2019/101 Supported by K NKONW A A WELDEGDEG E OL N ONTOET E S ESREI R E ISE S F OFRO R P R&A C T HTEH E NEENREGRYG Y ETX ITCREA C T I V E S G L O B A L P R A C T I C E THE BOTTOM LINE How Do Enterprises Benefit from Grid Connection? Many of the enterprises that make up Bangladesh’s dominant The Case of Rural Electrification in Bangladesh nonfarm sector (85 percent of GDP) are located in rural areas (World Bank 2015). Electrifying Does electrification increase productivity? How much progress has been made with rural them can be a major driving force Cross-country evidence suggests that it does electrification in Bangladesh? behind economic growth. This A dense network of rural electric cooperatives Rural electrification increases enterprise productivity in several ways: study is a product of the Status of Energy Access Report (SEAR), It allows businesses to operate well into the evening, giving them has been very successful in expanding access to more hours per day to generate sales and profits; it creates a more electricity an initiative of the Energy Sector comfortable work environment; and it allows enterprises to use Management Assistance Program Responsibility for providing electricity in rural areas falls under the machinery and tools that are more efficient and productive than their at the World Bank. By documenting purview of the Bangladesh Rural Electrification Board (BREB), a nonelectrical counterparts. the effects of energy access semi-autonomous government agency established in 1977 within In rural Bangladesh, output increased by 78 percent in grid-con- programs on beneficiaries’ welfare, the Ministry of Power. Based on the U.S. model of consumer-owned nected industries but only 8 percent in unconnected industries SEAR enables policy makers and rural electric cooperatives, BREB organizes prospective consumers over a five-year period (Barkat and others 2002). Grid-connected other stakeholders to understand into autonomous rural electric cooperatives called palli bidyut samitis industries generated 11 times more employment than unconnected better the linkage between energy that own, operate, and manage rural distribution systems within their industries; profits were higher (35 percent of sales in grid-connected interventions and benefits and to areas of jurisdiction. BREB monitors the operation, administration, industries and just 16 percent in unconnected enterprises); and make informed decisions about and financial management of the cooperatives. turnover at grid-connected retail shops was more than twice that of projects. The 80 cooperatives in Bangladesh are active in more than unconnected shops. Hussain Samad is a In rural Tanzania, grid connectivity increased the financial and 75,000 villages in 61 districts. Serving 12 million residential, 1.5 million consultant in the Energy physical assets of enterprise owners, and people in the community commercial, and more than 160,000 industrial customers, they have Sector Management been highly successful in expanding access to electricity (figure 1). benefitted from the use of electrical machinery (Maleko 2005). Assistance Program at the Average system loss is less than 12 percent (much less than the World Bank, where he acts as In small carpentry and tailoring shops in rural Kenya, adoption of electricity increased productivity per worker, lowered unit prices, national average), and the bill collection rate exceeds 99 percent.1 principal investigator of the Multi-Tier Framework Global Survey activities. and increased gross daily revenues (Kirubi 2006). Some agro-based industries also benefited from electrification. Maize processed by an Elisa Portale is a senior electric milling machine fetched a much higher price than unpro- energy specialist in the cessed maize, and electric-powered cold storage allowed farmers to 1 The cooperatives’ financial sustainability is not always guaranteed. During periods of Energy Sector Management Assistance Program at the preserve meat and milk longer and command prices that were four delayed tariff adjustments, operations and maintenance and debt service suffer. Commercial World Bank, where she is times higher over time. customers are customers engaged primarily in retail business (trade) or services. Industrial responsible for coordinating ESMAP’s customers are engaged in production and processing activities. How Do Enterprises Benefit from Grid Connection? knowledge products. Rural Electrification in Bangladesh 2 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh Figure 1. Annual and cumulative number of connections of BREB commercial and industrial customers, 2004/05–2016/17 Commercial customers Industrial customers 1,400 180 Number of connections Number of connections 1,200 160 140 (thousands) 1,000 (thousands) 120 800 100 Cumulative connections The 80 rural electric 600 80 Annual connections added 60 400 cooperatives in Bangladesh 200 136 146 40 68 59 27 27 18 37 20 32 68 20 12 26 have been highly 0 0 5 6 7 8 9 0 1 2 3 4 5 6 7 5 6 7 8 9 0 1 2 3 4 5 6 7 /0 /0 /0 /0 /0 /1 /1 /1 /1 /1 /1 /1 /1 /0 /0 /0 /0 /0 /1 /1 /1 /1 /1 /1 /1 /1 successful in expanding 04 05 06 07 08 09 10 11 12 13 14 15 16 04 05 06 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 access to electricity. Source: Information