The World Bank Asia Sustainable and Alternative Energy Program 81485 S o u t h A sia E n er g y S t u d ies Lighting Rural India: Load Segregation Experience in Selected States February 2014 Copyright © 2014 The International Bank for Reconstruction and Development/ The World Bank Group 1818 H Street, NW Washington, DC 20433 USA All rights reserved. First printing: February 2014 Manufactured in the United States of America. Photo credits: The World Bank The World Bank Asia Sustainable and Alternative Energy Program S o u t h A sia E n erg y S t u d ies Lighting Rural India: Load Segregation Experience in Selected States February 2014 Ashish Khanna, Mohua Mukherjee, Sudeshna Ghosh Banerjee, Kavita Saraswat, and Mani Khurana Copyright © 2014 The International Bank for Reconstruction and Development / The World Bank Group 1818 H Street, NW Washington, DC 20433, USA All rights reserved First printing: February 2014 Manufactured in the United States of America. Please cite this report as follows: ASTAE (Asia Sustainable and Alternative Energy Program). 2014. Lighting Rural India: Load Segregation Experience in Selected States. South Asia Energy Studies. Washington, DC: World Bank. Photo credits: The World Bank. The findings, interpretations, and conclusions expressed in this report are entirely those of the authors and should not be attributed in any manner to the World Bank, or its affiliated organizations, or to members of its board of executive directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. Contents Foreword ....................................................................................................................................vii Acknowledgments ....................................................................................................................viii Acronyms and Abbreviations ....................................................................................................ix Units of Measure ..................................................................................................................................................... ix Currency Equivalents ............................................................................................................................................... ix Executive Summary ...................................................................................................................xi 1 Introduction ............................................................................................................................1 Study Background and Objectives ............................................................................................................................1 Study Method ...........................................................................................................................................................2 Framework for Evaluation ..................................................................................................................................2 Selected States for Study .................................................................................................................................2 Structure of This Report ............................................................................................................................................4 2 Key Findings from Selected States ......................................................................................5 Institutional Evaluation: Project-Cycle Framework ....................................................................................................5 Conceptualization...............................................................................................................................................5 Execution...........................................................................................................................................................5 Evaluation...........................................................................................................................................................7 Financial Evaluation ..................................................................................................................................................7 Subdivision Findings ..........................................................................................................................................7 Aggregate Findings: Distribution Utility Level ...................................................................................................9 Aggregate Findings: State Level .....................................................................................................................10 Economic Evaluation ..............................................................................................................................................13 Quality of Supply .............................................................................................................................................13 Socioeconomic Benefits .................................................................................................................................15 Summary Remarks .................................................................................................................................................16 Appendix: Haryana’s Monitoring and Evaluation Experience .................................................................................17 3 Observations on Perceptions and Emerging Lessons ......................................................21 Perceptions and Observations ...............................................................................................................................21 Emerging Lessons ..................................................................................................................................................24 4 Toward Sustainable Models: Guidance Note ....................................................................25 Institutional Framework ..........................................................................................................................................25 Conceptualization and Planning ..............................................................................................................................26 Conceptualization ............................................................................................................................................26 Planning ...........................................................................................................................................................29 Execution ................................................................................................................................................................30 Monitoring and Evaluation ......................................................................................................................................30 Summary Remarks .................................................................................................................................................31 Appendix: Consultations Held in Gujarat, Rajasthan, and Madhya Pradesh ...........................................................32 iii iv Contents 5 The Way Forward .................................................................................................................37 Annex. State Profiles .................................................................................................................39 Boxes 1.1 Profile Highlights of the Selected States...........................................................................................................3 3.1 Advanced Metering Infrastructure...................................................................................................................21 Figures ES.1 Changes in Quality of Supply, Vinchiya Subdivision, Gujarat........................................................................... xiii ES.2 Changes in Quality of Supply, Bassi Subdivision, Rajasthan........................................................................... xiii ES.3 Proposed Institutional Framework for Rural Load Segregation........................................................................ xv 1.1 Institutional Evaluation Framework....................................................................................................................2 2.1 Financial Framework for Cost Benefit Analysis..................................................................................................7 2.2 Trends of Commercial Parameters, Vinchiya Subdivision...................................................................................8 2.3 Trends of Commercial Parameters, Bassi Subdivision.......................................................................................8 2.4 Trend of Commercial Parameters, PGVCL Rural Subdivisions...........................................................................9 2.5 Trend of Commercial Parameters, Jaipur District Circle...................................................................................10 2.6  Monthly Averages of Maximum and Minimum Demand, Showing Difference in Annual Averages, Gujarat................................................................................................................................. 11 2.7 Peak Demand Growth...................................................................................................................................... 11 2.8 Cash Profit/Loss and Subsidy Trend, Gujarat....................................................................................................12 2.9  Drivers of Financial Losses for Distribution Utilities, Rajasthan.......................................................................12 2.10 Economic Framework for Cost Benefit Analysis..............................................................................................13 2.11 Quality-of-Supply Changes Reported by Survey Respondents, Vinchiya..........................................................14 2.12 Quality-of-Supply Changes Reported by Survey Respondents, Bassi..............................................................14 2.13 Trend in Distribution Transformer Damage, Bassi.............................................................................................14 2.A1 Implementing the Management Information System, Haryana.......................................................................18 2.A2 Screenshot of MIS Application, Haryana..........................................................................................................18 2.A3 Screenshot of MIS-Generated Report, Haryana...............................................................................................19 4.1 Proposed Institutional Framework for Rural Load Segregation........................................................................25 A.1 Energy Input and Sales Trend, Andhra Pradesh................................................................................................40 A.2 Recent Trend in State Subsidies and Financial Losses, Andhra Pradesh..........................................................40 A.3 Consumer and Consumption Mix of Distribution Utilities, Andhra Pradesh.....................................................41 A.4 Energy Input and Sales Trend, Gujarat..............................................................................................................42 A.5 Recent Trend in State Subsidies and Distribution Profits, Gujarat....................................................................42 A.6 Consumer and Consumption Mix of Distribution Utilities, Gujarat..................................................................43 A.7  Agricultural Energy Consumption and GDP Post JGY......................................................................................43 A.8 Energy Input and Sales Trend, Rajasthan.........................................................................................................44 A.9 Recent Trend in State Subsidies and Financial Losses, Rajasthan...................................................................44 A.10 Consumer and Consumption Mix of Distribution Utilities, Rajasthan..............................................................45 A.11 Agricultural Energy Consumption and GDP after FRP Initiation.......................................................................45 A.12 Energy Input and Sales Trend, Haryana............................................................................................................46 A.13 Recent Trend in State Subsidy and Financial Losses, Haryana.........................................................................46 A.14 Consumer and Consumption Mix of Distribution Utilities, Haryana.................................................................47 Contents v Tables 1.1 Selected Rural Subdivisions and Distribution Utilities, by State.........................................................................4 1.2 Electricity Distribution of Surveyed Sample.......................................................................................................4 2.1 Data Highlights from the Institutional Evaluation...............................................................................................6 2.2 Payback Scenarios for Vinchiya and Bassi Subdivisions.....................................................................................9 2.3 Trend in Energy Sales to Agriculture Consumers, Jaipur District Circle...........................................................10 2.4  Recent Trend in Short-Term Power Purchase and Sale..................................................................................... 11 2.5  Change in Household Expenditure, Vinchiya....................................................................................................15 2.6  Economic Internal Rate of Return from Rural Load Segregation, Vinchiya.......................................................15 2.7 Change in Household Expenditure, Bassi........................................................................................................15 2.8  Economic Internal Rate of Return from Rural Load Segregation, Bassi...........................................................16 2.A1 Data Requirements for Agricultural Consumption Methods............................................................................17 4.1 Decision Matrix for Strategic Objectives..........................................................................................................27 4.2 Decision Matrix for Existing Technical and Operational Infrastructure..............................................................28 4.3 Decision Matrix for Physical and Socioeconomic Parameters.........................................................................29 Foreword Socioeconomic development of the rural populace is The study findings reveal that segregated systems can critical to India achieving its stated objective of inclusive be used to manage peak demand, identify and reduce growth. It is widely accepted that access to a reliable and losses previously hidden in agricultural consumption, sufficient power supply is a key enabler of rural economic improve power supply to rural domestic consumers, growth. Traditionally, India’s rural power supply has been and bolster socioeconomic development. Enabling the restricted by having feeders to villages serve both agri- segregated system with information technology (IT) can culture and household loads. Because agriculture power further improve monitoring and control and bring about supply is rationed by the distribution utilities, residential transparency and efficiency: Agricultural consumption on consumers often suffer from inadequate service. which the subsidy is based can be exactly determined, even without consumer metering, and data collected In 2003, some states in India began to separate their from the system can be used for strategic decision mak- agriculture and non-agriculture electricity loads in rural ing and operational improvement. areas to provide household consumers increased hours of power supply while restricting supply to agriculture That said, all of the many benefits of feeder segregation loads. This practice has led to significant improvement in may not necessarily be realized in every case. For this the overall socioeconomic status of the rural population, reason, the study recommends that each state design resulting in various states requesting financial assistance a rural power supply system customized to suit its local from the Ministry of Power to undertake similar pro- conditions and desired outcomes. The study further rec- grams. In response, the Ministry of Power is planning to ommends that a central knowledge hub be set up by formulate a large central scheme that will provide states the Government of India to assist states in undertaking funding to undertake feeder segregation. At the Minis- such improvement programs. I am pleased to note that try’s request, the World Bank carried out this study to the World Bank will provide the government support in assess the experience of states that have already under- setting up this knowledge hub, along with continued taken rural load segregation. The study’s significant find- strengthening of the design and implementation of rural ings and recommendations, documented in this report, power supply programs, ultimately benefiting the socio- can be used to improve the overall positive impacts of economic progress of 800 million rural people in India. the program. Onno Ruhl Jack Stein Country Director for India Sector Director South Asia Region Sustainable Development Department The World Bank The World Bank vii Acknowledgments This report summarizes the findings of a World Bank Rajasthan for their cooperation and generous support. It study on India’s recent experiences in rural feeder load also extends special thanks to counterparts in the Min- segregation, undertaken at the request of India’s Minis- istries of Power and Finance, who provided constructive try of Power. The study was led by Ashish Khanna and guidance throughout the study. The report was edited by Mohua Mukherjee, with a core team including Sudeshna Norma Adams and typeset by Laura Johnson. Ghosh Banerjee, Kavita Saraswat, and Mani Khurana of the Energy Sector Unit, South Asia Sustainable The team wishes to thank Jyoti Shukla, Senior Man- Development. ager, and peer reviewers Tushar Shah of the Interna- tional Water Management Institute and Mohinder Gulati The study benefited from a background study prepared and Pedro Antmann of the World Bank for their useful by staff of Pricewaterhouse Coopers Private Ltd, who insights during preparation of the study. Finally, the team interacted with the distribution utilities and shared data gratefully acknowledges the generous funding support and first-hand experiences with other key stakeholders, provided by the World Bank’s Energy Sector Manage- as well as a consumer survey led by MRS Private Ltd. ment Assistance Program (ESMAP), Asia Sustainable The team wishes to thank officials of the distribution utili- and Alternative Energy Program (ASTAE), and Office of ties and the respective Principal Secretaries of Energy the Chief Economist, South Asia Region. in the states of Andhra Pradesh, Gujarat, Haryana, and viii Acronyms and Abbreviations ABC Aerial Bunched Cables LT Low Tension AMI Advanced Metering Infrastructure M&E Monitoring and Evaluation AMR Automated Meter Reading MIS Management Information System AT&C Aggregate Technical & Commercial MKVVCL Madhya Kshetra Vidyut Vitran Company CAPEX Capital Expenditure Ltd DPR Detailed Project Report PGVCL Pashchim Gujarat Vij Company Ltd DT Distribution Transformer PMU Project Management Unit EIRR Economic Internal Rate of Return PPP Public Private Partnership FRP Feeder Renovation Program PwC Pricewaterhouse Coopers GUVNL Gujarat Urja Vikas Nigam Ltd REC Rural Electrification Corporation Ltd HT High Tension RMR Remote Meter Reading HVDS High Voltage Distribution System RVPNL Rajasthan Rajya Vidjut Prasaran Nigam Ltd IRR Internal Rate of Return SDT Special Design Transformer IT Information Technology UGVCL Uttar Gujarat Vij Company Ltd JGY Jyoti Gram Yojna VCB Vacuum Circuit Breaker JVVNL Jaipur Vidyut Vitran Nigam Ltd XPLE Cross-Linked Polyethylene Units of Measure Currency Equivalents GWh gigawatt hour Currency unit = Indian Rupee (Rs.) ha hectare Rs. 53 = US$1 (January 2012) km2 square kilometer kV kilovolt kVA kilovolt-ampere kW kilowatt kWh kilowatt hour mm millimeter MW megawatt ix Executive Summary Subsidizing electricity for irrigated agriculture in India dates Against this backdrop, the World Bank, at the request of India’s back to the advent of the Green Revolution in the late 1960s. Ministry of Power, undertook a study of India’s experience Decades later, those subsidies remain in place, and in rural load segregation. The study’s overall goal was to electricity tariffs for farmers amount to less than 10 per- draw lessons that can be applied to implementing such cent of the cost of power supply. Typically, farmers are schemes more broadly across the country. The study charged a flat tariff rate based on the horsepower per had four main objectives. The first was to compare load pump rather than the level of power actually used, which segregation approaches tried in various states to draw is not metered or otherwise recorded. Because mixed lessons that can be applied to future programs. The sec- feeders in villages supply both agriculture and non-agri- ond was to evaluate the financial viability of the schemes culture loads, the amount of power consumed cannot be based on financial and operational parameters. The third disaggregated by farm or non-farm use or by the amount was to gain a better understanding of the socioeconomic of power lost to technical inefficiencies or pilferage. High benefits of the schemes by conducting a primary survey system transmission and distribution losses, estimated of rural end-user consumers. The fourth was to develop at 35 percent, basically