16 This may suggest that electricity demand is not inelastic and significant. tariff levels impact the consumption levels of firms. 11 As of May 16, 2016, electricity tariffs for commercial users in Kuwait were raised from 2 fils/kWh (0.7 cents) to 25 fils/kWh. The percentage of firms using a generator, as reported by World Bank 12 Eleven of the 15 economies where commercial electricity tariffs are the highest are islands Enterprise Surveys data, is another measure of energy demand. It could be economies with less than one million in population. These are Antigua and Barbuda; Cape surmised that—where electricity services are unsatisfactory for businesses Verde; Dominica; Kiribati; Marshall Islands; Micronesia, Fed. Sts.; Samoa; Seychelles; (including reliability of electricity supply, tariffs and efficiency of connection Solomon Islands; Tonga; and Vanuatu. 13 The distance to frontier score is calculated. For each economy, the number of procedures, to the grid)—more firms rely on off-grid solutions; this is corroborated by time and costs are normalized to a common unit from 0 to 100, where 0 represents the regression results.17 For example, when income is controlled for, a 1% lowest performance and 100 represents the frontier. An average of these three scores is increase in the level of outages is associated with an 8 percentage point then computed and used for the analysis. increase in the share of businesses owning a generator. Similarly, a 1 cent 14 The correlation coefficient between the distance to the frontier ease of connection score (2014-2016 average) and SAIDI (2013-2015 average) is -0.52. This relationship is significant increase in electricity tariffs is associated with an increase of 0.6% in the at the 1% level. share of generator ownership, while a 1% increase in the efficiency of 15 The correlation coefficient between electricity tariffs (2014-2016 average) and SAIDI electricity connection is associated with a 0.26 percentage points decrease (2013-2015 average) is -0.10. in the share of firms owning a generator. More burdensome connection 16 While the IEA measures aggregate electricity consumption for households and businesses, procedures are associated with more firms having self-supply capacity. according to the US Energy Information Association commercial and industrial customers account for 60% of electricity consumption. Using electricity consumption as the dependent variable, outages (log of SAIDI) and electricity tariffs are found to be negatively Conclusion associated to consumption. These relationships are significant at the 1% level when controlling for income. A robust regression is used to minimize the impact of outliers. Using recent Doing Business data, this note provides an analysis of electricity 17 The regression uses three-year averages for Doing Business data as the survey year for the sector constraints across economies. The findings provide some interesting World Bank Enterprise Surveys varies from economy to economy. World Bank Enterprise Surveys data prior to 2010 are excluded and the weighted aggregates for each major insights. Not surprisingly, service unreliability is a significant factor in business city are selected. For outages, a logarithm of SAIDI is used to normalize data. The low-income economies, where power outages fluctuate significantly from results presented below show little change when SAIFI is used. A robust regression is used year to year. Furthermore, electricity tariffs are associated with an to minimize the impact of outliers. References Abeberese, A. B. 2016. “Electricity Cost and Firm Performance: Evidence from India.” Review of Grimm, M., R. Hartwig, and J. Lay. 2012. “How Much Does Utility Access Matter for the Performance Economics and Statistics. of Micro and Small Enterprises?” Available from http://bit.ly/1Ge4hlF. Adair-Rohani, H., K. Zukor, S. Bonjour, S. Wilburn, A. C. Kuesel, R. Hebert, and E. R. Fletcher. 2013. IEA (International Energy Agency). 2016. World Energy Outlook 2016. Paris, France: International “Limited Electricity Access in Health Facilities of Sub-Saharan Africa: a Systematic Review of Energy Agency. Data on Electricity Access, Sources, and Reliability.” Global Health: Science and Practice 1(2): Jewell, M. 2006. “Connecting the Dots: Wasted Energy Everywhere.” Energy & Power Management, 249-261. June, p.8. Alcott, H., A. Collard-Wexler, and S. O’Connell. 