obtained by authors from Chief Engineer’s Office, BREB. How are the benefits of electrification measured? • Inverse probability-weighted FE estimation (IPW-FE): IPW-FE is a two-step process that controls for time-varying unob- Several techniques were applied to gauge the gains served factors. In the first step, the probability of grid connectivity from electrification of each enterprise (called the propensity score) is calculated An impact evaluation compares outcomes between a treatment using the first-round data based on a wide range of control group (enterprises that are connected to the grid) and a control variables.3 The probability is measured for both the treatment and group (enterprises not connected to the grid), called the counter- the control groups. In the second stage, the propensity score is factual. Determining the counterfactual is at the heart of any impact used to create a weight, which is then used in the FE regression evaluation endeavor; different evaluation techniques make different (similar to the one described earlier) to estimate impacts.4 IPW-FE assumptions about how to construct it. provides more unbiased estimates of impacts than does simple Two evaluation techniques were used to measure the impact of FE, and thus is preferable to it. electrification in rural Bangladesh. • Fixed-effects (FE) method: In its simplest form, the FE The data for this study came from two panel surveys of commer- method consists of first measuring the difference in the outcome cial and industrial enterprises carried out in 2005 and 2010 in rural variables between two periods for the treatment and control Bangladesh. groups and then taking a second difference between the two More than 3,000 commercial enterprises were surveyed in 2005, differences measured earlier (a technique known as “difference of which about 2,000 were resurveyed in 2010. After data cleaning of differences” modeling). The second difference is the estimated 1,571 commercial units were deemed suitable for panel data impact. In a more advanced form, regression is used to control for factors other than the intervention that might have influenced 3 Control variables include sex of the enterprise owner, sole proprietorship, size of the enter- outcomes. FE controls for unobserved factors as long as they do prise, years of operation, enterprise capital, share of hired labor, operating cost, use of alternate not change between survey periods.2 sources of electricity (such as generator sets) or fuels by the enterprise (kerosene, for example), prices of alternate fuels, and so on. 2 Unobserved factors are factors that may influence the outcomes but that are not (or cannot 4 The weight is the inverse of the probability of grid connectivity. That is, it is given by be) measured. Examples include the entrepreneurship skills of the enterprise owner or foresight w = 1 , for grid-connected enterprises, and w = 1 for enterprises not connected to the grid, p (1–p) about market trends. where p is the probability of grid-connectivity. 3 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh Table 1. Electrification rate in commercial and industrial analysis. Similarly, about 1,600 industrial enterprises were surveyed enterprises (percent) in 2005, of which about 1,175 were resurveyed in 2010. After data cleaning, about 1,025 industrial enterprises remained for analysis. Commercial enterprises Industrial enterprises The surveys collected information on basic characteristics of Division 2005 2010 2005 2010 commercial and industrial enterprises (type of activity and owner- The electrification rate Barisal 55.6 67.5 88.6 90.1 ship, years of operation, months of operation per year, capital assets, increased between 2005 Chittagong 52.4 69.2 85.6 92.6 labor use, operating cost) as well as information on their energy use Dhaka 54.3 64.0 86.7 97.0 (source, type of use, and consumption of electricity). In addition, for and 2010, rising from grid-connected enterprises, information on reliability and quality Khulna 55.6 85.9 94.2 95.3 52 percent to 69 percent of service was collected. The surveys covered 287 villages from Rajshahi 49.5 67.1 88.9 93.3 for commercial enterprises, 40 districts in 6 divisions. Sylhet 41.2 57.1 80.9 83.8 and from 88 percent to The data revealed that the electrification rate increased between Total 52.0 69.0 88.2 93.4 2005 and 2010, rising from 52 percent to 69 percent for commercial 93 percent for industrial Source: BREB surveys 2005, 2010. enterprises, and from 88 percent to 93 percent for industrial enter- enterprises. prises (table 1). Table 2. Revenues, profits, and profit margins of grid-connected and nonconnected commercial enterprises, 2010 How did electrification affect enterprises in rural Statistical Bangladesh? Enterprises Enterprises significance with grid without grid of the Electrification effects on revenues, profits, and Item connection connection difference profit margins varied with the type and size of the Revenue (thousand Tk/year) Trading enterprises 1,097.1 753.4 *** enterprise and with the reliability of service Service enterprises 199.1 233.9 — We begin with a short review of descriptive statistics before moving All enterprises 982.0 689.4 *** on to a discussion of the results of impact