camouflage theft. The absence of a guidance note that policy makers can use to formulate metering, along with the extremely low cost of supply, a national action plan on rural feeder load segregation. means that rural agriculture consumers lack incentives to The study was divided into four parts aligned with these conserve and control their power use. objectives. Today, the agriculture sector accounts for 20–45 percent of The study used an institutional framework to evaluate key total power sale in most Indian states; in some states, the parameters across the project cycle, while financial and eco- agriculture power subsidy equals twice the expenditure on nomic frameworks were used to analyze costs and benefits. health or rural development. Empirical evidence suggests Four states—Andhra Pradesh, Gujarat, Rajasthan, and that large farmers have benefited disproportionately Haryana—were selected for the study. Selection crite- from the subsidy policy. Most small and marginal farmers ria included the amount of time elapsed since comple- lack access to electricity, instead depending on rainfed tion of the feeder segregation scheme, type of technical agriculture. Within an environment of chronic shortages intervention used, and coverage composition. In Andhra and wasteful consumption that threatens groundwater Pradesh, where data collection is ongoing, key lessons depletion, state utilities have sought to limit the subsidy will not be finalized until analysis for all selected states burden by restricting daily agriculture power supply to has been completed. Gujarat completed its feeder load 6–8 hours, often at night. But this practice compromises scheme in 2006, while Rajasthan’s program is ongoing. the quality and quantity of supply to non-farm consumers In Haryana, feeder segregation was completed in mid- connected to the same feeder, constraining their produc- 2010; only the first part of the study was applied since tive economic activities. commenting on the scheme’s impact was considered premature. One pragmatic solution tried by some states is rural feeder load segregation, which physically or virtually separates paid and Findings show that the states analyzed share a common goal nominally-paid rural feeder loads. By separating agriculture of ensuring adequate supply hours to farmers and non-agricul- and non-agriculture connections, utilities can attempt to ture rural consumers to support socioeconomic development. measure and limit the amount of power supplied free to Other common drivers of implementing load segrega- farmers for irrigation, while ensuring that non-agriculture tion are ensuring rural consumers good-quality, reliable consumers receive better-quality supply for longer peri- power supply, enhancing energy accounting and audit- ods throughout the day. To date, eight states have initi- ing leading to a restatement of agricultural consumption ated such schemes: Andhra Pradesh, Gujarat, Haryana, and thus loss levels, and enhancing load management Punjab, Karnataka, Maharashtra, Madhya Pradesh, and through supply rotation for agriculture consumers. Rajasthan. xi xii Lighting Rural India: Load Segregation Experience in Selected States The technical approach selected for feeder load segregation For all states, the study found that monitoring and evaluation has been unique to each state’s political thinking, regulatory (M&E) of project execution and outcomes had been negligible. policies, and state of the power sector. Both Andhra Pradesh For example, summary reports on agricultural consump- and Gujarat initially undertook virtual segregation and tion based on segregated load data had not been pre- later switched to physical segregation to eliminate theft pared for management decision making in Gujarat and and frequent power outages. To address groundwater Rajasthan even four and five years, respectively, after issues, Gujarat incorporated feeder segregation into its project completion. As mentioned above, baseline data integrated rural development program. Rajasthan chose a studies were not conducted prior to load segregation in virtual scheme that became part of its feeder renovation any of the states, and development of an MIS tool to program (FRP). Haryana adopted a physical segregation measure and monitor agricultural consumption in rural approach to tackle the problem of high distribution losses areas had not been envisaged as part of the schemes. in the power sector. Before-and-after analyses to assess the However, Gujarat undertook a post-scheme evaluation direct benefits from load segregation were not possible since through third parties, and Haryana recently designed an no baseline data collection and analysis had been undertaken. M&E framework to accurately estimate agricultural con- sumption post segregation and ensure transparency in Project costs varied by state, depending on the technical determining subsidy levels. approach adopted and whether the system architecture was stand-alone or integrated. The capital investment per Rural load segregation is expected to provide quantifiable feeder varied from US$64,150 (Rs. 3.4 million) in Andhra financial returns through two main channels: (i) increased Pradesh to $128,302 (Rs. 6.8 million) in Gujarat, while revenue accrued to the utility as a result of loss reduction capital investment per kilometer ranged from $377 ,358 and/or changes in the sales mix of the project area and (ii) (Rs. .20 million) in Andhra Pradesh to $624,528 (Rs. .33 reduced cost achieved through lower power-procurement million) in Haryana. Because project scopes differed by cost at the margin. Analysis of the financial benefits state, direct cost comparisons were not feasible. of rural load segregation was based on a detailed data assessment of two rural subdivisions: Vinchiya (Gujarat) No specific institutional frameworks had been put in place and Bassi (Rajasthan). Findings at the level of these sub- to execute the feeder load segregation schemes. With the divisions were then aggregated to all rural subdivisions at exception of Rajasthan, where the circle head was des- the level of the respective distribution utilities and finally ignated as project manager for the FRP , the segregation the states. schemes had been managed as part of routine business operations, with no dedicated project management Rural load segregation has been a critical factor in bringing units. In Andhra Pradesh, Rajasthan, and Haryana, the transparency to agricultural consumption, resulting in accu- states’ respective distribution utilities were considered rate estimates of distribution losses. In Vinchiya, sales to the the project owner—defined as the entity that initiates agriculture sector between 2006 and 2010 decreased the scheme and owns it through the implementation by more than half of total energy input, while distribu- phase—while the state government was the project tion losses increased by about as much. Revenue gain owner in the case of Gujarat. over the period can be attributed, in part, to growth in non-agricultural consumption. The utilities achieved trans- Andhra Pradesh, Gujarat, and Rajasthan decided to undertake parency in agricultural consumption and estimating dis- pilots before initiating statewide rollout of load segregation. tribution losses. The physical separation of agriculture Although Haryana did not undertake a pilot, official field supply from continuously supplied household and com- visits were made to Gujarat and Andhra Pradesh to gain mercial connections empowered the utility to plan load insights from these states’ experiences. The extent to rotation on agriculture-dominant feeders and improve which lessons from the pilots were integrated into final peak demand management. In Bassi, the ongoing FRP project designs is unclear since there is no formal docu- contributed significantly to reducing distribution losses, mentation on pilot results. None of the schemes included but there was no loss restatement as in Vinchiya. Utility components of remote meter reading or advanced metering sales to the agriculture sector post segregation grew by infrastructure, which are used to capture online metering about 8 percent of total energy input, while distribution data and prepare a user-friendly management informa- losses fell by 16 percent. Both metered and unmetered tion system (MIS) to undertake measurement and con- agriculture connections rose consistently. Despite loss trol of agricultural consumption. reductions achieved through virtual segregation, higher Executive Summary xiii Figure ES.1 Changes in Quality of Supply, Vinchiya Subdivision, Gujarat a. “Rare/Never”Power Outages b. Low-Voltage Problems 100 100 Before segregation 85.9 80.0 80 76.8 80 After segregation 68.2 69.1 60 60 Percent Percent 52.0 40 33.7 40 28.6 20 13.0 20 10.5 5.9 4.6 0 0 Households Agriculture Business Households Agriculture Business and industry and industry Source: MRS Private Ltd Survey 2011. sales to agriculture, along with a dwindling subsidy dis- 150 percent (figure ES.1a), while those reporting low- bursement, pressured the state’s fiscal viability. voltage problems decreased by about 74 percent (figure ES.1b). Similarly, the failure rate of distribution transform- Rural feeder segregation is envisaged to improve welfare ers for the utility fell from 24.18 percent in FY2006–07 outcomes through two direct channels: (i) better quality and to 19.79 percent in FY2008–09. In Bassi (Rajasthan), the reliability of electricity supply and (ii) socioeconomic ben- percentage of households that reported power outages efits. Better quality of supply comprises such issues as as “rare/never” increased by nearly 34 percent (figure voltage, outage frequency, and data on system balancing ES.2a), while those reporting low-voltage problems fell and failure rate of distribution transformers, while socio- by more than 71 percent (figure ES.2b). Results of focus economic benefits refer to non-agriculture consumers’ group discussions show that farmers were satisfied with increased hours of electricity supply. The study analyzed scheduled power supply with minimum interruptions. the welfare outcomes for these two variables in Vinchiya (Gujarat) and Bassi (Rajasthan). The financial and economic results of improved power sup- ply in Gujarat and Rajasthan have been mixed, suggesting the Survey findings show a marked improvement in quality and need for integrated analysis. Gujarat has managed to con- reliability of supply resulting from feeder segregation. The per- trol the subsidy and financial losses, while overall finan- centage of households in Vinchiya (Gujarat) that reported cial losses and subsidy in Rajasthan continue to increase. power outages as “rare/never” increased by more than To derive the maximum benefit from investments, Figure ES.2 Changes in Quality of Supply, Bassi Subdivision, Rajasthan a. “Rare/Never” Power Outages b. Low-Voltage Problems 100 100 Before segregation 86.7 81.3 79.6 81.1 80 80 After segregation 60 60 Percent Percent 42.8 40 33.9 40 28.8 20 20 18.1 9.0 10.25 10.1 0.4 0 0 Households Agriculture Business Households Agriculture Business and industry and industry Source: MRS Private Ltd Survey 2011. xiv Lighting Rural India: Load Segregation Experience in Selected States institutional and governance reforms need to accompany various other factors (e.g., load forecasting and distribu- technological changes. Also, in the case of higher agricul- tion network planning). tural consumption after segregation, subsequent study phases need to assess cross-sector linkages between Other common perceptions are that feeder segregation is a the socioeconomic cost of excessive groundwater one-time investment that can substitute for agriculture meter- extraction and the benefits of increased agricultural GDP . ing. As previously mentioned, baseline data was not Such an integrated analysis is critical to formulating a collected and M&E systems were not in place during comprehensive framework for designing and evaluating project development. As a result, it was not possible to rural load segregation schemes. evaluate subsequent benefits or monitor system param- eters post segregation. In reality, feeder segregation is Qualitative and quantitative survey data from Gujarat and Raj- an ongoing activity that requires setting up systems to asthan were analyzed to verify prevailing perceptions of rural continuously monitor and enforce discipline with regard load segregation. A common perception is that feeder to new connections. The energy input channeled into load segregation is the only solution to guarantee a con- feeders connected to agriculture consumers is available, tinuous electricity supply to non-agriculture connections. but it also includes technical losses and consumption by In reality, feeder load segregation is needed only in socio- unauthorized loads. Thus, metering is essential to obtain political situations where metering of agriculture loads is data on customer-specific consumption and implement not possible. Another common perception is that imple- direct subsidy-delivery mechanisms. In scenarios where menting load segregation will necessitate the creation consumer metering is not possible for socio-political rea- of new feeders and added transformers. But the study sons, meters can be installed at the level of the substa- findings show that existing infrastructure, which is aged tion or distribution transformer. But to ensure commercial and overloaded, requires augmentation irrespective of losses are excluded from agricultural consumption and segregation. Contrary to the assumption that agricultural related subsidy payments to utilities, it is important to consumption will be restated post segregation, lead- maintain consumer indexing and have IT-enabled meters ing to a restatement of loss levels for transmission and for remote data collection and automated data analysis distribution, the study found that estimated agricultural to detect unauthorized loads. consumption continues to be based on earlier estimates rather than data captured from the load segregation The emerging lessons suggest that a standard approach to scheme. rural power system design using load segregation in isolation is unlikely to achieve the various states’ desired outcomes. It is logical to assume that subsidy processes would be trans- Maximizing the benefits of load segregation schemes parent after segregation and that loss reduction, higher reve- requires accompanying institutional and governance nue, and an improved load factor would ensure financial return reforms at the utility level. At most substations, feeder on investment. However, the study could not establish the meters compatible with remote reading are already overall impact of load segregation on subsidy transpar- installed. Data from these meters needs to be automati- ency owing to the lack of a system for collecting data cally collected and analyzed. This will require setting up a at the feeder level (e.g., metering based on information data monitoring center dedicated exclusively to manag- technology [IT]). Also, the proportionate contribution of ing the data provided by the acquisition system and tak- load segregation to financial return on investment could ing action based on that information. Operators should not be established. be trained extensively in appropriate use of the system and supported by crews responsible for field inspections It is generally assumed that feeder load segregation will in potentially irregular situations detected with support result in improved incomes and livelihoods and better rural of the software. service standards. While the primary survey established that considerable socioeconomic development occurred It is vital to communicate the objective of load segregation to over the course of implementing the schemes, their pre- field staff and institutionalize a system to retain segregated cise contribution could not be quantified. However, even feeders while releasing new connections and modifying if just 5 percent of increased income could be attributed existing ones. To manage the switching of loads between to rural load segregation, the economic return would be feeders in cases of breakdown, the utility should set up a strong 15 percent. In terms of better rural service stan- and institutionalize a system to track such changes and dards, load segregation has resulted in improved rural assign consumption to the appropriate feeder. Feeder supply hours. But the strategy for achieving this includes segregation provides the “hardware” for a system Executive Summary xv capable of delivering differentiated service to farmers a multi-tiered, multi-skilled project management set-up, and non-agricultural rural consumers, along with manage- including a dedicated project management unit, project ment decision-making tools for effective monitoring. But managers appointed for the entire project cycle, and the eventual outcome in terms of better quality of supply third-party quality assurance. M&E is a dynamic process and sustainable operations is a function of the necessary comprising regular feedback on post-implementation “software;” that is, the simultaneous and integrated benefit estimates and sharing of results and experiences application of organizational changes, accountability sys- with key stakeholders and decision makers (figure ES.3). tems, and use of IT-based metering. This study demonstrates that there is no one-size-fits-all solu- The study developed a guidance note designed to enable tion to rural power supply improvement. Project proposals state governments and utilities to adopt the rural power sup- should be evaluated as part of each state’s broad stra- ply approaches that best fit their on-the-ground realities. The tegic program for improving rural power supply. Given guidance note offers the various conditions to consider, the enormous amounts of planned or already allocated along with key issues that must be addressed at each investments by various Indian states, there is an urgent stage of the project cycle, from conceptualization and need to establish centralized rules of engagement outlin- planning through execution and M&E. During project ing the principles that should underpin the design of any conceptualization, decision makers must identify the initiative to improve the sustainability of rural power sup- strategic objective, evaluate alternative models, and ply while maintaining techno-economic viability. decide on the best-fit solution. A decision-matrix exercise is provided to evaluate models by strategic objective and Based on consultations with India’s Ministry of Power, it key technical and socioeconomic parameters. Planning was decided that a central knowledge hub should be set up includes preparing a robust baseline, stakeholder com- to support states in undertaking rural power supply improve- munication strategy, and Detailed Project Report (DPR); ment programs. It was also advised that one or two states performing cost-benefit analyses; and adopting a suitable far along in implementing their feeder load segregation procurement strategy. The execution phase must ensure schemes create integrated data centers to collect and Figure ES.3 Proposed Institutional Framework for Rural Load Segregation Conceptualization and Planning Execution Monitoring and Evaluation Objectives and drivers Dedicated project management unit Third-party evaluation of financial and socioeconomic benefits Evaluation of existing technical and Multi-tiered, multi-skilled project operational infrastructure management team Communication of outcomes Physical and socioeconomic Engagement with regulator parameters in rural areas Third-party quality assurance Baseline data collection Use of IT-based systems Stakeholder communication strategy Sustainability of segregated system Financing arrangements Financial and economic appraisal Detailed Project Report Procurement strategy Source: Authors. Note: The institutional framework can be applied to an entire state, a particular distribution utility, or even selected business units within a utility. xvi Lighting Rural India: Load Segregation Experience in Selected States analyze data for such strategic purposes as ensuring essential starting points that could lead to a win-win situ- transparency in determining subsidies for distribution util- ation for all stakeholders. ities and improving operational efficiency. Furthermore, it was decided that one or two states on the threshold of The knowledge hub set up within the CEA should be respon- undertaking rural load segregation should be selected to sible for developing standard documentation templates for the work with the Central Electricity Authority (CEA) on con- DPR, project management and operational manuals, technical ceptualizing and designing improved rural power supply. specifications, and standard bidding documents. In addition, it should develop processes for project implementation, Setting rules of engagement and principles to ensure improved data management, and integrated operations to ensure rural power supply while maintaining techno-economic via- sustainability of the rural power supply system. The bility can be achieved using a common strategic framework. knowledge hub can assist states that desire to follow the If feeder load segregation emerges as the most optimal strategic framework with project design and implemen- solution, it should be amenable to the direct delivery of tation, including the use of AMR. Finally, the experiences subsidies to farmers. Improved measurement and reliabil- and outcomes of the demonstration projects should be widely ity of agricultural consumption data, utilizing automated disseminated by the knowledge hub so that lessons in suc- meter reading (AMR) and similar IT-based initiatives, are cess can be replicated across the country. 