2014. “How do Electricity Shortages Affect Productivity? Kirkpatrick, C., D. Parker, and Y. Zhang. 2002. “Electricity Sector Reform in Developing Countries: An Evidence from India.” NBER Working Paper W19977, National Bureau of Economic Research, Econometric Assessment of the Effects of Privatisation, Competition and Regulation.” Centre on Cambridge, MA. Regulation and Competition (CRC) Working Paper 30593, University of Manchester, Institute for Andersen, T. B., and Dalgaard, Carl-Johan. 2013. “Power Outages and Economic Growth in Africa”. Development Policy and Management (IDPM). Energy Economics 38 (2013): 19-23. Khandker, S. R., H. A. Samad, R. Ali, and D. F. Barnes. 2014. “Who Benefits Most from Rural Andres, L., J. L. Guasch, and S. Lopez Azumendi. 2008. “Regulatory Governance and Sector Electrification? Evidence in India.” The Energy Journal 35(2). Performance: Methodology and Evaluation for Electricity Distribution in Latin America.” Policy Research Working Paper 4494. World Bank, Washington, DC. Kirubi, C., A. Jacobson, D. M. Kammen, and A. Mills. 2009. “Community-Based Electric Micro-Grids Can Contribute to Rural Development: Evidence from Kenya.” World Development 73: 1208–21. Asiedu, E. 2013. “Foreign Direct Investment in Africa: The Role of Natural Resources, Market Size, Government Policy, Institutions and Political Instability.” SSRN Electronic Journal. Mutale, A. 2015. “Zambia: Hello darkness.” Financial Mail. May 21. http://www. financialmail.co.za/features/2015/05/21/zambia-hello-darkness. Calderón, C., Moral-Benito, E., & Servén, L. (2014). Is infrastructure capital productive? A dynamic heterogeneous approach. Journal of Applied Econometrics, 30(2), 177-198. OECD (Organisation for Economic Cooperation and Development). 2012. Financing SMEs and Entrepreneurs: Understanding and Developing an OECD Scoreboard. Paris, France: OECD Cubbin, J., and J. Stern. 2006. “The Impact of Regulatory Governance and Privatization on Electricity Publishing. Industry Generation Capacity in Developing Economies.” The World Bank Economic Review Advance Access. World Bank, Washington, DC. UN (United Nations). Commodity Trade Statistics Database (database). New York, NY, https://comtrade.un.org. Eifert, B., A. Gelb, and V. Ramachandran. 2008. “The Cost of Doing Business in Africa; Evidence from Enterprise Survey Data.” World Development 36(9): 1531-1546. Viera, X., Hochstetler, R., Mello, J.C., and L.A Barroso. (2016). Aligning Regulatory Incentives and Price Signals in the Brazilian Wholesale and Retail Electricity Market. Cigre. CE-025. Paris. Escribano, A., J. L. Guasch, and J. Pena. 2009. “Assessing the Impact of Infrastructure Constraints on Firm Productivity in Africa.” Working Paper 9, Africa Infrastructure Sector Diagnostic, World World Bank. Doing Business Indicators (database). Washington, DC (accessed February 22, 2017), Bank, Washington DC. http://www.doingbusiness.org/. Fedderke, J., and Z. Bogetic. 2006. “Infrastructure and Growth in South Africa: Direct and Indirect World Bank. Enterprise Surveys (database). Washington, DC (accessed February 22, 2017), Productivity Impacts of Nineteen Infrastructure Measures.” Policy Research Working Paper, https://www.enterprisesurveys.org/Data. World Bank, Washington DC. World Bank. World Development Indicators (database). Washington, DC (accessed February 23, Foster, V., and J. Steinbuks. 2010. “When do Firms Generate? Evidence on In-House Electricity Supply 2017), http://data.worldbank.org/data-catalog/world-development-indicators. in Africa.” Energy Economics 32(2010): 505-514. World Bank Group. 2015. Main Report. Independent Evaluation Group. Washington, DC: World Bank Geginat, C. 2009. Getting Electricity: A Pilot Indicator set From the Doing Business Project. Doing Group. http://documents.worldbank.org/curated/en/416421468196746577/Main-report. Business, Washington, DC: World Bank Group. WEF (World Economic Forum). 2016. The Global Competitiveness Report 2016-17. Geneva, Geginat, C., and R. Ramalho. 2015. “Electricity Connections and Firm Performance in 183 Countries.” Switzerland (retrieved February 22, 2017), https://www.weforum.org/reports/the-global- Global Indicators Group, World Bank Group, Washington, DC. competitiveness-report-2016-2017-1. 4 Global Knowledge & Research Hub in Malaysia