evaluations conducted Profit (thousand Tk/year) using the rigorous techniques discussed in the previous section. Trading enterprises 381.0 318.4 * The revenue and profit of grid-connected commercial enterprises were mostly higher than those of unconnected enterprises, and the Service enterprises 89.0 70.6 ** differences were statistically significant (table 2). In contrast, the All enterprises 343.5 287.9 * profit margin of unconnected commercial enterprises was higher Profit margin than that of grid-connected enterprises, although the difference was Trading enterprises 0.294 0.298 — not statistically significant. Service enterprises 0.438 0.486 — Among small and micro industrial enterprises, grid-connected All enterprises 0.312 0.320 — enterprises performed better than unconnected enterprises (table 3). Source: BREB surveys 2010. In contrast, for medium-size and large industrial enterprises, ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. unconnected enterprises performed better. Most of the differences — no statistical significance. 4 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh Table 3. Revenues profits, and profit margins of grid-connected Table 4. Impact of grid electrification on revenues, profits, and and nonconnected industrial enterprises, 2010 profit margins of commercial enterprises Enterprises Enterprises Statistical Item FE estimates IPW-FE estimates with grid without grid significance of Revenue (percent change) 25.3*** 28.0*** Item connection connection the difference Profit (percent change) 35.0** 39.1** Causality between grid Revenue (thousand Tk/year) Profit margin (change in — 14* Small and 1,313 697 *** electrification and the percentage points) microenterprises observed outcomes Medium-size and 8,840 12,976 ** Source: BREB surveys 2005, 2010. ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. can be ascertained only large enterprises — no statistical significance. All enterprises 2,933 1,948 *** through rigorous impact Profit (thousand Tk/year) evaluations. Impacts varied Table 5. Impact of grid electrification on revenues, profits, and Small and 653 263 *** by the type of activity and microenterprises profit margins of industrial enterprises the size of the enterprises. Medium-size and 6,012 9,979 — Item FE estimates IPW-FE estimates large enterprises Revenue (percent change) 35.7* 43.3* All enterprises 1,807 1,264 ** Profit (percent change) 23.1* 32.1* Profit margin Profit margin (change in — — Small and 0.30 0.23 ** percentage points) microenterprises Source: BREB surveys 2005, 2010. Medium-size and 0.34 0.43 — ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. large enterprises — no statistical significance. All enterprises 0.31 0.25 * Source: BREB surveys 2010. ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. — no statistical significance significant, whereas the IPW-FE estimates found that profit margins were 14 percentage points higher for grid-connected enterprises. The revenues and profits of industrial enterprises also increased sub- in outcomes between grid and unconnected enterprises were stantially after connection to the grid—by 43 percent and 32 percent, statistically significant. respectively based on IPW-FE estimates (table 5). Grid access did not While descriptive statistics, such as those reported in tables 2 affect the profit margins of industrial enterprises, however—possibly and 3, are indicative of the status of outcomes of grid-connected because of diminishing returns on profits as revenue grows. enterprises relative to unconnected ones, they do not imply causality Impacts varied by the type of activity and the size of the between grid electrification and the observed outcomes. Causality can enterprises. The IPW-FE estimation revealed that grid connectivity be ascertained only through rigorous impact evaluations, using tech- appeared to have had a greater effect on the revenue of trading niques such as those described in the previous section. Findings from enterprises (27.1 percent increase) than service enterprises evaluations using the FE and IPW-FE methods are discussed below. (17.6 percent increase) but that profits increased more for service Grid electrification raised the revenue of commercial enterprises enterprises (46.1 percent) than for trading enterprises (31.6 percent). by 25–28 percent and increased their profits by 35–39 percent The effect of grid connection on profit margins was the same for (table 4). For profit margins, the FE estimates were not statistically trading and services enterprises (table 6). 