1 Introduction Subsidizing electric power for irrigated agriculture in India ful consumption that threatens groundwater depletion, dates back to the advent of the Green Revolution; in the state utilities have sought to limit the subsidy burden by late 1960s, large government subsidies were put in place restricting daily agriculture power supply to 6–8 hours, to cover the energy costs of pumping groundwater for often at night. However, this practice compromises the increased irrigation to obtain higher crop yields. Decades quality and quantity of supply to non-agriculture consum- later, those subsidies are still in place, and electricity tar- ers connected to the same feeder, constraining their pro- iffs for farmers amount to less than 10 percent of the ductive economic activities. cost of supply. Typically, farmers are charged a flat tar- iff rate based on the horsepower per pump rather than the level of power actually used, which is unmetered. Study Background and Objectives Because feeders in villages are mixed,1 supplying both agriculture and non-agriculture loads, the amount of The ideal power-delivery approach for rural consumers power consumed cannot be disaggregated by farm or with diverse electricity needs would be a robust distribu- non-farm use or by the amount of power lost to techni- tion infrastructure, with an adequate ratio of high-tension cal inefficiencies or pilferage; indeed, high system trans- (HT) to low-tension (LT) conductors, using consumer mission and distribution losses, estimated at 35 percent, metering based entirely on information technology (IT).4 basically camouflage theft.2 The absence of metering, Such an approach is supported by studies conducted along with the extremely low cost of supply, means that by the Planning Commission of India, which has advo- all rural consumers lack incentives to conserve and con- cated for universal metering of agricultural consumers, trol their power use. with subsidies limited to specified amounts of initial con- sumption. But under the current institutional and socio- Power consumption in the agriculture sector accounts for political constraints, it is difficult to meter all agricultural 20–45 percent of total power sale in most states in India. connections, monitor their supply remotely, and maintain In some, the magnitude of the agriculture power subsidy infrastructure. is twice the annual budgetary expenditure on health or rural development. Empirical evidence suggests that large One pragmatic solution tried by some states, particu- farmers have benefited disproportionately from the sub- larly in agrarian areas, is feeder load segregation. Using sidy policy; most small and marginal farmers lack access either physical or virtual mechanisms to separate paid to electricity, instead depending on rainfed agriculture.3 and nominally-paid feeder loads, utilities can attempt to Within an environment of chronic shortages and waste- measure and limit the amount of power supplied free to farmers for irrigation, while ensuring that non-agriculture 1. A feeder is defined as an 11-kV wire emanating from a 33-kV or higher voltage substation connecting end-user electricity consumers through distribution transformers and a low-voltage network. 4. Automated meter reading (AMR) technology has matured, and its 2. L. Monari, “Power Subsidies,” Viewpoint No. 244, 2 (Washington, cost has fallen substantially. In rural India, AMR can be effectively DC: World Bank, 2002). used as the widespread penetration of mobile networks provides a 3. Ibid., 3. convenient communication platform for the system. 1 2 Lighting Rural India: Load Segregation Experience in Selected States consumers receive better-quality supply for longer Study Method periods throughout the day. To date, eight states have initiated rural load segregation schemes. The earliest pro- The study was divided into four parts, aligned with the gram was in Andhra Pradesh in 2001, followed by Guja- four main objectives described above. Lessons learned rat, Haryana, Punjab, Karnataka, Maharashtra, Madhya from achieving first three parts formed the basis for Pradesh, and Rajasthan. developing the guidance note. Against this backdrop, the World Bank, at the request of India’s Ministry of Power, undertook a study of Framework for Evaluation India’s experience in rural load segregation in 2011–12. The findings were structured across project conceptual- The study’s overall goal was to draw lessons that can ization, execution, and post-implementation evaluation, be applied to implementing rural feeder segregation with frameworks also provided for financial and eco- schemes more broadly. The study’s main objectives were nomic cost-benefit analyses. Figure 1.1 summarizes the fourfold. The first was to compare and contrast load seg- framework for institutional evaluation of specific param- regation approaches from past experiences across states eters across the project cycle. to draw lessons that can be applied to future programs. The second was to evaluate the financial viability of the schemes based on financial and operational parameters. Selected States for Study Given the difficulty of conducting the assessment at the Four states—Andhra Pradesh, Gujarat, Rajasthan, and level of utilities, the subdivision level was selected as the Haryana—were selected for the study (box 1.1). Selec- basis for the analysis. The third objective was to gain a tion criteria included the amount of time elapsed since better understanding of the socioeconomic benefits of completion of the feeder segregation scheme, type of the schemes by conducting a primary survey of rural end- technical intervention used, and coverage composition. user consumers. Finally, the study aimed to develop a In the case of Andhra Pradesh, data collection is ongo- guidance note for policy makers that can be used to for- ing, and key lessons will not be finalized until analysis for mulate a national action plan on rural feeder load segre- all selected states has been completed. Gujarat’s feeder gation; the note comprises a recommended approach for segregation scheme was completed in 2006, while Raj- developing a feasibility study and subsequent framework asthan’s program is ongoing. Though Haryana completed for the plan’s design, implementation, and monitoring feeder segregation in mid-2010, only the first part of the and evaluation (M&E). study was applied since commenting on the scheme’s Figure 1.1 Institutional Evaluation Framework Conceptualization Execution Evaluation Objectives and drivers Institutional framework Monitoring and Evaluation (M&E) framework Technical system architecture Project owner Execution hurdles and implementation deficits Technical approach Baseline data collection Financing arrangements Program planning Procurement strategies Source: Authors. Introduction 3 Box 1.1 Profile Highlights of the Selected States In Andhra Pradesh, India’s third largest state, agriculture accounts for more than one-quarter of state GDP . Most farmers depend on rainfed agriculture, while 28 percent of total cultivated land is irrigated. Since 2005, distribution losses have steadily declined, but subsidies and financial losses have risen. Agricultural consumers represent 14 percent of the utilities’ consumer base, accounting for about one-third of sales. After virtual segregation of all mixed rural feeder loads (2001–05), the state switched to physical segregation and pilots are under way. Key objectives are to boost small rural industries by providing non-agriculture customers a 24-hour, three-phase supply and improve agricultural productivity by extending agricultural consumers’ 7–9 hours of three-phase supply for pump sets. Gujarat state in northwest India has a per-capita electricity consumption of 1,354 kWh, nearly twice the national average. In 2003, Jyoti Gram Yojna (JGY), an innovative rural electricity scheme, introduced the physical separation of 11-kV feeder lines serving rural agriculture consumers and rural household and commercial connections. The program also installed meters on feeders to eliminate theft. Since JGY was completed in 2006, agricultural energy consumption has trended upward. Today Gujarat’s power distribution utility has a profitable balance sheet. In Rajasthan, India’s largest state in terms of land area, agriculture contributes more than one-quarter of state GDP (FY2010–11). Agriculture consumers represent just 12 percent of the utilities’ consumer base but 39 percent of sales. Rural load segregation, initiated in 2005 as part of the Feeder Renovation Program, aims to reduce losses of mixed rural feeders and improve rural households’ supply quality. Integrated with HVDS on agricultural feeders, DT meter- ing, and replacement of LT cables with ABC, this ongoing program adopted virtual segregation. To eliminate agri- cultural theft, a roster switch on existing feeders balances three-phase, agricultural supply hours with single-phase hours when households receive an unrestricted supply. Since 2005, distribution losses have steadily declined, yet higher agricultural consumption has more than offset the financial benefit. Haryana, located in northwest India, is among one of the most prosperous states in the country. In 2009–10, per capita income was estimated at US$1,486 (Rs. 78,781), and the literacy rate was 71.4 percent. Canals are the main source of irrigated water for cultivating diverse crops. Farmers comprise 11 percent of the electricity consumer base, accounting for 39 percent of total utility sales. Since the state government erected dedicated 11-kV feeders to separate agriculture loads from rural household connections in 2005, distribution losses have steadily declined. All of the 1,226 feeders erected were equipped with AMR compatible bulk meters. Over the same period, the utili- ties’ financial losses have increased. Delayed completion of the scheme was resolved in mid-2010 when the state government began regular monitoring. Detailed summaries of the state profiles are provided in the annex. impact was considered premature. Recently, Haryana rural business-unit levels. For each state, one distribution developed an M&E system at the segregated feeder utility was selected, based on the proportion of agricul- level and is in the process of implementing a manage- ture consumers served and the state’s agricultural con- ment information system (MIS) to enable a robust esti- sumption. For each utility, a corresponding business unit mation of agricultural consumption (chapter 2, appendix). or rural subdivision was identified for detailed evaluation of the scheme’s impact (table 1.1). To coordinate with the World Bank and facilitate data col- lection, the Ministry of Power appointed nodal officers in To evaluate the impact of feeder load segregation on the respective states (i.e., Andhra Pradesh, Gujarat, and socioeconomic outcomes, MRS Private Ltd led the Rajasthan), and the World Bank hired Pricewaterhouse implementation of a primary survey of electricity con- Coopers Private Ltd (PwC) and MRS Private Ltd to imple- sumers (table 1.2). Load segregation was carried out ment the activity. The staff of PwC interacted with the dis- statewide, and it was not possible to select “with” and tribution companies and other key stakeholders to share “without feeders” for establishing causality and quanti- data and first-hand experience. Throughout the study, the fying impacts. Thus, characteristics of the same house- World Bank team consulted with the nodal officers, as holds were studied before and after segregation phases well as with concerned officers at corporate-office and to determine socioeconomic changes. 4 Lighting Rural India: Load Segregation Experience in Selected States Table 1.1 Selected Rural Subdivisions and Distribution Utilities, by State State Rural subdivision Distribution utility Andhra Pradesh Warangal, Hasanparthy (mandal) Andhra Pradesh North Power Distribution Company Ltd., Warangal Gujarat Vinchiya Paschim Gujarat Vij Company Ltd, Rajkot Rajasthan Bassi Jaipur Vidyut Vitran Nigam Ltd, Jaipur Source: Authors. Table 1.2 Electricity Distribution of Surveyed Sample Consumer type (number) State Household Commercial Small industry Agriculture Total Andhra Pradesh 875 313 187 625 2,000 Gujarat 417 167 83 333 1,000 Rajasthan 417 167 83 333 1,000 Source: Authors. Structure of This Report from the study, based on evidence from the Gujarat and Rajasthan cases, and emerging lessons. Chapter 4 pro- This report is organized as follows. Chapter 2 presents vides decision makers guidance on steps to take at each key findings from evaluating the rural feeder segregation stage of the project cycle for the various models consid- approaches adopted by the selected states using insti- ered. Finally, chapter 5 summarizes the key lessons that tutional, financial, and economic frameworks for analy- emerged from the study and offers recommendations on sis. Chapter 3 summarizes perceptions and observations next steps. 2 Key Findings from Selected States Experiences of the selected Indian states in implement- splitters, unbalanced loads, failure of distribution trans- ing various approaches to rural feeder segregation were formers, and frequent interruptions due to power-system evaluated based on parameters within institutional, finan- faults. To address groundwater issues, Gujarat incorpo- cial, and economic frameworks. This chapter presents the rated feeder segregation into its integrated rural devel- key findings that resulted from these several analyses. opment program. In Haryana, where the power sector faced high distribution losses, a physical approach to load segregation was selected. Rajasthan chose a vir- Institutional Evaluation: tual scheme that became part of its feeder renovation Project-Cycle Framework program (FRP), which included various other system strengthening elements, including a high voltage distri- Following the criteria of the institutional evaluation frame- bution system (HVDS) on agricultural feeders, metering work (figure 1.1), the study collected information on the of distribution transformers, and replacement of low- four selected states across the project cycle (table 2.1). tension, bare overhead conductors with insulated aerial The subsections that follow describe the key findings bunched cables (ABC). that emerged from the data analysis, grouped by project phase. Because no baseline data collection and analysis was undertaken, as suggested in chapter 1, it was not pos- sible to conduct a before-and-after analysis to evaluate Conceptualization the benefits resulting directly from the feeder segrega- tion program. In terms of project cost, load segregation The study found that all states shared a primary objective: schemes varied across states, depending on the techni- ensuring supply hours to agriculture and non-agriculture cal approach and system architecture used (i.e., stand- rural consumers to support socioeconomic development. alone or integrated) (table 2.1). Because project scopes Other common drivers, variously prioritized by state, differed by state, direct cost comparisons were not were ensuring rural consumers good-quality, reliable feasible. power supply; enhancing energy accounting and audit- ing leading to a restatement of agricultural consumption and thus loss levels; and enhancing load management Execution through supply rotation for agriculture consumers. The study found that no specific institutional framework The technical approach selected for feeder load segre- was set up for implementing the feeder load segrega- gation has been unique to each state’s political thinking, tion schemes. With the exception of Rajasthan, where regulatory policies, and state of the power sector. For the circle head was designated as project manager for example, Andhra Pradesh and Gujarat both undertook the FRP , the schemes were managed as part of routine virtual segregation initially but later switched to physi- business operations, with no dedicated project manage- cal segregation owing to issues of theft using phase ment units (PMUs). In Andhra Pradesh, Rajasthan, and Haryana, the states’ respective distribution utilities were 5 6 Lighting Rural India: Load Segregation Experience in Selected States Table 2.1 Data Highlights from the Institutional Evaluation Factor Andhra Pradesh Gujarat Rajasthan Haryana Project description Agricultural feeders 8,878 1,904 8,126 1,226 (number) (physical segregation) Capital investment Total (Rs. billion) 3.01a 1.29 4.48 5.73 Per feeder (Rs. million) 3.40 a 6.80 5.50 4.70 Per kilometer (Rs. .20a .23 .20 b .33 million) Conceptualization Technical system Stand-alone Stand-alone Integrated (feeder segrega- Stand-alone architecture tion, renovation, and HVDS) Segregation approach Virtual (2001–05); pilots for Virtual (prior to 2003); physi- Virtual on all rural mixed Physical (2006–10) selected (investment physical (2010–ongoing) cal through JGY (2003–06) feeders (2005–ongoing) period) Financing Utilities funding pilots for Initially a government-local Financial institutions Financial institutions (REC), arrangements physical segregation; finaliza- body participatory scheme (e.g., REC); Indian banks with 10 percent equity from tion of funding for full-scale but mainly funded through (private-sector/public-sector state government. project under way. state-government grant. undertaking). Program planning All rural feeders for pilots DPR not prepared; scheme Feeder-wise DPRs prepared Subdivision-specific DPR selected from administrative estimates prepared at by in-house staff (22.7 per- prepared. Financial cost divisions (mandals) located respective subdivision cent IRR). Scheme initially benefit evaluated through near district headquarters; levels, and cost approved by prioritized high-loss feeders percentage gross returns; envisioned benefits not respective divisions. for pilot implementation, sample DPR studied showed quantified in draft DPRs. but later was extended to all 27.75 percent gross return rural feeders. for the scheme. Procurement strategy Physical segregation pilots Entire scheme initiated Typically on partial turnkey; Turnkey contractor for on partial turnkey; distribu- internally; all material pro- distribution companies turnkey works and in-house tion companies procure cured centrally by the former procure key materials, DTs, material management wing VCBs, DTs, and HT and LT Gujarat Electricity Board. VCBs, and HT and XPLE for labor contracts. conductors, while imple- conductors, while execut- menting contractor procures ing contractor procures the the balance. balance. Execution Institutional Pilots managed through No scheme-specific frame- Circle head of respective Distribution company’s cen- framework routine business operations; work during execution. operational areas designated tralized planning-and-design specific institutional frame- as project manager; junior cell responsible for planning work for complete rollout yet engineers of respective rural and awarding contract and to be decided. subdivisions designated as monitoring project execution. managers for specific sets of feeders. Project owner Distribution utility State government Distribution utility Distribution utility Evaluation M&E framework JGY cell set up at GUVNL M&E framework recently office and project evaluation designed and implemented agencies selected to moni- with an MIS (appendix). tor completed and ongoing work. Source: Authors. a. Estimates based on DPR values. b. Estimate based on conversion of LT length to HT at assumed LT:HT ratio of 1:0.6. Key Findings from Selected States 7 Figure 2.1 Financial Framework for Cost Benefit Analysis Cost-benefit Cost Benefits estimations Stand-alone Revenue improvements load segregation through loss reduction Scenario-based payback scheme Integrated load segregation Cost reduction through (with HVDS/other CAPEX) lower-cost power purchased at the margin Source: Authors. considered the project owner—defined as the entity that to determine subsidy levels for the distribution utilities initiates the scheme and owns it through the implemen- (appendix). tation phase—while the state government was the proj- ect owner in the case of Gujarat. Financial Evaluation In Andhra Pradesh, Gujarat, and Rajasthan, pilots were undertaken before initiating statewide rollout of feeder The rural load segregation scheme is expected to provide load segregation to decide on the most suitable technical quantifiable financial returns through two main channels: approaches and evaluate consumer response. The extent (i) increased revenue accrued to the utility as a result of to which lessons from the pilots were integrated into final loss reduction and changes in the sales mix of the project project designs is unclear since pilot results were not for- area and (ii) reduced cost achieved through lower power- mally documented. Although Haryana did not undertake procurement cost at the margin. Peak load shaving a pilot program before initiating its segregation scheme, resulting from more efficient management of agricultural it conducted official field visits to Andhra Pradesh and load may also contribute to cost reduction (figure 2.1). Gujarat to gain insights from these states’ experiences in load segregation. None of the schemes included IT- As previously mentioned (table 1.1), the analysis of finan- based metering (i.e., remote or automated meter read- cial benefits resulted from a detailed assessment based ing) used to capture online metering data and prepare a on data from two rural subdivisions, one from each of user-friendly management information system (MIS) on two states. The data was subsequently aggregated to all measurement and control of agricultural consumption. rural subdivisions at the level of the respective distribu- tion utilities and finally the states. Evaluation The study found that M&E of project implementation and Subdivision Findings outcomes was negligible in all states. As previously men- The two rural subdivisions selected for financial analysis tioned, baseline data studies had not been conducted were Vinchiya (Gujarat) and Bassi (Rajasthan). In Vinchiya, prior to load segregation. In addition, development of rural load segregation has been a critical factor in bring- an MIS tool to measure and monitor rural agricultural ing transparency to agricultural consumption, resulting in consumption had not been envisaged as part of the accurate estimates of distribution losses. From 2006 to schemes. However, Gujarat undertook a post-scheme 2010, sales to the agriculture sector decreased by more evaluation through third parties, with the Indian Institute than half of total energy input (from 75 percent to 23 of Management, Ahmadabad, Centre for Environmental percent), while distribution losses increased by nearly as Planning and Technology, and Institute of Rural Manage- much (from 16 percent to 64 percent). Revenue gain over ment, Anand appointed as project evaluation agencies. the period can be attributed, in part, to growth in non- Moreover, Haryana recently designed an M&E frame- agriculture consumption (from 9 percent to 15 percent) work to more accurately estimate agricultural consump- (figure 2.2). tion after segregation and establish a transparent system 8 Lighting Rural India: Load Segregation Experience in Selected States Figure 2.2 Trends of Commercial Parameters, Vinchiya Subdivision 70 Initiation of load segregation 1.0 0.88 Energy sales to 0.9 60 agriculture sector 0.73 0.8 Million kWh 50 8.45 Energy sales to 0.70 0.7 15.98% 0.56 44.40 41.65 non-agriculture sector 5.33 4.92 0.6 Rs./kWh 40 12.39% (67.79%) (63.75%) Distribution losses 4.77 32.86 41.14 0.5 30 (65.39%) (71.09%) 0.4 Assessment per kWh of energy input 20 39.49 0.3 32.92 7.95 9.84 5.96 5.92 0.2 10 11.44 10.81 13.16 13.84 0.1 0 0 2004–05 2005–06 2006–07 2007–08 2008–09 2009–10 Source: Office of Executive Engineer, Rural Division, Old Power House, Rajkot. Although Bassi did not have a significant loss restate- revenue assessment per unit of energy input from Rs. ment as in Vinchiya, the ongoing Feeder Renovation 1.55 to Rs.1.49 over the period.5 Program (FRP) has contributed significantly to reducing distribution losses (figure 2.3). Between 2006 and 2010, To calculate the payback period for the subdivisions, net losses fell by nearly one-third (from 42 percent to 26 per- revenue gain (conversion of distribution losses to energy cent), but increased agricultural consumption negated sales) was computed for the years after program comple- the financial benefit. Most of the loss reduction occurred tion. Incremental revenue for a given year was calculated in the agriculture sector, where sales increased by 14 percent of total energy input (from 40 percent to 54 per- cent). Over the same period, non-agriculture sales grew 5. The adverse financial impact of significantly higher agricultural con- sumption in Rajasthan leading to higher financial losses post feeder by just 3 percent (from 18 percent to 21 percent). Higher segregation requires a detailed third-party assessment, including sales to agriculture consumers reduced the annual parallel metering of sample agricultural feeders. Figure 2.3 Trends of Commercial Parameters, Bassi Subdivision Initiation of load segregation 1.55 250 1.6 1.48 1.49 Energy sales to 1.42 1.5 agriculture sector 200 53.88 Energy sales to 49.14 1.4 non-agriculture sector 53.88 Million kWh 150 (25.53%) Rs./kWh (31.36%) Distribution losses 51.59 39.52 1.3 55.86 100 (41.83%) (35.01%) Assessment per kWh 32.77 1.2 of energy input 24.62 50 28.73 85.16 103.84 1.1 53.05 67.05 0 1.0 2006–07 2007–08 2008–09 2009–10 Source: Office of Superintendent Engineer, Jaipur District Circle, Power House, Jaipur. Key Findings from Selected States 9 Table 2.2 Payback Scenarios for Vinchiya and were calculated, assuming that net annual revenue gain Bassi Subdivisions was due to multiple initiatives, including load segrega- tion. The payback ranges were 3–9 years for Vinchiya and Subdivision (number of years) 5–20 years for Bassi (table 2.2). Scenario (%) Vinchiya Bassi 100 3 5 Aggregate Findings: Distribution Utility Level 50 5 10 Findings for the Vinchiya and Bassi subdivisions were 25 9 20 aggregated to the level of the respective distribution Source: Authors’ calculations. utilities: Paschim Gujarat Vij Company Ltd. (PGVCL) and Jaipur District Circle. For PGVCL, the commercial param- eters of 25 primary rural subdivisions were analyzed for as the product of loss reduction in units and the average the post-segregation period (FY2006–07 to 2010–11), and revenue assessment per unit of energy sales for the cor- five primary rural subdivisions were selected from each responding year. For Vinchiya, 71.09 percent was taken operational circle. In the case of Jaipur District Circle, a as the highest loss level, which occurred the year of pro- purely rural circle with 21 rural subdivisions was analyzed gram completion (FY2007–08). It was assumed that, if over the same post-segregation period. not for the load segregation scheme, losses would have risen further. Similarly for Bassi, it was assumed that, For PGVCAL, rural load segregation has been a critical without the FRP , the highest loss level of 41.83 percent factor in bringing transparency to agriculture consump- (FY2006–07) would have continued upward. Bassi’s pay- tion, resulting in the accurate estimation of distribution back period was computed in the years post FY2008–09, losses. Over the post-segregation period, sales to the the year of program completion. agriculture sector decreased by 9 percent of total energy input (from 49 percent to 40 percent), while distribution The loss reduction in any year was calculated as the differ- losses increased by the same percentage (from 33 per- ence between the highest loss level (71.09 percent and cent to 42 percent) (figure 2.4).7 41.83 percent, respectively) and the actual loss recorded for the corresponding year. The net gain in revenue for For Jaipur District Circle, sales to the agriculture sector any year was the difference between the incremental over the post-segregation period grew by about 8 percent revenue and the annual operational cost, calculated at 1.5 of total energy input (from about 42 percent to 50 per- percent of total asset value.6 Scenario-based paybacks cent), while distribution losses fell by 16 percent (from 35 percent to 19 percent) (figure 2.5). Initial revenue gain in 6. The assumption of similar average revenue assessment is expected in a business-as-usual scenario, while in real terms, the 7. The per-unit revenue assessment data for the identified subdivi- average revenue assessment in Bassi decreased over the years due sions had not yet been completed at the time of this writing. to a higher proportion of sales to agriculture consumers. Figure 2.4 Trend of Commercial Parameters, PGVCL Rural Subdivisions 2,500 2,000 53.88 780 797 Energy sales to Million kWh 816 1,500 555 (41.34%) (40.84%) (33.43%) (42.70%) agriculture sector Energy sales to 1,000 286 358 344 339 non-agriculture sector Distribution losses 500 796 763 757 819 0 2006–07 2007–08 2008–09 2009–10 Source: Office of Deputy Chief Account Officer (R&C), PGVCL, Rajkot. 10 Lighting Rural India: Load Segregation Experience in Selected States Figure 2.5 Trend of Commercial Parameters, Jaipur District Circle Initiation of load segregation 3,000 3.0 Energy sales to 2,500 521 2.5 agriculture sector (19.30%) 513 Energy sales to 1.86 1.86 2.0 2,000 1.85 (22.18%) non-agriculture sector Million kWh 822 Rs./kWh 620 (30.27%) Distribution losses 1,500 1.59 641 685 1.5 (35.33%) Assessment per kWh 556 1,000 428 1.0 of energy input 1,358 500 873 1,114 0.5 745 0 0 2006–07 2007–08 2008–09 2009–10 Source: Office of Superintendent Engineer, Jaipur District Circle, Power House, Jaipur. the subdivision can be attributed to loss reduction, along Aggregate Findings: State Level with increased non-agriculture consumption, which rose The final step of the financial analysis was to aggregate by 6 percent (from 24 percent to 30 percent). findings from PVGCL and the Jaipur District Circle to the level of the respective states. In Gujarat, the presence of After initiating the FRP in the Jaipur District Circle, the mixed loads before rural feeder segregation had made it number of both metered and unmetered agriculture difficult for the utility to regulate agriculture power sup- connections rose consistently (table 2.3). The key rea- ply. The physical separation of agriculture supply from son could be the Indian government’s policy to release continuously supplied household and commercial con- a larger number of new agricultural connections, along nections through JGY empowered the utility to plan load with loss reductions in agricultural supply owing to the rotation on agriculture-dominant feeders and improve use of HVDS and ABC. For both FY 2007–08 and FY peak demand management. In Rajasthan, virtual segre- 2008–09, the number of new agriculture connections gation implemented under the FRP has also resulted in increased by more than half. rural loss reductions; however, increased sales to agricul- ture, along with a dwindling subsidy disbursement, has pressured the state’s fiscal viability. Table 2.3 Trend in Energy Sales to Agriculture Consumers, Jaipur District Circle Energy sold to agriculture consumers (100,000) Agriculture consumers (number) Metered Added connections metered added Fiscal year Metered Un-metered Total sales Metered Un-metered Total (number) 2006–07 1,966 5,483 7,449 241 42,235 53,939 96,174 3,834 2007–08 3,156 5,569 8,725 1,190 48,209 47,404 95,613 5,974 2008–09 4,637 6,501 11,138 1,480 58,343 42,956 101,299 10,134 2009–10 6,409 7,173 13,582 1,772 66,661 40,708 107,369 8,318 2010–11 6,517 6,709 13,226 108 70,580 38,971 109,551 3,919 Source: Office of the Chief Accounts Officer, JVVNL, Jaipur. Key Findings from Selected States 11  onthly Averages of Maximum and Minimum Demand, Showing Difference in Figure 2.6 M Annual Averages, Gujarat 1,600 1,400 1,200 FY 2007–08 1,000 FY 2008–09 MW 800 Annual average 600 FY 2007–08 400 Annual average FY 2008–09 200 0 April May June July Aug. Sept. Oct. Nov. Dec. Jan. Feb. March Source: Gujarat State Load Dispatch Center (http://www.sldcguj.com).  ecent Trend in Short-Term Power Table 2.4 R Figure 2.7 Peak Demand Growth Purchase and Sale 10,000 9,335 9,338 9,515 Power purchase or sale 2007–08 2008–09 2009–10 8,000 Peak demand Puchase share 6,000 MW Long-term (percent) 97.27 96.84 99.03 4,000 Short-term (percent) 2.73 3.16 0.97 2,000 Short-term purchase 0 2007–08 2008–09 2009–10 Million kilowatt hoursa 1,470 1,761 580 Source: Authors’ calculations. Rs. (billion) 3.22 10.55 4.03 Rs. per unit 2.2 6 6.9 Gujarat Short-term sale To better understand the impact of load segregation Million kilowatt hoursa 1,595 1,328 3,604 in Gujarat, monthly averages of the daily differences between maximum and minimum demand were plot- Rs. (billion) 7.54 9.38 17.80 ted for FY2007–08 and 2008–09. Results showed that Rs. per unit 4.7 7.1 4.9 second half of FY2008–09 exhibited a dip in the monthly Net short-term purchase/sale demand difference, compared to the same period the Million kilowatt hours -125 433 -3,024 previous year, suggesting a flattening of the load curve (figure 2.6). Rs. (billion) -4.33 1.16 -13.78 Source: GUVNL annual report for the respective years. This analysis is also supported by the recent trend in a. Includes unscheduled interchange from the grid. Gujarat’s energy input and peak demand growth. From FY2007–08 to 2009–10, the state registered 10.39 per- period,9 while sale on the short-term market increased cent growth in energy input, compared to only 1.93 (table 2.4). The net impact has been positive, enabling percent peak-demand growth (figure 2.7).8 Flattening of GUVNL to reduce the overall cost of power purchase. the load curve can lead to better power-purchase plan- ning. Indeed, the short-term power purchased by Gujarat Urja Vikas Nigam Ltd (GUVNL) declined over the same 9. GUVNL is an electricity services holding company set up in 1999 by the former Gujarat Electricity Board as part of power-sector 8. Based on tariff orders of the Gujarat Electricity Regulatory Com- restructuring. mission and load generation balance reports of the Gujarat State Load Dispatch Center and Central Electricity Authority for the respective years. 12 Lighting Rural India: Load Segregation Experience in Selected States Figure 2.8 Cash Profit/Loss and Subsidy Trend, Gujarat 15.0 10,000 9,873 Subsidy received 12.5 11.78 12.06 11.00 53.88 Cash profit/loss 9,605 11.00 9,500 Agriculture GDP at 10.0 Million kWh 9,213 current prices Billion Rs. 9,062 7.5 9,000 Agriculture energy consumption 4.84 5.0 3.89 4.75 3.15 8,500 3.41 2.5 2.08 2.14 2.22 0 8,000 2005–06 2006–07 2007–08 2008–09 Source: Power Finance Corporation performance reports, Central Statistical Organization. Post-segregation years have witnessed dramatic growth Figure 2.9 Drivers of Financial Losses for in agricultural energy consumption, along with a mod- Distribution Utilities, Rajasthan est rise in agricultural GDP. A decline in the government a. Widening Gap between Energy Cost and Revenue power subsidy received by the agriculture sector, con- 6.48 comitant with rising profits of the distribution utility, have 6 marked a positive trend (figure 2.8). 5.19 5 4.90 4.48 4 3.53 Rajasthan Rs./kWh 4.12 2.92 2.84 3 2.71 Since 2005, Rajasthan has experienced substantial finan- 2 cial losses, which have distressed the state’s overall 1 financial position. A key factor has been the widening 0 FY 2006–07 FY 2007–08 FY 2008–09 gap between the average cost of energy supply and rev- enue realization per unit of energy sales. Over a two-year Average cost of supply per kWh sales period (FY2006–07 to 2008–09), the average supply cost Average realization per kWh sales increased by Rs. 2.0 per unit (from Rs. 4.48 to Rs. 6.48), Per-kWh cost of power purchase while the average realization per unit of sales decreased Source: Audited accounts of distribution utilities, Rajasthan. by Rs. 0.21 (from Rs. 2.92 to Rs. 2.71). The increased cost of power purchase has been a primary component b. Diminishing Subsidy and Growing Cash Losses in the increasing cost of supply (figure 2.9a). 100 Other major factors that have influenced increasing finan- 77.13 80 cial losses are declining subsidy support from the state 60 38.02 government and higher interest and finance charges. In 40 Billion Rs. 20 16.29 10.12 17.4111.43 14.27 10.51 FY2008–09, less than 14 percent of the total subsidy 0 requirement was disbursed (Rs. 10.51 billion out of the -5.12 -2.99 -20 -21.67 required Rs. 77 .13 billion), compared to about 38 percent -40 -60 received the previous year (figure 2.9b). From FY2006–07 -80 -64.86 to 2008–09, interest and finance charges rose by nearly 2005–06 2006–07 2007–08 2008–09 45 percent (from Rs. 7 .19 billion to Rs. 16.12 billion). Subsidy booked Subsidy received Cash profit/loss Source: Power Finance Corporation performance reports, Central Statistical Organization. Key Findings from Selected States 13 Figure 2.10 Economic Framework for Cost Benefit Analysis Cost Socioeconomic benefits Cost-benefit analysis • Stand-alone load segregation • Supply quality and reliability • Calculation of various payback • Load segregation integrated - Reduced outage frequency and scenarios of feeder segregation with HVDS/other CAPEX DT failure rate schemes for impact on • Beyond power sector, cost of - Improved load balancing - increased household income and excessive groundwater extraction • Household benefits - reduced fossil-fuel expenditure (not envisaged in this study phase) - Increased income - Reduced fossil-fuel expenditure due to increased electricity use • Agricultural benefits - Increased income resulting from better power supply (links between power and agriculture income not covered in this study phase) Source: Authors. Economic Evaluation the PGVCL distribution utility fell from 24.18 percent in FY2006–07 to 19.79 percent in FY2008–09.10 The rural feeder segregation schemes are envisaged to improve welfare outcomes through two direct channels: Virtual load operation on agriculture-dominated feeders (i) improved quality and reliability of electricity supply and using special design transformers (SDTs) was checked (ii) socioeconomic benefits. Better quality of supply is for system balancing. A single day’s hourly loading data related to such issues as voltage, outage frequency, and for all 11-kV agriculture and JGY feeders in the Vinchiya data on system balancing and the failure rate of distribu- substation was collected and analyzed to determine the tion transformers (DTs). Socioeconomic benefits relate loading pattern during single- and three-phase feeder to non-agriculture rural households’ increased hours of operations. The extent of system unbalancing was higher electricity; these are estimated by analyzing household during single-phase, SDT operations. Typically, SDTs income and expenditure on fossil fuels in the absence of cater to a small proportion of single-phase load owing electricity (figure 2.10). to physical segregation already being in place; thus, the impact of system unbalancing was found to be limited. The subsections that follow discuss the welfare out- comes for the quality-of-supply and socioeconomic Post-segregation survey results in Bassi, though not as variables in Vinchiya (Gujarat) and Bassi (Rajasthan) sub- dramatic as those in Vinchiya,11 also highlight significant divisions, based on findings from the 2011 survey con- reductions in supply interruptions and a marked improve- ducted by MRS Private Ltd. ment in consumer voltage profile. The percentage of households reporting power outages as “rare/never” increased nearly 34 percent (figure 2.12a), while those Quality of Supply reporting low-voltage problems fell by more than 71 per- In Vinchiya, the survey findings suggest a marked cent (figure 2.12b). Results of focus group discussions improvement in quality and reliability of supply resulting from rural feeder segregation. The percentage of house- holds reporting power outages as “rare/never” increased 10. According to the tariff order of the Gujarat Electricity Regulatory by more than 150 percent (figure 2.11a), while those Commission for FY2009–10. 11. The comparatively better supply reliability results in Vinchiya can reporting low-voltage problems decreased by about 74 be attributed, in part, to Gujarat having completed its load segrega- percent (figure 2.11b). Similarly, the DT failure rates of tion program some years earlier. 14 Lighting Rural India: Load Segregation Experience in Selected States Figure 2.11 Quality-of-Supply Changes Reported by Survey Respondents, Vinchiya a. “Rare/Never”Power Outages b. Low-Voltage Problems 100 100 Before segregation 85.9 80.0 80 76.8 80 After segregation 68.2 69.1 60 60 Percent Percent 52.0 40 33.7 40 28.6 20 13.0 20 10.5 5.9 4.6 0 0 Households Agriculture Business Households Agriculture Business and industry and industry Source: MRS Private Ltd Survey 2011. Figure 2.12 Quality-of-Supply Changes Reported by Survey Respondents, Bassi a. “Rare/Never” Power Outages b. Low-Voltage Problems 100 100 Before segregation 86.7 81.3 79.6 81.1 80 80 After segregation 60 60 Percent Percent 42.8 40 33.9 40 28.8 20 20 18.1 9.0 10.25 10.1 0.4 0 0 Households Agriculture Business Households Agriculture Business and industry and industry Source: MRS Private Ltd Survey 2011. show that farmers were satisfied with scheduled power  rend in Distribution Transformer Figure 2.13 T supply with minimum interruptions. Damage, Bassi Three-phase (agriculture) In Bassi, single-phase DTs were installed to cater to non- One-phase (households) agriculture connections, while existing three-phase DTs supplied the agriculture load. In the first year of opera- 20 17.4 Feeder tion, the single-phase DTs recorded a significant damage Renovation 15 Work 12.3 rate, which increased in subsequent years. The reason Percent 9.7 for such high failure rates was inappropriate provision- 10 11.6 ing during the planning stage for unauthorized loads and 8.7 7.3 5 5.4 future load growth. For the three-phase DTs, the dam- age rate decreased after segregation primarily because 0 FY2006–07 FY2007–08 FY2008–09 FY2009–10 of reduced loading resulting from shifting non-agriculture load to single-phase DTs. However, the failure rate has shown signs of increasing owing to load growth and pos- sible imbalance in some agriculture feeders due to higher Source: Office of the Assistant Engineer, Bassi subdivision, Jaipur. currents over a sustained period of time (figure 2.13). Key Findings from Selected States 15 Socioeconomic Benefits  conomic Internal Rate of Return from Table 2.6 E In Vinchiya, household income is expected to rise, owing Rural Load Segregation, Vinchiya to more time spent on income-generating activities and Contribution to income higher income from the sale of agricultural produce. Dur- generation (percent) ing the post-segregation period, the average monthly household income registered an increase not entirely EIRR calculation 25 10 5 attributable to increased time spent on income-generat- Contribution to 50 55 38 31 ing activities. For this reason, the study conducted a sen- reduced fossil-fuel sitivity analysis on various percentages of this attribution 25 50 32 23 consumption (table 2.5). (percent) 10 47 27 18 5 46 26 15 The enhanced supply hours resulting from rural load seg- Source: Authors’ calculations. regation will lead to an uninterrupted supply of quality power for both agriculture and non-agriculture consum- ers. However, supply hours will also depend on the over- Table 2.7 Change in Household Expenditure, Bassi all availability of power from the distribution utility and the prescribed supply schedules. The enhanced supply hours will substitute for end-user consumers’ existing Pre- Post- Parameter, monthly average (Rs.)a segregation segregation use of fossil fuels for household lighting and diesel gen- erators for irrigation. Income 6,667 6,802 Fossil-fuel expenditure 104 90 It was assumed that improvements in quality of supply Electricity expenditure 222 241 would increase income generation and reduce fossil-fuel Source: MRS Survey, 2011 expenditure in rural areas. For this reason, the benefits Note: Rural load segregation in the Bassi subdivision was initiated from quality of supply were not quantified for the eco- in FY2006–07 , with major works concluded by FY2008–09. Midway nomic cost-benefit analysis. The economic benefits were through the project execution period (i.e., FY2007–08) was consid- estimated for various scenarios based on the impact of ered the base year for survey responses on pre-segregation values; all data from previous years is converted to FY2007–08 levels. load segregation on income generation and contribution a. At constant FY2007–08 prices. to reduced fossil-fuel expenditure. only on the subdivision level and cannot be extrapolated The economic internal rate of return (EIRR) was found to to the entire state.12 be robust across a wide range of assumptions—from 15 percent up to 55 percent (table 2.6). Estimates are based In Bassi, the post-segregation period has exhibited an increase in time spent on income-generating activi- Table 2.5 C  hange in Household Expenditure, ties, which has resulted in higher monthly household Vinchiya incomes. The average household monthly expenditure on electricity has increased, while the average expenditure Parameter, monthly average Pre- Post- on fossil fuels has decreased (table 2.7). (Rs.)a segregation segregation Income 4,392 5,346 Since load segregation was part of the larger FRP, its esti- Fossil-fuel expenditure 124 135 mated cost was limited to supply and installation of 11-kV double-break isolators (400 ampere) and supply and Electricity expenditure 181 209 erection of single-phase DTs. The actual cost per feeder Source: MRS Private Ltd Survey, 2011. amounted to 23 percent of total program feeder cost. Note: Rural load segregation in the Vinchiya subdivision was initiated in FY2005–06, with major works concluded by FY2007–08. Midway The same assumptions were used to extract the cost through the project execution period (i.e., FY2006–07) is considered the base year for survey responses on pre-segregation values; all data from previous years is converted to FY2006–07 levels. 