5 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh Table 6. Impact of grid electrification on revenues, profits, and Table 8. Average daily duration of power outages (hours) in profit margins of trading and services enterprises commercial and industrial enterprises, 2005 and 2010 Item FE estimates IPW-FE estimates Enterprise type 2005 2010 Revenue (percent change) Commercial 7.3 5.6 Trading enterprises 27.5*** 27.1*** Industrial 7.0 5.2 Grid electrification Service enterprises 20.2** 17.6** Source: BREB surveys 2005, 2010. raised the revenue of Profit (percent change) commercial enterprises Trading enterprises 37.0** 31.6** Among industrial enterprises, medium-size and large firms by 25–28 percent and Service enterprises 43.4 *** 46.1** gained more from grid electrification than did small and microen- increased their profits Profit margin (change in percentage points) terprises (table 7). This is not surprising. Medium-size and large by 35–39 percent. The Trading enterprises — 18* enterprises use higher-capacity machinery, consuming more power Service enterprises 18* 18* than smaller and microenterprises. One would therefore expect them revenues and profits of Source: BREB surveys 2005, 2010. to make more productive and efficient use of electricity. industrial enterprises also ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. Grid connectivity alone may not be enough to yield the maximum increased substantially — no statistical significance. benefits from electrification. Lack of reliability and service interrup- after connection to the tions increase the cost of using electricity, as enterprises are forced grid—by 43 percent and to rely on alternate sources of energy or back-up generators. So, it is Table 7. Impact of grid electrification on industrial enterprises, important to look at the effects of the reliability of electricity service 32 percent. by size on the outcomes. The duration of daily power outages declined between 2005 and Item FE estimates IPW-FE estimates 2010 (table 8). Still, firms faced more than five hours of outages a day. Revenue (percent change) As tables 9 and 10 show, improving the reliability of service can Small and microenterprises 37.6* — substantially increase the benefits of electrification for enterprises. Medium-size and large 46.9** 45.2** For each additional hour of power outage, revenues of commercial enterprises enterprises fall by 4.6 percent and profits by 6.0 percent. In the Profit (percent change) absence of power outages, grid electrification can increase com- Small and microenterprises 26.4* 25.7* mercial enterprises’ revenues by 48 percent and profits by almost Medium-size and large 22.7** 33.7** 60 percent (table 9). For industrial enterprises, the gains associated enterprises with reliability are even greater: 59 percent for revenue and 60 per- Profit margin (change in percentage points) cent for profit (table 10) Small and microenterprises — — Medium-size and large 0.12* 0.22* enterprises Source: BREB surveys 2005, 2010. ** and * indicate statistical significance of 5 and 10 percent, respectively. — no statistical significance. 6 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh Table 9. Impact of service reliability on commercial enterprises Table 10. Impact of service reliability on industrial enterprises Item FE estimates IPW-FE estimates Item FE estimates IPW-FE estimates Revenue (percent change) Revenue (percent change) Impact with no outage 56.8*** 47.5** Impact with no outage 66.6* 68.9* Impact of one hour of outage −4.2** −4.6** Impact of one hour of outage −6.7* −7.6** Improving the reliability of Profit (percent change) Profit (percent change) service can substantially Impact with no outage 64.4*** 59.5 Impact with no outage 65.2* 59.7* increase the benefits Impact of one hour of outage −4.7* −6.0** Impact of one hour of outage −7.4** −6.3** of electrification. In the Profit margin (change in percentage points) Profit margin (change in percentage points) absence of power outages, Impact with no outage — 19.5* Impact with no outage — — grid electrification can Impact of one hour of outage — −1.4** Impact of one hour of outage — — increase commercial Source: BREB surveys 2005, 2010. Source: BREB surveys 2005, 2010. ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. ***, ** and * indicate statistical significance of 1, 5, and 10 percent, respectively. enterprises’ revenues by — no statistical significance. — no statistical significance. 48 percent and profits by almost 60 percent. For industrial enterprises, What are the policy implications? Off-grid electrification should be promoted in areas where the gains associated with extension of the grid is not feasible. Bangladesh’s Infrastructure Policy makers can multiply the beneficial effects Development Company Limited (IDCOL) has been promoting solar reliability are even greater. of electricity by expanding grid coverage, increasing mini-grids to serve household and enterprise energy needs in grid reliability, and promoting off-grid electrification remote and isolated areas, especially on riverine islands. Seventeen where appropriate solar mini-grid projects are in operation, and IDCOL has a target of installing 200 solar mini-grids by 2025. In 2013, only about one million of the Bangladesh’s three million rural Energy from solar mini-grids is expensive—more than Tk 30/kWh, enterprises were connected to the electrical grid.5 This rate has to be compared with an average of Tk 8/kWh for grid energy. Instead of raised. connecting to a mini-grid, some enterprises can use large stand- Increasing grid reliability will require more transmission and alone solar panel systems, which are now feasible because of the distribution lines. During 2009–18, about 200,000 kilometers of sharp decline in the price of solar panels and storage technology. transmission line were installed, and 370 substations (with a total These systems can charge electric vehicles and power solar freezers capacity of about 5,500 MVA) were built. BREB has recently under- for dairy farms, agricultural produce storage, fish storage, and egg taken 15 projects to improve transmission and distribution. These are incubators. positive signs. 