12. The micro-level benefit estimation against reduction in DT failure rates and power-supply reliability were not considered in the EIRR a. At constant FY2006–07 prices. calculation. 16 Lighting Rural India: Load Segregation Experience in Selected States Table 2.8 E  conomic Internal Rate of Return likely the case in Rajasthan, where agricultural power from Rural Load Segregation, Bassi consumption has more than offset the benefit of loss reduction. Furthermore, the current study has not deter- Contribution to income generation mined the economic benefits of rural feeder segregation (percent) on the growth of state agricultural GDP , which would EIRR calculation 25 10 5 need to be analyzed as part of the M&E framework. Contribution to 50 47 43 41 reduced fossil- 25 31 26 25 fuel consumption Summary Remarks (percent) 10 19 13 11 5 14 7 4 The financial and economic results of improved power Source: Authors’ calculation. supply in Gujarat and Rajasthan have been mixed. In the case of Gujarat, a static subsidy and limited agriculture power supply were accompanied by double-digit agri- of virtual feeder segregation. Economic benefits were cultural growth rates. By contrast, in Rajasthan, higher estimated for various scenarios based on the impact of agricultural consumption and financial losses have not load segregation on income generation and contribution resulted in commensurate agricultural growth. To derive to reduced fossil-fuel expenditure. the maximum benefit from investments, institutional and governance reforms need to accompany technologi- With the exception of reducing the contributions of rural cal changes. In the case of higher agricultural consump- load segregation to 5 percent for both fossil-fuel reduc- tion post segregation, subsequent study phases need to tion and income generation, the EIRR calculations are assess cross-sector linkages between the socioeconomic robust across a range of assumptions—from 11 percent cost of excessive groundwater extraction and the benefits up to 47 percent (table 2.8). of increased agricultural GDP. Such an integrated analysis is critical to formulating a comprehensive framework for the As stated in the framework for economic cost-benefit design and evaluation of rural load segregation schemes. analysis, these EIRRs are based on the estimated impact The next chapter highlights observations on commonly of power-sector supply improvement, but do not consider held perceptions about load segregation to better inform the cost impact of excessive groundwater extraction that the design and development of future programs. may result indirectly from rural load segregation; this is Key Findings from Selected States 17 Appendix: Haryana’s Monitoring and Evaluation Experience Haryana state recently designed a monitoring and evalu- in the post-segregation situation. Results revealed sig- ation (M&E) framework, implemented with a manage- nificant variation between utility records on consumer ment information system (MIS). The objective was to information and actual connections; a relatively large create a system to more accurately estimate agricultural proportion of consumers were without working meters. consumption following rural feeder segregation and In addition, there was significant variation in actual con- establish transparency in determining subsidies for the nected load vis-à-vis utility records; longer hours of three- distribution utilities. While the findings from this exercise phase supply (8–14 hours) were made available. Finally, reflect on-the-ground realities in Haryana, the lessons peak load was higher than connected load, indicating the can be applied in other Indian states. presence of unauthorized load. Initially, four methods were considered for estimating Given that Haryana lacks consumer indexing,15 feeder agricultural consumption, defined as follows (table 2.A1): input is clearly the most suitable method. Typically, seg- regated feeders are equipped with downloadable meters • Feeder input: agriculture consumption = energy input that can remotely transmit consumption information to into agricultural feeders – technical losses13 – com- data centers in corporate offices. Agricultural consump- mercial losses.14 tion at the feeder level can be estimated by subtracting • Feeder load factor: agriculture consumption = con- the technical and commercial losses from the feeder nected load (kW) x annual average feeder load factor energy input. However, this approach assumes a minimal x 8,760 (number of hours per year). non-agricultural load on these segregated feeders. Until • Energy audit: agriculture consumption = unmetered an interface with the feeder metering system is devel- consumption + metered consumption, where oped and segregation is clearly differentiated between unmetered consumption = feeder input – technical agriculture and rural household loads, feeder load factor, losses – metered consumption of non-agriculture a less data-intensive method, can also be considered. and agriculture consumers. • Specific consumption: agriculture consumption = spe- Development of the MIS can be customized to any of the cific consumption of metered agriculture consumers four methods for estimating agricultural power consump- (kWh per kW) x total connected load (kW). tion, depending on the availability of information (figure 2.A1). A detailed pilot study of eight feeders was conducted to establish data availability and on-the-ground realities The MIS depicted here is a web-based application (fig- ure 2.A2). It can generate various reports as required by 13. Technical losses were estimated by running a pilot on eight senior management of the distribution utility, regulator, selected feeders in Haryana’s two distribution utilities and simulat- or other key stakeholders (figure 2.A3). ing the load flows after determining system parameters (e.g., size and configuration of conductors and consumer details) by conduct- ing a walk-through survey along feeder and consumer premises. 15. Consumer indexing (i.e., mapping consumers and their loads to 14. Commercial losses were estimated by reducing aggregate tech- respective feeders) could have been conducted during the load seg- nical and commercial losses with technical losses. regation exercise at low incremental cost to the utility. Table 2.A1 Data Requirements for Agricultural Consumption Methods Parameter Feeder input Feeder load Energy audit Specific consumption Technical losses may be Single-line diagram Requires single-line computed based on a Not required. Not required. information diagram of the feeder. sample. Consumer indexing Not required. Not required. Required. Not required. Requires billing data Requires billing data Billing data Not required. Not required. found to be in error. found to be in error. Estimated commercial Requires a rough Assumes system has Required. Not required. losses estimate. no commercial losses. Source: Authors. 18 Lighting Rural India: Load Segregation Experience in Selected States Figure 2.A1 Implementing the Management Information System, Haryana Estimation of Agricultural Power Consumption for the State of Haryana Prerequisites Activities Approvals Segregation of agricultural Data collection from field level— Discussion with and approval of feeders/availability of predominant at sub-station, subdivision levels agricultural power consumption agricultural feeders values by top management of distribution company Updating of the MIS with field data Installation of MIS at corporate at regular intervals—at circle levels office and circle level Finalization of agricultural consumption values by each Computation of agricultural power distribution company for ARR filing Buy-in from concerned stakeholders consumption—at corporate office on proposed methodology level for each distribution company Source: World Bank 2011. Figure 2.A2 Screenshot of MIS Application, Haryana Source: World Bank 2011. Key Findings from Selected States 19 Figure 2.A3 Screenshot of MIS-Generated Report, Haryana Source: World Bank 2011. 3 Observations on Perceptions and Emerging Lessons The experiences of Gujarat and Rajasthan, analyzed in reliable, assured power supply. In Gujarat, for example, the previous chapter, offer evidence of prevailing views distribution utilities have received requests from a group on rural feeder load segregation schemes. Based on the of potato cultivators who are willing to pay commercial quantitative and qualitative data gathered from the sur- rates for increased hours of an assured supply. In such vey sample, this chapter offers observations on 11 com- cases where metering is possible, reliable solutions monly held perceptions and emerging lessons from the based on information technology (IT) are available. One study. such IT-based solution is Advanced Metering Infrastruc- ture (box 3.1). Perceptions and Observations Perception 2. Load segregation creates the need for additional infrastructure. Perception 1. Feeder load segregation is the only Observation. The common perception is that implement- solution to guarantee a continuous electricity supply ing load segregation creates the need for new feeders to non-agriculture rural connections. and additional transformers. However, the study findings Observation. Feeder load segregation—whether virtual show that existing infrastructure is aged and overloaded, or physical—has its associated financial and human requiring augmentation irrespective of segregation. At resource requirements. It is needed only in socio-political the outset, the newly segregated assets may not be opti- situations where metering of agriculture loads is not pos- mally loaded; however, the annual 6–7 percent demand sible. However, the study findings reveal cases where growth rate across consumer categories suggests that farmers are willing to bear the costs associated with a the infrastructure will be fully utilized within the next few Box 3.1 Advanced Metering Infrastructure Advanced Metering Infrastructure (AMI) can be used to limit supply hours to farmers without segregation of loads. This IT-based system uses two-way communication and back-end software applications with energy meters to mea- sure, collect, analyze, and control energy use on a real-time or near real-time basis. It is quite effective in communi- cating and implementing supply rosters through remote connect/disconnect and handling detection of unauthorized consumption, thus reducing losses. The system can be designed to perform a wide range of functions. It can imple- ment varying tariffs for consumers connected on the same feeder, including time-of-day tariffs, which can encourage farmers to switch on pump sets during off-peak hours and thus eliminate the need for remote connect/disconnect of irrigation loads. It can perform energy audits, provide utilities accurate data, and reduce billing-cycle costs. The AMI technology has matured and is now commercially available at an affordable cost. Source: Authors. 21 22 Lighting Rural India: Load Segregation Experience in Selected States years. Had the infrastructure been robust and optimal Perception 6. Feeder load segregation schemes initially, the requirements would have been limited to a result in socioeconomic benefits in rural areas simple reorganization of loads with the addition of a few through improved livelihoods and income and better new feeders. health and educational outcomes. Observation. The primary survey established that consid- Perception 3. Agricultural consumption will be erable socioeconomic development occurred over the restated post segregation, leading to a restatement period of the respective rural load segregation schemes, of loss levels for transmission and distribution. although their precise contributions could not be quan- Observation. Contrary to the common perception, the tified. However, if just a 5 percent increase in income study findings show that estimated agricultural consump- could be attributed to rural load segregation, the eco- tion is not based on data captured from the rural load nomic return would be a strong 15 percent. Although segregation program. In the case of Gujarat, accounting the link between the power subsidy and agricultural for agricultural consumption improved at the subdivision growth is beyond the scope of this study, it is interesting level after segregation. But at the state level, agricultural to compare results in the cases analyzed. Over the past consumption in utility and regulatory accounts—even five years, Gujarat has experienced double-digit agricul- five years after program completion—is based on ear- tural growth at the macro-level despite limited subsidy lier estimates. Similarly, in Rajasthan, where agricultural and power supply. By contrast, Rajasthan’s increased consumption increased post segregation at the level of agricultural consumption and consequent higher finan- the distribution utility, agricultural consumption for both cial losses have failed to translate into commensurate utility and regulatory accounts continues to be based on growth in the state’s agricultural GDP. prior estimates. Perception 7. Load segregation will lead to better Perception 4. The booking, allocation, and delivery rural service standards. of subsidies post segregation will be a transparent Observation. Load segregation has resulted in improved process that earmarks agriculture supply. supply hours in rural areas. However, the strategy for Observation. The overall impact of load segregation on supplying better-quality power in rural areas should also agriculture subsidies could not be established by this include load forecasting and distribution network plan- study owing to the lack of an agreed system with the ning, including a needs assessment for HVDS to ensure regulator for feeder data collection. In Gujarat, the state that the HT:LT ratio is within the norm. Supply expansion government allocates a fixed subsidy for agriculture sup- should also take the availability of generation capacity ply; however, in Rajasthan, as well as Andhra Pradesh into account. and Haryana, subsidy booking and allocation have exhib- ited a rising trend (annex). Perception 8. Investments are based on baseline data of agricultural consumption, and monitoring Perception 5. Financial return on investments in and evaluation systems have been put in place. load segregation is possible through loss reduction, Observation. Baseline data on such factors as existing loss increased revenue, and an improved load factor. levels, agriculture consumption, connected load, number Observation. In Gujarat, the strengthening of agriculture of consumers, and number of functional feeders was not accounting led to loss restatement and subsequent loss collected. Moreover, monitoring and evaluation (M&E) reductions at the subdivision level (Vinchiya). Following systems were not put in place during project develop- segregation, revenue assessment and peak-load man- ment. As a result, it was not possible to monitor system agement improved, but the proportionate contribution parameters post segregation or evaluate subsequent of the JGY scheme could not be established. In Rajast- benefits. Remote metering systems are not in place to han, the FRP resulted in loss reduction; however, a cor- capture detailed data on 11-kV feeders, even years after responding increase in agricultural consumption at the completion of physical work. level of the Bassi subdivision and distribution utility have resulted in greater financial distress. Observations on Perceptions and Emerging Lessons 23 Perception 9. Feeder segregation is a one-time Perception 11. Without customer metering, IT-based investment. tools are not useful for measuring and monitoring power consumption. Observation. In reality, feeder segregation is an ongoing activity requiring the setting up of systems to continu- Observation. When consumer metering is not possible, ously monitor and enforce discipline with regard to new meters can be installed at the substation or DT level. connections; otherwise, one runs the risk that segregated In the case of virtual segregation, meters installed agriculture feeders will be reconverted into non-agricul- at feeder take-off points at substations record time- ture ones within a few years time. During execution of stamped energy consumption, providing data on agricul- the rural segregation scheme in Gujarat, farmers residing tural consumption.17 In the case of physical segregation, in farmhouses demanded power supply along the pattern feeder-level energy input directly estimates agricultural of rural non-agriculture consumers. In response, special consumption. But to ensure that commercial losses (i.e., design transformers were introduced on agriculture- unauthorized and under-reported loads) are excluded dominant feeders to provide single-phase supply when from agricultural consumption and related subsidy pay- three-phase supply for agriculture was switched off.16 In ments to utilities, it is important to implement consumer Haryana, connections had to be provided from agriculture indexing and institutionalize a system for its continuous feeders to one-room tenets, known as “dhanis, ” and, in updating when new consumer profiles are added and some cases, large farmhouses in agricultural fields. All existing ones are revised.18 utilities must deal systemically with similar demands, which may occasionally arise. Currently, most distribution utilities implement enter- prise resource planning (ERP) systems, along with a Perception 10. Feeder segregation is a viable restructured-accelerated power development reform substitute for agriculture metering. program (R-APDRP). In an ERP environment, it is fairly simple to implement solutions that automatically update Observation. For segregated feeder systems, meters system parameters when loads are changed and perform are generally provided at the substation level, and the energy audits to generate risk flags whenever a mis- amount of energy channeled into feeders connected match between a connected load and energy consump- predominantly to agriculture consumers is available. But tion occurs. Even without these systems, implementing this energy input also includes technical losses and con- an IT-based solution on feeder meters with consumer sumption by unauthorized loads. While technical losses indexing in place is a powerful tool for monitoring and can be calculated, there is no way to estimate commer- evaluating agriculture consumption in the absence of cial losses since most utilities lack robust energy audit consumer metering.19 Yet ensuring that the information systems and operational practices to handle changes in generated is acted upon requires supporting institutional connected loads. Metering of agriculture consumers is and governance reforms at the level of the utility. essential to obtain data on customer-specific consump- tion and subsequently implement direct subsidy-deliv- ery mechanisms. But metering alone may not lead to a 24-hour supply for non-agriculture consumers unless 17. Metering data can be correlated with the agriculture supply schedule to estimate agricultural consumption; it can also be used farmers are willing to pay cost-reflective rates for con- to detect unauthorized system loads by comparing connected loads sumption beyond their allocated quota. Unless sup- with actual consumption. ported by an effective M&E system, feeder segregation 18. Consumer indexing refers to mapping consumers to specific is a necessary but insufficient tool for ensuring transpar- feeders and maintaining an updated record of consumer loads; this system works successfully when integrated across commercial ent subsidy allocation to farmers. (new connections and billing) and operations departments. 19. IT-based metering (e.g., advanced metering infrastructure or remote meter reading) can automatically capture meter data, trans- fer it to centrally located servers that analyze it and generate reports for both strategic and operational use, prepare consumer bills, and help perform energy audits (box 3.1). These systems can also be designed to remotely connect/disconnect loads. Most importantly, automated meter reading provides accurate and regular data that 16. Based on information provided by the Office of the Managing can be used for operational and strategic planning, without placing Director, GUVNL. added burden on the utility’s typically scarce manpower resources. 24 Lighting Rural India: Load Segregation Experience in Selected States Emerging Lessons It is also critical to communicate the objective of the load segregation exercise to field staff and institutionalize The evidence presented in this chapter suggests that a a system to retain segregated feeders while releasing standard approach to rural power supply involving rural new connections and modifying existing ones. To man- load segregation alone is unlikely to meet the states’ vari- age the switching of loads between feeders in cases of ous strategic objectives. As mentioned above, maximiz- breakdown, the utility should set up and institutionalize ing the benefits of load segregation schemes requires a system to track such changes and assign consumption accompanying institutional and governance reforms at to the appropriate feeder. Feeder segregation provides the utility level. At most substations, feeder meters com- the “hardware” for a system capable of delivering dif- patible with remote reading are already installed. Data ferentiated service to farmers and non-agricultural rural from these meters needs to be automatically collected consumers, along with management decision-making and analyzed. This will require setting up a data monitor- tools for effective monitoring. But the eventual outcome ing center dedicated exclusively to managing the informa- in terms of better quality of supply and sustainable opera- tion provided by the acquisition system and taking action tions is a function of the necessary “software;” that is, based on it. Operators should be trained extensively in the simultaneous and integrated application of organiza- appropriate use of the system and supported by crews tional changes, accountability systems, and use of infor- responsible for field inspections in potentially irregular mation technology. situations detected with support of the software. 4 Toward Sustainable Models: Guidance Note Following Gujarat’s relatively successful Jyoti Gram the development of a sustainable rural power supply Yojana (JGY) program, various other Indian states have system. initiated or re-started rural load segregation programs. To enable decision makers to adopt the approaches Institutional Framework that best fit their states’ on-the-ground realities, this chapter offers state governments and distribution utili- The proposed institutional framework for rural load seg- ties a guidance note on the various feeder segregation regation covers the entire project cycle, from conceptu- models to consider, along with key issues that must be alization and planning through execution and monitoring addressed at each stage of the project cycle to ensure and evaluation (M&E) (figure 4.1). Figure 4.1 Proposed Institutional Framework for Rural Load Segregation Conceptualization and Planning Execution Monitoring and Evaluation Objectives and drivers Dedicated project management unit Third-party evaluation of financial and socioeconomic benefits Evaluation of existing technical and Multi-tiered, multi-skilled project operational infrastructure management team Communication of outcomes Physical and socioeconomic Engagement with regulator parameters in rural areas Third-party quality assurance Baseline data collection Use of IT-based systems Stakeholder communication strategy Sustainability of segregated system Financing arrangements Financial and economic appraisal Detailed Project Report Procurement strategy Source: Authors. Note: The institutional framework can be applied to an entire state, a particular distribution utility, or even selected business units within a utility. 25 26 Lighting Rural India: Load Segregation Experience in Selected States Conceptualization and Planning Currently, five load-segregation models are prevalent in India. These are physical segregation of consumers con- Conceptualization nected to the rural distribution network with or without high voltage distribution system (HVDS) integration, vir- During the conceptual phase of the rural load segregation tual segregation of consumers using logical controls at scheme, decision makers must: the grid substation with or without HVDS integration, and mixed feeder with IT-based metering initiatives; the latter • Identify the strategic objective of load segregation. model can be used with any of the other four options • Evaluate alternative models to achieve the strategic for transparent measurement of consumption and effec- objective. tive load management. The study conducted a decision- • If load segregation is selected as the most viable matrix exercise that separately mapped the strategic option, decide on the best-fit model based on the objectives and the two sets of state-specific parameters strategic objective and analysis of state-specific against the five load-segregation models. The results of parameters. the exercise are presented below. Identify the strategic objective. As previously discussed, the Best-Fit Models by Strategic Objective strategic objective of load segregation can vary widely, Table 4.1 presents the decision matrix for selecting the ranging from socioeconomic development of rural areas best-fit load segregation model available to the distribu- to achieving loss reduction and enhanced transparency tion utility by strategic objective. After obtaining a set of in estimating agriculture consumption. States may also “high,” “medium, ” and “low” results against the appli- be interested in improving non-agriculture consumers’ cable objectives, the utility can decide on the best fit. supply quality and reliability or enhancing agriculture con- sumers’ supply quality and quantity by creating additional In terms of loss reduction, physical segregation with network capacity and improving network infrastructure. HVDS is the most suitable model since it reduces the While most states are interested in all of these objec- LT network and enables agricultural metering using DTs. tives, prioritization will likely differ by state. For example, Virtual segregation without HVDS, which provides only in Haryana, loss reduction was the primary objective minimal reduction in the LT network and no assistance in of load segregation; Rajasthan’s main objective was metering agricultural consumption, is the least suitable strengthening power-supply infrastructure for farmers, option. The remaining models are moderately suitable while Gujarat desired that rural household connections when weighed against the loss-reduction objective. have power supplied 24 hours a day. If the strategic objective is ensuring transparency in agri- Evaluate alternative models. Given that load segregation is cultural consumption and determining subsidy levels, capital intensive, alternative models to achieve the stra- physical segregation with or without HVDS is highly suit- tegic objective should be evaluated. For example, if the able if appropriate systems are set up within the utility to strategic objective is limited to loss reduction, a more regularly collect and analyze the energy flow data from effective option may be strengthening energy audit prac- segregated feeders. IT-based AMI systems, equipped tices and performance accountability in the field, along with automated data collection and analysis functions, with using an IT-based system to prepare an MIS. If con- are highly suited to accurately calculating agricultural sumer metering is possible or if DTs are dedicated to consumption and subsidy estimates. In fact, IT-based rural loads, IT-based metering (i.e., AMI) at the consumer solutions enable regulators and other external stakehold- or DT level can be an effective way to implement supply ers to access consumption data on a real-time basis. rosters, measure consumption, and thus accurately esti- Virtual load segregation models are moderately suitable mate system losses without undertaking load segrega- owing to limitations in agriculture metering of end-user tion (box 3.1). consumers, even after segregation. Decide on the best-fit model. If load segregation is found to For supply-related objectives, physical segregation mod- be the most viable option, a best-fit model will need to els and IT-based AMI are quite a good fit since they be selected based on the strategic objective and analy- enable load management on both agriculture and non- sis of state-specific parameters of existing technical and agriculture connections. Virtual segregation models, operational infrastructure and the physical and socioeco- which have all rural consumers on the same network, nomic status in rural areas. are only moderately suitable owing their extensive load management requirements. Toward Sustainable Models: Guidance Note 27 Table 4.1 Decision Matrix for Strategic Objectives Virtual segregation Physical segregation Mixed feeder with IT-based AMI with remote meter reading Strategic objective Without HVDS With HVDS Without HVDS With HVDS and connect/disconnect Loss reduction Low Medium Medium High Medium Transparency in agriculture Medium Medium High High High consumption and subsidy Continuous, 24-hour supply to non-agriculture Medium Medium High High High consumers Quality and quantity of supply to agriculture Medium High Medium High Low consumers Source: Authors. Note: “High” = most suitable; “Medium” = moderately suitable; “Low” = least suitable. If the objective is ensuring agriculture consumers quality connections may pose implementation complexities and quantity of supply, releasing pending agriculture con- owing to consumer discontent with direct metering; nections and planning for future ones would require cre- thus, this option is the least suitable. That said, imple- ating additional distribution infrastructure. HVDS-based menting some type of IT-based solution to automatically load segregation models, whether virtual or physical, collect and analyze energy data at the substation level provide the added infrastructure required and thus are would be useful. the best fit, followed by virtual or physical load segrega- tion without HVDS. IT-based AMI creates no additional If the state distribution utilities have suffered significant infrastructure and thus is the least suitable choice. financial losses in recent years, it would not be feasible to implement capital-intensive projects. The best option Evaluating Existing Technical and Operational Infrastructure would be direct agriculture metering through IT-based Table 4.2 presents the decision matrix for evaluating the AMI, followed by virtual load segregation without HVDS. five segregation models against the existing technical The other three models score low against this business and operational infrastructure constraints of the distribu- scenario. ” “medium, tion utility. After obtaining a set of “high, ” and “low” results against the applicable parameters, the util- For a business scenario with high AT&C losses, the dis- ity in question can decide on the most techno-economi- tribution utility would require a metering infrastructure cally feasible solution. to accurately measure losses in the rural and agricultural networks. The best-fit options are agriculture metering States with a large proportion of LT networks are sus- using IT-based AMI or DT metering using HVDS. While ceptible to higher technical losses and power pilferage. physical segregation without HVDS would segregate the Both physical and virtual segregation models with HVDS agriculture distribution network, it would provide for con- provide for a higher HT:LT ratio and thus are the best fit. solidated metering of agricultural consumption; thus, this An IT-based AMI solution is least suitable since it fails to option is a moderately good fit. Virtual segregation with- provide for any change in network structure. The virtual out load segregation is the least suitable option since it and physical segregation models without HVDS are mod- does not support agriculture metering. erately suitable for this business scenario. States with inadequate power supply to meet electricity For states with a high proportion of unmetered agricul- demand would be unable to maximize benefits from load tural consumption, it is essential to create a segregated segregation; thus, the most suitable models are IT-based network that can support agriculture metering. Physical AMI and virtual segregation without HVDS. Physical load segregation models with or without HVDS and virtual segregation models have the lowest applicability since segregation with HVDS make it possible to provide for they would incur high capital costs but would be unable dedicated metering infrastructure for agriculture connec- to attain the benefits of continuous power supply to non- tions. By contrast, IT-based AMI on unmetered agriculture agriculture consumers. 28 Lighting Rural India: Load Segregation Experience in Selected States Control of agriculture supply hours helps to regulate suitable option. Though virtual segregation without HVDS groundwater usage. If the state distribution utility has may not require added infrastructure, it would have little the necessary skills set and consumer acceptability to effect on supply quality and load management and thus control the supply hours for agriculture connections, then is ranked low. physical load segregation models and IT-based AMI are the best choices since they enable improved control of Rural load segregation is expected to result in increased agriculture supply. Virtual segregation models are mod- agricultural consumption owing to improved supply hours erately suitable owing to their comparatively lower load and release of pending connections. In such a scenario, management capability. a limited agriculture subsidy would place added financial burden on the state distribution utility; thus, IT-based For states that exclude agriculture load from peak load AMI would be the best fit since it would not support management, load segregation can achieve the same increased agricultural consumption. Physical segrega- benefits. Based on the operational ease of peak load tion models are least appropriate, while virtual segrega- management, the physical segregation models and IT- tion models are moderately suitable, depending on the based AMI are the most suitable options, while virtual potential for increased agricultural consumption follow- segregation without HVDS is the least suitable. ing segregation. Evaluating Physical and Socioeconomic Status For rural areas with low levels of customer service, Table 4.3 presents the decision matrix for evaluating HVDS-based load segregation models are the most suit- the five segregation models against the physical char- able options, given their ability to improve supply qual- acteristics and socioeconomic status of the rural areas ity. Based on this criterion, segregation models without considered. HVDS are a moderately good fit, while IT-based AMI is ranked low. For states with a large geographical spread, segrega- tion would entail a rural distribution network of signifi- If states require intensive communication with farmers cant length, requiring large capital expenditure. Given before initiating rural schemes, the best options are phys- its cost-effectiveness, IT-based AMI would be the best- ical segregation models, which entail such communica- fit solution, followed by virtual segregation with HVDS tion. Virtual segregation models are a moderately good as a medium fit. Because of the added infrastructure fit, while IT-based AMI is least suitable. that may be required, physical segregation is the least Table 4.2 Decision Matrix for Existing Technical and Operational Infrastructure Virtual segregation Physical segregation Mixed feeder with Business scenario of IT-based AMI with technical and operational remote meter reading and infrastructure Without HVDS With HVDS Without HVDS With HVDS connect/disconnect Low HT:LT ratio Medium High Medium High Low Low metering of Medium High High High Low agriculture sales High financial losses in Medium Low Low Low High utility High AT&C losses Low High Medium High High Inadequate power supply High Medium Low Low High Ability to control supply hours to agriculture Medium Medium High High High consumers (with ground- water implications) Wide scope for peak load Low Medium High High High management Source: Authors. Note: “High” = most suitable; “Medium” = moderately suitable; “Low” = least suitable. Toward Sustainable Models: Guidance Note 29 Table 4.3 Decision Matrix for Physical and Socioeconomic Parameters Virtual segregation Physical segregation Mixed feeder with Business scenario of IT-based AMI with physical, social, and remote meter reading and economic parameters Without HVDS With HVDS Without HVDS With HVDS connect/disconnect Large geographical spread Low Medium Low Low High Large number of agriculture connections Medium Medium Low Low High with limited subsidy Low customer service Medium High Medium High Low levels in rural areas Intensive communication Medium Medium High High Low with farmers Source: Authors. Note: “High” = most suitable; “Medium” = moderately suitable; “Low” = least suitable. Remarks Develop a stakeholder communication strategy. The com- This decision-matrix exercise has considered the three munication strategy should aim at identifying appropri- sets of parameters in isolation; yet in reality, most are ate target audiences (e.g., public authorities, lending interlinked. In practice, states and distribution utilities agencies, regulators, and consumer representatives) and should identify the most critical parameters for their requisite communication channels, ranging from public unique situations and utilize the results of the exercise. media to direct presentations. It is critical that a consul- Ultimately, they will need to review all of the relevant tative forum be used by the distribution utility to inform parameters as a whole before deciding on the best-fit key stakeholder groups (e.g., regulators, consumers, load segregation model. and village administrative bodies [panchayats]) about the project’s rationale and benefits. The consultative forum can be used to inform consumers about the cost of the Planning segregation scheme and help them to appreciate the Planning the design and implementation of the rural seg- value of national resources deployed for irrigation. The regation scheme includes the following key steps: forum can also be used to obtain consumer feedback to develop an optimum project design. This approach can • Prepare a robust baseline. inculcate a sense of ownership among stakeholders and • Develop a stakeholder communication strategy. thus ensure the project’s success. • Decide on financing arrangements. • Undertake financial and economic cost benefit Decide on financing arrangements. Prior to project imple- analyses. mentation, it is important to decide on the financing • Prepare a Detailed Project Report (DPR). arrangements. For example, given a project’s potential • Adopt an appropriate procurement strategy. socioeconomic benefits, the state government may decide to contribute all or part of the required funding. Prepare a robust baseline. The proposed project requires a robust baseline of technical, financial, and economic Undertake financial and economic cost benefit analyses. parameters (e.g., consumer profile, consumption pat- Financial benefits, calculated against system costs, tern, collection efficiency, hours or supply, and loss lev- include increased revenue from loss reduction and reduc- els) through a third-party assessment. So that project tion in power purchase cost through peak load reduction targets can be set and benefits evaluated, the baseline (figure 2.1). The payback period or internal rate of return parameters should be monitored from the scheme’s (IRR) should be calculated accordingly. Economic bene- conceptualization phase through evaluation. If the utility fits, calculated against power-sector costs and water-sec- has chosen to implement physical load segregation, con- tor extraction, include improved quality of power supply sumer indexing can also be conducted with low incre- through reduced outage and improved voltage, increased mental cost to the utility. household income, and higher agricultural output (figure 2.10); ancillary socioeconomic benefits resulting from 30 Lighting Rural India: Load Segregation Experience in Selected States increased hours of rural non-agricultural supply include institutionalizing a review system of senior management better health and educational outcomes and increased or state government, and facilitating experience-sharing employment opportunities in villages. between the PMUs. At the middle-management level, a cross-functional team should be set up. At the field Prepare a Detailed Project Report (DPR). All analyses level, dedicated PMUs with technical and management should be captured in a streamlined, standardized DPR expertise should be established to ensure effective proj- template. The DPR should include the baseline data, ect implementation. The PMUs are responsible for con- expected project outcomes, financing plan, financial and tract management, inspections, progress review, and economic appraisals of the investment, implementation resolving day-to-day issues. Dedicated project managers arrangements, stakeholder communication strategy, and should be appointed for the entire project cycle. Given monitoring and evaluation (M&E) framework. the limited resources of distribution utilities, third-party, quality-assurance consultants should be hired on a com- Adopt an appropriate procurement strategy. The strategy petitive basis to conduct field supervision and factory selected by the implementing distribution utility depends inspections. on the mode of execution (e.g., in-house; turnkey engi- neering, procurement, and contracting [EPC]; or more Institutionalize the principles of load segregation. Feeder evolved public private partnership [PPP] mechanisms that load segregation is not a one-time activity. Longer-term outsource system operation and maintenance). Deciding sustainability requires developing and institutionalizing on in-house execution should take into account the util- the system to ensure that the principles of load segrega- ity’s past experience in materials procurement. Similarly, tion are followed across the organization (e.g., when new deciding to partially or fully outsource on a turnkey basis connections are released or feeder-failure events occur). should consider past performance of the prospective turnkey contractors in the project implementation area; Ensure regulatory acceptance. From the project outset, turnkey EPC is recommended since it would ensure par- keeping state regulators informed of all key decisions ticipation of reputed industry players, leading, in turn, to is critical to facilitating required regulatory investment implementation efficiency and quality assurance. If the approvals. Incorporating regulatory expectations into the utility decides to procure