5 Some unconnected establishments have access to solar home systems or solar mini-grids. The actual electrification rate is therefore higher than the grid connectivity rate. 7 H o w D o E n terpr i ses B e n e f i t f r o m G r i d C o n n ect i o n ? R u ra l E l ectr i f i cat i o n i n B a n g l adesh References Maleko, Godwin C. 2005. “Impact of Electricity Services on Micro- MAKE FURTHER Barkat, A., M. Rahman, S. Zaman, A. Podder, S. Halim, N. Ratna, Enterprise in Rural Tanzania.” University of Twente, Department of CONNECTIONS Energy and Sustainable Development, Enschede, Netherlands. M. Majid, A. Maksud, A. Karim, and S. Islam. 2002. “Economic World Bank. 2015. “Bangladesh: More and Better Jobs to Accelerate and Social Impact Evaluation Study of the Rural Electrification Live Wire 2014/9. “Tracking Access Shared Growth and End Extreme Poverty. A Systematic Country to Electricity,” by Sudeshna Ghosh Program in Bangladesh,” Report to the National Rural Electric Diagnostic.” Discussion Draft. Banerjee and Elisa Portale. Cooperative Association International. Dhaka: Human Development Research Centre. Live Wire 2014/21. “Scaling Up The authors thank Raihan Elahi (Lead Energy Specialist, GEE01), and Zubair Access to Electricity: The Case of Kirubi, C. 2006. “How Important Is Modern Energy for Micro- K. M. Sadeque (Senior Energy Specialist, GEE08) for reviewing drafts of this Bangladesh,” by Zubair Sadeque, enterprises? Evidence from Rural Kenya,” Thesis, Master of Live Wire, as well as S. M. Zafar Sadeque (BREB Board Member) for providing Dana Rysankova, Raihan Elahi, and Science in Energy and Resources, University of California, information on the expansion of Bangladesh’s network of rural electric Ruchi Soni. Berkeley. cooperatives. Live Wire 2014/32. “Tracking Progress Toward Sustainable Energy for All in South Asia,” by Elisa Portale and Joeri de Wit. Live Wire 2017/76. “Increasing the Potential of Concessions to Expand Rural Electrification in Sub-Saharan Africa,” by Richard Hosier, Morgan Bazilian, and Tatia Lemondzhava. Live Wire 2017/80. “Promoting Productive Uses of Electricity in Rural Electrification Programs: Experience from Peru,” by Janina Franco, V. Susan Bogach, Inés Pérez Arroyo, and Maite Lasa. Live Wire 2018/93. “Making a Difference in People’s Lives? Rural Electrification in the Lao People’s Democratic Republic,” by Hussain Samad and Elisa Portale. Live Wire 2019/97. “Investing in Mini Grids Now, Integrating with the Main Grid Later: A Menu of Good Policy and Regulatory Options,” by the Global Facility on Mini Grids. Find these and the entire Live Wire archive at https://openknowledge. worldbank.org/handle/10986/17135. Get Connected to Live Wire Live Wire briefs are The Live Wire series of online knowledge notes, an initiative of the World Bank Group’s designed for easy reading Energy and Extractives Global Practice, offers rich insights from project and analytical work on the screen and for done by the World Bank Group. downloading and self-printing “Live Wire is designed in color or black and white. Every day, Bank Group experts apply their knowledge and expertise to solve practical problems in for practitioners, policy client countries. Live Wire captures the rich insights gained in the field, allowing authors to share For World Bank Group their findings with other practitioners, policy makers, and planners. employees: Professional makers, and planners printing can be done on a inside and outside the Shouldn’t you be connected to Live Wire? customized basis for meetings World Bank Group. and events by contacting Since 2014, the 80 briefs in the series have dealt with vital topics such as energy demand and GSDPM Customer Service It is a resource to supply; renewable energy; energy efficiency; energy policy; economic growth; environmental Center at (202) 458-7479, or protection; climate change mitigation; power systems; rural and urban development; access to share with clients, energy; infrastructure economics; private sector participation; access to finance; and regulation. sending a written request to cgsdpm@worldbank.org. colleagues, and • Topic briefs offer technical knowledge on key energy issues. counterparts.” • Case studies highlight lessons from experience in implementation, often with insights from private sector engagement. • Briefs on global trends provide analytical overviews of key energy data and developments. • Bank views portray the Bank Group’s energy and extractives sector activities. The format is accessible, rigorous, and concise enough to be easily shared. The 4–12 pages of each brief make ample use of graphics. Briefs are peer-reviewed by seasoned practitioners within the World Bank Group and professionally edited and produced. While their main channel of dissemination is online, Live Wires are available in print-ready files for specific client needs. Please visit the World Bank Group’s Open Knowledge Repository to browse the Live Wire collection and download the issues important to you: www.worldbank.org/energy/livewire