materials separately, third-party project’s documented objectives can enhance regulatory quality-assurance agencies should be engaged to con- buy-in on project results and future regulatory engage- duct inspections. ment when, based on the project results, agricultural and rural consumption norms are revisited. Execution Use IT-based solutions. IT systems should be used to sup- port automated data collection and analysis and M&E. Subsequent to project conceptualization and planning is Depending on the load segregation model that the state field-level implementation. The key steps that comprise adopts, installations can be at the feeder, DT, or consumer the scheme’s execution phase are as follows: level. IT-based solutions can help utilities overcome the dearth of available manpower for remote meter reading, • Ensure a multi-tiered, multi-skilled project man- manual errors in data recording, and lack of detailed con- agement organization, including a dedicated proj- sumption data for future analysis. Because these solu- ect management unit (PMU), project managers tions represent a paradigm shift from business as usual, appointed for the entire project cycle, and third-party substantial change management is required to imple- quality assurance.20 ment institutional and governance reforms to maximize • Institutionalize the principles of load segregation. the benefits of load segregation investments. • Ensure regulatory acceptance. • Use IT-based solutions. Monitoring and Evaluation Ensure a multi-tiered, multi-skilled project management set- up. At the corporate level, this means setting up a spe- After project execution, the final phase of the cycle is cialized cell responsible for centralized project monitoring monitoring and evaluation (M&E)—a dynamic process and taking requisite corrective and preventive measures, comprising regular feedback on post-implementation benefit estimates and results-sharing with key stake- holders and decision makers. The main steps of the M&E 20. The project management structure and allocation of responsibili- process are as follows: ties should be decided on during the project’s conceptualization and planning phase. Toward Sustainable Models: Guidance Note 31 • Appoint a creditable third party for quality-assurance lessons learned should be captured and used to improve M&E. the design and implementation of future feeder load seg- • Set up a knowledge hub on rural power supply. regation schemes. • Disseminate results and share experience and les- sons learned with stakeholders and utilities. Summary Remarks Appoint a creditable third party. The implementing utility or state government or regulator should appoint a credit- In 2011, the findings of this study were shared with six able third party to evaluate both the financial and socio- states through consultations with relevant ministries and economic benefits of the rural load segregation scheme distribution utilities. In Gujarat and Rajasthan, the sample and document improvements against the baseline. Eval- cases for this study, the findings were broadly endorsed uating the socioeconomic benefits requires analyzing a (appendix). Utility officials in Gujarat expressed a desire selected sample from the primary survey of end-user to move ahead with a pilot on direct subsidy delivery consumers. The sample selection must be based on rel- and an agriculture power supply system based on 100 evant statistical techniques and should include all con- percent consumer metering for willing customers (i.e., sumer classes impacted by load segregation. The benefit potato growers). In Rajasthan, effective load manage- evaluation should not be initiated until at least one year ment post segregation has meant restricting increased after project completion to ensure that significant post- peak demand to about 67 percent against a 130 percent execution data is included. increase in connected load of agriculture consumers. Madhya Pradesh is about to start a rural feeder segrega- Set up a knowledge hub. A knowledge hub or center of tion program based on the Gujarat model. Uttar Pradesh excellence should be established within the CEA to has formulated a program to set up separate substations undertake the design and implementation of the rural for rural areas and sub-district towns, and Bihar intends power supply system. to provide dedicated rural feeders for agriculture supply only. The diverse results of these state consultations Disseminate results and share experience and lessons reinforce the general finding that drivers of rural power learned. Results of the project evaluation should be dis- supply differ markedly across Indian states; thus, solu- seminated to all concerned stakeholders, including tions for improving rural power supply must be custom- the utilities. The implementing utility’s experience and ized to fit each state’s requirements. 32 Lighting Rural India: Load Segregation Experience in Selected States Appendix: Consultations Held in Gujarat, Rajasthan, and Madhya Pradesh Gujarat: Minutes of Meeting with the Principal Secretary of Energy and Petrochemicals Meeting topic Sharing findings of World Bank study on rural feeder load segregation Date/location October 4, 2011/Gandhinagar Attendees Gujarat World Bank team D. J. Pandian, Principal Secretary, Ashish Khanna Secretary of Energy and Petrochemicals Kavita Saraswat S. B. Raval, Managing Director, PGVCL Pricewaterhouse Coopers (PwC) C. L. Sharma, Chief Engineer, UGVCL Mohammad Saif N. Srivastava, Managing Director, UGVCL Suresh Shahdadpuri, Chief Nodal Officer (e-Urja and RAPDRP-A), UGVCL Agenda items Sharing findings with and receiving feedback from the Gujarat power-sector team on the rural feeder load segregation study. Discussing state-specific issues in preparation for the proposed national guidance note on rural load segregation. Components Discussion points Field visits Prior to the meeting, the World Bank study team visited the UGVCL Energy Management Center, UGVCL Area Load Dispatch Center, Rupal village (to interact with rural consumers covered under the JGY), and the Randheja substation. The team noted that UGVCL has implemented automatic data acquisition at the feeder level and is conducting effective energy management and auditing. Load scheduling implemented through the Area Load Dispatch Center allows the utility to restrict its peak demand to about 3,000 MW against a connected load of 7,300 MW (figures quoted by the utility). Sharing of study The World Bank team presented the study background, approach, and key findings. The audience agreed with findings the study findings specific to Gujarat. The UGVCL Managing Director noted that the agriculture subsidy, as included in the study, pertains only to unmetered agricultural consumption. It was suggested that the study report highlight the total subsidy against agricultural consumption. The requisite data was subsequently requested from the concerned official. State-specific Scheme data not utilized for agriculture approvals. Gujarat still uses fixed consumption norms for queries and unmetered consumers as part of regulatory submissions. It was agreed that the actual agricultural data inputs for should be shared with the regulator to finalize the annual unmetered consumption. national guidance Marginal investments in remote meter reading to enable web-based disclosure of agricultural data. note While individual agricultural consumption might not be disclosed publicly, feeder/DT-level data can be publicly shared for transparency. The Principal Secretary suggested that, for all states that have successfully segregated rural load, the central government should provide financial support for installing RMR infrastructure on segregated feeders. Moving from load segregation to individual metering to enable direct subsidy delivery. Gujarat is ready to evaluate the direct subsidy model on a pilot basis; however, it was agreed that the pilot would require installing advanced metering infrastructure to extract the requisite data from end-user consumers. Key JGY lessons to carry forward to other states interested in similar initiatives: • A strong procurement system (including material, quality, and vendor management systems) must be in place before initiating such state-level schemes. • The key challenge is the demand for 24-hour supply by greenhouse consumers supplied through segregated agriculture feeders. • The Managing Directors’ view was that the model to be followed eventually is 100 percent consumer metering supported by AMR, which can contribute to direct-subsidy delivery in rural areas by implementing time-of-day tariffs. A group of agricultural consumers is willing to pay the market price for reliable, good-quality power. These initiatives could be started with this group. Toward Sustainable Models: Guidance Note 33 Rajasthan: Minutes of Meeting with the Secretary of Energy Meeting topic Sharing findings of World Bank study on rural feeder load segregation Date/location October 12, 2011/Jaipur Attendees Rajasthan World Bank team Shailendra Agarwal, Chairman and Managing Ashish Khanna Director, RVPNL Kavita Saraswat Y. K. Raizada, Technical Director, RVPNL Pricewaterhouse Coopers (PwC) Naresh Pal Gangwar, Secretary of Energy, Mohammad Saif Government of Rajasthan A. K. Gupta, Technical Director, JVVNL Finance Director, JVVNL Superintendent Engineer, Jaipur City, JVVNL T. S. Sharma, Superintendent Engineer/MIS, JVVNL Agenda items Sharing findings with and receiving feedback from the Rajasthan power-sector team on the rural feeder load segregation study. Discussing state-specific issues in preparation for the proposed national guidance note on rural load segregation. Components Discussion points Field visits Before the meeting, the World Bank study team visited the Bassi subdivision selected for the load segregation study. The team was informed that agriculture consumers are connected through HVDS, and meters are provided on all transformers; thus, consumers are metered on a 100 percent basis. However, some meters are defective. No IT-based solution has been implemented to read and process the data. Sharing of study The World Bank team presented the study background, approach, and key findings. The audience raised the findings following key points: • Increased agriculture connections post completion of the FRP were due primarily to the release of all pending connections for the past 10–15 years, increased supply hours to agriculture consumers, and the rise in connected load per consumer due to groundwater depletion. The team indicated that the study report would be updated to reflect this information. • Rajasthan was regulating the supplied agriculture connections even before the FRP was initiated; thus, the FRP might not result in the flattening of the demand load curve as in Gujarat. It was agreed that the requisite data substantiating this statement would be shared with the study team for further analysis and validation. • Because subdivisions in Rajasthan differ significantly, the study report should include a caveat highlighting that the detailed study was limited to a sample subdivision; this caveat was subsequently highlighted in the report. State-specific Primary objectives of load segregation. The primary objective was improving the quality and hours of queries and supply for rural consumers. inputs for Scheme data not utilized for agriculture approvals. Unmetered agriculture consumers are billed using national guidance standard consumption norms prescribed by the regulator. While unmetered consumers have individual note meters on DTs, they are not read regularly. It was agreed that actual consumption data can be used for approvals from the regulator. Marginal investments in remote meter reading to enable web-based disclosure of agricultural data. Most existing DT metering of agriculture lacks a communication facility. To date, the data has not been compiled owing to the large capital investments required and equipment security concerns. Key FRP lessons to carry forward to other states interested in similar initiatives: • The new FRP scheme proposed will integrate the installation of energy-efficient pump sets as a demand- side mechanism (DSM) measure. • Agri-products should be linked to the subsidy. • Smaller transformers (10 kVA) should be used in HVDS. 34 Lighting Rural India: Load Segregation Experience in Selected States Madhya Pradesh: Minutes of Meeting with the Distribution Utilities Meeting topic Sharing findings of World Bank study on rural feeder load segregation Date October 19, 2011 Attendees Madhya Pradesh World Bank team Nitesh Vyas, Managing Director, Madhya Pradesh Ashish Khanna MKVVCL Pricewaterhouse Coopers (PwC) Feeder separation project team members from the Mohammad Saif three Madhya Pradesh distribution utilities Representatives of the REC Agenda items Sharing the study findings from Gujarat and Rajasthan on rural feeder load segregation. Discussing the proposed institutional framework for load segregation and its applicability to the Madhya Pradesh distribution utilities. Components Discussion points Sharing of study The World Bank team presented the study background, approach, and key findings in the context of their findings relevance to the Madhya Pradesh rural load segregation scheme. The Madhya Pradesh team shared the following information. Madhya Pradesh distribution utilities have a total of 75 segregation schemes, averaging about US$9.4–11.3 million (Rs. 50–60 crores) per scheme. A third-party project management consultant has been hired to manage the planning, design, execution, and monitoring of the scheme. Nine months have passed since the contract for execution was awarded; no pilots were undertaken. Execution of the scheme will be outsourced using turnkey contracts. The distribution utilities have decided to adopt the Gujarat model of physical feeder segregation and aim to restate their distribution losses post segregation. The East distribution utility has prepared a separate scheme for installing feeder/DT-level automated metering infrastructure on the segregated feeders. The decision to connect agriculture or non-agriculture load to the new feeder created as a result of physical segregation will depend on the respective load proportions on a given feeder. Institutional A general framework for undertaking load segregation was prepared using the study findings presented. The framework for rural load segregation scheme of Madhya Pradesh was mapped onto the framework. The key highlights were load segregation as follows: • Strategic objective. In Madhya Pradesh, the primary objective of feeder segregation is continuous supply to non-agriculture rural consumers. The approach is physical segregation without HVDS, except where agriculture connections have high connected load. Feeder segregation is planned for selected state districts where feeders have been selected, based on existing loss levels. • Conceptualization – Technical and operational evaluation. Key positives for the proposed approach to load segregation are low metering of agriculture connections in the state, large scope for peak load management, and ability to control supply hours on agriculture feeders. However, high financial losses of state utilities and existing inadequate supply are deterrents to the proposed scheme. – Socioeconomic status in rural areas. The physical separation may be comparatively costly owing to the state’s large geographical spread. However, the scheme would require limited communication with farmers compared to other load segregation models. – Execution and evaluation phases. The three project teams of the distribution utilities were updated on basic project requirements to ensure a robust and successful execution of the scheme and its subsequent evaluation. These include (i) preparation of a robust baseline and subsequent financial and technical targets; (ii) procurement strategy that optimizes existing state practices through a cell with technical and project management expertise; (iii) tier-based project management for robust monitoring; (iv) continuous regulatory engagement from the outset of the evaluation phase; (v) use of advanced metering infrastructure to support data collection, monitoring, and evaluation of the scheme; (vi) third-party inspections during and after project execution, and (vii) engaging a creditable third party to evaluate both the financial and social costs and benefits of the scheme vis-à-vis the documented baselines. Toward Sustainable Models: Guidance Note 35 Madhya Pradesh: Minutes of Meeting with Water Resource Department Meeting topic Sharing findings of World Bank study on rural feeder load segregation with the Principal Secretary of the Water Resource Department, Madhya Pradesh Date October 19, 2011 Attendees Madhya Pradesh World Bank team R. S. Julaniya, Principal Secretary, Water Resource Ashish Khanna Department Pricewaterhouse Coopers (PwC) Mohammad Saif Agenda items Sharing the study findings from Gujarat and Rajasthan on rural feeder load segregation. Discussing the impact of Madhya Pradesh’s upcoming rural load segregation scheme on the state’s groundwater reserves. Components Discussion points Sharing of study The Principal Secretary was updated on the study undertaken in Gujarat and Rajasthan. Details of the findings ongoing load segregation scheme by the distribution utilities in Madhya Pradesh were also shared. Impact of load The impact of rural load segregation on the state’s groundwater levels was discussed at length. The views segregation on shared by the Principal Secretary were as follows: groundwater • Every three years, the state water department conducts a study of the depleted blocks. According to the reserves results of these studies, there is a 5 percent marginal increase in the depleted blocks. The groundwater situation appears alarming due to variations in the block classification criteria adopted in various studies; thus, feeder segregation is not expected to have a large impact on the state’s groundwater levels. • Feeder segregation must be primarily an energy-sector scheme; inter-linkages with other sectors would make it difficult to coordinate across departments. • The agriculture business does not receive requisite returns, which explains the need for subsidy to farmers. The agriculture subsidy can be removed if the agriculture sector can be assured of needed financial returns. 5 The Way Forward This study has demonstrated that there is no one-size- At the central level, setting rules of engagement and fits-all solution to improving rural power supply. Project principles for ensuring improved rural power supply while proposals should be evaluated as part of each state’s maintaining techno-economic viability can be achieved broad strategic program for improving rural power sup- using a common strategic framework for designing the ply. Given the enormous amounts of planned or already most optimal system. If feeder load segregation emerges allocated investments by various states in India, there is as the best solution, it should be amenable to the direct an urgent need to establish centralized rules of engage- delivery of subsidies to farmers. Improved measurement ment outlining the principles that should underpin the and reliability of agricultural consumption data are essen- design of any initiative to improve the sustainability of tial starting points that could lead to a win-win situation rural power supply while maintaining techno-economic for all stakeholders. Automated meter reading (AMR) and viability. similar initiatives based on information technology should be an integral part of rural power system design. The final results of this study were presented to India’s Ministry of Power in April 2012. Based on these consulta- The recommended knowledge hub set up within the tions, several key recommendations were made. First, CEA should be responsible for developing standard doc- it was decided that a knowledge hub or center of excel- umentation templates for the Detailed Project Report lence should be established within the Central Electricity (DPR), as well as project management and operational Authority (CEA) to undertake the design and implemen- manuals, technical specifications, and standard bidding tation of the rural power supply system. It was also documents. In addition, it should develop processes for advised that one or two states far along in implementing project implementation, data management, and inte- their feeder load segregation schemes create integrated grated operations to ensure the sustainability of the rural data centers to collect and analyze data for such strategic power supply system. The knowledge hub can assist purposes as ensuring transparency in determining sub- states that desire to follow the strategic framework with sidies for distribution utilities and improving operational project design and implementation, including the use efficiency. Furthermore, it was decided that one or two of AMR. Finally, the experiences and outcomes of the states on the threshold of undertaking rural load segrega- demonstration projects should be widely disseminated tion should be selected to work with the CEA on the con- by the knowledge hub so that lessons in success can be ceptualization and design of improved rural power supply. replicated across the country. 37 Annex State Profiles This annex provides summary profiles of the four Indian Andhra Pradesh Electricity Reform Act, enacted in 1998, states included in the rural feeder load segregation study: provided for the constitution of the Electricity Regulatory Andhra Pradesh, Gujarat, Rajasthan, and Haryana. Each Commission and restructuring of the power industry. In summary highlights key information, including dates of early 1999, the APSEB was unbundled into APGENCO power sector reform, technical and financial trends of and APTRANSCO. The following year, APTRANSCO was the distribution utilities and consumer mix, state climate further unbundled so that its management was limited to and crop characteristics, and elements of the segrega- transmission, while four utility companies were formed tion scheme. to manage distribution in the state’s central, eastern, northern, and southern zones, respectively.24 Andhra Pradesh Technical and Financial Trends Andhra Pradesh is India’s third largest state, with an In recent years, Andhra Pradesh has witnessed a con- area of 276,754 km2.It accounts for 8.4 percent of the tinuing decline in distribution losses (figure A.1). At the country’s territory and has its longest coastline (972 km). same time, the subsidies booked by the distribution utili- The state is endowed with a variety of physiographic fea- ties have been less than those received from the state tures, ranging from high hills to undulating plains and a government, causing financial losses to increase (figure coastal deltaic environment.21 Andhra Pradesh ranks fifth A.2).25 in terms of population, at about 84.7 million.22 In 2010, state GDP was US$77 .6 billion (Rs. 4.1 trillion), account- ing for 7.63 percent of the country’s GDP . In FY2008–09, Consumer Profile agriculture and agriculture-related services contributed The Andhra Pradesh distribution utilities have 14 percent 26.34 percent of state GDP , compared to 10.73 percent representation of agriculture consumers in the consumer for manufacturing industries and 47 .81 percent for the base, and 32 percent share of agricultural consumption, services sector.23 The state’s literacy rate is 67 .66 per- the same percentage as industrial sales (figure A.3). Con- cent, about 6 percent lower than the national average. sumption against unmetered agriculture use is higher than against metered agriculture use.26 Power Sector Reform The Andhra Pradesh State Electricity Board (APSEB), formed in 1959, was responsible for all three power-sector functions: generation, transmission, and distribution. The 24. As reported by the People’s Monitoring Group on Electricity Regulation (PMGR). 25. Power Finance Corporation (PFC) performance report. 21. Government of Andhra Pradesh official portal. 26. Projected values for FY2009–10, according to the Andhra 22. 2011 census. Pradesh Electricity Regulatory Commission (AERC) tariff order for 23. Details are available at http://www.indiastat.com. the state distribution utilities. 39 40 Lighting Rural India: Load Segregation Experience in Selected States Figure A.1 Energy Input and Sales Trend, Andhra Pradesh Energy input Energy sales 80,000 50 Distribution losses 70,000 68,851 62,295 40 60,000 56,980 59,063 53,501 53,165 Percent losses 50,000 47,831 48,398 Million kWh 44,847 30 39,652 40,000 30,000 17.10 10.10 20 15.06 14.66 14.22 20,000 10 10,000 0 0 2005–06 2006–07 2007–08 2008–09 2009–10 Source: PwC. Agriculture Profile: Climate and Crops Irrigation intensity—the ratio of gross irrigated area to net irrigated area—is 1.35 (FY2007–08 figure).27 Andhra Pradesh is characterized by a generally hot and humid climate. A reasonably long coastal belt accounts for the state’s relatively mild winters. Agricultural produc- Rural Load Segregation tion depends heavily on seasonal rainfall distribution. The two most important sources of rainfall are the southwest Initially, a virtual-segregation scheme was applied for all and northwest monsoon winds, occurring in June–Sep- mixed rural feeders, and was completed in 2005. How- tember and October–December, respectively. The net ever, this approach limited the three-phase supply avail- area cultivated comprises 39.4 percent of the total land able to non-agriculture rural consumers. In response, area (2007–08 figure). physical segregation has been planned and pilots are under way. The key objectives of physical load segrega- Rice is the state’s major staple crop. Various other tion are to (i) make three-phase supply available to rural important crops include jowar, bajra, maize, small millet, consumers 24 hours a day in order to boost small rural pulses, tobacco, cotton, and sugarcane. Cropping inten- industries and thus improve socioeconomic benefits and sity—the ratio of gross area sown to net area sown—is (ii) extend the daily 7–9 hours of three-phase supply avail- 1.26 (FY2007–08 figure). In FY2005–06, total irrigated able for pump sets. area represented 27 .8 percent of total cultivated area. 27. Details are available at http://agri.ap.nic.in/rainfallsenario.htm. Figure A.2 Recent Trend in State Subsidies and Financial Losses, Andhra Pradesh 100 79.80 80 Cash profit 60 28.66 46.15 Subsidies booked Billion Rs. 24.08 40 13.03 15.33 19.57 13.03 14.83 18.41 Subsidy received 20 0 1.51 5.81 4.66 8.46 -20 -40 -25.29 2004–05 2005–06 2006–07 2007–08 2008–09 Source: PwC. Annex. State Profiles 41 Figure A.3 Consumer and Consumption Mix of Distribution Utilities, Andhra Pradesh Consumers Consumption 1% 1% 8% 14% 9% 32% Agriculture 32% Households Business Industry Other 76% 21% 6% Source: PwC. Currently, there are about 13,731 feeders across the Power Sector Reform state’s four distribution companies; of these, 8,878 are The Gujarat Electricity Industry Reorganization and Regu- mixed rural feeders. The draft Detailed Project Report lation Act, passed in 2003, restructured the electricity (DPR) on segregating all agriculture feeders, prepared industry with the goal of improving the efficiency of in August 2010, estimates the total cost of segregating management and delivery of consumer services. Under mixed rural feeders at about US$568.7 million (Rs. 30.14 the Act’s provisions, the government framed the Gujarat billion). Electricity Industry Reorganization and Comprehensive Transfer Scheme of 2003, whereby the Gujarat Electric- ity Board was reorganized into seven successor compa- Gujarat nies with separate responsibilities for electricity trading, generation, transmission, and distribution (effective April Gujarat state, located in northwest India, is the country’s 1, 2005).31 seventh largest state, with a total land area of 196,077 km2. The state is bordered by Rajasthan to the northeast, Madhya Pradesh to the east, and Maharashtra and the Technical and Financial Trends Union territories of Diu, Daman, Dadra, and Nagar Haveli The FY 2007–08 socioeconomic review of Gujarat’s to the south. The Arabian Sea borders the state to both Directorate of Economics and Statistics estimated the the west and southwest.28 Gujarat’s population, at 60.38 state’s per capita electricity consumption at 1,354 kWh, million, accounts for about 5 percent of the country’s which is much higher than the national average. In recent total population.29 In FY2010, the state GDP was US$71.7 years, energy input and sales have increased,32 while dis- billion (Rs. 3.8 trillion),30 accounting for 6.89 percent of tribution losses have declined somewhat (figure A.4). India’s total GDP . Agriculture and agriculture-related ser- vices contribute more than 14 percent of state GDP , com- Gujarat is one of the few Indian states where the distri- pared to about 31 percent for manufacturing industries bution utilities have a profitable balance sheet. For most and 41.12 percent for the services sector (FY2008–09 years, the subsidies booked by the utilities, based on the figure). The state’s literacy rate is 79.31 percent, 5.27 government’s prior consent, and the subsidies received percent higher than the national average. are equal (figure A.5). 28. Government of Gujarat official portal (http://www.gujaratindia. com). 31. Official website of Gujarat Urja Vikas Nigam Ltd (GUVNL). 29. 2001 census. 32. Aggregate revenue requirement (ARR) tariff order for the respec- 30. Official Gujarat state portal, Department of Industry. tive distribution utilities. 42 Lighting Rural India: Load Segregation Experience in Selected States Figure A.4 Energy Input and Sales Trend, Gujarat 55,000 53,018 50 49,762 50,000 48,027 Energy input 44,544 40 45,000 Percent losses Energy sales Million kWh 40,012 40,000 26.69% 27.29% 36,867 30 Distribution losses 34,922 35,000 25.91% 24.53% 32,654 20 30,000 10 25,000 20,000 0 2006–07 2007–08 2008–09 2009–10 Source: PwC. Figure A.5 Recent Trend in State Subsidies and Distribution Profits, Gujarat 12.37 14 11.78 12.06 11.00 11.00 11.78 12 11.00 11.00 Cash profit 10 Billion Rs. Subsidies booked 8 6 Subsidy received 4 3.15 2.22 2.08 2.14 2 0 2005–06 2006–07 2007–08 2008–09 Source: Power Finance Corporation (PFC) performance report. Consumer Profile Gujarat is India’s largest producer of castor, tobacco, and isabgul (psyllium) and its second largest producer In FY2008–09, Gujarat’s distribution utilities had a total of sesame seed, cotton, and groundnut. Most crops consumer base of approximately 10 million, of which sown in winter and harvested in spring (rabi) and all sum- agriculture consumers accounted for 8 percent. Over the mer crops require irrigation; however, most crops sown same period, sales to agriculture consumers totaled 32 during the monsoon (kharif) are rainfed. Double-crop, percent (figure A.6). irrigated annual plantings have an estimated cropping intensity of 103–120 percent, averaging about 113 per- Agriculture Profile: Climate and Crops cent. The agriculture sector consumes a large percent- age of surface water and groundwater for irrigation. Even Well over half of Gujarat’s total land area can be charac- so, gross irrigated area accounts for only 31.8 percent of terized as arid and semi-arid (i.e., 24.94 percent in the the gross area sown. arid zone and 33.66 percent in the semi-arid zone). Aver- age annual rainfall varies widely throughout the state, ranging from just 300 mm in the western part of Kutch Rural Load Segregation to 2,100 mm in the southern part of Valsad district and In Gujarat, rural load segregation was introduced as part the Dangs. The monsoons arrive in mid-June and recede of the Jyoti Gram Yojna (JGY), with the goal of supply- by late-September. About 95 percent of total annual rain- ing continuous quality rural power supply. Introduced in fall is received during these months, with the maximum 2003–04, JGY covered more than 18,000 electrified villag- number of rainy days occurring in July and August. es.33 All mixed rural feeders were physically segregated The state features a wide diversity of major crops, includ- ing wheat, bajra, rice, maize, groundnut, mustard, ses- 33. Energy and Petro Chemicals Department, Government of ame, pigeon pea, green gram, sugarcane, and cotton. Gujarat. Annex. State Profiles 43 Figure A.6 Consumer and Consumption Mix of Distribution Utilities, Gujarat Consumers Consumption 2% 0% 5% 11% 8% 32% Agriculture 43% Households Business Industry Other 79% 15% 5% Source: Gujarat Electricity Regulatory Commission order on annual performance review for FY2009–10 (http://agri.gujarat.gov.in/). Figure A.7 A  gricultural Energy Consumption (Rs. 12.9 billion),34 largely funded through a state grant. and GDP Post JGY Since the JGY program ended, both agricultural energy consumption and agriculture GDP have generally trended 320 9,700 upward following a slight dip in FY2005–06 to 2006–07 312.80 310 (figure A.7). 9,600 300 9,605 9,500 Million kWh 290 Rajasthan Billion Rs. 9,400 280 273.54 9,300 270 265.62 Located in northwest India, Rajasthan is India’s largest 9,213 260 9,200 state in terms of land area, with a total of 342,269 km2. The state is bordered on the west and northwest by 250 9,100 9,062 Pakistan, on the north and northeast by Punjab, Haryana, 240 9,000 and Uttar Pradesh, on the east and southeast by Uttar 2005–06 2006–07 2007–08 Pradesh and Madhya Pradesh, and on the southwest by Agriculture GDP at Gujarat.35 Its population, about 68.62 million,36 accounts constant prices (1999–2000) for 5.67 percent of the country total. For FY2010–11, Agriculture energy state GDP was projected at US$57 .2 billion (Rs. 3.03 consumption in the state trillion), with a 9.69 percent growth rate. The agriculture sector contributed 26 percent of state GDP , compared to Source: PwC. 27 percent for manufacturing industries and 47 percent for the services sector.37 In 2001, the state ranked ninth in terms of the human development index. Its literacy rate is 67 .06 percent, about 7 percent below the national into JGY and agriculture-dominated feeders. JGY feed- average.38 ers provide rural households, commercial, and industrial users a continuous power supply, while ag-dominated feeders supply fixed duration power for agriculture use. Before JGY, virtual segregation was prevalent. During proj- ect execution, the concept of special design transform- 34. According to Gujarat’s JGY cell. ers was introduced, whereby farmers in remote areas 35. Official website of Government of Rajasthan. would receive a continuous single-phase supply through 36. 2011 census. ag-dominated feeders. Implemented from FY2002–03 to 37 . 2001 census, website of Directorate of Economics and Statis- tics, Government of Rajasthan. 2005–06, the scheme resulted in the creation of 1,904 38. Official website of Department of Industries, Government of JGY feeders, at a total cost of about US$243.4 million Rajasthan. 44 Lighting Rural India: Load Segregation Experience in Selected States Figure A.8 Energy Input and Sales Trend, Rajasthan 50,000 50 [[??]]% 40,000 40,844 40 Energy input [[??]]% 36,301 34,484 Percent losses 31.40% Energy sales Million kWh 30,195 30,634 30,633 30,000 26,649 30 Distribution losses 23,656 20,035 26.59% 25.08% 20,000 18,056 20 10,000 10 0 0 2005–06 2006–07 2007–08 2008–09 2009–10 Source: PwC. Power Sector Reform Consumer Profile A year after passage of the 1999 Power Sector Reform Act, Rajasthan’s distribution utilities have a total consumer the Rajasthan State Electricity Board was unbundled into base of approximately 5.3 million, of which agriculture five successor entities with separate functions for power consumers account for 12 percent (FY2009–10 figures). generation, transmission, and distribution. As a result, For FY2011–21, the utilities projected agriculture con- three independent distribution companies were formed: sumer sales at 39 percent (figure A.10).40 Jaipur, Ajmer, and Jodhpur Vidyut Vitran Nigam Ltd. Agriculture Profile: Climate and Crops Technical and Financial Trends Rajasthan’s climate is characterized, in large part, by arid Since 2005, distribution losses in Rajasthan have steadily and semi-arid conditions, with erratic and uneven rainfall declined (figure A.8)39 But the subsidy requirements pro- distribution. The state depends heavily on rainfed agricul- jected by the distribution utilities have risen, while the ture. Of the 21.6 million ha in cultivated area, assured subsidies received have remained flat (figure A.9). irrigation covers only 6.4 million ha. Even so, Rajasthan is 40. Consumer business plan (FY2010–11 to 2014–15) and consump- tion-ARR and retail tariff petitions (FY2011–12) of Rajasthan distribu- 39. Audited account statements of Rajasthan’s distribution utilities. tion utilities. Figure A.9 Recent Trend in State Subsidies and Financial Losses, Rajasthan 100 80 76.55 Cash profit 60 Subsidies booked 38.02 40 Subsidy received 20.16 16.29 17.41 Billion Rs 20 10.51 10.46 11.43 14.27 10.12 0 -5.71 -5.12 -2.99 -20 -21.67 -40 -60 -64.86 -80 2004–05 2005–06 2006–07 2007–08 2008–09 Source: PwC. Annex. State Profiles 45 Figure A.10 Consumer and Consumption Mix of Distribution Utilities, Rajasthan Consumers Consumption 2% 1% 12% 10% 9% 39% Agriculture 26% Households Business Industry Other 76% 19% 6% Source: PwC. a leading producer of coarse cereal. Wheat and barley are cultivated over large areas, as are pulses, sugarcane, and  gricultural Energy Consumption Figure A.11 A oilseeds. Cotton and tobacco are major cash crops. The and GDP after FRP Initiation state is among India’s largest producers of edible oils and is its second largest producer of oilseeds.41 350 326.10 12,000 313.02 296.07 9,786 300 267.80 10,000 8,145 Rural Load Segregation 250 8,000 Million kWh 6,658 Billion Rs. Load segregation in Rajasthan was initiated in 2005 5,623 as part of the Feeder Renovation Program (FRP). The 200 6,000 scheme’s main objective was to reduce distribution 100 4,000 losses on mixed rural feeders and provide increased supply to non-agriculture rural households. The FRP was 50 2,000 integrated with other system strengthening elements, including HVDS on agricultural feeders, DT metering, and 0 0 2005–06 2006–07 2007–08 2008–09 replacement of LT cables with ABC. Agriculture GDP at Rajasthan adopted virtual segregation, whereby single- constant prices (1999–2000) phase DTs were installed on existing rural feeders for Agriculture energy household and non-household loads. A roster switch was consumption in the state used to balance three-phase, agricultural supply hours Source: PwC. with single-phase hours when households received an unrestricted supply. The FRP covered a total of 8,126 feeders, with a total planned outlay of approximately US$846.2 million (Rs. 44.85 billion). Since 2005–06, agri- Himachal Pradesh to the north, and Rajasthan to the cultural energy consumption and agriculture GDP have south. The national capital territory of Delhi juts into consistently trended upward (figure A.11). the state. Its total geographic area is 44,212 km2. Hary- ana is home to more than 25.35 million people, with a population density of 573.4 persons per km2. State GDP Haryana is estimated at about US$40.8 billion (Rs. 2.2 trillion) (FY2009–10 figure). Agriculture and agriculture-related Haryana, one of India’s most prosperous states, is situ- services account for 16.1 percent of state GDP , com- ated in the northwestern part of the country. It is bor- pared to 30.5 percent for manufacturing industries and dered by Uttar Pradesh to the east, Punjab to the west, 53.4 percent for the services sector. Per capita income is 41. Details are available at http://www.rajasthankrishi.gov.in/. 46 Lighting Rural India: Load Segregation Experience in Selected States Figure A.12 Energy Input and Sales Trend, Haryana 35,000 32 30.97% 30,000 31,095 30 Energy input 29.15% 26,145 25,379 Percent losses 23,514 28.06% Energy sales Million kWh 22,042 22,868 25,000 19,321 28 Distribution losses 16,660 18,258 26.46% 15,216 26.10% 20,000 26 15,000 24 5,000 22 2005–06 2006–07 2007–08 2008–09 2009–10 Source: PwC. estimated at Rs 78,781 (FY2009–10 figure).42 The state’s year, two distribution companies—Uttar Haryana Bijli Vit- literacy rate is 71.40 percent, about 3 percent below the ran Nigam Ltd (UHBVNL) and Dakshin Haryana Bijli Vitran national average. Nigam Ltd (DHBVNL)—were separated from HVPNL to focus exclusively on distribution and retail supply in their respective geographic areas. Power Sector Reform Subsequent to the Haryana Reform Act of 1998, the state Technical and Financial Trends government unbundled the Haryana State Electricity Board into two independent companies: Haryana Power In recent years, energy input and sales in Haryana have Generation Corporation Ltd (HPGCL) and Haryana Vidyut trended upward, as distribution losses have generally Prasaran Nigam Ltd (HVPNL). HPGL was put in charge of declined (figure A.12). However, the subsidies required electricity generation, while HVPNL was responsible for by the distribution utilities have exceeded the amounts transmission, distribution, and retail supply. The following received, causing financial losses to increase (figure A.13). 42. http://www.esaharyana.gov.in. Figure A.13 Recent Trend in State Subsidy and Financial Losses, Haryana 50 41.51 40 30 26.38 32.89 Cash profit 23.39 Billion Rs. 26.37 20 17.97 22.76 Subsidies booked 13.09 Subsidy received 15.44 10 12.89 0 -2.68 -4.04 -10 -7.85 -14.85 -20 -16.91 2004–05 2005–06 2006–07 2007–08 2008–09 Source: PwC. Annex. State Profiles 47 Consumer Profile (rabi) (e.g., chili, bajra, jawar, pulses, and vegetables) and rainy season (kharif) (e.g., sugarcane, groundnut, maize, Distribution utilities in Haryana have a total consumer and paddy). Irrigation, a major source of water for cultiva- base of approximately 4.5 million. Of this number, agri- tion, depends on the various canals operating throughout culture consumers account for just 11 percent, yet repre- the state (e.g., Western Yamuna, Gurgaon, Jui, Jawahar- sent 39 percent of total sales (figure A.14).43 lal Lal Nehru, and Bhakra).44 Agriculture Profile: Climate and Crops Rural Load Segregation Haryana’s climate varies from arid to semi-arid. Contrast- Haryana’s scheme for rural load segregation, initiated in ing seasons feature hot summers, with temperatures FY2005–06, aimed to regulate supply to rural consumers rising up to 50°C in May and June, and cold winters, by providing rural households urban patterns of power with temperatures falling to as low as 1°C in December supply, improving voltage, stabilizing the distribution sys- and January. Average annual rainfall is 455 mm; about tem, and reducing system losses. The scheme separated 70 percent occurs between July and September and the rural domestic load from agriculture load by erecting remainder in December–February. Rainfall distribution is dedicated 11-kV feeders. Under the scheme, 1,226 new uneven, with the Shivalik Hills region receiving the larg- feeders were erected with a total length of 17 ,308 km. All est amounts and Aravali Hills the least. feeders were equipped with bulk meters. Eight-six percent of Haryana’s total land area of 4.4 mil- Project implementation relied mainly on turnkey con- lion ha is arable (i.e., 3.8 million ha); of this amount, 96 tracts awarded through the distribution utilities’ planning percent is under cultivation. The state has a net cropped and design units. In FY2006–07 , multiple agencies were area of 3.62 million ha, with a cropping intensity of 177 awarded contracts with planned completion dates within percent. The northwest climate zone is suitable for grow- two years of the project starting date. Segregation work ing rice, wheat, vegetables, and temperate fruits, while was delayed, but was completed in mid-2010 when sub- the southwest is suitable for high-quality produce, tropical jected to regular monitoring by the state government. fruits, exotic vegetables, and herbal and medicinal plants. Crop production can be broadly group into winter season 43. Based on ARR petitions of DHBVNL and UHBVNL for FY2011–12. 44. More details are available at http://agriharyana.nic.in/. Figure A.14 Consumer and Consumption Mix of Distribution Utilities, Haryana Consumers Consumption 2% 0% 8% 10% 11% 39% Agriculture 27% Households Business Industry Other 77% 7% 19% Source: PwC. The World Bank The World Bank Group Asia Sustainable and Alternative Energy Program 1818 H Street, NW Washington, DC 20433 USA www.worldbank.org/astae